(Perspectives On Individual Differences) Robert T. Brown (Auth.), John A. Glover, Royce R. Ronning, Cecil R. Reynolds (Eds.) - Handbook of Creativity-Springer US (1989) PDF

Handbook of
Creativity
PERSPECTIVES ON INDIVIDUAL DIFFERENCES
CECIL R. REYNOLDS, Texas A&M University, College Station

ROBERT T. BROWN, University oj North Carolina, Wilmington
Recent volumes in the series
THE CAUSES AND CURES OF CRIMINALITY

Hans J. Eysenck and Gisli H. Gudjonsson
HANDBOOK OF CREATIVITY
Edited by John A. Glover, Royce R. Ronning, and Cecil R. Reynolds
HANDBOOK OF MUL TIVARIATE EXPERIMENTAL

PSYCHOLOGY
Second Edition
Edited by John R. Nesselroade and Raymond B. Cattell
HISTORICAL FOUNDATIONS OF EDUCATIONAL

PSYCHOLOGY
Edited by John A. Glover and Royce R. Ronning
THE INDIVIDUAL SUBJECT AND SCIENTIFIC PSYCHOLOGY

Edited by Jaan Valsiner
LEARNING STRATEGIES AND LEARNING STYLES

Edited by Ronald R. Schmeck
METHODOLOGICAL AND STATISTICAL ADVANCES IN THE

STUDY OF INDIVIDUAL DIFFERENCES
Edited by Cecil R. Reynolds and Victor L. Willson
THE NEUROPSYCHOLOGY OF INDIVIDUAL DIFFERENCES

A Developmental Perspective
Edited by Lawrence C. Hartlage and Cathy F. Telzrow
PERSONALITY DIMENSIONS AND AROUSAL

Edited by Jan Strelau and Hans J. Eysenck
THEORETICAL FOUNDATIONS OF BEHAVIOR THERAPY

Edited by Hans J. Eysenck and Irene Martin
A Continuation Order Plan is available for this series. A continuation order will bring
delivery of each new volume immediately upon publication. Volumes are billed only upon
actual shipment. For further information please contact the publisher.
Handbook of
Creativity
Edited by
John A. Glover
Teachers College
Ball State University
Muncie, Indiana
Royce R. Ronning
University of Nebraska
Lincoln, Nebraska
and
Cecil R. Reynolds
Texas A&M University
College Station, Texas
Springer Science+Business Media, LLC

ISBN 978-1-4419-3212-9 ISBN 978-1-4757-5356-1 (eBook)
DOI 10.1007/978-1-4757-5356-1
© 1989 Springer Science+Business Media New York

Originally published by Plenum Press, New York in 1989
Softcover reprint of the hardcover 1st edition 1989
All rights reserved

No part of this book may be reproduced, stored in a retrieval system, or transmitted
in any form or by any means, electronic, mechanical, photocopying, microfilming,
recording, or otherwise, without written permission from the Publisher
This volume is dedicated to Dr. E. Paul Torrance
E. Paul and Pansy Torrance

Contributors
Bonnie B. Armbruster Roger Bruning

Center for the Study of Reading Department of Educational Psychology
University of Illinois at Urbana-Champaign University of Nebraska
Champaign, Illinois 61820 Lincoln, Nebraska 68588-0641
Patricia Bachelor Linda J. Carey

Department of Psychology Center for the Study of Writing
California State University at Long Beach Carnegie-Mellon University
Long Beach, California 90840 Pittsburgh, Pennsylvania 15238
Carolyn A. Ball John Clement

Department of Psychology Scientific Reasoning Research Institute
University of Missouri University of Massachusetts
Columbia, Missouri 65211 Amherst, Massachusetts 01003
Michael Basseches
Massachusetts School of Professional Carolyn A. Colvin
Psychology, and Department of Teacher Education
Clinical Developmental Institute San Diego State University
Belmont, Massachusetts 02178 San Diego, California 92182
Suzanne Benack E. Thomas Dowd

Department of Psychology Department of Educational Psychology
Union College Kent State University
Schenectady, New York 12308 Kent, Ohio 44242
Bruce K. Britton Karen Fitzgerald

Department of Psychology Department of Psychology
University of Georgia University of Missouri
Athens, Georgia 30602 Columbia, Missouri 65211
Robert T. Brown Linda Flower

Department of Psychology Center for the Study of Writing
University of North Carolina at Wilmington Carnegie-Mellon University
Wilmington, North Carolina 28403-3297 Pittsburgh, Pennsylvania 15238
vii
viii CONTRIBUTORS
John H. Flowers Colin Martindale

Department of Psychology Department of Psychology
University of Nebraska University of Maine
Lincoln, Nebraska 68588 Orono, Maine, 04469-0140
Calvin P. Garbin Linda F. Mattocks

University of Nebraska University of Georgia
Lincoln, Nebraska 68588 Athens, Georgia 30602
Shawn M. Glynn Mary L. Means

Department of Educational Psychology Department of Psychology
University of Georgia Learning Research and Development Center
Athens, Georgia 30602 University of Pittsburgh
Pittsburgh, Pennsylvania 15260
Elizabeth M. Goetz
The Edna A. Hill Child Development Laboratory William B. Michael
Preschools Department of Educational Psychology
Department of Human Development University of Southern California
University of Kansas Los Angeles, California 90089
Lawrence, Kansas 66045 David Moshman
Department of Educational Psychology
Patricia A. Haensly University of Nebraska
Department of Educational Psychology Lincoln, Nebraska 68588
College Station, Texas 77843 K. Denise Muth
Department of Elementary Education
John R. Hayes University of Georgia
Department of Psychology Athens, Georgia 30602
Carnegie-Mellon University
Pittsburgh, Pennsylvania 15238 John A. O'Looney
Department of Language Education
P. Paul Heppner University of Georgia
Department of Psychology Athens, Georgia 30602
University of Missouri
Columbia, Missouri 65211 Robert Prentky
Research Department
Dennis Hocevar The Massachusetts Treatment Center
Department of Educational Psychology Bridgewater, Massachusetts 02324; and
University of Southern California Department of Psychiatry
Los Angeles, California 90089-0031 Boston University School of Medicine
Boston, Massachusetts 02118
Carolyn A. Jones Cecil R. Reynolds
University of Missouri Texas A&M University
Columbia, Missouri 65211 College Station, Texas 77843
Leslie E. Lukin Lyle F. Schoenfeldt

Department of Educational Psychology Department of Management
University of Nebraska Texas A & M· University
Lincoln, Nebraska 68588 College Station, Texas 77843
CONTRIBUTORS ix
Margaret Semrud-Clikeman James F. Voss

Department of Educational Psychology Department of Psychology
University of Georgia Learning Research and Development Center
Athens, Georgia 30602 University of Pittsburgh
Pittsburgh, Pennsylvania 15260
Barry S. Stein
Department of Educational Psychology
Tennessee Technological University
Richard W. Woodman
Cookeville, Tennessee 38505
Department of Management
Thomas Swan College Station, Texas 77843
Department of Psychology
Union College
Schenectady, New York 12308
Claudia R. Wright
P. E. Vernon Department of Educational Psychology and
Late of the Department of Psychology Administration
University of Calgary California State University
Calgary, Alberta, Canada T3A 2E3 Long Beach, California 90840
Preface
The motivation underlying our development of a "handbook" of creativity was different from what
usually is described by editors of other such volumes. Our sense that a handbook was needed sprang not
from a deluge of highly erudite studies calling out for organization, nor did it stem from a belief that the
field had become so fully articulated that such a book was necessary to provide summation and reference.
Instead, this handbook was conceptualized as an attempt to provide structure and organization for a field of
study that, from our perspective, had come to be a large-scale example of a "degenerating" research
program (see Brown, Chapter 1).
The handbook grew out of a series of discussions that spanned several years. At the heart of most of
our interactions was a profound unhappiness with the state of research on creativity. Our consensus was
that the number of "good" works published on creativity each year was small and growing smaller.
Further, we could not point to a journal, text, or professional organization that was providing leadership for
the field in shaping a scientifically sound framework for the development of research programs in
creativity. At the same time, we were casting about for a means of honoring a dear friend, E. Paul
Torrance. Our decision was that we might best be able to honor Paul and influence research on creativity by
developing a handbook designed to challenge traditional perspectives while offering research agendas
based on contemporary psychological views.
The contributors to our volume were selected with extreme care, because our intent was not to put
together merely a collection of' 'name" individuals in creativity. Instead, we chose to solicit chapters from
people of extraordinary ability who we believed could step beyond the context of the literature on creativity
and address aspects of the field from the broader perspective of contemporary psychology. It is our
contributors who must be credited for the positive features of this volume. In our judgment, our contrib-
utors wrote chapters of uncommon erudition, chapters, we believe, will "make a difference." The
scholars represented in this volume made every effort asked of them and taught us a great deal about the
meaning of scholarship. Any omissions or errors are our responsibility alone.
We had three major goals for this volume. First, we wanted to provide a critique of the level of
development of research in creativity. Second, we wanted to articulate a series of research agendas that
could lead to "progressive" rather than "degenerating" research programs. Third, we hoped to provide
structure for the field either through intellectual leadership or, alternatively, by providing a target for
researchers who wished to throw intellectual rocks at somebody. The first two goals are part and parcel of
each chapter in this volume. The level of criticism ranges from questions about the utility of the concept
itself (Brown) to the observation that there has been no research examining the role of perception in the
creative process despite a great amount of talk about the issue (Flowers and Garbin, Chapter 8). The
authors of each chapter also describe research agendas growing out of the contents of their chapters,
agendas we believe could lead to fruitful, reinvigorated programs of research. Whether or not we have
attained our third goal remains to be seen.
The handbook is organized into four parts. Part I, "The Nature ofthe Beast," consists of six chapters
xi
xii PREFACE
that examine basic issues related to the definition of creativity and how it is to be measured. Robert T.
Brown begins with his chapter, "Creativity: What Are We to Measure?" which sets the tone for the
volume as he analyzes the scientific utility of the global concept of creativity as well as the specific
definitions of creativity that have appeared in the literature. The second chapter, prepared by Michael and
Wright, is entitled "Psychometric Issues in the Assessment of Creativity," and follows directly from
Brown's chapter but focuses entirely on issues often neglected in studies of creativity-the psychometrics
involved in assessing the construct. In Chapter 3, "A Taxonomy and Critique of Measurements Used in
the Study of Creativity," Hocevar and Bachelor build on the previous two chapters and provide an
exhaustive analysis of currently available measures of creativity. Chapter 4, written by Woodman and
Schoenfeldt, is entitled "Individual Differences in Creativity: An Interactionist Perspective." Their
analysis of individual differences is followed by P. E. Vernon's chapter, "The Nature-Nurture Problem in
Creativity. " The final chapter in this part, "Creativity and Intelligence," was prepared by Haensly and
Reynolds, who, in their chosen topic, return to the basic questions addressed in the first chapter concerning
the nature of creativity.
Part II of the handbook, "Cognitive Models of Creativity," is devoted simply to that topic-the
development and articulation of cognitively oriented models of creative thought. It opens with Hayes's
chapter, "Cognitive Processes in Creativity. " This general perspective then is followed by a focus on
perception in the chapter written by Flowers and Garbin, "Creativity and Perception." Stein then empha-
sizes memorial processes in his chapter, "Memory and Creativity," and the focus shifts to how people
think about their thinking in Armbruster's chapter, "Metacognition in Creativity." The last two chapters
in this part of the book emphasize a developmental perspective. Moshman and Lukin examine the
relationship between the development of reasoning and creative abilities in their chapter, "The Creative
Construction of Rationality: A Paradox?" followed by Benack, Basseches, and Swan who discuss "Di-
alectical Thinking and Adult Creativity. "
Part III of the volume is entitled "Personalogical Variables and Creativity." This section, featuring
four different perspectives on creativity based on personality dimensions, opens with Martindale's chapter,
"Personality, Situation, and Creativity," which provides a general perspective on personalogical ap-
proaches to creativity. Next, Dowd explores the construct of "self" and its relationship to creativity in
"The Self and Creativity: Several Constructs in Search of a Theory." Prentky analyzes an extremely
interesting topic in his chapter, "Creativity and Psychopathology: Gamboling at the Seat of Madness."
Finally, Heppner, Fitzgerald, and Jones examine the role of creativity in therapeutic settings in "Examin-
ing Counselors' Creative Processes in Counseling."
The fourth part of the volume, "Applications," examines a broad array of issues involved in
enhancing creative abilities. Three chapters emphasize the development of creativity in writing. Carey and
Flower introduce the section with their chapter, "Foundations for Creativity in the Writing Process:
Rhetorical Representations of Ill-Defined Problems," followed by one written by O'Looney, Glynn,
Britton, and Mattocks, "Cognition and Writing: The Idea Generation Process." The last of the three
chapters centered on writing takes a particularly effective applied approach as Colvin and Bruning
examine "Creating the Conditions for Creativity in Reader Response to Literature. "
Three chapters focus on science and social science. These chapters are organized by Clement's
"Learning via Model Construction and Criticism: Protocol Evidence on Sources of Creativity in Science."
A more applied approach can be seen in the chapter developed by Glynn, Britton, Semrud-Clikeman, and
Muth, "Analogical Reasoning and Problem Solving in Science Textbooks." The last chapter in this
sequence was developed by Voss and Means and is entitled "Toward a Model of Creativity Based upon
Problem Solving in the Social Sciences."
The final two chapters in the volume are highly applied and emphasize the facilitation of creativity at
different stages oflife. Goetz's chapter, "The Teaching of Creativity to Preschool Children: The Behavior
Analysis Approach," summarizes the last several years of her research program. The emphasis shifts to
adults in Britton and Glynn'S chapter, "Mental Management and Creativity: A Cognitive Model of Time
Management for Intellectual Productivity. "
A very large number of people were involved in putting this volume together-far too many for us to
list in this brief space. Very grateful thanks, however, must be extended to some scholars who were
particularly helpful. We thank Steve Benton and Barbara Plake for their erudite reviews. We thank Mike
PREFACE xiii
Shaughnessy, former president of the National Association of Creative Adults and Children and editor of
their journal, for frank discussions concerning the state of the field. John Zimmer, a man who needs no
introduction, was his usual cogent and helpful self when asked to analyze difficult issues. We thank E. Paul
Torrance, to whom this volume is dedicated, for his unflagging support and insistence on rigorous analyses
ofthe area. Finally, we must thank Eliot Werner, our editor at Plenum, who encouraged the planning and
development of the volume and was extremely helpful in shaping our thinking about the nature of this
volume.
As is the case in any edited volume, the handbook did not turn out exactly as it was planned. Illnesses,
family emergencies, job changes, and other circumstances sometimes defeat the best of plans. Conse-
quently, chapters focusing on creativity in mathematics, creativity in music, and the development of
cognitive skills in students were not included. Our greatest losses, though, were far larger and more deeply
felt. Don MacKinnon, who was working on a chapter focused on the criterion problem, died last year. In
addition, Philip E. Vernon died shortly after finishing his chapter for this volume. They will be sorely
missed by their friends, family, and indeed all psychologists.
This volume does not represent a final effort. Hindsight shows us gaps and problems we should have
foreseen but did not. In particular, we hope one day to be able to compile a more complete set of works
focused on domain-specific creativity. We also hope to be able to examine the influence of journals and
organizations on the quality of research in the area. In the meantime, we do very much hope that our readers
find the chapters in this volume to be as interesting and exciting as we did.
JOHN A. GLOVER
ROYCE R. RONNING
CECIL R. REYNOLDS
Contents
PART I THE NATURE OF THE BEAST
Chapter 1 Creativity: What are We to Measure? 3

Robert T. Brown
What is Creativity? .............................................................. 3

Historical Antecedents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Creativity and Scientific Methodology ............................................... 6
Testing for Intelligence and Testing for Creativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Problems with the Literature ............................................. . . . . . . . . . . 9
Process and Product: The Criterion Issue ............................................. 11
Guilford's Views on Creativity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Other Divergent Thought Approaches: Creativity as an Associative Process ................. 19
Overall Evaluation of Creativity as Divergent Thought. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Problem Finding as Creativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Evaluation as a Creative Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Should We Just Leave It All to Chance? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Amabile's Componential/Social Psychological Approach. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Conclusions ............................................................ . . . . . . . . 28
References ..................................................................... 30
Chapter 2 Psychometric Issues in the Assessment of Creativity 33

William B. Michael and Claudia R. Wright
Overview ............................................................. . . . . . . . . . 33
Alternative Methods for Measurement of Creativity .................................... 33
Concerns Relating to Validity ...................................................... 34
Concerns Relating to Reliability .................................................... 44
Concerns Relating to Scoring of Creativity Measures ................................... 46
Concerns Relating to the Use of Norms .............................................. 49
Test Administration Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Concluding Statement ............................................................ 50
References ...... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
xv
xvi CONTENTS
Chapter 3 A Taxonomy and Critique of Measurements Used in the Study of Creativity. . . . . . 53

Dennis Hocevar and Patricia Bachelor
A Taxonomy of Creativity Measurements ............................................ 53

Discussion and Critique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Conclusions .................................................................... 62
Agenda for Future Research on Measurement Issues .................................... 64
Summary...................................................................... 69
References ..................................................................... 70
Chapter 4 Individual Differences in Creativity: An Interactionist Perspective 77

Richard W. Woodman and Lyle F. Schoenfeldt
Theoretical Explanations for Differences in Creative Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

An Interactionist Model of Creative Behavior ......................................... 79
Factors Underlying Individual Differences in Creativity ................................. 80
Summary and Conclusions: An Interactionist Approach to Individual Differences in Creativity .. 87
References ..................................................................... 89
Chapter 5 The Nature-Nurture Problem in Creativity 93

P. E. Vernon
Genetic Theory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Definitions ..................................................................... 94
Distribution of Creativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Sociology of Creativity ........................................................... 95
Cognitive and Motivational Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
The Relation of General Intelligence and Creativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Other Cognitive Indices of Creativity: Divergent Thinking ............................... 98
Evidence for Genetic Factors in Creativity: Rarity of Genius ............................. 99
Neurological Factors ............................................................. 100
Consistency, or Predictability, from Early Childhood to Adulthood. . . . . . . . . . . . . . . . . . . . . . .. 101
Sex or Gender Differences .................................................... , . . . . 10 1
FamilyResemblance ............................................................. 103
Upbringing and Environment ...................................................... 104
Trainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 105
Summary and Conclusions ........................................................ 106
Future Developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 108
References ..................................................................... 108
Chapter 6 Creativity and Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

Patricia A. Haensly and Cecil R. Reynolds
The Relationship between Creativity and Intelligence ................................... 111
Intelligence as a Component in Theories and Assessment of Creativity ..................... 114
Creativity as a Component in Theories and Assessment of Intelligence ..................... 120
Problems of Restricted Range in Studies of Creativity and Intelligence ..................... 124
The Synergistic Effect between Creativity and Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 127
Integration: A Model for Creativity and Intelligence .................................... 129
References ..................................................................... 130
CONTENTS xvii
PART II COGNITIVE MODELS OF CREATIVITY
Chapter 7 Cognitive Processes in Creativity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 5

John R. Hayes
Characteristics of Creative People . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

What Cognitive Processes Are Involved in Creative Acts? ............................... 138
Summary ...................................................................... 143
References ..................................................................... 144
Chapter 8 Creativity and Perception. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

John H. Flowers and Calvin P. Garbin
Why Have Perceptual Psychologists Had Little to Say about Creativity? .................... 147
Is Perceptual Organization Incompatible with Creative Thought? .......................... 148
The Importance of Executively Controlled Processing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
A Configural Conception of Creativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 150
Impairment of Sensory Processes ................................................... 150
Environmental Constraints on Sensory Input .......................................... 151
Altered and Transitional States-Decoupling Perceptual Mechanisms from Sensory Input. . . . . . 152
Assessing Individual Variation in Perceptual Organization Ability. . . . . . . . . . . . . . . . . . . . . . . .. 154
Individual Variation in Intramodal Visual Code Transformation. . . . . . . . . . . . . . . . . . . . . . . . . .. 156
An Example of the Interplay of Involuntary and Executively Driven Creative Processes:
Synesthesia and Cross-Modal Representations ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 157
A Flow Diagram of the Creative Process and Its Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
References ..................................................................... 161
Chapter 9 Memory and Creativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

Barry S. Stein
The Effects of Memory on Creativity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 164

Methods for Investigating Creative Transfer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 165
Creating Opportunities for Transfer ................................................. 171
The Effects of Problem Definitions on Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Implications for Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 174
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 175
References ..................................................................... 175
Chapter 10 Metacognition in Creativity 177

Bonnie B. Armbruster
What is Metacognition? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 177

Concluding Remarks ............................................................. 181
References ..................................................................... 181
Chapter 11 The Creative Construction of Rationality: A Paradox? 183

David Moshman and Leslie E. Lukin
Lori: The Creative Construction of Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

The Development of Reasoning .................................................... 185
xviii CONTENTS
Empiricist and Nativist Accounts of the Development of Reasoning. . . . . . . . . . . . . . . . . . . . . . .. 188

A Constructivist Account of Rationality .............................................. 193
The Subjective Creation of Objectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 195
References ..................................................................... 197
Chapter 12 Dialectical Thinking and Adult Creativity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199

Suzanne Benack, Michael Basseches, and ThomaJ Swan
Adult Cognitive Development: The Move from Formal to Postformal Operations . . . . . . . . . . . .. 199
Dialectical Thinking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 200
Characteristics of the Creative Process ............................................... 203
Dialectical Ontology and Creativity ................................................. 204
Dialectical Epistemology and Creativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 205
Metasystematic Thinking and Creative Work: The Evolving Systems Approach. . . . . . . . . . . . .. 206
Summary and Implications for Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 207
References ..................................................................... 208
PART III PERSONALOGICAL VARIABLES AND CREATIVITY
Chapter 13 Personality, Situation, and Creativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 211

Colin Martindale
Introduction .................................................................... 211

Prerequisites for Creativity ........................................................ 212
The Creative Process ............................................. . . . . . . . . . . . . . . .. 214
Theories of the Creative Process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 216
The Creative Personality .......................................................... 219
Situational Variables ............................................................. 226
Research Questions .............................................................. 228
Conclusions .................................................................... 228
References ..................................................................... 228
Chapter 14 The Self and Creativity: Several Constructs in Search of a Theory 233
E. Thomas Dowd
Definition of Creativity ........................................................... 233

Definition of the Self . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 233
Creativity and the Self-Concept. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 234
Creativity and Self-Actualization ................................................... 235
Creativity and Locus of Control .................................................... 235
Creativity and Divergent Thinking .................................................. 236
Creativity and Intrinsic/Extrinsic Motivation .......................................... 237
Creativity and Openness to Inner Experiences ......................................... 237
Creativity and Complex Thinking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 237
Creativity and Autonomy/Independence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 237
Critique and Suggestions for Future Research ......................................... 238
Fostering Creativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 239
References .......................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 241
CONTENTS xix
Chapter 15 Creativity and Psychopathology: Gamboling at the Seat of Madness . . . . . . . . . . .. 243

Robert Prentky
An Historical Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ., 244

Research on the Personality Correlates of Creativity .................................... 252
Research on Cognitive Aspects of Creativity .......................................... 256
Concluding Thoughts and Speculations .............................................. 261
References ..................................................................... 266
Chapter 16 Examining Counselors' Creative Processes in Counseling . . . . . . . . . . . . . . . . . . . .. 27 1

P. Paul Heppner, Karen Fitzgerald, and Carolyn A. Jones
The Counseling Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 272

Counselor Creativity within Counseling .............................................. 272
Variables Affecting Creativity within Counseling ...................................... 275
Final Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 278
References ..................................................................... 279
PART IV APPLICATIONS
Chapter 17 Foundations for Creativity in the Writing Process: Rhetorical Representations

of Ill-Defined Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 283
Linda J. Carey and Linda Flower
Introduction: An Approach to Studying Creativity in Writing . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 283

Creativity and Ill-Defined Problems ................................................. 284
The Effect of Task Representation on Originality ...................................... 288
Integrating Topic and Rhetorical Knowledge in Initial Planning . . . . . . . . . . . . . . . . . . . . . . . . . .. 290
Developing and Applying Problem-Solving Strategies for Revision ........................ 295
References ..................................................................... 302
Chapter 18 Cognition and Writing: The Idea Generation Process ....................... 305
John A. O'Looney, Shawn M. Glynn, Bruce K. Britton, and Linda F. Mattocks
Human Information-Processing Model ............................................... 306

Models of Writing: Relationship of Idea Generation to Other Writing Subprocesses .. . . . . . . . .. 308
Cognitive Constraints Imposed on Idea Generation When Writing ......................... 312
A Framework for Methods ofIdea Development. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 314
Questions to Prompt Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 317
An Illustrational Plan for Future Research ............................................ 318
References ..................................................................... 319
Chapter 19 Creating the Conditions for Creativity in Reader Response to Literature 323
Carolyn A. Colvin and Roger Bruning
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 323
Reader Response: Responding to Reading by Writing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 324
xx CONTENTS
Creative Dimensions of Reader Response. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 327

Enabling Conditions for Creativity in Reader Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 331
Teaching Methods in Reader Response. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 334
Issues for Research in Reader Response .............................................. 335
Summary ...................................................................... 336
References ..................................................................... 337
Chapter 20 Learning via Model Construction and Criticism: Protocol Evidence on Sources
of Creativity in Science . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 341
John Clement
Introduction .................................................................... 341

Background Questions from Philosophy of Science. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 342
Some Possible Views of Hypothesis Formation Processes in the Individual Scientist .......... 345
Evidence from Thinking Aloud Protocols on Model Construction Cycles Using Analogies ..... 348
Eureka or Accretion? The Presence of Insight in S2's Protocol. . . . . . . . . . . . . . . . . . . . . . . . . . .. 365
Creative Mental Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 369
Darwin's Theory of Natural Selection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 373
Features of Creative Thinking and Implications for Future Research ....................... 374
EducationalImplications .......................................................... 376
Summary ...................................................................... 377
References ..................................................................... 380
Chapter 21 Analogical Reasoning and Problem Solving in Science Textbooks ............. 383
Shawn M. Glynn, Bruce K. Britton, Margaret Semrud-Clikeman, and K. Denise Muth
Analogical Reasoning and Comprehending Science Text ................................ 383

The Teaching-with-Analogies (TWA) Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 389
Creative Function of Analogies . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 392
Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 396
References ..................................................................... 397
Chapter 22 Toward a Model of Creativity Based upon Problem Solving

in the Social Sciences .................................................. 399
James F. Voss and Mary L. Meam
On the Solving of Ill-Structured Problems ............................................ 400

Prior Knowledge, Search, and Creativity ............................................. 403
Value, Affect, and the Creative Process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 407
Creativity in the Social Sciences ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 408
Instruction and Creativity in the Social Sciences ....................................... 409
Summary ...................................................................... 409
References ..................................................................... 410
Chapter 23 The Teaching of Creativity to Preschool Children:

The Behavior Analysis Approach ........................................ 411
Elizabeth M. Goetz
Characteristics of Behavior Analysis and Traditional Approaches . . . . . . . . . . . . . . . . . . . . . . . . .. 411

CONTENTS xxi
Cohorts of Early Training. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 412
Definition of Creativity ........................................................... 413
Determination of Originality ....................................................... 414
Validation Measures ............................................................. 415
Unresolved Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 415
Individual Analysis Experimental Designs ............................................ 416
Behavior Principles .............................................................. 417
Characteristics of Effective Teaching of Creativity to Preschool Children ................... 418
Development of a Focused Environment for Creativity of Preschool Children . . . . . . . . . . . . . . .. 424
Summary and Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 426
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 426
Chapter 24 Mental Management and Creativity: A Cognitive Model of Time Management

for Intellectual Productivity ............................................ 429
Bruce K. Britton and Shawn M. Glynn
Introduction .................................................................... 429

The Time Manager. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 430
Special Properties of Time Management for the Intellectually Productive ................... 434
Summary ...................................................................... 437
References ..................................................................... 440
Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 441
PART!
The Nature of the Beast
This first part of the handbook attempts to define creativity, to examine issues related to its measurement,
and to consider creativity in relation to other abilities. Of the six chapters comprising this part of the
volume, Brown opens with an analysis of the concept of creativity and its value as a heuristic in psychol-
ogy. Michael and Wright then examine psychometric issues involved in attempts to assess creativity. Their
chapter is followed by Hocevar and Bachelor's critique of instruments designed to measure creativity.
After this emphasis on measurement, the focus shifts next to Woodman and Schoenfeldt's analysis of
individual differences in creativity, and then to Vernon, who considers the nature-nurture issue with
regard to creativity. The first part closes with a return to questions centering on the nature of creativity in
the chapter written by Haensly and Reynolds, as they seek to determine the relationship of creativity to
intelligence.
CHAPTER 1
Creativity
WHAT ARE WE TO MEASURE?
Robert T. Brown
Creativity consists of at least four components: (1) tests suffices, but focus is increasingly on actual
the creative process, (2) the creative product, (3) creative performance. From the standpoint of em-
the creative person, and (4) the creative situation pirical testing of a theoretical creativity construct,
(MacKinnon, 1970; Mooney, 1963). It has been anchoring in performance is mandatory. This ne-
studied from so many frequently incompatible the- cessity, however, raises the issue of the criterion for
oretical perspectives, each with its own assump- creative performance, which is itself difficult and
tions, methodologies, biases, and even meta- controversial. This chapter will discuss the issue as
theoretical view, that coverage of all in a single it pertains to creativity itself, but in an admittedly
chapter is not possible. This chapter concentrates superficial way. Because some creativity theories
on the divergent thinking approach to the study of are tied closely to particular creativity tests, criteri-
the creative process-an approach that has the most on validity of the tests is important in evaluating
explicitly developed theoretical base, underlies these theories and will be considered in this chapter.
most creativity tests, and has generated the most That coverage will necessarily result in some over-
empirical research. Along the way, information on lap with other chapters in this handbook, particu-
some other components and theoretical approaches larly that of Hocevar and Bachelor.
will be presented. Most other approaches are cov-
ered in detail in other chapters in this handbook, and
general presentations of creativity theories are in What Is Creativity?
Albert (1983), Bloomberg (1973), Busse and
Mansfield (1980), and Vernon (1970). The question of what wnstitutes creativity needs
A discussion of underlying processes does not to be addressed on two quite different levels: First,
exist in a vacuum; after all, one needs an observable what individual and/or contextual factors or pro-
guide to the unobservable. With creativity, unfortu- cesses lead to a creative product? The answer to that
nately, even those guides are problematic, leaving psychological issue is the subject of much of this
questions about adequate anchor variables unre- book. Fair warning: If you expect to find the answer
solved. For some people, performance on creativity in this chapter, skip it and try elsewhere. The best
that you will get from me is a description and eval-
Robert T. Brown • Department of Psychology, University uation of some approaches to the question. That
of North Carolina at Wilmington, Wilmington, NC 28403-3297. each approach has different answers with different
3
4 PART I • THE NATURE OF THE BEAST
implications for assessment and application reflects performance on IQ tests (Zigler & Butterfield,
one of my own laws: When intelligent and informed 1968). As Zigler and Butterfield pointed out, in-
people of good will disagree widely about some- terpreting changes in IQ as necessarily reflecting
thing, chances are that nobody really knows what's changes in intelligence is unwarranted. Indeed,
going on. they documented that increases in IQ consequent to
Second, what is the status of creativity as a con- nursery-school attendance reflected changes in var-
cept and what implications does that status have for iables other than intelligence.
theory development and evaluation? Most theorists In relation to creativity, as Wallach (1971, 1986)
explicitly or implicitly view creativity as both an has detailed, researchers have often concluded that
intervening variable (MacCorquodale & Meehl, they have increased creativity itself on the basis of
1948; Turner, 1968), not directly observable but evidence that some manipulation increased perfor-
used to explain relations between stimuli and re- mance on a creativity test. This would be true if crea-
sponses, and a trait, something that a person "pos- tivity tests measured creativity and only creativity-
sesses" and that varies among people. Since inter- a highly dubious contention as we shall see.
vening variables are useful only under certain
conditions and traits have certain general charac- Traits
teristics, we need to determine whether creativity
meets the criteria of either. Traits, one type of intervening variable, are inter-
nal predispositions toward certain behaviors that
are generally inferred through tests or attitude
Intervening Variables scales. Traits should be stable over some consider-
In his classic analysis, Miller (1959) demon- able period of time, show reliable individual dif-
strated that an intervening variable is efficient only ferences, have some presumed or demonstrated ge-
if it mediates between multiple antecedents and out- netic loadings, and predict particular behaviors in a
comes. Although the cognitive revolution leaves us variety of settings. Again, general intelligence, as
relatively free to pack constructs into the organism measured by IQ tests, is a good example: It shows
of interest, we should at some point pause to consid- reliable individual differences, is fairly stable from
er whether they serve an efficient explanatory pur- age six years into adulthood, appears to have sub-
pose. Miller demonstrated that thirst was a useful stantial genetic loading, and predicts overall school
intervening variable only when it could bridge at achievement fairly well and occupational level and
least three antecedent and three consequent vari- income less well (Eysenck, 1979).
ables. With fewer than three, the intervening vari- Whether creativity as measured by standard
able used more connections than did simply directly creativity tests actually meets the criteria of a trait
linking antecedent and consequent variables. Given appears doubtful. Individual differences are clear,
the many antecedent and outcome conditions to but stability over time is not, and at least one study
which intelligence can be related, its status is se- of genetic influences (Pezzullo, Thorsen, &
cure, however much we still argue over what the Madaus, 1972) found no evidence for heritability of
underlying construct actually is (e.g., Eysenck, items from the Torrance Tests of Creative Thought.
1979; Sternberg & Detterman, 1979, 1986). Does Of particular importance, people appear not to be
creativity mediate between sufficient antecedent more or less creative generally but more or less
and outcome variables to justify its status as an creative in relatively narrow areas. The creativity
intervening variable? Evidence that we can vary equivalent of Spearman's g, or general intellectual
different antecedents which may be presumed to factor, appears to be the Holy Grail of creativity
affect a creative process, which, in tum, affects a researchers.
variety of creative products, is not readily
available.
An additional problem, to which I will return Historical Antecedents
later, is multiple intervening variables that may af-
fect a given outcome, leading to conceptual confu- At least four approaches can be traced back to the
sion. For example, informational content (factual early twentieth century, with creativity viewed as
knowledge the individual has acquired) and moti- (a) an aspect of intelligence; (b) a largely uncon-
vational-emotional factors (such as anxiety or scious process; (c) an aspect of problem solving;
achievement motivation) may facilitate or impair and (d) an associative process.
CHAPTER 1 • WHAT ARE WE TO MEASURE? 5
Creativity as an Aspect of Intelligence sciously. Thus, for Poincare, conscious but unsuc-
cessful effort to solve a problem sets in motion an
In their 1896 paper on mental testing, Binet and
unconscious process that leads to a random com-
Henri (as described in Freeman, 1924) proposed
bination of ideas, one of which may emerge as an
that 10 mental functions, including imagination, be
appropriate creative solution.
measured by tests. Binet asked children to describe
what they saw in an inkblot as a measure of imag-
ination, but he and his colleagues could not develop Creativity as an Element of Problem Solving
a reliable scoring system and discarded the test.
Several early problem-solving models have in-
Binet and Simon's 1905 scale included three open-
fluenced both creativity theories and applied pro-
ended items that appeared to tap creativity-give
grams. As summarized in Table 1, each model
rhyming words, complete sentences, and construct
views problem solving as a linear set of discrete
sentences containing three given words. Creativity
steps. Incubation is the term used for unconscious
researchers in the early twentieth century widely
generation of potential solutions. For Wall as
adopted such tests, which Guilford (1967) later
(1926), incubation was more structured and guided
used as measures of divergent thought. However,
than for Poincare, and preparation, which included
Binet dropped the items in later editions of his test,
the individual's previous education, was a general
and the predictive success and interest in analysis of
orientation toward problem solving as well as the
standard intelligence tests seem to have diverted
consideration of the problem at hand. A more recent
attention from the potential value of open-ended
investigation of scientific creativity (Busse &
tests.
Mansfield, 1980; Mansfield & Busse, 1981) led to
the conclusion that scientists go through a series of
Creativity as an Unconscious Process steps similar to those of Rossman (1931) and Wal-
In an often quoted passage, Poincare (1913) re- las (1926): (1) selection of a problem, (2) extended
counted one of his discoveries: effort to solve the problem, (3) setting constraints
on a solution to the problem, (4) changing the con-
Just at this time I left Caen, where I was then living, to go on a straints (transformation or restructuring), and (5)
geologic excursion under the auspices of the school of mines. verification and elaboration.
The changes of travel made me forget my mathematical work.
Having reached Coutances, we entered an omnibus to go some
place or other. At the moment when I put my foot on the step the Creativity as an Associative Process
idea came to me, without anything in my former thoughts seem-
ing to have paved the way for it, that the transformations I had Spearman (1931) claimed that the generation of
used to define the Fuchsian functions were identical with those of novel ideas could be explained by three "neo-
non-Euclidean geometry. I did not verify the idea; I should not genetic processes," those capable of generating
have had time, as, upon taking my seat in the omnibus, I went on new mental content: (1) the Principle of Experi-
with a conversation already commenced, but I felt a perfect
certainty. On my return to Caen, for conscience' sake I verified
ence- "A person tends to know his own sensa-
the result at my leisure. (pp. 387-388) tions, feelings, and strivings" (p. 16); (2) the Prin-
ciple of Relations- "When two or more items
Poincare then proposed an influential theory on (percepts or ideas) are given, a person may perceive
the generation of creative ideas: the" appearance of them to be in various ways related" (p. 18); and,
sudden illumination [is] a manifest sign of long, most importantly, (3) the Principle of Correlates-
unconscious prior work" (p. 389). During a period "When any item and a relation to it are present to
of apparent rest, "great numbers of combination mind, then the mind can generate in itself another
[of ideas are] blindly formed by the subliminal item so related" (p. 24). Spearman's basic model,
self" (p. 391). Most are useless and remain uncon- shown in Figure 1, involves an active process in
scious, but particularly' 'harmonious," "useful," which associations with an initial idea can be freed
and "beautiful" ones may break into con- from their relation to it and thus lead to something
sciousness. Poincare further suggested that initial wholly new:
intense prior conscious work on the problem is nec-
essary to "unhook" relevant ideas from fixed posi- The final act in creativity must be assigned to the third neogenetic
process; that of displacing a relation from the ideas which were
tions so that they are free to join during the uncon- its original fundaments to another idea, and thereby generating
scious process. Finally, the solution that has the further idea which is correlative to the past named, and which
emerged from the unconscious is evaluated con- may be entirely novel. (p. 83)
Table 1. Summaries of Problem-Solving Models a

Rossman (1931) Wallas (1926) Dewey (1910)
Invention Creative production Problem solving
I. Need or difficulty observed I. A felt difficulty (problem
found)
2. Analysis; problem defined 1. Preparation (problem stated; 2. Definition and location of
information obtained; attitude difficulty (problem formulated)
set toward appropriate solution)
3. Information surveyed; possible
occurrence of incubation
4. Many possible solutions 3. Suggestion of possible
formulated solution(s)
5. Critical evaluation of solutions; 4. Development of implications of
sustained and ongoing 2. Incubation (unconscious solution(s) through reasoning
incubation, particularly in generation of potential
complex problems solutions)
6. Formulation of new ideas, 3. Illumination ("Eureka!" or
"inventions," and solutions "Aha!"-idea emerges from
unconscious)
7. Evaluation and refinement of 4. Verification (evaluation of 5. Experimental corroboration of
most promising solution; solution) conjectural solution
acceptance of final solution
aBased on and revised after Michael (1977).
These processes were conscious ones-Spearman proaches to creativity should also be mentioned. As
suggested that preliminary work induced fatigue, reported by Taylor and Ellison (1975), late in his
which masked mental developments that were re- career L. L. Thurstone and his colleagues began a
tained and appeared as if by a miracle (through study of creativity based on his factor-analytic
some putative incubation process) when the fatigue model of intelligence and using open-ended tests.
had passed. In one of these projects, Taylor (1947) isolated two
Although rarely cited, Spearman's Principle of factors, fluency of ideas and fluency of ex-
Correlates is similar to several more recent con- pressions, which appeared to be related to creativity
cepts, including Guilford's (1967) transfer recall, and were reflected in subsequent work on divergent
Mednick's (1962) remote associations, Koestler's thought.
(1964) "bisociation of matrices," and Rothen-
berg's (1979) lanusian thinking, although Koestler
Creativity and Scientific Methodology
,.- . 09
returned the process to the unconscious.
A more recent antecedent to cognitive ap-
Although much of the discussion in this section is
common knowledge, a brief review of some basic
II '\ characteristics of scientific data and theory may be
\ : helpful as a referent against which to describe and
D"~--~:D '1•_____
-----1...1
I
evaluate positions on creativity. Because scientists
and philosophers of science still argue over how
science does or should work, what constitutes a
"good" theory, and how data affect theory (e.g.,
Principle of Principle of Principle of
experience relations correlates Hacking, 1981; Harding, 1976; Lakatos & Mus-
grave, 1970), this presentation will be not only
Figure 1. Spearman's three principles of creativity. oversimplified but at odds with some positions.
Theories, as explanations of observed phe- the discovery of hitherto unknown novel facts"
nomena, generally consist of a set of statements or whereas degenerating "theories are fabricated only
models about processes that are themselves not di- in order to accommodate known ones" (Lakatos,
rectly observable. This set, the theory's "internal 1978, p. 5). Beyond that,
principles," is linked directly or indirectly to the
observables by "bridge statements" (Hempel, the hallmark of empirical progress is not trivial verifications [;]
"refutations" . . . are not the hallmark of empirical failure. . .
1966). Thus, "ideational fluency" may be seen as since all programmes grow in a permanent ocean of anomalies.
a bridge between particular observations and the What really count are dramatic, unexpected, stunning predic-
concept of transfer recall in Guilford's theory of tions: a few of them are enough to tilt the balance; where theory
creativity. At a minimum, theories must meet two lags behind the facts, we are dealing with miserable degenerating
criteria: explanatory relevance and testability. A research programmes. (Lakatos, 1978, p. 6)
theory has explanatory relevance if it provides good
grounds for believing that the phenomenon to be From this perspective, Freudian theory would be
explained actually occurs (Hempel, 1966). The ex- viewed as a degenerating program.
planation, then, should relate directly to the Other criteria for evaluating theories include
explained. power, the extent to which the theory accounts for
A theory is testable if it makes clear and specific what is known, and simplicity or parsimony, that
predictions about empirical phenomena, obviously other things being equal, the fewer explanatory
about events unknown at the time. A dominant statements used, the better. Criteria for selecting
twentieth century view of an adequate theory has the simpler of two theories are not necessarily clear,
been Popperian falsificationism (e.g., Popper, however (Hempel, 1966).
1935): For a theory to be scientific it must be falsi- Scientific data must also have certain charac-
fiable, and, in principle, a disconfirming crucial teristics (e.g., Hempel, 1966). First, by definition,
experiment should lead scientists to abandon the they must be empirically verifiable and available
theory. A theory that does not make explicit predic- for observation in principle by anyone. Thus, the
tions cannot be falsified and is not scientific. Thus, frequent example: My dreams cannot be offered as
Popper's well-known pronouncement that Freudian data since they cannot be observed by anyone but
theory was essentially pseudoscience was based on me, whereas my dream reports are acceptable since
the apparent impossibility of falsifying it-any re- anyone can read them. The distinction needs to be
sult could be accounted for. But most philosophers maintained; one should not slide from the report to
of science now agree that falsification of a theory the dream as data. Further, scientific data should be
does not lead to it being discarded. Confronting an reliable; repeated observations should yield similar
anomalous result, scientists will either modify the results. Statements about "who does what" and
theory or simply ignore the anomaly. Theories are "who does what more than someone else" need to
discarded not on the basis of falsified predictions be based on appropriate samples and meet appropri-
but in favor of a better theory (e.g., Conant, 1947; ate standards of probability. Bridge terms and all
Lakatos, 1970). manipulatory and measurement terms should be op-
Lakatos (1970, 1978) proposed an increasingly erationally defined so that one can find one's way
popular view of change in scientific theories that into the theory and back out. Problems arise if oper-
avoided problems associated with the falsifica- ational definitions for theoretical terms overlap in
tionism of Popper (1935) and the sociologically ori- their observable referents. Without clear distinc-
ented scientific revolutions of Kuhn (1962). tions, an observation could reflect the operation of
Lakatos dealt with broad "research programmes" one of several theoretical constructs, leading to
that appear almost equivalent to Kuhnian para- conceptual confusion.
digms. Although with creativity we are dealing Finally, to bridge the theoretical construct-em-
with a narrower and lower-level set of concepts and pirical referent gap, operational definitions should
research strategies, Lakatos's approach is still rele- be relevant to the phenomena of interest, not just the
vant, although he might object to its use. Lakatos construct. If the operational referents of a construct
contrasted "progressive research programmes" refer only to that construct or its bridge statements
with "degenerating" ones and suggested that sci- and not to the phenomena to be explained by the
entists tend to join the former and abandon the lat- construct, then the system may be internally co-
ter. The benchmark for differentiating between the hesive but have no reality contact. The theory may
two is that a progressive research program' 'leads to then lack relevance.
as immediate memory of numbers, mental images,

Testing for Intelligence and Testing for attention, and comprehension. These tests tapped
Creativity skills more directly related to what is commonly
understood as intelligent behavior than did those of
Comparing the development of intelligence tests Galton and Cattell. Of particular interest, Binet
with that of creativity tests may help to explain why compared performance of bright and dull children,
the intelligence-testing movement, for all its flaws, attempting to develop tests that would differentiate
has led to a better understanding of intelligence than between them. When asked in 1904 by the Minister
the creativity-testing movement has of creativity. of Public Instruction to develop tests to identify
As is well known, Francis Galton and James children who needed special instruction, Binet al-
McKean Cattell both developed intelligence tests ready had been trying to relate test performance to a
based on Galton's theory that individual differences criterion-school performance. With the develop-
in giftedness arose from differences in mental ener- ment of intelligence tests that were standardized
gy and psychophysical discriminability. While he against a meaningful criterion, theories about intel-
was at Columbia University, Cattell in 1890 pub- ligence, such as Spearman's two-factor theory,
lished in Mind a set of physical and psychophysical could be evaluated. Thus, Boring's (1923) classic
tests that was designed to determine stability of statement, "intelligence is what intelligence tests
mental faculties over time, their interrelationships, test" had real meaning. The success of intelligence
and their variation across situations. The tests mea- tests and research on intelligence rests, I suggest,
sured such qualities as strength of hand, visual and on the initial establishment of criterion validity that
auditory acuity, pitch and weight discrimination, enabled later theoretical constructs to be tied to ob-
discrimination of two points on skin, free associa- servable behavior. Indeed, generally the most suc-
tion to words, and memory. Note that the tests arose cessful psychological tests may be those initially
from a theory of intelligence, not from the goal of developed to meet a practical need and validated
predicting intelligent behavior. against a specific criterion.
A classic study by Wissler (1901, as reported in Consider now the development of creativity
Freeman, 1924) demonstrated the inadequacy of tests. In his seminal American Psychological Asso-
Cattell's tests and led to the abandonment of the ciation (APA) presidential address on creativity,
"mental energy" theory of intelligence (but see which is detailed below, Guilford (1950) described
Eysenck, 1979, for a contemporary resurrection). several presumed components of creative thought
Wissler administered several tests to students and and potential tests. He subsequently developed his
obtained information about their academic stand- theoretical multifactor structure of intellect and
ing, both overall and in individual courses. Thus, used specialized tests to identify the factors, partic-
he looked at both the internal consistency of the ularly those associated with creative thinking. But
tests and their prediction of a presumed correlate of the tests had not been validated against any external
mental ability-college grades. Mental tests measure of creative productivity. In time, other re-
showed little intercorrelation (r = .10- .20), indi- searchers developed creativity tests according to
cating that they were not tapping a common ability, their particular theoretical orientations, also with-
and virtually no correlation with average class out establishing criterion validity.
standing. But standing in a given class correlated (r With little evidence that the creativity tests actu-
=.50-.60) with standing in others, suggesting that ally measured creative production in addition to
a common factor, but one not tapped by the tests, some putative theoretical concept of creative
underlay performance in different academic areas. thought, professionals began to use them for re-
In today's terms, Wissler's results demonstrated search and assessment. As indicated by Wallach
that Cattell's tests had neither construct nor criteri- (1986), the tests even became the evaluative criteri-
on validity. on for programs to increase creativity. We should
At about the same time, Binet and his co-workers not be surprised that disillusionment began early
were attempting to develop mental tests in Paris. (e.g., Hudson, 1966) and increased with the ap-
Although Binet had a theory of intelligence and pearance of studies indicating that most tests were
many of the tests were similar to those of Galton and not internally consistent and that none reliably pre-
Cattell, even his early work reflected a more prac- dicted actual creativity. The basic problem seems to
tical approach. A list of proposed tests published in be that creativity tests had only apparent construct
1895 included memory of a short paragraph as well validity and certainly not criterion validity-the
same problem that doomed Galton and Cattell. Un- creativity affected their reports. See Weisberg
fortunately, no Binet of creativity has appeared. (1986) for a critical discussion.
The one test (as far as I know) that purports to
measure directly the creative process suffers from
Problems with the Literature the same problem. Using the Creative Process
Check List, Ghiselin, Rompel, and Taylor (1964)
Several problems in the research and theoretical asked scientists to check words from two lists that
literature contribute to the difficulty in elucidating applied to them before, during, and after they had
the process(es) of creativity. Some disappear from grasped a new insight, solved a problem, or in some
notice as initial reports become summarized in sec- way brought "a new order into being within the
ondary sources, but their effects may linger as con- field of [this] consciousness" (pp. 21-22). The
tributions to the multiple inconsistencies in the word lists were adjectives describing "States of
area. Others seemingly are tolerated because of the Attention" (focused, diffused, vague) and "States
nature of the process studied and typical meth- of Feeling" (tense, happy, uneasy). Selection of
odologies used. Reasoning that would not be ac- the words by Ghiselin et at. (1964) was based on
cepted in other areas abounds in the creativity liter- assumptions about the effect of favorable or un-
ature; basic rules governing scientific data and favorable states and feelings on creativity ..The data
theorizing are routinely ignored. Thus, some clearly were introspective and, as the authors ad-
"findings" may be more apparent than real. mitted, subjective. Further, the way in which re-
plies would speak to the causal issue of process as
opposed to correlates is unclear.
Use of Introspective and Retrospective Data
By their nature, self-reports and reports of events
Use of Inappropriate Comparison Groups
or people in the past are not open to empirical ver-
ification. Individuals tend to report what they think Because high- and low-creative groups may dif-
is relevant and can report only what they remember. fer on many dimensions other than those relevant to
Such data are open to alternative interpretations- creativity, comparisons between them may not be
rarely witp. any possibility of testing among them. clearly interpretable. Problems are compounded,
This problem applies to virtually.the entire psycho- when investigators' methods result in actual bias or
analytic literature and is a major barrier to the psy- confounding.
choanalytic approach as a scientific one. (However, Consider a rarely mentioned confounding in
Ericsson and Simon, 1984, have provided evidence MacKinnon's (1962, 1965) classic description of
for the validity of their protocol analysis methods personality differences among more and less cre-
for studying concurrent and immediately retro- ative architects. His 40 highly creative architects
spective self-reports.) came to the University of California at Berkeley in
Other approaches suffer from the same limita- groups of 10, for a three-day interview and assess-
tion. One of the first books that expressly dealt with ment; the two groups ofless creative architects indi-
the process issue in creativity (Ghiselin, 1952) con- vidually completed, in about six or seven hours at
sists largely of autobiographical and biographical home, a selection of materials from the total bat-
accounts of how various people came to create vari- tery. The extent to which differences in physical
ous things, frequently through some spontaneous or and social settings between the groups during as-
unconscious process. Although such reports may be sessment produced effects that were attributed to
valuable as sources for hypotheses for collection of differences in creativity cannot be determined. One
future scientific data, we err in treating them as would expect, however, that the special attention
scientific data themselves, as somehow sufficient to would itself have significantly affected the creative
justify postulating something like "incubation." architects' behavior.
The stories of Poincare's sudden realization of the
proof of Fuchsian equations, Kekule' s daydream(?)
Inferring Causality from Correlations
of the benzene ring, and Coleridge's drug-induced
dream of "Kubla Khan," however colorful and Virtually every undergraduate psychology major
compelling, are simply stories whose literary merit learns that correlational data rarely can be used to
outweighs their scientific value. The reporters, infer causality because of third-variable and direc-
after all, were highly creative people, and that very tionality problems. Obviously one must do the best
one can with data, but creativity research seems ally use multiple measures and/or intercorrelate
blithefully oblivious to these concerns. All too many measures. In describing differences among
often, differences of some kind, say, inX, between his three groups of architects, MacKinnon (1962)
"high-creative" and "low-creative" groups are only occasionally reported any statistical tests,
inferred to cause differences in creativity with leaving in doubt which differences actually were
rarely a thought that they may result from dif- significant. Routinely, an alpha-level of .05 used
ferences in creativity. Next, someone develops a with multiple tests or correlations is set, although
program to train X as a way to increase creativity! . the likelihood of chance effect being "significant"
Further, if a test purportively measures an underly- is increased as the number of tests increase. Two
ing creativity trait, say, ideational fluency, then recent studies by well-respected researchers in stan-
scores on that test are used as an index of that trait, dard journals of creativity and giftedness are exam-
regardless of the possible contribution of some ples of failure to adjust alpha levels. One used an
other trait, say, obsessive-compulsiveness. This alpha level of .05 in a correlation matrix of 78 cor-
problem will be discussed in more detail later. Con- relations; the other used an alpha level of .10 to
cerns about predictive validity are not reduced with determine significance of analyses of variance of
reports that a particular test successfully predicted five measures of unreported degree of indepen-
creative .production when the creative production dence! One should not be surprised that studies
actually occurred some years before the tests were using such procedures produce conflicting find-
administered, as has been the case in several ings. See Mansfield and Busse (1981) for related
studies. criticisms.
Inconsistent and Contradictory Findings Failure to Describe Methodology in Detail

Intercorrelations among the same creativity tests Authors of studies comparing groups that sup-
and between those tests and measured intelligence posedly differ in tested intelligence and/or creativ-
vary widely from study to study, perplexing in- ity occasionally do not give actual test scores either
teractions are common, and failures to replicate are of their groups or of the overall sample from which
routine. Defenders of almost any position on any they were drawn. Further, they do not provide inter-
issue can find ample supporting citations. Those correlations among their tests. Absence of such in-
analyzing the creativity literature themselves need formation prevents the reader from determining
several personality characteristics commonly at- how different the groups really were.
tributed to creative people-resistance to frustra-
tion and high tolerance for ambiguity and chaos, in Prevalence of Abstruse and Cryptic Writing
particular .
Perhaps because of the apparently almost' 'mys-
tical" qualities of the process under consideration,
Failure to Adhere to Principles of Statistical otherwise sensible writers seem to lose their grip on
Inference linguistic reality when dealing with creativity. As
Research, even some classic studies, on corre- Toscanini purportedly said of Bruno Walter, "He's
lates of high- and low-creative people commonly an excellent conductor, but when he comes to a
reflects many problems of statistical inference. really beautiful piece of music, he melts." Fox
Some errors which generally make Type I errors (1963, pp. 123-124) suggested that failure to use
more likely include (a) failure to conduct overall critical faculties in the study of creativity "has re-
analyses of variance before making multiple paired sulted in a great deal of verbiage which would have
comparisons; (b) conversely, inference of a signifi- been better left unprinted . . . the field has been
cant difference between two groups on the basis of inundated with a flood of words with little mean-
an overall significant difference among many ing." One's concern is hardly lessened when a
groups without making paired comparisons; (c) well-known researcher explicitly introduces into
failure to cite any statistical analysis at all; and (d) his theory of creativity an unmeasurable and super-
failure to adjust alpha level when using multiple natural "cosmic" dimension which "may be re-
measures or correlations. garded as the original creative life force, whose
The last points warrant some amplification. power and action manifest in all that is, the grand
Studies of creativity and creative individuals gener- source of all creativity, products of which are but
specific instances of it" (Khatena, 1984, p. 57).

Many authors could be brought up on charges of
premeditated obfuscation. eft06K£ S'·\ls.l..OS
Process and Product: The Criterion

Issue
Recently, I asked the students at the beginning of

one of my classes to draw the model of creative
problem-solving that was presented by Albrecht
(1980). The model is a fairly typical one: Divergent
thought precedes convergent thought in a set of
steps beginning with "Problem Finding" and end-
ing with "Evaluating Results." Usually conscien-
tious, one of my students, Beth, obviously had not
studied the model. However, after a short pause,
she drew what appears here as Figure 2, certainly a
most unexpected and novel response.
But does the novelty of the "model" qualify it as
a creative product and by inference endow Beth
with a degree of something called creative process?
Ultimately, concepts of the creative process must
be linked to creative products. After going so far as
to say that "there is no unique entity identifiable as
the creative process" (p. 134), Fox (1963) con-
cluded that "creativity has no meaning except in
relation to the creative product" (p. 140). As those
who are familiar with the vagaries of criticism in the Figure 2. Beth's pop quiz "problem-solving model." Figure is
arts well know, identifying such products is diffi- labeled "Brooke Shields" and is saying, "Hi! I'm a problem-
solving model! I have many problems to solve! Who to date?
cult and controversial. The criterion issue is of such What to wear? How to handle my money and my mom? Whether
importance (e.g., Taylor & Holland, 1964) to an to keep on being a virgin?" Reprinted by permission of Elizabeth
understanding of process that some coverage, how- Riffe.
ever brief and inadequate, must be provided. Its
complexity may be seen in the Criterion Committee
Reports of the Utah Conferences on Creativity such as Fox (1963), that novelty alone is not
(e.g., Taylor & Barron, 1963). enough: "Somehow the mere oddities must be
In an influential conceptual paper relating criteria weeded out" (p. 4). Appropriateness is a crucial
for creative products to characteristics of creators conjoint criterion to unusualness. A product must
and the responses of others, Jackson and Messick fit the demands of the situation and needs of the
(1967) gave novelty primacy: creator, and with complex products, the individual
parts must form a cohesive whole. Was Beth's
No matter what other positive qualities it might possess, we product appropriate? Internally, it is certainly con-
generally insist as a first step that a product be novel before we sistent, but it hardly meets the explicit demands of
are willing to call it creative. Indeed, the conjoining of novelty
and creativeness is so deeply ingrained in our thinking
the situation. On the other hand, Beth knew both
that . . . novelty often comes to be used as the most common that a blank sheet of paper would have been worth-
and, in some of our current paper-and-pencil tests of creativity, less and that I rewarded last gasp efforts. So, to me,
the only measure of a product's creativeness. (p. 4) her drawing also met criteria of appropriateness;
that decision is subjective, however, and others
Beth's product meets that criterion-in hundreds of may have reacted quite differently. If appropri-
quizzes on the problem-solving model no one else ateness had been judged by the audience at the pre-
has come close. mier of Stravinsky's "Le Sacre du printemps,"
But Jackson and Messick stated, as have others after all, a second performance would never have
taken place. Obviously, artists, scientists, and in- Amabile (e.g., 1983a,b), and Sternberg (1985)
ventors may suffer neglect throughout their lives have all successfully used groups of judges to eval-
only to be "discovered" after they have died. Their uate products in terms of creativity. What is of im-
products' novelty was apparent throughout-per- portance, however, is that such procedures gener-
haps overwhelmingly so-but appropriateness was ally involve experts using their own standards of
seen only in retrospect. The criterion of appropri- creativity, thus avoiding the thorny issue of objec-
ateness, then, however appropriate, is easier to de- tively defining specific criteria.
scribe than to apply, particularly in the short term. Realizing the importance of a unified view, Jack-
Jackson and Messick added two further criteria to son and Messick (1967) considered the possible
determine the quality and level of creative products. cognitive characteristics of creators, reactions on
"Transformation of constraint" involves combin- the part of viewers, and standards used in judg-
ing elements in a way that breaks through tradition ments, for each of their four criteria: "The creative
and leads to a new perspective or way of viewing person, his product, and the world's response to it
reality. It occurs when a product is qualitatively combine to form the drama of human invention"
different from its predecessors and may be inferred (p. 18). A modification of their formulation is in
in part from the effect that the product has on the Table 2. Jackson and Messick's description of their
viewer. "Condensation" or "coalescence of four criteria as being interrelated and of increasing
meaning" is present in products which "do not complexity implies that products fall on a con-
divulge their total meaning on first viewing" (p. tinuum from less to more creative and that people
10). Such products require repeated examination may have the characteristics in Table 2 to a lesser or
for complete understanding, allow for multiple in- greater extent. Creativity, then, would be dis-
terpretations, and may raise new questions while tributed continuously like intelligence, as assumed
answering an old one. In Beth's "model," perhaps also, for example, by Guilford (1950) and Amabile
transformation is present but not condensation. My (1983a,b). But as others have noted, a continuous
conclusion is that her product was indeed creative in distribution of creativity is indeed an assumption.
the way that successful cartoons are (no small An alternative proposal may be seen in Ghiselin
achievement), but that certainly nothing qualita- (1963), who apparently subscribed to the Robert
tively new was apparent. Benchley school of organization: "The world is
Jackson and Messick realized that criteria for divided into two groups of people , those who divide
creativity become more complex and ambiguous, the world into two groups and those who do not. "
going from novelty to condensation, and also that Ghiselin described creative products as being of
agreement on judgment of transformation and con- two qualitatively different types: lower-level, sec-
densation would be more difficult. The fact that ondary creativity simply extends some known con-
judges of creative products can agree, however, is cept into a new area of application, whereas higher-
seen in the work of those using consensual tech- level, primary creativity "alters the universe of
niques. Although their methods differ, MacKinnon meaning itself" (p. 42) and dramatically changes
(1962), Getzels and Csikszentmihalyi (1976), the way in which we view reality. A requisite for
Table 2. Jackson and Messick's View of Relationships among Aspects of Creativitya

Others' responses to
Creative people creative products
Personal Judgmental Esthetic
Property of creative products Cognitive style quality standard response
Unusualness (novelty) Tolerance of incongruity Original Norms Surprise

and inconsistency
Appropriateness Analytic and intuitive Sensitive Context Satisfaction
Transformation Openminded Flexible Constraints Stimulation
Condensation Reflective and spon- Poetic Summary Savoring
taneous power
aBased on and revised after Jackson and Messick (1967).
primary creativity was true' 'production of insight" in this volume are devoted to issues raised by
rather than "reproduction or copying of insight in Guilford, I will describe his address in some detail.
any degree whatever" (p. 38). People who produce Then I will present Guilford's well-known "struc-
primary creativity are, then, qualitatively different ture of intellect" , his concept of creativity as diver-
from those who produce only secondary creativity. gent thought, and his less well known problem-
As Nicholls (1972) emphasized, studying "creativ- solving model of creativity.
ity" test results of normal people who will never
produce anything actually creative may not help us
General Characteristics of Guilford's (1950)
understand the creative processes of those relatively
few truly creative individuals. The issue of distribu- Position
tion of creativity has not been resolved. As his title 1. Creativity is a set of traits. Creativity is a
suggested, Ghiselin's criteria were indeed "ulti- "pattern of traits that are characteristic of creative
mate" and have not been translated into operational persons" (p. 444). Until recently, the trait theory
terms. has dominated the study of creativity.
Giving quantum theory and the theory of rela- 2. Creativity should be stable. By suggesting that
tivity as examples of primary creativity, Ghiselin creative potential might be discovered in child-
seems to have applied Kuhn's (1962) view of hood, Guilford indicated that underlying traits
change in sciences to the entire creative process. As should show some long-term stability.
is well known, Kuhn proposed that science worked 3. Reliability of creativity tests will be low.
in two ways: "Normal" science involved the ex- Considerable variations in actual creative produc-
tension of a theory along the lines of its particular tivity within people will lead to low reliability,
paradigm, whereas "revolutionary" science in- which, although Guilford did not specifically say,
volved the overthrow of one paradigm and its represents considerable measurement problems.
placement with a radically different one calling for 4. Completion tests are needed to measure
a new view of the world. In the same volume as creativity. Tests of creativity should at least partly
Ghiselin's chapter, Kuhn (1963) suggested that be open ended, allowing responders to generate
normal science actually involved considerable con- their own answers instead of identifying a correct
vergent thought whereas revolutionary science one.
called for the flexibility and open-mindedness that 5. Creativity-test scores will show little correla-
characterize the divergent thinker. Whether more tion with intelligence-test scores. Abilities tapped
than coincidental similarity exists between their by standard intelligence tests are relatively unim-
views is difficult to tell since neither referred to the portant for creative behavior and those underlying
other's chapter. creativity are not tapped by intelligence tests. This
We will see in numerous sections in this chapter presumed independence of creativity and intel-
and in other chapters in this volume that the issues ligence anticipated the distinction between con-
of what does and does not objectively qualify as vergent and divergent thought in Guilford's struc-
creative and the way in which creativity is dis- ture of intellect and stimulated much research on
tributed have not been solved and remain as major creativity and creativity tests.
problems for the field. As indicated above, the con- 6. Creative performance depends on more than
sensual approach provides an operational, but not creativity. Motivational and temperamental traits
conceptual, definition of the creative product. As determine whether an individual with creative abili-
my colleague Andy Jackson suggested to me, we ties actually performs creative behavior.
are left with a version of the Potter Stewart ap- 7. Creative abilities are continuously dis-
proach to pornography ("I can't define pornogra- tributed. "Whatever the nature of creative talent
phy, but I know it when I see it. ")-We cannot may be, those persons who are recognized as cre-
define a creative product, but we know it when we ative merely have more of what all of us have" (p.
see it. 446). We should therefore be able to study
creativity in normal people.
Guilford's Views on Creativity
Specific Creative Abilities
Guilford's (1950) APA presidential address is
generally viewed as the foundation of much con- Guilford (1950) hypothesized that at least eight
temporary research on creativity. As entire chapters primary abilities underlay creativity. For most, he
suggested some possible tests; one example will be "symbolic structures must often be broken down
given for each ability. before new ones can be built" (p. 453).
1. Sensitivity to problems. Creative people see 7. Complexity. Possibly related to synthesizing,
problems where others do not, an ability possibly complexity refers to the "numbers of interrelated
related to curiosity. Test: List things that are wrong ideas an individual can manipulate at once" (p.
with, or could be improved in, common household 453). (Guilford did not suggest tests for abilities 5-
appliances. 7.)
2. Fluency. Those people who produce large 8. Evaluation. At some point, the value of new
numbers of ideas are more likely to have significant ideas must be determined. Test: Rank in order of
ideas. Test: State as many consequences as possible excellence several correct solutions to a problem.
to a hypothetical situation, such as: "A new in- Guilford, then, saw creativity as a result of the
vention makes it unnecessary for people to eat" (p. action of several more-or-Iess independent traits.
452). However, he and most others came to focus on
3. Novel ideas. Creative people have unusual but fluency, flexibility, and to a lesser extent, novelty
appropriate ideas. Test: Note the frequency of re- as the crucial aspects of creativity. Other factors,
mote verbal items (only those indirectly linked by such as evaluation, were admittedly underplayed
mediators to the original item) in a word-associa- (Guilford, 1967). Given the general theoretical ori-
tion test. entation of the period, we should not be surprised
4. Flexibility. Creative people should be able that Guilford ignored possibly important roles of
easily to change set. Test: Note the variety of types situational factors. After all, traits were assumed to
of answers to completion questions. operate broadly across situations.
5 and 6. Synthesizing and analyzing abilities. Of importance, Guilford (1950) explicitly stated
"Creative thinking requires the organizing of ideas that the first step in validating his tests should be
into larger, more inclusive patterns" (p. 453) and factorial validity, a form of construct validity, de-
OPERATION:
Evaluation - - -......
Convergent production ~
Divergent production~~
Memo~
Cognition
~~
-0
Units----~.........
u
Classes-----..
tJ::l Relations ---~
a
~ Systems------
a..
Transformations --..........
CONTENT:
Figural----...J
Symbolic - - - - - - '
Semantic---------' C
Behavioral---------'
Figure 3. Guilford's "Structure of Intellect" model of intelligence. From The Nature ofHuman Intelligence by J. P. Guilford, 1967.
New York: McGraw-Hili Book Co. Copyright 1967 by McGraw-Hili. Reprinted by permission.
termined by factor analysis of test answers. Only ford, who viewed divergent thought as a fairly gen-
after identification of factors and development of eral process underlying creative production; and (c)
appropriate tests would we be "justified in taking "P. S." Guilford, who developed a complex
up the time of creative people with tests" (p. 453). "structure of intellect problem-solving (SIPS)"
Thus, criterion validity would follow construct va- model, which incorporated all five operations of the
lidity-a position that had great impact on the way SI. "P. S." Guilford, less well known than either
in which most researchers studied creativity "S. I." and "D. T.", may well now be more in
thereafter. line with general trends in creativity research and
theory. I shall describe each aspect of Guilford in
some detail.
The Three Faces of J. P. Guilford
Guilford's subsequent writings (e.g., 1967, Structure 0/ Intellect (SI). Guilford (1956,
1977) suggest that he developed three psychologi- 1959, 1967) represented his tridimensional struc-
cal personalities. Although not as independent as ture of intellect, which presented intelligence as
Eve's, they make difficult a single description of his much broader than would be assessed by standard
position on creativity. Brief profiles of these intelligence tests, as the cube shown in Figure 3.
"faces" are: (a) "S. I." Guilford, psychometri- Operations are the five ways in which humans can
cian and factor analyst, who developed the process four different kinds of informational con-
structure of intellect (SI), which consisted of 120 tents that can lead to six different types of products
virtually independent factors; (b) "D. T." Guil- (see Table 3 for brief descriptions). Guilford (1982)
Table 3. Brief Definitions of the Three Dimensions in Guilford's Structure of Intellect a

Operations Contents Products
Major intellectual processes; differ- Types of information with which Types of outcome of information
ent ways in which humans deal humans deal processing
with information
Figural (F). Concrete visual, audi- Units (U). Mutually exclusive
Cognition (C). Discovery or recog- tory, or other sensory forms items of information; "things"
nition of different forms of in-
Symbolic (S). Denotative signs Classes (C). Groupings of sets of
formation; comprehension
(letters, numbers, words, etc.) items based on their common
Memory (M). Storage and potential without consideration of mean- properties
availability of information in its ing or form
Relations (R). Connections be-
original form
Semantic (M). Meaning of words tween items based on variables
Divergent production (D). Genera- and occasionally pictures; impor- applying to them (e. g., relative
tion of variety and amount of in- tant in verbal thinking and size.)
formation, based on given communication
Systems (S). Organized, interrelat-
information; most involved in
Behavioral (B). Nonverbals ining, or interacting groups of
creative potential
volved in human interactions, items or parts
Convergent production (N). Gener- particularly concerning such
Transformations (T). Various types
ation of conventionally accepted things as moods, desires, and in-
of changes in existing
single best answer to a given tentions
information
problem
Implications (I). Extrapolations or
Evaluation (E). Making judgments
elaborations of information in
concerning extent to which a
terms of, e.g., consequences or
particular piece of information
expectancies
meets given criterion (adequacy,
suitability, etc.)
aBased on and condensed from Michael (1977).
later split figural content into auditory and visual, these factors include several types of fluency, origi-
resulting in five contents and 150 factors. Because nality, flexibility, and elaboration. Test of DP,
auditory and visual appear not to have been well then, should be useful as tests of creative productiv-
delineated, the earlier model is presented here. ity, although neither anyone test nor anyone over-
Guilford claimed that the 120 factors arising from all creativity score is feasible as a predictor
this 5 X 6 X 4 matrix were essentially independent (Guilford, 1975). Guilford (1975) suggested, how-
of one another; as opposed to Spearman, he pro- ever, that higher-order factors related to divergent
posed that intelligence had no even higher-level thought (DT) might emerge.
general factor. In successive reports, Guilford
claimed to have located increasing numbers of SI Structure of Intellect Problem-Solving (SIPS)
factors, eventually over two-thirds of them. The Model. Guilford (e.g., 1967, 1977) presented a
degree to which he viewed the factors as actually conceptual model for problem solving that incorpo-
independent is unclear, however, given his view rated the five operations and four contents of the SI.
on creative thought. As can be seen in Figure 4, Guilford synthesized
many of the steps in the earlier problem-solving
Divergent Production (DP) and Creative Think- models of Dewey (1910), Wallas (1926), and
ing. "The greatest importance of divergent-pro- Rossman (1931), the feedback-based TOTE (test-
duction abilities is in connection with creative operate-test-exit) model of Miller, Galanter, and
thinking, where many alternative ideas need to be Pribram (1960), and basic elements of SI into an
brought to light with ease. Since creative thinking is information-processing model involving evaluation
an important aspect of problem solving, these abili- at each step. One of the relatively few summaries of
ties are also important in that connection" SIPS is in Michael (1977), on whose detailed pre-
(Guilford, 1977, p. 108). Thus, the 24 DP factors sentation much of this section is based. In the three
arising from combinations of DP with various con- inputs, "E" refers to information from the external
tents and products should be different elements of environment and "S" to that from the internal en-
creativity. As shown in the examples in Table 4, vironment or soma in terms of motivational, emo-
Table 4. Examples of Specific Divergent-Production Factors and Tests a

Factor Label Definition Sample of tests
Semantic units DMU Appropriate meanings given to stim- Consequences of a given event (What
ulus term (ideational fluency) if people didn't need to sleep?); util-
ity test (uses for a brick or a coat
hanger)
Semantic relation- DMR Appropriate relationships given to stim- Synonyms and/or antonyms of given
ships ulus term (associational fluency) words
Symbolic units DSU Words of appropriate surface character- Words rhyming with a given word;
istic given to stimulus term (word words with given first and last letter
fluency)
Semantic systems DMS Complex ideational systems given to Sentences of four words with first let-
stimulus term (expressional fluency). ter of each word specified
Semantic transfor- DMT Unusual, remotely related, or clever re- Rare associations to words; remote
mations sponses to stimulus term (origi- consequences of a given event
nality)
Semantic classes DMC Different classes of categories given to Different types of uses for a brick
stimulus term (flexibility).
Semantic implica- DMI Going beyond information given in Steps specified to make a given plan
tions terms of e.g., detail or consequences work
(elaboration)
aIn all cases, (a) factors should be read "Divergent production of ... " as in "Divergent production of semantic units"; (b)
definitions are in terms of ability to produce indicated information; and (c) tests are scored in terms of number of specified
responses given. Based on information in Guilford (1967, 1977).
EXIT
I
FILTERING COGNITION PRODUCTION COGNITION PRODUCTION

Attention Problem Answers New I-.....L.-l New answers 1-......._
aroused and sensed and generated information generated
directed structured obtained
EVALUATION EVALUATION EVALUATION EVALUATION

Input and Answers New tests New answers
cognition tested of problem tested
tested structure
f
p
EJ
MEMORY ~ Marie 5+7 -12 Beer
®
ijr;:::. )~
q
STORAGE p=>q
~ H2O
~
~ 6 Y-a/b
USA
.r11 Truth
VISUAL - FIGURAL SYMBOLIC SEMANTIC BEHAVIORAL

INFORMATION INFORMATION INFORMATION INFORMATION
Concrete: perceivable Signs Verbally meaningful Psychological
Figure 4. Guilford's "Structure of Intellect Problem-Solving (SIPS) Model." From The Nature of Human Intelligence by J. P.
Guilford, 1967, New York: McGraw-Hill Book Co. Copyright 1967 by McGraw-Hill. Reprinted by permission.
tional, and temperamental tendencies. The latter information within the memory storage in relation to the de-
influence the effort that will be expended in solving mands of a clearly defined problem. (Michael, 1977, p. 159)
the problem. The array of the four contents dis-
played in memory is illustrative only; all types can An active search, then, can free information in
be searched at each step, and new information can memory from the context and cues with which it
also be stored in memory during problem solving. was remembered so that it can appear as a novel
The various exits indicate that the individual may solution to a problem. New combinations of present
decide the problem is trivial and not worth pursuing and past information may presumably also be asso-
(Exit I) or give up at various later points. Also, if the ciated, allowing for truly novel ideas. Although not
problem is not solved after the second production cited by Guilford, Spearman's (1931) principle of
stage, in which novel or creative solutions are sug- correlates appears highly similar to transfer recall.
gested, the entire process can be repeated as in For Guilford (1967), transfer recall was an impor-
TOTE. tant feature of divergent production, but he also
New solutions are generated through what explicitly stated that SIPS does not distinguish be-
Guilford called "transfer recall," in which infor- tween convergent and divergent production: "The
mation is retrieved production stages in the model actually represent
both" (p. 316). Guilford appears to have designed
from a partial set of cues in the memory storage and the using of
this information in a new context and in novel ways. This trans- SIPS to show how SI would function in real world
formed use of retrieved information is effected largely by flexi- problem-solving situations, but the collapse of con-
bly reclassifying, reinterpreting, or redefining well-organized vergent thought (CT) and DT into "production"
seems to damage the presumed independence of multaneous test of the independence of the 120 fac-
those processes in SI and leads to conceptual tors was never undertaken, a task which Eysenck
confusion. (1979) estimated would have called for 192,000
test-taking hours from 2,400 subjects!
But more serious from the standpoint of the sci-
Evaluation of Guilford's Views on Creativity
entific status of Guilford's theory, the SI is essen-
An overall evaluation of Guilford's system is be- tially an encapsulated system, dealing with analysis
yond the scope of this chapter; however, because of of tests, but rarely the actual behavior they sup-
their implications for his views on creativity, some posedly measure. A problem common to virtually
general criticisms must be presented. (For a more all trait-test approaches to creativity is that the hy-
detailed evaluation, see Eysenck, 1979.) From the pothetical creativity constructs were not clearly an-
initial publication of The Nature of Human Intel- chored to observable real-world creativity. As oth-
ligence (1967), reviewers have criticized Guil- ers (e.g., Eysenck, 1979; Wallach, 1986) have
ford's factor-analysis procedures. Of course, some- commented, creativity test results became identical
one will criticize anyone's factor-analysis pro- with creative production itself. The issue became
cedures, a situation that led Detterman (1979) to his even more confused when educators or those in the
private sector developed programs to increase
Law VIII. Neverfactor analyze anything . ... First, it is impossi- creativity and then used performance on creativity
ble to conduct a factor analysis on data which are completely
tests as a criterion to evaluate performance, which
suitable .... Second, even if it were possible to extract factors
correctly, determining an acceptable rotation has never been as Wallach (1986) has documented, is common
accomplished by anyone in the history of Western civilization. practice. The obvious difficulty, as indicated
(p. 169, italics in original) above, is that creativity tests have construct validity
but not criterion validity. Paraphrasing Guthrie's
Guilford's approach is particularly troublesome. criticism of Tolman: Guilford leaves the person
For one thing, he appears largely to have first con- poised in divergent thought.
structed SI and then set about verifying it. Although Particularly troublesome for Guilford's view of
some theoretical orientation is necessary for scien- DT as creativity, other than its theoretical confusion
tific data gathering, Guilford seems to have com- with CT in the SIPS, is disconfIrmation of particu-
mitted the error about which Sherlock Holmes lar predictions about its distinctiveness. For exam-
warned Doctor Watson: "It is a capital mistake to ple, according to SI, Divergent Production of Se-
theorize before one has data. Insensibly, one begins mantic Units (ideational fluency) should be
to twist facts to suit theories, instead of theories to essentially independent of Divergent Production of
suit facts." As many (e.g., Carroll, 1968; Hom & Semantic Transformations (originality). Hocevar
Knapp, 1973; Humphreys, 1962) have commented, (1979) administered three of Guilford's DT tests,
Guilford's Procrustean method of targeted rotations which can be scored for both ideational fluency and
in factor analysis is so subjective as almost to guar- originality (Alternate Uses, Plot Titles, and Conse-
antee identification of the desired structure. l Using quences) to college students. The originality scores
targeted rotations, Hom and Knapp (1973) identi- showed high internal consistency and significant
fied factors as accurately by determining targets convergent validity, as reflected in correlations be-
randomly as by use of SI! tween tests (that is, for example, originality on Al-
Additional problems further weaken the integrity ternate Uses correlated with that on Consequences).
of SI. Many of the tests had low reliability, such Hqwever, partialing out the effects of ideational
that failure of results of one test to correlate with fluency from the originality scores virtually elimi-
those of others could reflect error variance (Ey- nated both reliability and validity of originality,
senck, 1979), and many of the factors were identi- leading to serious questions about the actual con-
fied using small numbers of tests on small numbers struct validity of originality (Hocevar, 1979).
of subjects from restricted populations. A full si- Results of analyses of the dimensionality of CT
and DT are as troublesome. In an approach reminis-
cent of Wissler's study of Cattell's intelligence
I "Procrustes, in Greek legend, was a robber of Attica, who tests, Thorndike (e.g., 1963), Wallach and Kogan
placed all who fell into his hands on an iron bed. If they were
longer than the bed he cut off the redundant part, if shorter he
(1965), and Wallach (1970, 1971) suggested that
stretched them till they fitted it; he was slain by Theseus" Guilford's distinction between the dimensions of
(Brewer's, 1962, p. 731). divergent and convergent thought should be de-
monstrable correlationally: DT tests should corre- development of a variety of applied programs for
late more with each other than with standard tests of creative problem solving. His emphasis on the need
CT, such as IQ tests, which should also intercorre- for a theory of tasks (Sternberg, 1979) has had con-
late. In more concrete terms, creative behaviors tinuing influence. Finally, Guilford's conception of
should be coherent and should be distinguishable intelligence as multifaceted is reflected in contem-
from other sets of behaviors (Thorndike, 1963). porary theories as divergent themselves as those of
Thorndike found that six CT tests showed moderate Sternberg (1986) and Gardner (1983).
intercorrelations; but DT tests showed no more con-
sistency among themselves than they did with the
CT tests. Other studies, reviewed in Wallach
(1970), similarly indicate that Guilford's concept of
DT is not empirically distinct from standard tests Other Divergent Thought Approaches:
of CT. Creativity as an Associative Process
The SI, then, was not a progressive research pro-
gram that led to important confirmed predictions. Many approaches to creativity, including Spear-
Because of this, and the difficulty of distinguishing man's, are implicitly associationistic: A creative
empirically among different factor analytic ap- idea results from the novel combination of two or
proaches, interest in SI in particular and factor- more ideas that have been freed from their normal
analytic models in general declined with the correlates. Mednick (1962; Mednick & Mednick,
emergence of other approaches (e.g., Sternberg & 1964) offered an explicitly associationistic theory
Detterman, 1979, 1986). Unfortunately, the em- of creativity based on introspective accounts of
piricallimitations of Guilford's approach have not creativity (e.g., Ghiselin, 1952). Suggesting that
prevented some (e.g., Meeker, 1969) from urging creative solutions could occur through serendipity,
its application to classroom, reflecting once again similarity, or, most importantly, mediation, Med-
the tendency to go directly from theory to applica- nick (1962) defined the creative thinking process as
tion, bypassing adequate empirical confirmation.
the fonning of associative elements into new combinations
With all the criticisms, one may all too easily lose
which either meet specified requirements or are in some way
sight of Guilford's lasting contributions. Re- useful. The more mutually remote the elements of the new com-
gardless of his quixotic tilt at the windmill of Spear- bination, the more creative the process or solution. (p. 221)
man's "g," Guilford nevertheless focused atten-
tion on aspects of cognitive processing other than For Mednick (1962), creativity was a function of
traditional intelligence that underlie adaptive be- people's "associative hierarchy, " which is the way
havior. Delineating between divergent and con- in which they produce associations to words or
vergent thought led to a reconsideration of pro- problems. As can be seen in Figure 5, these hier-
cesses involved in problem solving and the archies resemble generalization gradients of differ-
HIGH
W SLOPE OF NONCREATIVE INDIVIDUAL
!;i:
~
~
~
z
W
~
Figure 5. Mednick's concept of associative Law~~~~~ _____________________________
strength and gradients. Hypothetical gradients ca..1MON UNIQUE
are shown for steep and shallow degrees of as- ASSOCIATES ASSOCIATES
sociative strength to a stimulus word. After
Wallach and Kogan (1965). ASSOCIATES IN ORDER OF POTENTIAL EMISSION
ent degrees of steepness, with associations to word reports emphasized the importance of having a per-
or problems ranging from common to unique. Indi· missive or playful set. They defined the creative
viduals with relatively steep gradients tend to give process as involving two elements: "first, the pro-
common associations at high strength but few or no duction of associative content that is abundant and
uncommon associations. In extreme, their re- that is unique; second, the presence in the associator
sponses may be rigid and stereotyped. Those with of a playful, permissive task attitude" (p. 289). The
relatively flat gradients also initially give more divergent thought component largely focused on
common responses, but, as shown to the right in ideational fluency, and led to the construction of a
Figure 5, will then more likely make uncommon or battery of verbal and visual tests, many similar to
unique associations-remote from the initial those of Guilford. Each test had several items; one
stimulus. example of each will be given. The verbal tests
Mednick (1962) attempted to deal with a still asked for (a) instances ("Name all the round things
serious problem concerning creative products: dif- you can think of"), (b) alternate uses ("Tell me all
ferentiating between the actually creative and the the different ways you could use a newspaper"),
merely bizarre. As discussed earlier, open-ended and (c) similarities ("Tell me all the ways in which
tests involve no evaluation of appropriateness of a potato and a carrot are alike"). The visual tests
solution. Mednick suggested that a product's actual were (a) abstract patterns (e.g., five small circles
creativeness depends on the number of require- around a vertical straight line) and (b) straight or
ments it meets and thus brought convergent thought curved lines. In each case, subjects were to tell all
back into a theory of creativity. the things the pattern or line made them think of. Of
In order to measure ability to make remote asso- importance for Wallach and Kogan's theoretical
ciations and evaluate their appropriateness, Med- views and the topic of much subsequent research
nick (1962) developed the Remote Associates Test and controversy (see Wallach, 1971, 1986, and
(RAT). The RAT consists of 30 sets of three more Kogan, 1983, for summaries), the tasks were pre-
or less independent words sharing a common but sented in a gamelike setting with no time limit.
remote association. The subject's task is to locate a Each test was scored for both number and
particular experimenter-defined mediator that links uniqueness of responses.
the words. Thus, the three words "rat, blue, cot- In a study of fifth graders, Wallach and Kogan's
tage" all share the mediating association (1965) test battery proved to have two important
"cheese," and "out, dog, cat" all share "house." characteristics: First, the 10 creativity scores inter-
Mednick (1962) reported some very promising reli- correlated strongly for both boys and girls, the aver-
ability and validity data: (a) RAT scores generally age correlation being .41. Second, scores on their
correlated negatively with grades and not at all with battery essentially were independent of a battery of
measured intelligence; (b) high RAT scorers were 10 traditional measures of aptitUde and achieve-
more flexible, had more atypical views, and had ment, the average correlation being only .09. Fur-
interests similar to those of artists, psychologists, ther, the aptitude and achievement scores also
physicians, and authors; (c) RAT scores correlated showed high intercorrelations, averaging .51.
positively with number of associations to a list of Thus, Wallach and Kogan's two test batteries ap-
words and quantity and originality of anagram con- peared to tap two unitary and coherent but indepen-
struction; and (d) RAT scores correlated with the dent thought processes, one convergent, compara-
originality score of the Institute of Personality As- ble to traditional concepts of intelligence and the
sessment and Research (lPAR) Questionnaire in other divergent, loading heavily on ideational flu-
MacKinnon's study of eminent architects. ency. Wallach and Kogan (1965) concluded that
However, more recent reviews of the RAT in their fluency tests measured creativity, and com-
Buros (1972, 1978) indicate that RAT scores corre- pared on a variety of measures the four groups of
late with standard intelligence test scores but only children who scored high or low on the intelligence
poorly, if at all, with actual creative productivity. and/or creativity tests. The groups differed on cog-
Then, at least as measured by the RAT, the concept nitive, affective, and school behavior variables.
of associative hierarchies appears to be an intrigu- Further, several interactions among the intelligence
ing but questionable basis for creative thought. and creativity variables occurred. Although the re-
Wallach and Kogan (1965) adopted Mednick's sults were not entirely consistent, the creativity test
basic theoretical view that creativity is an asso- battery tapped a dimension that related in important
ciative process but suggested that creative people's ways to the children's behavior. For example, the
high creativity-high intelligence children were said thought may affect performance on DT tests. Thus,
to "exercise within themselves both control and scores on DT tests mayor may not reflect DT. Since
freedom," whereas the low creativity-high intel- open-ended DT tests generally are scored in terms
ligence children were "'addicted' to school of number of associations, any process increasing
achievement" (p. 303). productivity in general will increase scores on DT
Wallach (1970) considered his earlier research tests. Obviously, such a process may also increase
(Wallach & Kogan, 1965) and Mednick's (1962) actual creative productivity, which itself may result
work in relation to Maltzman's (1960) approach to from the sheer amount of ideas generated (e.g.,
training of creativity. He concluded that "what Campbell, 1960). But in this case, neither DT-test
matters most is the generating of associates; once performance nor actual creativity would be due to
produced, the evaluation of their relevance and ap- DT but to some other process. Indeed, authors (see,
propriate action in the light of this judgment seems Eysenck, 1979; Hocevar, 1980; Wallach, 1971,
to pose little difficulty" (pp. 1254-1255). He fur- 1986) have nominated several possibilities: extro-
ther suggested that the process actually underlying version, obsessiveness, compliance with experi-
the generation of many associates was attention de- menter demands, suggestibility, overachievement,
ployment. Creative individuals can attend to many persistence, and, from my perspective, need for
aspects of a given stimulus and thus produce more achievement or approval. Kogan (1983), sum-
and more varied associations. marizing evidence for the relative independence of
A number of studies have confirmed that Wal- quantity and quality of associations, concluded that
lach and Kogan's test battery shows internal con- most confounding factors would either affect quan-
sistency and relative independence from measures tity only or interfere with quality. However, several
of intelligence (see Wallach, 1970, 1971, 1986 for studies by Hocevar (e.g., 1979) indicate that idea-
reviews). However, several authors, including tional fluency, in terms of quantity, underlies origi-
Wallach, have expressed concern about what con- nality scores. Hocevar used college students, and
struct DT tests actually tap, a topic dealt with in the perhaps different processes operate at different
next section. ages, but, at present, differentiation between ef-
fects of DT itself and other factors on DT test per-
formance appears difficult at best.
The crucial issue is whether DT tests actually
Overall Evaluation of Creativity as predict creative production. As early as 1971, Wal-
Divergent Thought lach raised doubts about the extent to which tests of
ideational fluency predict real-world creativity.
Guilford's initial conception of a large number of With acute reasoning and after a close evaluation of
more or less independent DT factors was sup- the literature, he questioned the use of tests of idea-
planted primarily by a focus on ideational fluency tional fluency which by name became "creativity
as a general associative process, a sort of "g" fac- tests, " but did not have criterion validity, either to
tor, underlying virtually all types of creativity. This study the process of creativity or to evaluate pro-
theoretical shift, in tum, led to the development of grams designed to increase creativity. At the same
creativity tests that were narrower in focus than time, Cattell (1971), a persistent critic of creativity
Guilford's battery, more unitary, and designed to tests, concluded that
be independent of traditional tests of intelligence.
In spite of Kogan's (1983) suggestion that DT tests the verdict that a test measures creativity is only a projection of
can be used to study divergent thought regardless of the test constructor's personal view about what creativity is.
Thus in the intellectual tests developed by Guilford's students
their criterion validity, the tests are so closely relat- and many others who have worked on creativity in this decade,
ed to particular theoretical viewpoints that criterion creativity has finished up by being evaluated simply as oddity or
and construct validity appear to be greatly inter- bizarreness of response .... This indeed comes close to mistak-
woven. Evaluation of divergent thought as the ma- ing the shadow for the substance. (pp. 408-409)
jor process underlying creativity then itself be-
comes intertwined with evaluation of DT tests as Unfortunately, the issues regarding the importance
predictors of creativity productivity. of setting in which testing occurs (gamelike vs.
The problems become further complicated with testlike), relative value and independence of
consideration of the fact that, as others have sug- number and uniqueness of answers as criteria for
gested, hypothetical constructs other than divergent creativity, and predictive validity of ideational flu-
ency tests remain unresolved. Inconsistent and con- ing. " Their conclusion that' 'some divergent think-
tradictory findings allow almost any conclusion ing tests, administered under some conditions and
(see Kogan, 1983, and Wallach, 1971, 1986, for scored by some sets of criteria, do measure abilities
some opposing views). Here again is the problem of related to creative achievement and behavior in
a construct adrift and unanchored to meaningful some domains" (p. 447) hardly inspires confidence
criteria. in divergent thought as the important factor in
Even individual studies with apparently clear re- creativity. Increasingly, the validity of DT tests is
sults present difficulties. Consider Wallach and being questioned (e.g., Cronbach, 1984; Nicholls,
Wing's (1969) study of "predictive" validity of 1972; Simonton, 1984). Amabile (1983a,b) opted
intelligence and creativity tests. They mailed infor a consensual definition and Sternberg (1985)
coming freshman volunteers at Duke University suggested that they "capture at best only the most
four tasks from the Wallach and Kogan (1965) bat- trivial aspects of creativity" (p. 618).
tery, asking them to give as many responses as Barron and Harrington's (1981) concern about
possible, and also to list a variety of creative accom- the convergent-divergent thought distinction also is
plishments in high school. Wallach and Wing worth repeating:
scored the task results for number (ideational pro-
ductivity) and uniqueness (ideational uniqueness) Divergent thinking in fact goes hand in glove with convergent
of associations, and used students' SAT scores to thinking in every thought process that results in a new idea. The
aha! comes when the process reaches a conclusion. But process
measure intelligence. They then classified students
is precisely what is invisible in the usual DT tests used in
into the top and bottom third of scorers in terms of creativity research. A problem is set, and a written answer is
SAT, ideational productivity, and ideational obtained. What happens in between is anybody's guess, except
uniqueness. Although the upper third of all three the respondent's, who hasn't been asked. (p. 443)
groups had higher high school grades than did the
lower third, SAT scores predicted freshman grades Perhaps, but only with the great benefit of
at Duke better. Ideational productivity and hindsight, we can see why the initial promise of
uniqueness showed only slight and inconsistent re- divergent thought has not been fulfilled. Implicitly
lationship to grades. Of primary interest to the au- or explicitly, creativity theorists viewed divergent
thors, students in the top third on both ideational thought as a fairly general process that would ac-
measures generally reported having significantly count for a variety of creative activities. But several
more creative accomplishments in high school than lines of research and theory (e.g., Albert, 1983;
did those in the lower third, whereas students of Amabile, 1983a,b; Feldman, 1986; Gardner, 1983;
high and low intelligence reported essentially no Hocevar, 1981; Wallach, 1986) are converging on
significant differences in high school creativity. the conclusion that talent and creativity are domain
The commonly drawn conclusion that ideational specific whether by dint of "natural" proclivity,
fluency tests significantly predicted creative pro- extensive training, and/or education. As Barron
ductivity is, however, problematic. Feldman and Harrington (1981) suggested, more domain-
(1970) noted several technical problems with statis- specific aspects of divergent thought may underlie
tical analyses, representativeness of their sample (V creative productivity.
+ Q SAT = 1264!), and overlapping representation As a further problem, however, divergent
of given students in their three measures. Beyond thought and other approaches treat creativity as an
those issues is a serious logical flaw: The ideational intrapersonal trait. Although traits have explanato-
fluency tests said to "predict" creative productivi- ry value (e.g., Eysenck & Eysenck, 1985), since
ty were given after the creative productivity had the pioneering work of Mischel (1968), psychol-
occurred, producing a bit of a directionality prob- ogists have become increasingly aware of the im-
lem. Actually, creative productivity predicted per- portance of situational factors in determining be-
formance on tests of ideational fluency. In all fair- havior. The importance of such factors was seen by
ness, one must note that Wallach later changed his Getzels and Csikszentmihalyi (1976) in their study
views on the value of creativity tests. of artists, appears in Feldman's (1980, 1986) ac-
In their review, Barron and Harrington (1981, p. count of the importance of mentors in the develop-
447) stated that even after 80 years, we do not have ment of prodigies, and is given prominence in
a satisfactory answer to "the vitally important Amabile's (e.g., 1983a,b) model and research pro-
question of whether divergent thinking tests mea- gram. The role of sheer chance (Austin, 1978;
sure abilities actually involved in creative think- Campbell, 1960) also needs to be considered.
At the very least, as Wallach (1971, 1986) and conceptualized the creative process as problem
others (e.g., Nicholls, 1972) have emphasized, the solving consisting of "formulation of a problem,
use of DT tests as a criterion for, or classifying adoption of a method of solution, and the reaching
measure of, creativity is unwarranted, although pa- of solutions." They differentiated between "pre-
pers defending such use still appear (e. g. , McCrae, sented problem situations" in which one can apply
1987). As Nicholls (1972, p. 724) concluded, di- known answers to all three steps and "discovered
vergent thinking tests should not be called tests of problem situations" in which one must both identi-
creativity. Divergent thinking is of research in- fy the problem and develop new ways to solve it. In
terest, but not as "constrained by preconceptions the latter situation, "the problem solver must be-
about creativity." come a problem finder" (p. 81). Although difficult
to interpret at times because of its complexity and
numerous correlated measures, the Getzels and
Problem Finding as Creativity Csikszentmihalyi study deserves detailed descrip-
tion.
Some writers in recent years have reemphasized Getzels and Csikszentmihalyi administered tests
the importance of problem finding, in addition to of intelligence, personality, and fluency to art stu-
problem solving, as a crucial element of true dents at the Art Institute of Chicago and obtained
creativity. Related to Guilford's (1950) notion of teacher ratings of the students' artistic potential.
"sensitivity to problems" and the everyday mean- They then presented fine arts students with 27 ob-
ing of curiosity, creative individuals' problem find- jects (e.g., a small manikin, a bunch of grapes, a
ing actively refers to their looking for discrepancies brass hom) and asked them each to select whatever
or something they do not understand. Darwin, ac- ones they wished, arrange them into a composition,
cording to Gruber (1981), demonstrated "pur- and then complete a drawing. The investigators re-
posefulness" in seeking out variability in species as corded each student's behavior throughout the pro-
well as ways to account for it. Feldman's (e.g., cess, took several photographs, and interviewed
1980, 1986) six prodigies initially were recognized each student after he or she completed the drawing.
when they showed great-and frequently per- Getzels and Csikszentmihalyi scored each stu-
sistent-interest in some particular area. My infor- dent on problem finding at the formulation stage
mal interviews with creative scientists consistently (which and how many objects the student handled
reveal an early preoccupation with some observa- and how thoroughly the student examined them),
tion or reading that troubled them-something problem finding during problem solution stage
wanted further investigation. In some cases, that (time each student took to develop the structure of
preoccupation became a long-term career interest. the drawing, whether objects were rearranged or
Gestalt psychologists (e.g., Henle, 1962, 1974; drawing paper was changed, and the extent to
Wertheimer, 1959) in particular have emphasized which the student transformed objects in drawing
the importance of posing the correct question. Hen- them), and overall concern for problem solving
le (1974) argued that the perception of "dynamic (based on the interview). These were combined into
gaps" incites the creative process: "And yet posing a total problem-finding score. Groups of judges
the right question may be the most creative part of (artists, art teachers, mathematics students, and
the whole process. As James Stephens once re- business students) evaluated the drawings on over-
marked, 'A well-packed question carries its answer all aesthetic value, originality, and craftsmanship.
on its back as a snail carries its shell' (1920, p. 68)" Of major interest, artist-critics' and art teachers'
(p. 25). Henle described six conditions leading to evaluations of the drawings correlated with stu-
the perception of gaps: (1) contradictions of all dents' problem-finding scores. Thus, problem find-
kinds; (2) unexpected similarities between one phe- ing had at least short-term criterion validity as one
nomenon and an apparently unrelated one; (3) element of artistic creativity. Further, the behav-
strange, new phenomena; (4) in science, a hypoth- ioral protocols and interviews indicated that stu-
esis derived from theory; (5) again in science, diffi- dents with low problem-finding scores "saw" their
culties with prevailing explanation, and of great drawing's final form much sooner than did those
importance; and (6) the presence of a "welcoming with high problem-finding scores. Getzels and
mind," actively looking for those gaps. Csikszentimihalyi (1976, p. 154) suggested that
In a unique longitudinal study of prospective art- low scorers adopted a known solution whereas high
ists, Getzels and Csikszentmihalyi (1976, p. 79) scorers attempted to discover something new in the
objects: "In a sense, the one looks on the problem gifted individuals have exceptional problem-find-
as familiar, the elements as new; the other looks on ing ability.
the elements as familiar, and the problem as new."
It seems only a slight translation to suggest that the
low scorers coalesced early on convergent problem
solving whereas the high scorers maintained a di- Evaluation as a Creative Process
vergent approach.
Although apparently not the subject of much re-
Also of interest is what did not correlate with
search, evaluation of potential solutions may be an
success in school. Although instructors rated high
important aspect of creativity, as suggested ?y
achievers in art courses (B or above) higher on artis-
Mednick (1962). It certainly is an element of VIr-
tic potential than low achievers (C+ or below), the
tually all models of problem solving. My students'
two groups showed essentially no differences on
attempts at creative problem solving are often f~s
tests of fluency and flexibility (e.g., unusual uses
trated by their persistent failure to evaluate potential
test) or IQ. In fact, on the majority of tests, low
solutions. As a result, incorrect solutions are of-
achievers had higher fluency and flexibility scores.
fered frequently-and confidently. For example,
Seven years after initial testing, Getzels and
given the classic problem, "How can you make
Csikszentmihalyi located as many former subjects
four equilateral triangles out of six sticks of equal
as they could and rated the artistic success of each
length?" about half the students submit drawings of
on the basis of judgments of a gallery director, an two sticks as diagonals inside a square made of the
art critic, and the information gained in the follow-
other four. When queried, the students indicate that
up itself. A number of family factors influenced
they had not checked to see whether any of .their
success-successful artists were more likely to be
triangles were equilateral or whether the two dIago-
first-born and come from wealthy families, for ex-
nals actually reached the comers. Their premature
ample. But neither fluency, intelligence, nor values
acceptance of an obvious but incorrect solution pre-
tested the scores, and only one personality factor
cluded further search that might have yielded a
(low concern with social approval) correlated with
more creative and appropriate solution.
success. Teachers' evaluations of artistic potential Evaluation in the form of "selective retention"
were slightly correlated with success, but academic
is a crucial aspect of Campbell's (1960) theory of
grades were inversely related. . creativity, which will be discussed more fully in the
Of major interest, problem-finding total scores m next section. A variety of potential solutions is of
the drawing exercise significantly predicted success little value if "there is not the precise application of
as an artist seven years later (r = .30), demonstrat- a selective criterion which weeds out the over-
ing long-term predictive validity fo~ the pro?lem~ whelming bulk of inadequate trials" (pp. 391-
finding concept. Getzels and CSIkszentmlhalYI 392). Campbell went on to suggest that solutions
(1976) pointed out that their findings m~y ~ot ~pply may be evaluated against a number of criteria, and
to truly great artists, bringing up the cnteno? Issue that more creative thinkers may be able to keep
again. But their project not only reveals an Impor- track of and apply more criteria at one time.
tant role of problem finding, but also represents a Given the research devoted to other elements of
scientifically sound and potentially progressive, al-
creative problem-solving models, some might well
beit complex and time-consuming, research probe invested in evaluations of evaluation.
gram.
Unfortunately, their procedures do not allow for
clear separation among a number of the components
of creativity described in Guilford's (1950) original Should We Just Leave It All to Chance?
paper. One can infer, for example, that evaluation
was an ongoing process among all the students d~r Most recent psychological accounts suggest that
ing the drawing exercise, but that the more creative the creative process is an intrapersonal trait and is
ones avoided premature closure. In any event, the goal-directed. In the writings of philosophers and
processes underlying question asking app~ar to be scientists, however, one finds considerable em-
important ones. At least one program .desIgn~~ to phasis on internal and external chance factors.
stimulate creativity in children emphaSIzes trammg Campbell (1960) quoted a variety of mainly ni.n~
in question asking (Landau, 1985), and Sternberg teenth-century views: "With reference to ongl-
and Davidson (1986) have recently concluded that nality ... there is an active tum, a profuseness of
energy, put forth in trials of all kinds on the chance anecdotal accounts of the role of chance. In fair-
of making lucky hits" (Alexander Bain); "Ie prin- ness, Austin did not claim to offer a theory, but
cipe de l'invention est Ie hasard" [the principle of wanted to show the reality of research "in all its
discovery is chance] (Sourioux); the title of a major haphazard, unpredictable complexity" (p. xii).
address by Ernst Mach, "On the Part Played by In a fascinating if quirky paper, Campbell (1960)
Accident in Invention and Discovery"; and the ex- put the role of chance inside the individual. He
cerpt from Poincare cited earlier. proposed that the generation of multiple potential
The impact of external chance happenings or en- solutions to problems through' 'thought trials" was
counters is obvious enough to serve as a "What determined by chance:
if ... " parlor game: "What if Semmelweis's col-
league had not contracted fatal puerperal-feverlike Real gains in knowledge must have been the products of explora-
tion going beyond the limits of foresight or prescience, and in
symptoms after being nicked by a student with an this sense, blind. In the instances of such real gains, the suc-
autopsy-dirtied knife?" "What if the young cessful explorations were in origin as blind as those which failed.
Charles Ives had not heard two bands simul- (p. 381)
taneously playing different tunes?" "What if Pi-
casso had not seen African art?" and "What if Campbell claimed that a "blind-variation-and-se-
Alexander Fleming didn't really like to play water lective-retention" process was fundamental to all
polo?" Austin (1978), a neurologist, has used a cases where induction leads to a creative advance in
variety of case studies to document the role of exter- knowledge. The process is similar in its essentials
nal chance factors and to support his own basic to Darwinian evolution, in which mutations are not
position: "We are all creative, but often only to the guided by any purposive or teleological force but
extentthat we are lucky" (p. xii). He described four natural selection preserves those which have advan-
types of "good luck" and personality traits re- tageous effects. In considering individual examples
quired to capitalize on them: Chance I is simple of great creativity, Campbell says that our tendency
"blind" luck resulting from an accident; nothing is to "see marvelous achievements rooted in equally
attributable to the individual. Chance II (seren- marvelous antecedents" (p. 390) is no more than
dipity) results from general exploratory behavior the "fallacy of accident" and "post hoc, ergo
and follows what Austin terms the Kettering Princi- propter hoc." Indeed, Campbell suggests that if a
ple: Chance favors those in motion. Individuals dozen equally brilliant people proposed guesses
high in general curiosity and persistence may have about something unknown and if one guess should
"happy accidents. " Chance III is attributed to sa- be correct, it would inform us about the unknown
gacity and involves the Pasteur Principle: Chance but not about the "greater genius" of the correct
favors the prepared mind. Individuals with great guesser, who "just happened to be where lightning
specialized knowledge and a particular sensitivity struck" (p. 390). No more than we impute "fore-
will be likely to form and see the value of certain sight" to a successful mutant allele should we "ex-
new associations. Chance IV results from person- pect marvelous consequents to have had equally
alized action, what Austin terms, "altamirage," or marvelous antecedents" (p. 390).
"the quality of prompting good luck as a result of Individual differences in creativity will exist, ac-
personally distinctive actions" (p. 203), particu- cording to Campbell, in terms of people's accuracy
larly those "far removed from the area of the dis- and details of representation of the world, which
covery" (p. 78). It follows from the Disraeli Princi- reduce to differences in information and intel-
ple: Chance favors the individualized action. As an ligence; the number and variety of thought trials
example of Chance IV, Austin suggests that Flem- they can produce with a correct solution more likely
ing might never have discovered penicillin ifhe had by chance with greater frequency and variety; and
not liked to play water polo! The London hospital differences in evaluative criteria, discussed above.
where Fleming chose to work had a swimming In a sense, blind-variation could be viewed as a
pool. As a result, it was damp and fostered bacterial form of divergent thought except that it is usually
cultures. Austin used other elements of Fleming's given a goal-directed quality explicitly absent from
story to buttress his position that the greatest dis- Campbell. Similarly, Newell, Shaw, and Simon's
coveries occur when all four aspects of chance (1962) maze/model of creative problem solving ex-
operate. plicitly includes heuristics for choosing subsets of
From a scientific standpoint, Austin's account is all possible solutions to explore. Crovitz's (1970)
of limited value since it offers only post hoc and relational algorithm could be described as an orga-
nized blind-variation. One might suggest that actual creative productivity. In addition to empha-
Campbell has failed to consider the importance of sizing diligence, a motivational characteristic,
explicitly remote associations and differences in Welsh (1986) has amassed considerable support for
ability to make them. He seems to have difficulty in the importance of two personality variables, which
accounting both for the consistently creative person he calls origence and intellectance. Based on his
and for the highly productive but noncreative per- review, Nicholls (1972) suggested that intrinsic, or
son, who should by chance alone have a hit once in task, motivation may be an important and over-
a while. But if Campbell leaves too much to chance, looked factor. This chapter will conclude with a
he serves as a valuable reminder of the error of brief description of a recent interactive model of
leaving too little to it. creativity that features intrinsic motivation as one of
Simonton (e.g., 1984) has most explicitly incor- its main components.
porated Campbell's (1960) basic concept that cre- Amabile (1983b, p. 358) argued that "creativity
ative thought results from the combination of truly is best conceptualized not as a personality trait or a
chance combination of ideas with retention only of general ability but as a behavior resulting from par-
the useful ones. He extends the model to actual ticular constellations of personal characteristics,
creative productivity by emphasizing the impor- cognitive abilities, and social environments." As a
tance of creators' leadership skills in determining working framework of the components involved in
whether or not their products have been actually creative productivity, Amabile (1983a,b) proposed
accepted. the model shown in Figure 6. Not only do the three
Indeed, in addition to stating that creators are major components interact, but each itself results
leaders, Simonton (1984, p. 2) goes so far as to from an interplay of several internal and external
subsume creativity under leadership:" ... cre- factors. The subprocesses traditionally viewed as
ativity becomes a variety of leadership. " Using a being involved in the entire creative process appear
"historiometric" approach, Simonton (1984) has in both the domain-relevant and creativity-relevant
used archival data to generate a number of quan- components.
titative relationships, for example between Domain-relevant skills are the most basic; one
creativity and age and birth order. Although Simon- cannot be truly creative unless one knows a great
ton's approach is itself novel and intriguing, he has deal about a particular area, has the skills necessary
not in detail clearly defined some important con- to produce in that area, and has "talent," which
cepts including creativity and genius. Of course, as Amabile also puts in quotation marks, for that par-
with all essentially psychohistorical approaches, reticular area. Such skills comprise "the individual's
lationships, no matter how quantified, are limited complete set of response possibilities" (1983b, p.
by the archival sources that are available and actu- 363) from those involved in synthesis of the new
ally used. Such relationships are thus always open response to those involved in evaluation, and in-
to alternative interpretations. But Simonton's clude all knowledge about the domain, including
(1984) emphasis on leadership is itself an important facts, paradigms, esthetic criteria, and technical
contribution to considerations of what underlies skills.
creative productivity. Creativity-relevant skills are those cognitive and
personality characteristics that have traditionally
Amabile's Componential! Social been viewed as underlying generation of potentially
Psychological Approach creative responses. Briefly, cognitive style skills
are: breaking perceptual set; breaking cognitive set,
Given the general failure of divergent thinking or trying new problem-solving strategies; keeping
and cognitive trait models to account for creative response options open, or delaying closure; sus-
productivity, the role of other factors should be pending judgment; using "wide" categories; re-
considered. Numerous authors (e.g., Barron, 1968; membering large amounts of information accu-
Barron & Harrington, 1981; Cattell, 1971; Mac- rately; and breaking out of performance scripts or
Kinnon, 1965, 1970; Nicholls, 1972; and Welsh, algorithms. As indicated in Figure 6, also in this
1975, 1986) have suggested that personality and component are knowledge and use of heuristics, or
motivational factors underlie much of the indi- sets of rules, for generating novel responses. Con-
vidual differences among people in creativity. Cat- ducive workstyles include abilities to concentrate
tell ( 1971) proposed that other than intelligence and for long periods of time and to abandon fruitless
specific talents, personality factors most determine strategies. Relevant personality characteristics are
I 2 3
QOMAIN-RELEVANT SKILbS CREATIVITY-RELEVANT SKILLS TASK MOTIVATION
INCLUDES: INCLUDES: INCLUDES:
-- -
---
KNOWLEDGE ABOUT THE DOMAIN APPROPRIATE COGNITIVE STYLE ATTITUDES TOWARD THE TASK
TECHNICAL SKILLS REQUIRED IMPLICIT OR EXPLICIT - PERCEPTIONS OF OWN MOTIVATION
SPECIAL DOMAIN-RELEVANT KNOWLEDGE OF HEURISTICS FOR FOR UNDERTAKING THE TASK
GENERATING NOVEL IDEAS
"TALENT"
DEPENDS ON:
- CONDUCIVE WORK STYLE
DEPENDS ON:
- INITIAL LEVEL OF INTRINSIC
- INNATE COGNITIVE ABILITIES

DEPENDS ON:
- -
MOTIVATION TOWARD THE TASK
- INNATE PERCEPTUAL AND
MOTOR SKILLS --
TRAINING
EXPERIENCE IN IDEA GENERATION
PRESENCE OR ABSENCE OF
SALIENT EXTRINSIC CONSTRAINTS
IN THE SOCIAL ENVIRONMENT
- FORMAL AND INFORMAL
EDUCATION
PERSONALITY CHARACTERISTICS
- INDIVIDUAL ABILITY TO
COGNITIVELY MINIMIZE EXTRINSIC
CONSTRAINTS
Figure 6. Amabile's components of creativity (see text for explanation). From The social psychology of creativity: A componential
conceptualization by T. M. Amabile, 1983, Journal of Personality and Social Psychology, 45, 357-376. Copyright 1983 by the
American Psychological Association. Reprinted by permission of the author and the American Psychological Association.
those commonly reported as correlates of creative given task determines overall creativity on that
people, including self-discipline, ability to delay task.
gratification, perseverance, and absence of con- The framework deals with the entire creative pro-
formity. cess, from problem finding to evaluation, and in-
The final component, task motivation, is the volves feedback from Step 5. Briefly, problems are
focus of Amabile's own research. Her emphasis has presented in Step 1, either from the individual's
been on the role of intrinsic motivation, that "gen- intrinsic interest or from someone else, but the indi-
erated by the individual's reaction to intrinsic prop- vidual will be more motivated to solve problems
erties of the task" (l983b, p. 364), and which in- that arise intrinsically. Step 2 is preparatory to actu-
volves engaging in an activity for its own sake. An al response generation and reactivates or acquires
individual's task motivation is specific to a particu- information relevant to the task. As Amabile sug-
lar task and has two basic elements: baseline at- gests, Step 2 may involve a considerable invest-
titude toward the task, determined by an appraisal ment of time if the individual has to acquire the
of the degree to which the task matches the person's information afresh. In Step 3, novel potential solu-
own interests, and perceptions of motivation for tions are generated through use of both cognitive
undertaking the task, largely determined by exter- problem-solving pathways and potentially useful
nal social! environmental factors. In particular, elements in the environment. Evaluation of poten-
Amabile is concerned about extrinsic factors that tial solutions in Step 4 occurs through use of do-
may control the person's performance, such as remain-relevant skills. Finally, a decision is made in
wards. She has adopted the "overjustification prin- Step 5 that determines whether or not the process
ciple" from attribution theories, which states basi- can be terminated or must be rerun from Step 1.
cally that external constraints on an individual's Feedback from success or failure may increase or
involvement in a task are inversely related to intrin- decrease task motivation for future involvement
sic motivation. Thus, "extrinsic constraints will, with similar problems.
by impairing intrinsic motivation, have detrimental The framework is admittedly tentative and ap-
effects on creative performance" (1983b, p. 365). pears to have some shortcomings. Although feed-
A possible schematic framework representing in- back is explicitly involved, problem solving is pre-
teractions among the components in actual creative sented as an essentially linear process in which
production is shown in Figure 7. As with the com- failure or partial success leads to a return to Step 1.
ponents themselves, the framework synthesizes Given close evaluation of shortcomings of a poten-
earlier approaches with Amabile's own focus. For tial solution, however, one might well return to
example, as Amabile stated, Steps 1-5 at the top Step 2 or even 3 rather than Step 1. The mUltiple
are clearly based on earlier models of problem solv- feedback loops in Guilford's SIPS model may more
ing. An individual's level of each component on a accurately depict actual problem solving. Also, the
2 3 4 5
PROBLEM OR RESPONSE RESPONSE
TASK PRESENTATION PREPARATION GENERATION VALIDATION OUTCOME - - - - - - - - - - - - -,
I
I EXTERNAL
STIMULUS
I BUILDING UP AND/OR
REACTIVATING
STORE OF
SEARCH MEMORY
ANO IMMEDIATE
ENVIRONMENT
TEST RESPONSE
POSSIBILI T Y
AGAINST FACTUAL
iCOMPLETE ATTAINMENT
GOAL
or~
:---+ END
I
f-o ~:FL~V~~JION AND f-oo

OR TO GENERATE ~ (SUCCESSI
I I to
RESPONSE ~ ~~~~~E~~~:R~~
-
OR
INTERNAL RESPONSE POSSIBILITY
STIMULUS ALGORITHMS REASONABLE RESPONS~~
'\ ,, - -- POSSIBILITIES GENERATED END
- - --
~
1 (FAILUREI
,, 1 I
I
\
, 1 OR
----
\ I
\ 'I, , --I SOME PROGRESS TOWARD ~
,, r- RETURN TO I I
-
\ I GOAL
\
I
I
\ I
\
\
\
\
DOMAIN-RELEVANT CREATIVITY-
TASK MOTIVATION
SKILLS RELEVANT SKILLS
LEARNING
SET - BREAKING
I
~-------------------------------------------~
INCREASE OR DECREASE
Figure 7. Amabile's componential framework of creativity. Solid arrows indicate sequence of steps in process; broken arrows
indicate influences among factors. From The social psychology of creativity: A componential conceptualization by T. M. Amabile,
1983, Journal of Personality and Social Psychology, 45, 357-376. Copyright 1983 by the American Psychological Association.
Reprinted by permission of the author and the American Psychological Association.
role of specific goals for the process is not clear. proach and indicate the importance of so-
Without such goals, response validation is itself cial/contextual factors on creative productivity: (a)
problematic. Further, each of the three components for both children and adults, external evaluation or
influences directly only one or two of the steps, expectation of evaluation lowers creative produc-
whereas they might well affect virtually all. For tivity on verbal and artistic tasks; (b) external re-
example, task motivation stops directly guiding the wards generally decrease creative productivity; (c)
process at Step 3, but as has been noted earlier in choice in whether or how to engage in a particular
this chapter, problem solving frequently breaks problem activity increases creativity; and (d) ex-
down at the evaluation step, with a failure to compressed interest in an activity is positively related to
pare a proposed solution with the requirements of creative performance. These results support the in-
the task. This failure would be particularly likely to trinsic-motivation approach, although Amabile ap-
occur on tasks induced through extrinsic motivation pears not to have dealt with the crucial issue of the
in which the individual has little interest. origin of that intrinsic motivation. Of greater in-
Criticisms aside, however, Amabile has pro- terest from the perspective of this chapter, Amabile
vided a model explicit and clear enough to be both has provided a novel answer to the question of what
rich in testable hypotheses, some of which are de- we should measure about creativity, focusing on
tailed in Amabile (1983a,b), and easily modified to observable and manipulable determinants of actual
account for new information. Her research has led creative products rather than on underlying process.
to an impressive body of information on the role of Her approach has led to a number of confirmed
situational factors in the production of creative so- predictions and can synthesize a variety of previous
lutions. She has adopted a functional approach to research. As such, it represents a particularly pro-
creativity, varying antecedent and consequent en- gressive research program.
vironmental factors and measuring their effect on
creative productivity. Creativity of her subjects'
products has been evaluated by a consensual tech- Conclusions
nique, using groups of informed judges. A brief
summary of some of her results, largely from 1. Much theorizing about creativity has not been
Amabile (1983a), will show the power of the ap- clearly linked to observable antecedent and conse-
quent conditions. Constructs hang suspended in the spective case histories are generally unverifiable.
ether, being tied to reality by neither adequate Further, no control comparisons are possible in
bridge statements nor clear and distinctive opera- which the effect of presence or absence of given
tional definitions. One cannot determine specifical- factors can be assessed. Generally accepted princi-
ly how they relate to creative productivity or can ples of statistical inference should be consistently
lead to predictions. Such theories may be ques- followed.
tioned on the basis of both relevance and testability. 7. All too often, authors have followed Detter-
2. Several theoretical approaches, including the man's (1979, p. 174) Law XXI. "Lackingreliabili-
psychoanalytic and factorial, may be described as ty and or validity, theorize." The dictum, "Thou
degenerating and of questionable value. Interest in shalt not infer causality on the basis of correla-
Guilford's structure of intellect among researchers, tions," is frequently ignored by theorists trying to
if not educators, has greatly diminished, reflecting explain creativity. As a result, much explanation
an awareness of its limitations. The fact that some falls into the logical trap of post hoc, ergo prop-
professionals continue to offer psychoanalytic exter hoc.
planations is itself a phenomenon worthy of study, 8. The possibility that individual differences in
perhaps in terms of dissonance reduction. creativity reflect differences in personality and mo-
3. Clear operational discriminability needs to be tivation rather than cognitive variables needs to be
established among concepts. As informational con- reexamined in light of the current status of cog-
tent, emotional, and motivational factors influence nitive-trait approaches. The importance of such fac-
performance on intelligence tests, they also may tors has, of course, been emphasized by numerous
influence performance on creativity tests, particu- researchers.
larly fluency tests. Unfortunately, we cannot now 9. In studying creativity, many psychologists ap-
easily determine whether fluency tests tap creativity pear to have committed the "fundamental attribu-
or some other construct. One might suggest that, tion error" (e.g., Jones, 1979; Ross, 1977), under-
given the general stability of cognitive processes, if estimating the role of situational factors and
short-term programs increase creative productivity, overestimating the role of unique intrapersonal fac-
the effect is probably on a motivational-emotional tors. The role of situational in addition to intraper-
variable rather than on creativity itself. sonal factors in influencing creative productivity
4. The search for a "g" factor of creativity in needs to be considered more consistently in re-
terms of a divergent process is problematic; it does search. Both Stein (e.g., 1953) and Torrance (e.g.,
not seem to involve a progressive research pro- 1961, 1975) persuasively and persistently argued
gram. Creativity appears much more domain-spe- the importance of situational and cultural factors in
cific than intelligence and likely to consist of a determining actual creative productivity. Ama-
number of processes. Although it may consist of bile's (1983a,b) experimental-social and Feld-
essentially the same set of steps across domains, man's (1980, 1986) developmental-process ap-
individuals highly creative in one area will not like- proaches are but two examples that examine
ly be in others. The ability to generate remote asso- situational factors. Consideration of creativity as
ciations appears in so many accounts of creativity resulting from person-situational interactions
that its role should be pursued, but in domain-spe- brings the area more in accord with trends in psy-
cific ways. The appearance of an occasional "Re- chology as a whole.
naissance man" may reflect nothing more than A general implication is that we should, however
chance combination of domain-specific factors. reluctantly, question the presumed existence of any
5. Within domain, creativity is multi factor and such thing as creativity. I cannot improve on Clar-
consists of more than ability to generate associa- idge's observation (1987, p. 134) concerning an-
tions. Guilford's (1950) original description of the other construct, except to substitute the word
entire creative process, from problem finding to "creativity" for "arousal": "I have often felt that
evaluation, appears more accurately to reflect what as an explanatory concept in psychology, 'creativ-
leads to a creative product. As such, creativity may ity' has many of the qualities of a difficult but per-
best be seen as an aspect of general problem solving suasive lover, whom reason tells one to abandon yet
(e.g., Newell et al., 1962) in which all steps, and who continues to satisfy an inescapable need." Fi-
not just that involving production, are important. nally, as an indication of our current level of under-
6. Creativity research needs a more reliable and standing, I offer in Figure 8 a "Confusional Model
verifiable data base. Self-report data and retro- of Creativity. "
UNSOLVED CREATIVE CREATIVE

PROBLEM PROCESS PRODUCT
"DISCOVERED" OR A MIRACLE UNUSUALNESS Figure 8. Brown's "Confusional Model of

"PRESENIED" PROBLEM OCCURS HERE APPROPRIATENESS Creativity." As in Sidney Harris's classic cartoon,
CONDENSATION on which it is obviously based: "We need to be more
TRANSFORMATION specific in step 2. "
ACKNOWLEDGMENTS. The author thanks William Cattell, R. B. (1971). Abilities: Their structure, growth, and
B. Michael, Lee A. Jackson, Cecil R. Reynolds, action. Boston: Houghton Mifflin.
and especially John A. Glover for many useful dis- Claridge, G. (1987). Psychotism and arousal. In J. Strelau & H.
J. Eysenck (Eds.), Personality dimensions and arousal. New
cussions about creativity and suggestions for this York: Plenum Press.
chapter and Martha Jo Clemmons and Lydia R. Conant, J. B. (1947). On understanding science. New Haven,
Woodard for assistance in editing. He also thanks CT: Yale University Press.
George S. Welsh for conversations some years ago Cronbach, L. J. (1984). Essentials of psychological testing (4th
ed.). New York: Harper & Row.
that helped to focus attention on issues concerning Crovitz, H. F. (1970). Galton's walk. New York: Harper &
the study of creativity. Row.
Detterman, D. K. (1979). Detterman's laws of individual dif-
ference research. In R. J. Sternberg & D. K. Detterman
References (Eds.), Human intelligence (pp. 165-175). Norwood, NJ:
Ablex.
Albert, R. S. (Ed.). (1983). Genius and eminence. New York: Dewey, J. (1910). How we think. Boston: D. C. Heath.
Pergamon. Ericsson, K. A., & Simon, H. A. (1984). Protocol analysis.
Albrecht, K. (1980). Brain power. Englewood Cliffs, NJ: Pren- Cambridge, MA: M.LT. Press.
tice-Hall. Eysenck, H. J. (1979). The structure and measurement ofintel-
Amabile, T. M. (l983a). The social psychology of creativity. ligence. Berlin: Springer-Verlag.
New York: Springer-Verlag. Eysenck, H. J., & Eysenck, M. W. (1985). Personality and
Amabile, T. M. (l983b). The social psychology of creativity: A individual differences. New York: Plenum Press.
componential conceptualization. Journal of Personality and Feldman, D. H. (1970). Faulty construct-ion. Contemporary
Social Psychology, 45, 357-376. Psychology, 15, 3-4.
Austin, J. H. (1978). Chase, chance, and creativity. New York: Feldman, D. H. (1980). Beyond universals in cognitive develop-
Columbia University Press. ment. Norwood, NJ: Ablex.
Barron, F. (1968). Creativityandpersonalfreedom. Princeton, Feldman, D. H. (1986). Nature's gambit. New York: Basic
NJ: Van Nostrand. Books.
Barron, F., & Harrington, D. M. (1981). Creativity, intel- Fox, H. H. (1963). A critique on creativity in science. In M. A.
ligence, and personality. Annual Review of Psychology, 32, Coler (Ed.), Essays on creativity in the sciences (pp. 123-
439-476. 152). New York: New York University Press.
Bloomberg, M. (Ed.). (1973). Creativity: Theory and research. Freeman, F. N. (1924). Mental tests. New York: Houghton
New Haven, CT: College and University Press. Mifflin.
Boring, E. G. (1923). Intelligence as the tests test it. New Re- Gardner, H. (1983). Frames of mind. New York: Basic Books.
public, 35, 35-37. Getzels, J. W., & Csikszentmihalyi, M. (1976). The creative
Brewer's dictionary ofphrase andfable. (1962). (6thed.). Lon- vision: A longitudinal study of problem solving in art. New
don: Cassell. York: Wiley.
Buros, O. K. (Ed.). (1972). The seventh mental measurement Ghiselin, B. (Ed.). (1952). The creative process. New York:
yearbook. Highland Park, NJ: Gryphon Press. Mentor.
Buros, O. K. (Ed.). (1978). The eighth mental measurement Ghiselin, B. (1963). Ultimate criteria for two levels of creativity.
yearbook. Highland Park, NJ: Gryphon Press. In C. W. Taylor & F. Barron (Eds.), Scientific creativity: Its
Busse, T. V., & Mansfield, R. S. (1980). Theories of the cre- recognition and development (pp. 30-43). New York: Wiley.
ative process: A review and a perspective. Journal ofCreative Ghiselin, B., Rompel, R., & Taylor, C. W. (1964). A creative
Behavior, 14,91-103, 132. process check list: Its development and validation. In C. W.
Campbell, D. T. (1960). Blind variation and selective retention Taylor (Ed.), Widening horizons in creativity (pp. 19-33).
in creative thought as in other knowledge processes. New York: Wiley.
Psychological Review, 67, 380-400. Gruber, H. E. (1981). Darwin on man (2nd ed.). Chicago: Uni-
Carroll, J. B. (1968). [Review of The nature of human intel- versity of Chicago Press.
ligence by J. P. Guilford]. American Educational Research Guilford, J. P. (1950). Creativity. American Psychologist, 5,
Journal, 73, 105-112. 444-454.
CHAPTER I • WHAT ARE WE TO MEASURE? 31
Guilford, J. P. (1956). The structure of the intellect. Lakatos, I. (1978). Philosophical papers. Vol I: The meth-
Psychological Bulletin, 53, 267-293. odology of scientific research programmes (J. Worrall & G.
Guilford, J. P. (1959). Three faces of intellect. American Psy- Currie, Eds.). Cambridge: Cambridge University Press.
chologist, 14, 469-479. Lakatos, I., & Musgrave, A. (Eds.). (1970). Criticism and the
Guilford, J. P. (1967). The nature of human intelligence. New growth of knowledge. Cambridge: Cambridge University
York: McGraw-Hill. Press.
Guilford, J. P. (1975). Creativity: A quarter century of progress. Landau, E. (1985). Creative questioning for the future. In J.
In I. A. Taylor & J. W. Getzels (Eds.), Perspectives in Freeman (Ed.), The psychology of gifted children (pp. 379-
creativity (pp. 37-59). Chicago: Aldine. 392). New York: Wiley.
Guilford, J. P. (1977). Way beyond the IQ. Buffalo, NY: Cre- MacCorquodale, K., & Meehl, P. E. (1948). On a distinction
ative Education Foundation. between hypothetical constructs and intervening variables.
Guilford, 1. P. (1982). Cognitive psychology's ambiguities: Psychological Review, 55,95-107.
Some suggested remedies. Psychological Bulletin, 89, 48- MacKinnon, D. W. (1962). The personality correlates of
59. creativity: A study of American architects. In S. Coopersmith
Hacking, I. (Ed.). (1981). Scientific revolutions. Oxford: Ox- (Ed.), Personality research, Vol. 2 of G. S. Nielsen (Ed.),
ford University Press. Proceedings of the XIV International Congress of Applied
Harding, S. G. (Ed.). (1976). Can theories be refuted? Boston: Psychology (pp. 11-39). Copenhagen: Munksgaard.
Riedel. MacKinnon, D. W. (1965). Personality and the realization of
Hempel, C. G. (1966). Philosophy of natural science. creative potential. American Psychologist, 20, 273-281.
Englewood-Cliffs, NJ: Prentice-Hall. MacKinnon, D. (1970). Creativity: A multi-faceted phe-
Henle, M. (1962). The birth and death of ideas. In H. E. Gruber, nomenon. In J. D. Roslansky (Ed.), Creativity: A discussion
G. Terrell, & M. Wertheimer (Eds.), Contemporary ap- at the Nobel conference (pp. 17-32). Amsterdam: North-
proaches to creative thinking (pp. 31-62). New York: Holland.
Atherton. Maltzman, I. (1960). On the training of originality.
Henle, M. (1974). The cognitive approach: The snail beneath the Psychological Review, 67, 229-242.
shell. In S. Rosner & L. E. Abt (Eds.), Essays in creativity Mansfield, R. J., & Busse, T. V. (1981). The psychology of
(pp. 23-44). Croton-on-Hudson, NY: North River Press. creativity and discovery. Chicago: Nelson-Hall.
Hocevar, D. (1979). Ideational fluency as a confounding factor McCrae, R. R. (1987). Creativity, divergent thinking, and open-
in the measurement of originality. Journal of Educational ness to experience. Journal ofPersonality and Social Psychol-
Psychology, 71, 191-196. ogy, 52, 1258-1265.
Hocevar, D. (1980). Intelligence, divergent thinking, and Mednick, S. A. (1962). The associative basis of the creative
creativity. Intelligence, 4, 25-40. process. Psychological Review, 69, 220-232.
Hocevar, D. (1981). Measurement of creativity: Review and Mednick, S. A., & Mednick, M. T. (1964). An associative
critique. Journal of Personality Assessment, 45, 450-464. interpretation of the creative process. In C. W. Taylor (Ed.),
Hom, J. L., & Knapp, J. R. (1973). On the subjective character Widening horizons in creativity (pp. 54-68). New York:
of the empirical base of Guilford's structure of intellect model. Wiley.
Psychological Bulletin, 80, 33-43. Meeker, M. (1969). The structure of intellect: Its interpretation
Hudson, L. (1966). Contrary imaginations. London: Methuen. and uses. Columbus, OH: Merrill.
Humphreys, L. G. (1962). The organization of human abilities. Michael, W. B. (1977). Cognitive and affective components of
American Psychologist, 17, 475-483. creativity in mathematics and the physical sciences. In J. C.
Jackson, P. W., & Messick, S. (1967). The person, the product, Stanley, W. C. George, & C. H. Solano (Eds.), The gifted and
and the response: Conceptual problems in the assessment of the creative: A fifty-year perspective. Baltimore: Johns
creativity. In J. Kagan (Ed.), Creativity and learning (pp. 1- Hopkins University Press.
19). Boston: Houghton Mifflin. Miller, G. A., Galanter, E., & Pribram, K. H. (1960). Plans and
Jones, E. E. (1979). The rocky road from acts to dispositions. the structure of behavior. New York: Holt.
American Psychologist, 34, 107-117. Miller, N. E. (1959). Liberalization of basic S-R concepts: Ex-
Khatena, J. (1984). Imagery and creative imagination. Buffalo, tensions to conflict behavior, motivation and social learning.
NY: Bearly Limited. In S. Koch (Ed.), Psychology: A study of a science (Vol. 2,
Koestler, A. (1964). The act of creation. New York: Macmillan. pp. 196-292). New York: McGraw-Hill.
Kogan, N. (1983). Stylistic variation in childhood and adoles- Mischel, W. (1968). Personality and assessment. New York:
cence: Creativity, metaphor, and cognitive styles. In P. H. Wiley.
Mussen, J. H. Flavell, & E. M. Markman (Eds.), Handbook Mooney, R. L. (1963). A conceptual model for integrating four
of child psychology (4th ed., Vol. 3, pp. 630-706). New approaches to the identification of creative talent. In C. W.
York: Wiley. Taylor & F. Barron (Eds.), Scientific creativity: Its recogni-
Kuhn, T. S. (1962). The structure of scientific revolutions. tion and development (pp. 331-340). New York: Wiley.
Chicago: University of Chicago Press. Newell, A., Shaw, J. c., & Simon, H. A. (1962). The processes
Kuhn, T. S. (1963). The essential tension: Tradition and innova- of creative thinking. In H. E. Gruber, G. Terrell, & M.
tion in scientific research. In C. W. Taylor & F. Barron Wertheimer (Eds.), Contemporary approaches to creative
(Eds.), Scientific creativity: Its recognition and development thinking (pp. 63-119). New York: Atherton Press.
(pp. 341-354). New York: Wiley. Nicholls, J. G. (1972). Creativity in the person who will never
Lakatos, I. (1970). Falsification and the methodology of reproduce anything original and useful: The concept of
search programmes. In I. Lakatos & A. Musgrave (Eds.), creativity as a normally distributed trait. American Psychol-
Criticism and the growth of knowledge (pp. 8-101). ogist, 27, 717-727.
Cambridge: Cambridge University Press. Pezzullo, T. R., Thorsen, E. E., & Madaus, G. F. (1972). The
heritability of Jensen's level I and II and divergent thinking. In I. A. Taylor & J. W. Getzels (Eds.), Perspectives in
American Educational Research Journal, 9. 539-546. creativity (pp. 191-223). Chicago: Aldine.
Popper, K. R. (1935, trans. 1959). The logic of scientific discov- Taylor, C. W., & Holland, J. (1964). Predictors of creative
ery. New York: Basic Books. performance. In C. W. Taylor (Ed.), Creativity: Progress and
Poincare, H. (1913). Thefoundationsofscience. Lancaster, PA: potential (pp. 15-48). New York: McGraw-Hill.
Science Press. Thorndike, R. L. (1963). Some methodological issues in the
Ross, L. (1977). The intuitive psychologist and his shortcom- study of creativity. In Proceedings of the 1962 Invitational
ings: Distortions in the attribution process. In L. Berkowitz Conference on Testing Problems (pp. 40-54). Princeton, NJ:
(Ed.), Advances in experimental social psychology (Vol. 10, Educational Testing Service.
pp. 174-220). New York: Academic Press. Torrance, E. P. (1961). Factors affecting creative thinking in
Rossman, J. (1931). The psychology of the inventor: A study of children: An interim research report. Merrill-Palmer Quar-
the patentee. Washington, DC: Inventors Publishing Co. terly of Behavior and Development. 7. 171-180.
Rothenberg, A. (1979). The emerging goddess: The creative Torrance, E. P. (1975). Creativity research in education: Still
process in art, science, and other fields. Chicago: University alive. In I. A. Taylor & J. W. Getzels (Eds.), Perspectives in
of Chicago Press. creativity. Chicago: Aldine.
Simonton, D. K. (1984). Genius. creativity. and leadership: Turner, M. B. (1968). Psychology and the philosophy of sci-
Historiometric inquiries. Cambridge, MA: Harvard Univer- ence. New York: Appleton-Century-Crofts.
sity Press. Vernon, P. E. (Ed.). (1970). Creativity. Harmondsworth, Mid-
Spearman, C. (1931). Creative mind. New York: Appleton. dlesex, England: Penguin Books.
Stein, M. I. (1953). Creativity and culture. Journal of Person- Wallach, M. A. (1970). Creativity. In P. H. Mussen (Ed.),
ality, 36, 311-322. Carmichael's manual of child psychology (3rd ed., pp. 1211-
Sternberg, R. J. (1979). Intelligence research at the interface 1272). New York: Wiley.
between differential and cognitive psychology: Prospects and Wallach, M. A. (1971). The intelligence/creativity distinction.
proposals. In R. J. Sternberg & D. K. Detterman (Eds.), New York: General Learning Press.
Human intelligence (pp. 33-60). Norwood, NJ: Ablex. Wallach, M. A. (1986). Creativity testing and giftedness. In F.
Sternberg, R. J. (1985). Implicit theories of intelligence, D. Horowitz & M. O'Brien (Eds.), The gifted and talented:
creativity, and wisdom. Journal of Personality and Social Developmental perspectives (pp. 99-123). Washington, DC:
Psychology, 49, 607-627. American Psychological Association.
Sternberg, R. J. (1986). A triarchic theory of intellectual gifted- Wallach, M. A., & Kogan, N. (1965). Modes of thinking in
ness. In R. J. Sternberg & J. E. Davidson (Eds.), Conceptions young children. New York: Holt, Rinehart, & Winston.
of giftedness (pp. 223-243). Cambridge: Cambridge Univer- Wallach, M. A., & Wing, C. W., Jr. (1969). The talented stu-
sity Press. dent. New York: Holt, Rinehart, & Winston.
Sternberg, R. J., & Davidson, J. E. (1986). Cobnitive develop- Wallas, G. (1926). The art of thought. New York: Harcourt
ment in the gifted and talented. In F. D. Horowitz & M. Brace.
O'Brien (Eds.), The gifted and talented: Developmental per- Weisberg, R. W. (1986). Creativity: Genius and other myths.
spectives (pp. 37-74). Washington, DC: American Psycho- New York: W. H. Freeman.
logical Association. Welsh, G. S. (1975). Creativity and intelligence: A personality
Sternberg, R. J., & Detterman, D. K. (Eds.). (1979). Human approach. Chapel Hill, NC: Institute for Research in Social
intelligence. Norwood, NJ: Ablex. Science.
Sternberg, R. J., & Detterman, D. K. (Eds.). (1986). What is Welsh, G. S. (1986). Positive exceptionality: The academically
intelligence? Norwood, NJ: Ablex. gifted and the creative. In R. T. Brown & c. R. Reynolds
Taylor, C. W. (1947). A factorial study of fluency in writing. (Eds.), Psychological perspectives on childhood excep-
Psychometrika. 12, 239-262. tionality: A handbook (pp. 311-343). New York: Wiley.
Taylor, C. W., & Barron, F. (Eds.). (1963). Scientific Wertheimer, M. (1959). Productive thinking (enlarged ed.).
creativity: Its recognition and development. New York: New York: Harper & Row.
Wiley. Zigler, E., & Butterfield, E. C. (1968). Motivational aspects of
Taylor, C. W., & Ellison, R. L. (1975). Moving toward working changes in IQ test performance of culturally deprived nursery
models in creativity: Utah creativity experiences and insights. group school children. Child Development. 39. 1-14.
CHAPTER 2
Psychometric Issues in the

Assessment of Creativity
William B. Michael and Claudia R. Wright
Overview creativity. A cursory exploration of a few sel~ct~d

issues in the administration of tests of creatIvIty
In this first section, an overview of the remaining follows in the seventh section, and the eighth sec-
seven divisions of the chapter is presented. The tion contributes a concluding statement.
second section affords a brief description of cate-
gories of instrumentation to provide the read~r
with a foundation within which the psychometnc Alternative Methods for Measurement
issues in the assessment of creativity can be of Creativity
viewed. In the third section, psychometric con-
cerns pertaining to construct validity, content va-
Considerable diversification exists in the types of
lidity, and criterion-related validity are addressed.
measures used in the assessment of creativity. Al-
Subsequent to an abbreviated review of the mean-
though pencil-and-paper tests probably account for
ing of reliability, concerns regarding optimal ap-
the majority of measures employed in research
proaches to the estimation. of relia.bility of crea- studies, other techniques are commonly used. In a
tivity measures are consIdered III the fourth
comprehensive review and critique of currently
section. The impact upon reliability and validity of
available techniques for the measurement of
scoring procedures is examined in the fifth section
creativity, Hocevar (1981) arrived at a taxonomy of
with particular emphasis on the presence of ~uen
10 categories. He evaluated these alternative ap-
cy as a confounding factor in the. int~rpretatlO~ of
proaches to assessment of creativity .in terms of
scores of measures of divergent thmking. The sIxth
reliability, discriminant validity, dimensionality,
section provides a survey of a number of t~e diffi-
and convergent validity. These 10 categories in-
culties encountered in establishing normatIve data
cluded (a) tests of divergent thinking, (b) attitude
for the understanding of scores on measures of
and interest inventories, (c) personality inventories,
(d) biographical inventories, (e) teacher nomina-
tions, (f) peer nominations, (g) supervisor ratings,
WiUiam B. Michael • Department of Educational Psychol- (h) judgments of products, (i) eminence, and (j)
ogy. University of Southern California, Los Angeles, CA
90089. Claudia R. Wright • Department of Educa-
self-reported creative activities and achievements.
tional Psychology and Administration, California State Univer- Hocevar concluded that an inventory of self-re-
sity, Long Beach, CA 90840. ported creative endeavors and achievements is the
33
most defensible approach for identifying creative Construct Validity

individuals. In Chapter 3 of this volume, Hocevar
and Bachelor have given a detailed critique of in- In this discussion, the topic of construct validity
struments designed to assess creativity. as a psychometric concern in the assessment of
creativity is considered, followed by five major is-
sues that are presented in subsequent sections: (1)
construct validity is defined in general terms with
Concerns Relating to Validity reference to theoretical as well as statistical per-
spectives; special attention is given to definitions of
Most psychometricians as well as theorists who creativity which emphasize several constructs, such
have worked in the area of creativity probably as fluency, originality, flexibility, and elaboration;
would agree that validity of creativity assessment is (2) a review is presented of the techniques of ex-
the single most important consideration. As Wolf ploratory factor analysis and confirmatory factor
(1982) has so aptly pointed out, the validity of a test analysis employed in the study of construct validity;
is anchored to three concerns: (a) what a test is (3) the application of multitrait-multimethod ap-
supposed to be measuring, (b) what the score de- proaches to the problem of construct validity is con-
rived from the administration of a test means, and sidered in general, with specific attention given to
(c) how the score of an individual on a measure is its application to the domain of creativity; (4) the
related to other observable facts regarding that indi- use of three other methodologies comprising ex- .
vidual. Wolf has indicated that this third concern is perimental, longitudinal, and computerized imag-
the essence of the validation of the interpretation of ing studies in the process of validating constructs of
a test score-a concern that in many instances creativity is examined; and (5) some pertinent ques-
could be as significant as, if not more important tions are posed central to inquiry about construct
than, the validation of the test itself. validity.
Within this broad context, three major types of
validity-construct, content, and criterion-relat-
Definition and Rationale
ed-are examined, especially as they relate to the
assessment of creative endeavor. These three types Construct Validity Viewed from a Theoretical
of validity are based on Standards for Educational Orientation. Underlying the process of construct
and Psychological Testing prepared by the Com- validation is the development of a conceptual link-
mittee to Develop Standards for Educational and age or network among selected variables identified
Psychological Testing of the American Educational within any given behavioral domain (such as locus
Research Association, the American Psychological of control, intelligence, or creativity) which when
Association, and the National Council on Measure- taken together form a composite or construct con-
ment in Education which was published in 1985 by sidered to hold promise as an indicator of the man-
the American Psychological Association. Several ifestation of the target behavior. It is assumed that
general texts dealing with problems of psychologi- the process of construct validation requires an in-
cal and educational measurement also have treated ferential leap from observed patterns of response
these three types of validity on the basis of content consistency (to which some meaning is attached) to
in an earlier edition entitled Standards for Educa- a specific construct which purports to account for
tional and Psychological Tests and Manuals the observed consistency. As Messick (1975) point-
(American Psychological Association, 1974). A ed out, typically two major requirements, based on
representative sampling of writings that present in- the pioneer work of Campbell and Fiske (1959),
formation about standards or criteria for validity exist for establishing construct validity. The first
would include contributions by Anastasi (1982), requirement is convergent validation, the demon-
Campbell (1976), Crocker and Algina (1985), stration that a selected measure of a given behavior
Cronbach (1971, 1980), Nunnally (1978), Thorn- is related to other measures linked to the same con-
dike (1982), and Wolf (1982). In the context ofthe struct as well as to other criterion-related variables
assessment of creativity, validity issues have been which have a theory-based relationship to the con-
treated by a number of writers (e.g., Golann, 1963; struct. The second requirement involves discrim-
Guilford, 1971; Thorndike, 1963; Treffinger & inant validation which shows that the selected mea-
Poggio, 1972; Treffinger, Renzulli, & Feldhusen, sure is not linked to other measures and variables
1971). which theoretically support distinctly different con-
CHAPTER 2 • PSYCHOMETRIC ISSUES 35
structs. These requirements are also reflected in the eign context (Guilford, 1963, 1967, 1968, 1970,
technical standards related to construct validity list-1971; Guilford & Hoepfner, 1971; Guilford & Ten-
ed in the previously cited Standards for Educa- opyr, 1968; Michael, 1977; Torrance, 1966).
tional and Psychological Testing (American Psy- The first four constructs rely heavily on concep-
chological Association, 1985). Additional contri- tualizations associated with divergent thinking abil-
butions concerned with the integration of evidence ities. In the instance of fluency of verbal ex-
for construct validity may be found in the writings pression, the ability is conceptualized as the
of Cronbach (1971, 1980), Cronbach and Meehl divergent production of semantic units, relations,
(1955), Fiske (1987), Tenopyr (1977), and Wolf and systems within the SOl model. It usually is
(1982). operationalized in terms of the quantity of elements
that can be generated. Originality (adaptative flexi-
Construct Validity Viewed from a Psychometric bility) has been defined in terms of measures indi-
Orientation. Statistical procedures central to psy- cating statistical infrequency, that is, those re-
chometric methods often have been used to afford a sponses considered to be a rare occurrence either in
relatively objective basis for arriving at the selec- comparison of a given response of a subject to all of
tion of measures to represent response consistencies his or her previous responses, or in comparison of
that are conceptually linked to a construct. The sta- that given response to all responses from different
tistical procedures most frequently employed to subjects (a within vs. a between comparison). The
support the construct validation process include construct of flexibility generally applies to that of
correlational and factor-analytic approaches, such spontaneous flexibility in which a shift from one
as (a) the simple reporting of coefficients obtained class of responses to another is involved as in the
between total scores or subscores of different mea- listing of alternative uses of an object, such as a
sures selected on the basis of similarity in opera- shoe. In divergent thinking, elaboration may be
tional definitions and/or conceptualizations to indi- measured in terms of the number of occupations
cate the strength of interrelationships among the that might be associated with a given symbol, such
measures; (b) multitrait-multimethod analyses for as a key. In contrast to these four constructs indica-
addressing convergent and discriminant validity tive of divergent production, the fifth construct in-
concerns with particular attention applied to disen- volving the product of transformations arises fre-
tanglement of method and trait variance (Campbell quently in the determination of a unique solution to
& Fiske, 1959; Fiske, 1987); (c) exploratory factor a problem with many constraints as in mathematics
analysis (Gorsuch, 1983; Nunnally, 1978); and (d) or invention-a problem often involving a novel or
confirmatory factor analysis (Gorsuch, 1983; ingenious use of a familiar object or process in a
Joreskog, 1969; Joreskog & Sorbom, 1981), which new or foreign context.
can be subsumed under causal modeling (Bentler, Treffinger et al. (1971) have emphasized that the
1980). Despite the usefulness of various psycho- measures of most common creativity constructs
metric procedures to facilitate evidence of the valid- have been based on simple quantitative rather than
ity of a construct, it should be noted, as Cronbach qualitative dimensions. This quantitative emphasis
(1971) has stated, that "one validates, not a test, may run the risk of leading the investigator to ignore
but an interpretation of data arising from a specified in one hypothesized construct qualitative charac-
procedure" (p. 447). teristics in behaviors that actually reflect other con-
structs in need of operational specification. Defini-
Defining Constructs Related to Creativity. If tion of these other constructs in the form of new
creativity is organized as a complex of interrelated measures would permit an initial determination of
constructs, the problem of construct validity must the place of the original ('onstruct within an ordered
focus upon the operationalization of key theoretical network of hypothesized and measurable constructs
constructs which are thought to underlie creative in an expanded domain of human behavior. More-
behaviors. Among the most commonly cited con- over, it should be emphasized that a simple numer-
structs are those originally proposed by Guilford in ical count of frequency of responses to reflect a
his structure-of-intellect (SOl) model. They include construct of originality could overlook the occur-
fluency, originality, flexibility, and elaboration in rence of two or three highly significant responses on
divergent thinking and transformations involving the part of one examinee that qualitatively would be
the application of convergent thinking to the arrival worth a hundred fairly mundane responses of an-
at unique solutions, often in an unfamiliar or for- other examinee. Thus, the individual with the low
quantitative score could be unfairly penalized in his counting for smaller and smaller amounts of the
manifestation of what would be judged truly origi- variability left to be explained. As the initial factors
nal behavior. generated in the mathematical procedure are usu-
ally difficult to interpret, a procedure known as the
Factor-Analytic Approaches rotation of factor axes (the axes constituting mathe-
matical representations of initially derived dimen-
The use of factor-analytic approaches to con- sions to describe the intercorrelations among the
struct validation often constitutes an attempt to ex- variables in a correlation matrix) takes place to im-
plain the intercorrelations among test (item) prove the psychological meaningfulness and in-
variables in terms of a small number of psychologi- terpretability of the dimensions. As mentioned ear-
cally meaningful dimensions. An attempt is made lier, the variables loaded or weighted on the rotated
to interpret each dimension as an operational repre- factor axis are examined for common charac-
sentation of a latent trait or construct in terms of teristics to provide a basis for definition or descrip-
shared or common process and content charac- tion of the factor as a psychological construct. Ad-
teristics of these measures that serve to define or ditional information regarding factor-analytic
describe the dimension. An indication of the degree procedures may be found in standard texts, such as
of relationship of a test or item variable to a dimen- those by Gorsuch (1983) and Harman (1976).
sion is often expressed as a coefficient of correla-
tion, usually termed a factor loading or factor
weight. Confirmatory Factor-Analytic Techniques. Ex-
Frequently, hypotheses regarding the nature of ploratory factor analysis traditionally has made use
cognitive abilities or affective characteristics are of arbitrary methods not only for estimating com-
tested through use of factor-analytic techniques by munalities (sum of common factor variance for
choosing in advance three or four measures thought each test or item variable), which are inserted in the
to represent a common construct. If in the resulting diagonal of a correlation matrix prior to factor ex-
factor analysis the test or item variables hypoth- traction, but also for rotating factor axes. The re-
esized to represent the construct are loaded or sults frequently have led to a lack in uniformity and
weighted on a factor but are not correlated substan- replicability of the dimensionality of a specified
tially with any other factor, evidence is viewed as group of correlated variables. In contrast, confir-
being affirmative for that particular hypothesis. matory maximum likelihood factor analysis affords
This form of evidence for the validity of a construct a comparatively objective approach of providing
can occur in either exploratory or confirmatory fac- evidence to support or to disconfirm not only hy-
tor analyses if adequate conceptualization has oc- potheses obtained from exploratory factor analyses
curred and if corresponding care has been exercised (especially in the instance of a new data base), but
in the selection of measures to portray the con- also conceptually based hypotheses derived from a
structs hypothesized. researcher's theory. Specifically, the method of
confirmatory maximum likelihood factor analysis,
Exploratory Factor-Analytic Techniques. Ex- which has been made feasible in light of the avail-
ploratory factor-analytic approaches provide not ability of a set of computer programs in LISREL V
only a statistical method for tentative testing of ini- (Joreskog & Sorbom, 1981), furnishes a means
tial hypotheses derived from a rather tentative the- through which the consistency of observed patterns
oretical conceptualization, but also a strategy for of intercorrelations among measurable variables
trying to effect some degree of parsimony in under- can be compared with the configuration of intercor-
standing a large complex of intercorrelated vari- relations predicted from a group of hypothesized
ables in a domain of inquiry about which the re- latent variables (constructs). These hypothesized
searcher has limited knowledge or insight. variables that appear within a covariance structure
Generally, the overall goal in exploratory factor (causal model) are related to one another in a man-
analysis is to identify factors that will explain the ner harmonious with a researcher's theory. After
greatest amount of covariance among the variables. the theoretical constructs, or latent variables, and
Depending upon the statistical technique em- their indicators (e.g., test measures) have been
ployed, the first mathematically extracted factor specified, the interrelationships among the various
typically accounts for most of the covariance in the constructs can be indicated to constitute what is
data, with each subsequently extracted factor ac- called a structural model, and the relationship of
each construct to its measure (sometimes referred to likelihood factor analysis, were completed of cor-
as the measurement modef) can be indicated. relational data involving several creativity mea-
Constituting the major components of a causal sures from Guilford's Aptitudes Research Project,
model, sets of regression, or structural, equations which for many years was centered at the Univer-
are formulated that express the relationship of mea- sity of Southern California (Khattab et ai., 1982;
surement variables (dependent variables) to latent Khattab & Michael, 1986; Mace, Michael, &
variables (independent variables). These equations Hocevar, 1985). The resulting analyses yielded
designate the hypothesized effects associated with somewhat mixed results, partly as a function of
specific predictor variables upon one or several cri- which tests were incorporated within the correlation
terion variables. Understandably, a major statistical matrices. When a representative balance of ability
goal in the model becomes that of estimating its measures intended to reflect the content, process,
parameters as accurately as is feasible and of deter- and product elements of hypothesized first-order
mining (generally, through employment of a chi- and second-order SOl factors was included, reason-
square test) how closely the proposed model fits the able support for hypothesized constructs of the SOl
data acquired from the measured variables for the theory was forthcoming. Additional factor-analytic
particular group of subjects studied. If the judgment work with this comparatively new methodology
is reached that the model does not afford an appro- needs to be carried out to substantiate the validity of
priate fit to the data, the investigator may reject it as the constructs in the SOl model that have been
a plausible representative of the hypothesized thought to reflect creative behaviors.
causal structure describing the measured variables.
On the other hand, if application of significant tests
does not result in the rejection of the model (the null Concerns and Limitations in the Use of Factor-
hypothesis being tenable), then the model may be Analytic Approaches. Several conceptual and
considered as providing a reasonable representation methodological concerns and limitations arise in
of the hypothesized (causal) structure. Ordinarily, the use of factor-analytic techniques in the study of
in confirmatory maximum likelihood factor analy- a given psychological domain, such as creativity.
sis, a given factor model is evaluated against nu- Probably the one most important concern is that of
merous alternative factor models to provide a having as sound and carefully conceptualized the-
means of determining which model furnishes the oretical framework as is possible so that appropriate
most nearly adequate psychological interpretation steps can be taken to design a study that will yield
of the data as well as the optimal degree offit from a meaningful outcomes. Depending upon the stage of
statistical viewpoint in reproducing the entries in a development of a given area of psychological inqui-
given correlation model (Khattab, Michael, & ry, an integrated theoretical framework is necessary
Hocevar, 1982). for the appropriate selection of variables to con-
stitute the basis for operationalizing the hypotheses
Applications of Confirmatory Maximum Like- to be tested and to contribute to the design of the
lihood Factor Analysis to Guilford's SOl Model. correlation matrix that subsequently will be ana-
The published works of Guilford and his associates lyzed. It is important that marker or anchor vari-
(e.g., Guilford, 1967; Guilford & Hoepfner, 1971) ables defining factors that already have been pains-
have made use of various factor-analytic techniques takingly established in the research of other
that would be classified as exploratory. Many re- investigators be included so that (a) well-estab-
searchers (e.g., Harris & Harris, 1971; Hom & lished factors are not given new names or labels, (b)
Knapp, 1973; 1974), however, have criticized variance associated with previously defined dimen-
Guilford's factor-analytic solutions as not provid- sions is separated or removed from that associated
ing strong evidence in support of the SOl model with new constructs, and (c) an integration of new
largely because objective criteria were not em- research findings can be made with those of pre-
ployed. In fact, Hom and Knapp declared that the viously conducted studies. In contrast to carefully
Procrustean factor-analytic methods that Guilford planned analyses, the so-called shotgun approaches,
had employed could afford almost as strong support in which variables are indiscriminately introduced
for theories generated by random sampling pro- within a correlation matrix without reflective consid-
cedures as for the SOl model itself. erations of their impact upon the solution, may well
To overcome some of these objections of subjec- lead to psychologically meaningless or nonsensical
tivity, reanalyses, using confirmatory maximum resuIts-a circumstance resembling the familiar ex-
pression in computer work of" garbage in-garbage realization of numerous statistical assumptions,
out" (GIGO). such as normality, linearity, homoscedasticity, and
A number of psychometric and statistical con- level of measurement (e.g., ordinal or interval).
cerns also arises in planning factor-analytic studies Variations in the extent or manner in which these
that hopefully will have the potential to yield defini- several methodological features have been inte-
tive results. In the process of selecting variables, grated into the factor-analytic approach can lead to
attention obviously needs to be given to construct- rather different solutions that are open to conflicting
ing and/or selecting measures of as high reliability interpretations. (For additional information perti-
as possible so that significant amounts of systematic nent to issues in factor analysis as a research tech-
variance can be identified as being associated with nique, the reader is referred to Chapters 17 and 18 in
newly isolated factor dimensions. Pilot studies may Gorsuch, 1983.)
well be necessary to establish reasonably high reli-
abilities of the measures to be used and sufficient
variance to permit the emergence of correlation co- Multitrait-Multimethod Techniques
efficients that can be interpreted as reflecting genu- A useful procedure for operationally estimating
ine levels of association rather than inflated indices construct validity between two or more traits or
that have capitalized upon chance. In addition to the behavioral characteristics considered to be related
minimization of measurement error, sampling error to creativity is the multitrait-multimethod ap-
also needs to be controlled by taking groups both proach introduced by Campbell and Fiske (1959).
representative- of the popUlation of interest and ade- This correlational technique provides for the simul-
quate in size so that the factor-analytic results do not taneous testing for convergent and discriminant va-
reveal the capitalization upon random sampling er- lidities of intercorrelated variables through the anal-
ror. Moreover, care should be exercised not to in- ysis of two or more theoretically related behaviors
clude too many variables for the number of subjects (traits) assessed by two or more methods of mea-
in the sample. Although investigators differ in the surement. The basic idea of demonstrating con-
criteria that they employ, we have found that the vergent validity is that if two or more tests or mea-
ratio provided by the division of the total number of sures, each employing different methods, provide
subjects in the sample by the total number of vari- relatively similar estimates of a target behavior,
ables in the correlation matrix typically should be at then one can have greater confidence in the strength
least 15 to 20 if the measures are highly reliable (in of that inferred behavioral trait or characteristic and
excess of .80). can maintain that evidence of convergent validity
Several reservations need to be exercised in the (that the multiple methods converge on a single
interpretation of the outcomes of factor-analytic behavioral characteristic) has been achieved. The
studies in terms of the particular features of meth- concern of discriminant validity rests on the ability
odology employed: (a) the number of factors ex- of the measures to discriminate among different be-
tracted; (b) the type of estimate for the diagonal havioral traits. Thus, one can anticipate finding a
entries of the correlation matrix (unities, reliability relatively low degree of association between differ-
coefficients, or communalities) depending upon the ent behaviors measured by the same test.
factor-analytic model selected; (c) the method of To facilitate understanding of these principles,
rotation to be employed-orthogonal or oblique Figure 1, based upon fictitious data, is provided to
(often a function of the theoretical position taken); illustrate the basic organization of the multitrait-
(d) the selection of the particular statistical pro- multimethod matrix as it may apply to a construct-
cedure for effecting either an orthogonal or oblique validity problem in the area of creativity. In this
solution; (e) the choice of analyzing the intercor- figure, the matrix is composed of the intercorrela-
relations of individual items or of composites of tions among three creativity traits (originality, flex-
items (subtests) or of whole tests; (f) the setting of ibility, and fluency) which are measured by three
lower bounds in the magnitudes of factor loadings different creativity tests (Guilford Test, Torrance
for meaningful interpretation to offset the impact of Test, and Michael Test). To demonstrate construct
sampling and measurement error; (g) the optimal validity from application of this methodology, at
arrangement of the data presentation (as in trying to least two conditions must be addressed: (a) the first
demonstrate the presence of a hierarchical solution condition is a demonstration of convergent validity,
in harmony with the hypothesized configuration of meaning that a test should correlate highly with
constructs within a particular theory); and (h) the different tests which purport to measure the same
Method 1 Method 2 Method 3
(Guilford Test) (Torrance Test) (Michael Test)
Traits
Originality .92)
Method Flexibility
1 Fluency
Originality r#.__~:~===~~r~~-----·-ii61 (.94)
Method Flexibility ! .23 , ............... 52 ............:.1°1

,Mi'
I . . . . . . . . ... I
2 Fluency • 12----__ ::.:::.....
:L______________________
• 09
Originality ~ :9
~------------------
........< .. .091
...
! .......
~----------------------,
"-,J.! ----__ .39
........
.27 :
:
(.96)
: .22 ----__ .53----___ 14 !
I.... ...... I
: • 41 ------.,J.!i ----__ .31 !
_ _ _ _ ~~~~~~~~:~~-5J
Method Flexibility .67 (.90)
3
Fluency l~_~~ l.:..~__________·_~~~::=~~:~l7J .55 .58 ( .80)
Figure 1. A demonstration of convergent validity. The main diagonal of values in parentheses comprises the estimates of reliability.
The validity diagonal is represented by the underlined values. The heterotrait-monomethod comparisons are set off by the solid-line
triangles and each of the heterotrait-heteromethod comparisons is provided in the broken-line triangles.
theoretical construct, and (b) the second involves a associated with establishing construct validity re-
demonstration of discriminant validity-a condi- quire additional analyses provided by discriminant
tion in which a test is not statistically related to other validation techniques. Again, employing the multi-
measures from which it has been determined that trait-multimethod matrix, the first criterion for
the test should be conceptually and theoretically demonstrating discriminant validity which must be
different. In the problem of demonstrating con- addressed requires that the correlations between
vergent validity, the entries in the validity diagonal similar traits measured by different methods (the
represented by the underlined values in Figure 1 convergent validities) exceed the correlations be-
would be statistically significant and sufficiently tween different traits measured by different meth-
large to justify a validity investigation. Additional ods. (As noted previously, this comparison also
support for convergent validity is found if higher supports the demonstration of convergent validity.)
correlations exist for the validity diagonal values The second discriminant validity criterion consid-
than for those values associated with different traits ered is that the validity coefficients for each trait
measured by different measures (referred to as the exceed the correlations between that trait and differ-
heterotrait-heteromethod triangles which are ent traits measured by the same method. (This crite-
shown as broken-lined triangles). Thus, in order to rion is depicted by the heterotrait-monomethod
demonstrate convergent validity, the validity val- values designated by the solid-line triangles in Fig-
ues for a given trait, such as originality (e.g., .62, ure 1.) If high intercorrelations are obtained across
.59, and .61) should be statistically significant, different traits (e.g., between originality and fluen-
meaningful, and generally higher than the correla- cy) which are hypothesized to be conceptually and
tions obtained between originality and any other statistically unrelated employing the identical
two traits (e.g., flexibility and fluency) measured method or test, then a method effect is said to exist.
by different methods. Marsh and Hocevar (1983) have summarized the
According to Campbell and Fiske (1959), al- major shortcomings of the multitrait-multimethod
though convergent validation procedures empha- approach to include (a) the lack of objective criteria
sizing independence of methods provide key infor- for defining satisfactory results; (b) the discrepancy
mation for most validity problems, the demands between the conceptualization of the discriminant
validity criteria with particular reference to the con- creative behaviors. It is interesting to note in
cept of trait covariation-the correlation between Michael's review of research studies that conclu-
different traits that does not depend upon method; sions derived from intensive inquiry of life-history
(c) the risk of realizing rather markedly an assessed variables (Roe, 1951), based on in-depth interviews
method effect in one set of measures but not sub- concerning the personal characteristics of creative
stantially in another set because of differential lev- individuals, were similar to those identified from
els of correlation that arise as a function of system- use of standardized personality and ability mea-
atic variation in the corresponding magnitude of sures chosen to reflect creativity (Guilford, 1963;
reliabilities in one set or both sets of measures; and MacKinnon, 1962, 1965).
(d) the chance of obtaining inflated convergent va-
lidities also artificially increased by shared method Computerized Imaging of Cerebral Activity:
variance among the traits. With reference to the Creativity Constructs and Physiological Correlates.
second criticism (b), the existence of trait covaria- Over the past decade, a revolution has taken place
tion for two of Campbell and Fiske's (1959) criteria in the technological marketplace. Today, thanks to
can lead to conflicting interpretations. In one crite- advances in medical technology which employ
rion, the occurrence of covariation is viewed as computerized imaging methods, such as comput-
demonstrating a lack of discriminant validity, erized axial tomography (CAT)-sometimes re-
whereas the presence of co variation in the other ferred to as the CAT-scan-procedures are avail-
criterion (a criterion for patterns of correlations) is able for the mapping of cerebral activity which
interpreted as a demonstration of discriminant provide opportunities for studying the generality or
validity. specificity of brain functions associated with
human behavior. In a recent investigation on the
relationship between neurological correlates and
Use of Three Other Methodologies
creative thought, Hudspith (1985) studied subjects
Experimental Studies. Relatively few experi- who had been classified as either high-creative (n =
mental investigations have been undertaken to dem- 10) or low-creative (n = 10) based on their standing
onstrate the validity of constructs associated with on three measures of creativity-the Remote Asso-
creativity. The basic methodology involves the ciates Test (RAT) (Mednick & Mednick, 1962,
comparison of an experimental group that has been 1967) and two subtests from the Structure ofintel-
exposed to an activity hypothesized to be associated lect Learning Abilities Test (SOILAT) (Meeker &
with some aspect of creativity with a control group Meeker, 1975), both of which were designed to
that has not been. A prediction would be made that reflect operations of divergent thinking applied to
the experimental group would perform at a higher figural units and to semantic units. Each subject
level on a logically appropriate dependent variable, then received examinations of brainwave activity
the standing on which would be anticipated to be using the CAT apparatus during three experimental
elevated by participation in the planned interven- conditions-rest, word association, and spatial
tion. One such investigation reported by Parnes association. In each of the verbal and spatial condi-
(1964) indicated that students who had completed a tions, "mental" manipulation, rather than a pencil-
creative problem-solving course performed at a and-paper task, was required. For example,
higher average level on six out of eight creative instructions for the word-association condition re-
ability measures than did those students who had quired the subject, with closed eyes, to listen to a
not been exposed to the course. pair of words and then to think of a word which has
a common association with the pair. For the spatial
Longitudinal Studies. In a number of bio- condition, subjects were instructed to observe a
graphical and longitudinal studies of creative or em- three-dimensional model and then to imagine with
inent mathematicians and scientists summarized by their eyes closed the model folding open along a
Michael (1977), personality characteristics, demo- selected axis of the model.
graphics, and home-related variables were identi- Statistically significant outcomes were reported
fied. The primary goal of such longitudinal and between the high-creative and low-creative groups
archival investigations is not only to provide de- in terms of mean and differential measures of elec-
scriptions of salient characteristics of individuals troencephalograph (EEG) recordings provided by
classified as creative, but also to generate a pool of the CAT output. In general, subjects classified as
possible criterion variables from which to predict high-creative in relation to those identified as low-
creative tended to produce lower occipital (visual judgments of what cognitive processes and abili-
cortex) amplitudes as well as a combination of ties reflect creative behaviors?
lower occipital and higher prefrontal amplitudes in These questions suggest a number of potential
the three conditions of associational thinking inves- areas for research for clarifying the nature of cre-
tigated. This study, although preliminary in nature, ative behaviors and the context in which they occur.
points to the use of new methods for exploring cre- The reader is urged to study the paper by Fiske
ative activity and invites new hypotheses that will (1987) again and again for thoughtful stimulation
test neurological correlates of creativity constructs. and for its unlimited heuristic value.
Some Pertinent Questions in Construct

Creativity versus Intelligence
Validation
The distinction between creativity and general
A central area of concern does exist in the ac- intelligence has been a source of considerable con-
curacy of inferences regarding the manifestation of troversy in the research literature (e.g., Dellas &
creative behavior in relation to the context in which Gaier, 1970; Getzels & Jackson, 1962; Golann,
it occurs. Addressing this important issue, Fiske 1963; Guilford, 1967, 1968; Guilford & Hoepfner,
(1988) has introduced the termprotocoi, which per-
1971; Thorndike, 1963; Torrance, 1959). Although
tains to the conditions under which a study is con-
this concern about the distinction between creati-
ducted. His thoughtful paper has led to a number of
vity and general intelligence is primarily a concep-
questions that pertain to how method effects can
tual one beyond the scope of this chapter, mention
contribute to construct invalidity. In the context of
should be made of the fact that the realization of
the study of creativity, the authors have taken the
construct validity of the measures studied is essen-
liberty of rephrasing and, in some instances, para-
tial to the demonstration of the relative indepen-
phrasing a number of his concerns in the form of the
dence or dependence of these two concepts. Clear-
following sampling of questions (which is by no
ly, the appropriate application of psychometric
means exhaustive):
procedures and the fulfillment of the assumptions
1. Is the creative behavior serendipitous-does it
underlying them are necessary but not sufficient
happen by chance?
conditions for providing evidence regarding
2. Does the creative behavior occur under a struc-
whether creativity constructs are independent of
tured situation during testing or does it manifest
general intelligence or constitute a subset of abili-
itself in an unstructured situation? Stated another
ties that can be subsumed under the rubric of gener-
way, does the creative behavior take place under
al mental ability.
specified demand characteristics that are deter-
mined in part by restricted materials or delimiting
problem statements, or does the creative behavior
Closing Thoughts
arise in response to an open-ended question or
situation? The problems associated with establishing con-
3. Does the behavior occur following exposure to struct validity are pervasive regardless of the area of
e!xamples provided by models or in an apparently behavioral research. Special attention must be
spontaneous manner from unidentifiable internal given, however, to the study of such behavioral
processes? domains as creativity because of the apparent com-
4. What is the minimum amount of information plexity and potential interrelatedness of the con-
that can be given to a subject to evoke creative structs involved. Moreover, an improved under-
behavior? standing of what creativity is, on the assumption
5. Does the subject recognize his or her own that it can be defined and understood, may lead to
product as creative? the design of teaching strategies and to the planning
6. How useful, if not essential, is the variable of of educational environments so that maximal op-
time in formulating constructs about creative portunities exist for creative behaviors to emerge in
behaviors? a technological and scientifically oriented society.
7. What are the relative merits of change scores Such a society will require the best creative talents
over time or single scores at one point in time in the that are available not only in the sciences, but also
assessment of creative behaviors? in the arts and humanities for the enrichment of the
8. To what extent do measures rely on a priori quality of life.
Content Validity practical situation (process at the application level).

This same methodology could be employed to de-
In the familiar context of achievement testing, velop items for a test of creative behavior in which
content validity typically refers to the extent to several sets of items could represent different levels
which the behaviors associated with specific objec- of complexity (a hierarchy) of a problem-solving
tives of instruction or learning (both subject matter process. Additional information concerning both
and process objectives) are representative samples the preparation of grids to portray instructional ob-
of a relatively clearly defined broad domain or pop- jectives and the common techniques for writing
ulation of behaviors. Although an understanding of items to constitute the intended sampling emphasis
the construct or constructs underlying process ob- within the grid can be found in standard textbooks
jectives is essential to the realization of content va- in measurement and evaluation (Bloom, Hastings,
lidity of tasks involved in school learning, the im- & Madaus, 1971; Ebel, 1979; Gronlund, 1976;
portance of construct interpretation and hence of Mehrens & Lehmann, 1984; Sax, 1980; Thorndike
construct validity is probably relatively greater in & Hagen, 1977).
the assessment of creative endeavor than it is in the
evaluation of most cognitive tasks encountered in
Criterion-Related Validity
the classroom. Content validity would be relevant
in the assessment of creativity primarily to the de- Criterion-related validity of a test may be defined
gree to which a teacher has devised an instructional as the degree of accuracy with which the test is
unit specifically directed toward achieving certain predictive of the behavior of an individual or group
instructional objectives reflecting original, unique, of individuals in a clearly defined situation typically
or flexible types of behaviors. Such instructional amenable to measurement (customarily referred to
units, in the authors' experience, have been rela- as a criterion variable) (American Psychological
tively rare at all levels of the educational process Association, 1985; Thorndike & Hagen, 1977). For
unless specifically funded as a special project, per- example, scores on a scholastic aptitude test often
haps because they constitute a threat to the well- are employed to forecast standing on a criterion
established status quo of student conformity or to measure, such as grade-point average earned during
the expectations of the instructor, who regards in- the freshman year in college. Criterion-related va-
novative activities as either a nuisance to a well- lidity is customarily differentiated into concurrent
established routine or a challenge to his or her au- validity and predictive validity. In the instance of
thority or status. concurrent validity, the time lapse between the ad-
If a teacher is interested in bringing about the ministration of the antecedent, or predictor, mea-
manifestation of creative behaviors, the steps that sure, such as a scholastic aptitude test, and the at-
would be taken to realize content validity would tainment of criterion data is minimal, if not virtually
parallel those associated with the evaluation of the simultaneous. On the other hand, predictive valid-
nature and amounts of learning that have taken ity implies the administration of a test to be used for
place in almost any instructional unit. Essentially, a selection, placement, or classification or for man-
two-dimensional grid is prepared in which topical ifestation of a certain form of behavior (e.g., alco-
or subject matter objectives are listed as headings in holism, delinquency) often several weeks, months,
several adjacent columns (or rows) and the process or even years in advance of the collection of follow-
objectives associated with various levels of think- up information that is intended to reflect the ade-
ing or cognitive activity are cited as headings along quacy of performance or the adjustment of an indi-
a series of adjacent rows (or columns). The inter- vidual as represented by his or her standing on one
section of a category from one dimension with that or more criterion measures.
from a second dimension forms a cell for which one Typically, the degree of criterion-related validity
or more test items will be constructed to reflect the is indicated by a correlation coefficient that summa-
emphasis placed upon that particular content-pro- rizes the amount of relationship between scores on a
cess objective. For example, to find the cost to predictor variable with scores on the criterion mea-
cover a kitchen floor with linoleum at a certain sure. An alternative procedure that often precedes
number of dollars per square yard would reflect in the calculation of the correlation coefficient is the
the grid the crossing of the topic of the determina- preparation of a scatter diagram in which each dot
tion of areas of geometric figures (subject-matter or point represents the simultaneous standing of an
content) from the use of an appropriate formula individual on the predictor and criterion variables.
along with the introduction of a cost coefficient in a Such a scatter diagram also can be cast into the form
of an expectancy chart when the data have been ly, would be the occurrence of criterion contamina-
grouped into intervals on each variable. In an ex- tion in which the observer or evaluator of creative
pectancy chart, one can determine for those indi- behaviors uses consciously or unconsciously infor-
viduals falling within a given interval of scores on mation gained from previously administered tests
the predictor the probability of placing at or above as a basis of making a judgment of the adequacy of a
the lower limit of any specified interval on the crite- subject's performance in one or more criterion mea-
rion measure. Both scatter diagrams and expectan- sures (such as rating scales of creative behaviors). It
cy charts can provide the researcher with an indica- cannot be sufficiently emphasized that any judg-
tion of whether the trend of the data is linear or ment regarding the amount and or quality of
curvilinear and thus can afford evidence regarding creativity represented by criterion measures must
the appropriateness of applying alternative statis- be made independently of any knowledge of stand-
tical procedures for estimating the degree of rela- ing on predictor variables.
tionship between the two variables and for making A third area of concern, which is to be treated
predictions of placement on one variable from somewhat more extensively in the next section, is
knowledge of standing on the other. that of the degree of reliability of both the predictor
Several limiting or modifying circumstances and especially the criterion measure as a potentially
arise in conjunction with the meaningful interpreta- attenuating element of the magnitude of any valid-
tion of validity coefficients, particularly in the in- ity coefficient. In most circumstances, the reliabili-
stance of creativity measures. Probably the most ty of the criterion measure is likely to be lower than
important concern in the realization of concurrent that associated with the test. Essentially, if either
or predictive validity is that of identifying a truly the criterion variable or test variable is not stable or
satisfactory relevant criterion measure. In many inconsistent within itself, it cannot be expected to be
stances, creative behaviors are likely to be multidi- correlated highly or even moderately with any other
mensional, complex, and difficult to define in oper- measure. Painstaking care needs to be exercised in
ational terms. Almost inevitably, professional the construction of measures so that their reliability
judgment based on rational considerations and ex- can be maximized through the selection and reten-
tensive experience will be needed to determine the tion of the most discriminating items. Both sea-
appropriateness of creative activities in an educa- soned professional judgment and empirical evi-
tional, working, or job-related situation. Decisions dence derived from pilot studies involving
will need to be made concerning whether an observ- extensive use of item analyses may be required to
able product or an inferred process or a combination achieve an acceptably high level of reliability.
of these two components will be considered in the A fourth potentially limiting factor in the estima-
formation of criterion variables, because both prod- tion of a correlation coefficient is that of restriction
uct and process are inherent in the dimensionality of of range, a topic which Yamamoto (1965) consid-
most creative acts. ered at length. His paper was concerned with the
A comprehensive understanding of alternative influence of restriction of range as well as of lack of
theoretical orientations regarding the nature of test reliability on the amount of correlation between
creativity will help the investigator to select those measures of intelligence and creative thinking. As
constructs that are psychologically meaningful and, groups become more homogeneous and thus show
at the same time, reasonably amenable to measure- limited variability in scores on either the predictor
ment. One approach to ascertaining the dimension- or criterion measure, the potential for realizing a
ality of several criterion measures chosen to reflect high correlation coefficient is substantially re-
creative behaviors would be that of the previously duced. As many investigations in the study of
mentioned factor-analytic techniques that are used creativity are carried out in college or university
to identify the constructs of creativity. Specifically, settings in which the range of talent is quite re-
criterion measures would be factor analyzed along stricted because of selective admissions policies,
with numerous psychological tests, the factorial the obtained correlation coefficients will be consid-
composition of which has been reasonably well es- erably smaller than those found for members of a
tablished. To forecast creative behaviors in a job- popUlation without such restriction. Statistical cor-
oriented situation, one would select those tests that rections do exist for estimating what the correlation
duplicated the factors (empirical representations of between two variables would have been had there
constructs) identified in the criterion measures. been no restriction in range (Guilford & Fruchter,
A second circumstance that can influence the in- 1978; Gulliksen, 1950).
terpretation of a validity coefficient, often adverse- A fifth circumstance, which also is to be consid-
ered in the fifth section dealing with scoring, is that to the total variance in obtained scores-a total
of cultural differences in groups that may be stud- which is the sum of the variance in true scores and
ied. Validity coefficients may fail to register statis- that of the independent (uncorrelated) component
tical significance simply because either the predic- of error scores. The standard deviation of the error
tor variable or the criterion task intended to scores is referred to as the standard error of mea-
emphasize creative activity may be so foreign to the surement, which indicates the region within which
subject that he or she fails to understand the require- approximately 68% of obtained scores for an indi-
ments of either task relative to the expectations of vidual would fall above or below his or her true
significant others who may be from a quite different score (a true score being the mean of an infinite
cultural background. In short, an insufficient number of obtained scores on a test form for an
know ledge base on the part of the examinee or sub- individual with practice and fatigue effects con-
ject or a paucity of shared values between the eval- trolled). The higher the degree of reliability, the
uator and the subject may exist. Moreover, what less the standard error of measurement (Guilford,
may be judged to be creative in one culture may not 1954; Guilford & Fruchter, 1978; Nunnally, 1978;
be judged as creative in another. Thorndike & Hagen, 1977).
A sixth concern is a highly practical one of cost. Three broad approaches exist for the estimation of
In the judgment of management personnel either in test reliability: (a) the test-retest approach, which is
a school or in a business or service organization, the the correlation between scores on the same form of a
enormous expense in research and development test administered at different points in time-a
that may be required to devise appropriate predictor coefficient yielding an indication of stability of test
and criterion measures may not be warranted. The scores, although practice effects and correlated
additional benefits in productivity associated with errors may occur; (b) the equivalent or alternate
creativity (particularly in a routine or repetitive type (often called parallel) forms, the scores from each of
of work situation) may not be viewed to be worth which are correlated to yield an estimate of stability
the time and cost in dollars necessary to develop for different samplings of items-a procedure which
predictors and to monitor the acquisition of criteri- controls in part for memory of items and correlated
on data supposedly indicative of creative behaviors. error; and (c) the internal-consistency approach,
It should be mentioned that relative to other mea- which is employed to estimate the reliability of a
sures of cognitive ability or even measures of affec- single test form administered once with the assump-
tive characteristics, those of creativity have been tion being made that the items are relatively homoge-
used quite infrequently in the prediction of standing neous in content and process characteristics-a
on a criterion variable. Cattell and Butcher (1968) method essentially revealing the ratio of the
have devoted a number of chapters to the prediction covariance of items to the total variance of the test.
of achievement and creativity but have given rela- Coefficients of internal-consistency reliability are
tively little emphasis to the prediction of creative found in a number of ways, including use of the
behaviors from tests that in themselves were Kuder-Richardson formula 20 or 21 (Kuder & Rich-
intended to be indicative of creative activities. ardson, 1937), Cronbach's (1951) alpha, the split-
What work they did in the choice of predictors was half technique in which scores on odd-numbered
anchored toward personality variables rather than items are correlated with those on even-numbered
toward cognitive measures. The need for research items and then adjusted by the Spearman-Brown
to identify predictors for forecasting creative be- formula (Guilford & Fruchter, 1978) to afford an
haviors is a crucial one which may be expected to be estimate of the reliability of the total test (twice as
quite difficult in view of the problems associated long as either half), and Hoyt's (1941) analysis of
with defining operationally criterion measures invariance technique, especially useful in the deter-
dicative of relevant creative behaviors. mination of interrater reliability. Additional infor-
mation regarding these approaches may be found in a
number of texts, including those by Guilford and
Concerns Relating to Reliability Fruchter (1978), Gulliksen (1950), Nunnally
(1978), Thorndike (1982), and Thorndike and
As indicated earlier, reliability refers to the sta- Hagen (1977).
bility or consistency of scores on a given measure. In the estimation of reliability for tests of diver-
In classical test theory, reliability is defined as the gent thinking that tend to be open-ended and often
ratio of the variance of true scores (free of any error) complex, if not quite heterogeneous in the con-
structs that they represent, probably the most defen- such difficulties is to evaluate a series of behaviors
sible procedure would be to use alternate forms. in one category or the characteristics of one aspect
Many researchers who have only one form of a of a product at a time across all individuals or spec-
fluency measure frequently employ a test-retest ap- imens (in a manner analogous to scoring one essay
proach-a procedure far more defensible than that question at a time across all examinees). It also is
of internal consistency. In addition to the possibility important that the observer or observers have re-
that the characteristics of item responses may be ceived comparable amounts of training so that they
quite different (heterogeneous), the often speeded are prepared to yield judgments based on equivalent
·nature of measures of divergent thinking (especially levels of competency, although differences in val-
fluency) is likely to lead to a spuriously high esti- ues still can be expected to influence evaluations. It
mate if an internal-consistency approach is em- also has been found that the pooled judgments of
ployed-a circumstance that is well known to in- several evaluators (typically three to five) will be
vestigators who have used tests of the perceptual considerably more reliable than those of one, with
speed factor. In such tests, there is a tendency for the estimates of reliability increasing very much in
examinees to work at fairly constant rates over sets accordance with the well-known increments associ-
of items and to earn comparable scores on these sets ated with the Spearman-Brown formula (Guilford,
because of the relatively low level of difficulty of 1954; Guilford & Fruchter, 1978; Nunnally, 1978).
most items. A host of miscellaneous factors may be cited that
In the instance of those tests of convergent think- can influence the magnitudes of reliability coeffi-
ing that often resemble fairly traditional multiple- cients. As in the instance of validity, a marked re-
choice measures in problem-solving tasks, the is- striction in range of scores will be associated with
sues of estimating reliability are highly comparable attenuated estimates of reliability of either a criteri-
to those found for most alternate choice examina- on measure or its predictor. Likewise, if the level of
tions in which there is one correct answer. If the test difficulty of the vocabulary in the directions for
portrays one construct or if each of the items in it is taking a test or in the items to which subjects are
represented by a comparable factor structure (e.g., responding is too great, reliability can readily ap-
the presence of the same constructs to about the proximate zero, as the scores may be indicative of
same degree from item to item), internal-consisten- chance responses. Unpredictable influences upon
cy approaches may be quite appropriate. the size of reliability coefficients may occur if great
For such other kinds of measures as a self-report differences exist in the motivational levels of indi-
inventories, which may have several subscales in- viduals who are taking measures of creativity or
dicative of personality and interest traits, internal- who are being observed for creative acts in a school
consistency approaches to the estimation of reliabil- or occupational setting. The maturity of a subject,
ity are commonly used. One needs to be alert to the either in terms of chronological age or level of so-
possible presence of response sets, such as social cial awareness, also can lead to the realization of
desirability or acquiescence, that may lead to un- differential magnitudes of reliability estimates, not
wanted systematic variance that contributes to an only as a result of the degree of attentiveness and/or
inflated estimate of instrument reliability. commitment to the task at hand, but also as a func-
With respect to measures derived from observation of the extent of cognitive development, es-
tions of evaluators, as in the case of rating scales of pecially in the instance of children or young adoles-
creative behaviors or check lists for assessing cre- cents. Substantial heterogeneity in the cultural
ative merits of products as, for example, in the arts, backgrounds of members of a group can lead to the
internal-consistency approaches once again are introduction of erratic and unpredictable amounts
commonly employed to estimate reliability-es- of error variance in creative ability test scores, self-
pecially the analysis of variance technique initially report data, and observational judgments. Associ-
suggested in generalized form by Hoyt (1941) and ated with this heterogeneity, differences in role ex-
applied by Ebel (1951) specifically to ratings. In pectations on the part of members of different cul-
any type of evaluation involving observation, but tural groups or of males and females within a given
especially in the instance of ratings, exaggerated cultural group may contribute to unanticipated
estimates of reliability can occur because of the modality characteristics in distributions of scores
introduction of systemic variance attributable to re- and to differential patterns of response. The result-
sponse errors (sets), such as the halo effect, lenien- ing scores need to be analyzed separately for each
cy, severity, or central tendency. One solution to identifiable group rather than for a total sample. In
other words, in the interpretation of test scores, response-the less frequent or more rare, the great-
attention should be directed to the existence of dif- er the amount of implied creativity. Once again, a
ferentiallevels of reliability for subgroups that may judgment still must be made concerning whether an
be substantially different in cultural background, infrequent response is a bizarre or capricious one or
age, and previous opportunity for training. one actually indicative of a creative thought. In a
The need for additional research in conjunction subsequent part of this section, consideration is
with reliability concerns in the assessment of given to an overriding concern in the scoring of tests
creativity would not appear to be anchored to the of divergent thinking-namely, the confounding of
development of new psychometric or statistical ideational fluency with scores on each of several
methodology, but to the improved application of dimensions of creative endeavor.
existing procedures. Specifically, greater care In a few instances, objective tests of the multiple-
should be exercised in the construction of measures choice or matching format are employed in which
and in the monitoring of their use-particularly in the examinee is required to select the most creative
the instance of rating scales and related observa- response cited in relation to an activity described in
tional activities. A need continues to exist in any the stem of the item. As mentioned earlier, such
test-taking or observational situation to motivate tests may be particularly appropriate for problem-
both the subject and the observer (evaluator) to take solving tasks in mathematics and the sciences for
seriously the task and to do one's best. which only one solution or process involved in ar-
riving at the desired answer can be judged as being
both relevant and creative in terms of the collective
Concerns Relating to Scoring of wisdom of the test makers who constructed the key.
As has been stated, the psychometric issues involv-
Creativity Measures ing scoring such a test are basically analogous to
those found for almost any multiple-choice exam-
General Considerations
ination in any academic discipline. As indicated
In the assessment of creative endeavor, tests of previously, such tests are likely to reflect the prod-
divergent thinking are the ones most often used in uct that has been named by Guilford (Guilford,
educational settings and in educational and psycho- 1967; Guilford & Hoepfner, 1971; Michael, 1977)
logical research. As the responses are open-ended, as transformations, in which the examinee often
considerable subjectivity often occurs in evaluating has to find or devise a new use for a familiar object
just how creative, original, or ingenious the an- in a foreign context. Frequently, in higher mathe-
swers posed to problem situations may be. In other matics, it is assumed that the solution of a difficult
circumstances, a judgment is formulated concern- mathematics problem, in itself, requires a highly
ing how creative an observed product, as in visual creative act to generate the unique solution-a form
art, or behavior in solving a complex social problem of thinking often termed as convergent.
is thought to be. The scoring of responses is often Perhaps even more interesting than the unique
differentiated along a continuum to reflect their solution is the process by which the examinee ar-
being indicative of higher or lower levels of rives at the answer. There may be many alternative
creativity. approaches to reaching a unique solution with the
One of the difficulties in formulating judgments element of creativity being on just how ingenious
on the part of two or more evaluators is that of the process is in obtaining the answer. Once again,
obtaining a reasonable degree of agreement in the considerable subjective judgment may be required.
scoring procedure-a circumstance similar to that Moreover, those individuals who have worked ex-
found in scoring essays in creative writing. In other tensively in the psychology of creativity may be the
words, judges or raters often have great difficulty in least capable in making a judgment concerning just
reaching a consensus regarding what is a truly cre- how ingenious a process employed in reaching the
ative or original response. Frequently, indices of solution really is. Specialists in the field of creative
interrater reliability, such as intrac1ass correlation endeavor frequently are called upon to render an
coefficients, are employed to give some indication evaluation in much the same way that distinguished
of the level of agreement. scholars on editorial boards decide on the creative
Another method occasionally used to estimate merits of manuscripts submitted for publication.
the degree of manifested creativity is largely a nor- Although checklists, outlines of desired solutions
mative one based on the statistical frequency of a or processes for reaching them, rating scales, and
other psychometric devices may facilitate the real- served that, in the nine studies in which it was feasi-
ization of improved degrees of objectivity, the ble to apply the multitrait-multimethod standards
overall process of scoring or judging is likely to proposed by Campbell and Fiske (1959), the crite-
remain a highly subjective one open to the idio- ria for discriminant validity could not be met. In
syncracies and preferences of the evaluator, who addition, factor-analytic studies of intercorrelations
too often may not be too creative himself or herself. of fluency and originality scores on two or more
In a not totally unrelated matter, creativity in the tests have yielded dimensions unique to each of the
arts poses many unique problems of assessment. In tests of originality rather than common factors of
the performing arts, such as acting, singing, con- fluency and originality cutting across the originality
ducting, and playing a musical instrument, the on- measures (Kazelskis, 1972; Plass, Michael, &
going process, which may result in a final recorded Michael, 1974; Yamamoto & Frengel, 1966).
product, is what needs to be evaluated. Fortunately, In an attempt to address these difficulties,
the use of videotapes with appropriate sound sup- Hocevar (1979) found for a sample of 60 college
plementation may provide a permanent record students the intercorrelations of fluency and origi-
which numerous evaluators may judge on repeated nality scores of three of Guilford' s tests of divergent
occasions to arrive at some kind of consensus. Ac- thinking-Alternate Uses (Christensen, Guilford,
companying this approach might be rating scales on Merrifield, & Wilson, 1960), Plot Titles (Berger &
which observations can be recorded as well as com- Guilford, 1969), and Consequences (Christensen,
parative samples of parallel acts that can be em- Merrifield, & Guilford, 1958). It was hypothesized
ployed as relevant standards (criterion measures). that when variance associated with fluency scores
In any event, subjectivity in scoring will remain a would be statistically controlled (partialed out of
central issue as will the degree of competence of the originality scores), the internal-consistency esti-
evaluators. mates of reliability would show marked decrements
Perhaps somewhat easier to evaluate are final and also that the intercorrelations of originality
products, such as paintings or sculptured works, in scores (based on use of subjective judgment and on
which process is no longer the major consideration. application of an objective criterion representing
In the instance of products, other products of more rarity or statistical infrequency of response) would
or less agreed upon creative merit can be used as drop substantially if not approximate zero. As con-
standards against which to evaluate the newly com- siderable support was found for these two hypoth-
pleted works. In a supplementary fashion, rating eses with respect to either a subjective or an objec-
scales can be employed along with critical reviews tive determination of scores, it was concluded that
to arrive at some consensus. History has revealed both the reliability and convergent validity of origi-
only too often that what critics of one generation nality scores largely depend on individual dif-
have judged as a mediocre work turns out in a later ferences in ideational fluency rather than on
generation as a classic that may endure for centuries originality.
as an example of a truly creative contribution to In a related investigation, Hocevar and Michael
world culture. (1979) carried out two multitrait-multimethod
studies to examine the effects of use of two scoring
formulas on the discriminant validity of tests of
Fluency as a Confounding Factor divergent thinking. The first study involved 39 ele-
mentary school pupils in two fifth-grade classes
Fluency Scores as a Contaminating Influence
whose teachers administered the Torrance Tests of
An area of concern in the scoring of divergent Creative Thinking, Verbal Form A (Torrance,
thinking tests has been the potential confounding of 1974). Scoring for fluency, flexibility, and origi-
scores on tests of originality and flexibility with the nality made use of two procedures: (a) a summing
factor of ideational fluency that may have inflated formula, in which differential weights correspond-
spuriously both reliability estimates and correlation ing to responses of examinees were simply added
coefficients indicative of convergent validity (Clark according to criteria in the manual, and (b)
& Mirels, 1970; Hocevar, 1979). For example, in a percentage scores, requiring the flexibility and
review of 18 research studies in which 89 fluen- originality score for each item to be divided by the
cy / originality coefficients were cited, Hocevar total number of responses (fluency score). The sec-
(1979) noted an average coefficient of .69 with 82% ond study in whieh 60 college juniors and seniors
of the coefficients exceeding .50. Moreover, he ob- participated employed the three tests of divergent
thinking previously cited (Alternate Uses, Plot Ti- Suggested Ways for Controlling for Fluency
tles, and Consequences), which were scored for
To ?vercome mistaken conclusions regarding the
fluency ~nd originality. Use of the percentage score
eff~ctIveness of training programs in creativity, in
~es~lted In enhanced discriminant validity of sub-
WhIC? ~he ~cores on measures (dependent variables)
Jectl~el~ determined originality scores, although
of ongInahty and flexibility of those in experimen-
apphcatlOn of an objective scoring procedure based
tal groups may have been inflated by ideational
on . th~ .statistical .rarity was associated with alpha
fl~e~c~ in relation to those in control groups, and to
rehabIhty coefficIents approximating zero.
~llmmIZe ~h~ occurrence of spuriously high correla-
tIon coeffICIents among creativity measures simi-
larly confounded by ideational fluency, Hocevar
Fluency as a Possible Confounding Factor (.1979) has made a number of suggestions. In addi-
tlon to the use of partial correlation techniques, one
The investigation by Clark and Mirels (1970) on
could score each response to a flexibility or an origi-
fluency as a pervasive factor in the assessment of
nality item and calculate an average for all re-
creativity also was concerned with the estimation of
sponses rather than a total raw score. In mentioning
the extent to which fluency scores might account for
Clark and Mirels' (1970) approach to scoring an
the relatively low correlation sometimes reported
equal number of responses for each examinee,
?etw~en measures of creativity and those of general
H~c~var (1.979) has indicated the possibility of re-
Intelhgence. Using 93 art school students ranging in
qumng a. gIven number of responses to every ques-
age from 9 to 15 years, Clark and Mirels found the
tIon. He Illustrated this approach by suggesting that
intercorrelations of five different scores intended to
ex~mine~s be directed to cite the three most original
reflect creativity from a slightly modified version
or Ingemous uses that they can think of for a spec-
f~om To~a~ce's test concerned with figure comple-
ified object, such as a shoe. Fluency would be con-
tIon (ThInking Creatively with Pictures) (Torrance,
trolled in that each subject would present an equal
1966) and scores on the Henmon-Nelson Tests of
number of responses. One difficulty with this ap-
Me~t.al Ability (Houghton Mifflin, 1931-1974). In
proach, however, would be that the reliability of the
addItIon to appearing in raw score form, the re-
measures could be quite low. Such a statement
sponses on the creativity measures were corrected
would also hold for the Clark and Mirels (1970)
or adjusted for fluency. Whereas the uncorrected
suggestion of scoring an equal number of re-
creativity scores exhibited a mean intercorrelation
sponses. Both in this method and in the Hocevar
o~ .45, the corrected ones displayed coefficients
suggestion (Hocevar, 1979), the more fluent re-
wIth a mean of .08. The correlational data failed to
spondent could be penalized by being denied the
pro~ide definitive information regarding the possi-
opportunity to give a comparatively large number
ble Influence of fluency upon the correlation be-
of relevant responses indicative of originality or
tween scores on creativity measures and those on a
flexibility. To provide for a way of controlling for
test of general intelligence. For uncorrected
flu~ncy without sacrificing originality or flexibility
creativity scores, the mean correlation with the in-
estlmates, an alternative compromise approach
telligence measure was .09; for corrected scores,
may be useful. The respondent would be directed to
.l3. It could well be that the figural nature of the
gener~te as many possible uses for an object, such
material in the Torrance test simply portrayed con-
as a hght bulb, and then be asked to choose a pre-
structs quite different from those found in a measure
sel~cted number of the generated items (e. g., three)
of general scholastic aptitude within which both
WhICh the respondent considers to be the most
semantic and symbolic content would be somewhat
ingenious.
predominant. As Hocevar (1979) has pointed out,
the tendency of several investigators to use a strat-
egy to reveal that tests of creativity are highly inter-
Needed Research
co~elated and, at the same time, factorially distin-
gUIshable may be misleading because ideational The various suggested ways for controlling for
fluency may be the component that contributes sub- fluency in conjunction with establishing the con-
stantially to the high intercorrelations of creativity vergent and discriminant validity of creativity tests
measures and to a factorial structure suggesting that should be the subject of numerous empirical studies
these measures are distinct from intelligence tests. with a variety of individuals differing in test-taking
experience, maturity level, and intelligence. Ex- not creativity, would be correct relative to that per-
perimental studies incorporating as treatment vari- son's level of maturity. It would seem especially
ables differing instructions to examinees regarding important to observe the process by which this child
how they should respond to a fluency test and alter- arrived at the answer, as this process may provide
native modes of scoring would be expected to pro- the evaluator with valuable information concerning
vide some evidence that would afford a partial reso- whether an ingenious approach was employed in
lution of several of the concerns raised in the arriving at the solution or whether possibly the child
previous paragraphs. had been given the answer by another or had been
advised as to where to find the answer.
In the presentation of any set of norms, care must
be taken to indicate clearly the population for which
Concerns Relating to the Use of Norms the norms are applicable according to age, cultural
background, socioeconomic status, prior educa-
The development of usable norms for the assess- tional attainments, and existence of any special cir-
ment of creativity constitutes a difficult task, as cumstances, such as the participation by the nor-
many individuals perceive creativity to be a process mative group in a gifted program or in a summer
often resulting in a unique product with qualitative workshop directed toward the enhancement of
characteristics that may defy quantification. In the creativity. Norms also should be provided for each
arts, in particular, a given work may be so original subtest in a given battery of measures of creativity so
in its characteristics that there is no norm against that a profile of strengths and weaknesses in various
which it can be immediately compared. At a philo- components of creativity can be identified. One also
sophicallevel, perhaps the issue is one of the idi- needs to pay attention to scaling problems associated
ographic (individual) versus nomothetic (nor- with ceiling and floor effects, as one or more mea-
mative) characteristics of creativity-that is, the sures may be either too easy or too difficult relative to
uniqueness or universality of creative acts or prod- the general maturity level or experiential level of the
ucts. Certainly, when an artist is given a commis- group being studied.
sion to create a new work-whether it be a sym- Research efforts may be directed toward the eval-
phony, a sculpture, or a painting-one cannot uation of alternative methods of presenting norms. In
declare in advance what the product or the accom- particular, attention could be given to how norms
panying process will be. Reliance on evaluation of may be affected by the sequence in which particular
the merits of this work will depend upon the judg- tests within a battery are administered, by the level of
ment of critics or peers in a similar area of endeav- comprehensiveness of instructions for test taking
or. Even these individuals may find the product so including practice exercises, by the shortening or
foreign to their own background that they cannot lengthening of time limits especially for groups of
make a meaningful evaluation, especially if this culturally disadvantaged, and by the personality
creative work foreshadows a new school of thought characteristics of those administering the tests.
or the emergence of a new form of art (as in the
introduction of modem electronic technology in the
composition of music). Test Administration Issues
In the somewhat more common arena of school
learning in which expectations are at least broadly Issues in test administration are closely related to
defined for the use of educational or psychological several ofthose cited in the previous section dealing
tests specifically devised to give some indication of with norms. If one is interested in a formal assess-
creativity as in measures of flexibility, fluency, or ment of creativity so that comparative data can be
originality, it would appear reasonable to establish realized, it is essential that a number of factors be
preliminary normative data. Care needs to be exer- reasonably well controlled: (a) the time allowed for
cised in interpreting the performance of the student the examinee to respond to questions or time allo-
in relation to what would be anticipated for his or cated to observers to make their judgments; (b) the
her age group. If in a given problem-solving task in difficulty level of words and the mode of presenta-
science or mathematics, an 8-year-old child gives a tion of instructions, including comparable amounts
response typical of an individual 12 to 14 years of of illustrative or practice material for all examinees
age, it would appear that a judgment of precosity, if (such as warm-up exercises); and (c) distracting in-
fluences, such as noise, inappropriate lighting, leg- tional attainments, and existence of any special cir-
ibility of test format, and disturbing climatic factors cumstances, such as the participation by the nor-
(high humidity or extreme temperatures). More- mative group in a gifted program or in a summer
over, through pilot studies, a determination can be workshop directed toward the enhancement of
made of optimal time allocations and of the most creativity. Norms also should be provided for each
effective sequence of administration of tests if more subtest in a given battery of measures of creativity
than one is being given. so that a profile of strengths and weaknesses in
In conjunction with the preceding section regard- various components of creativity can be identifIed.
ing norms, several suggestions were made for re- One also needs to pay attention to scaling problems
search that would be applicable to conditions of test associated with ceiling and floor effects, as one or
administration. In particular, controlled research is more measures may be either too easy or too diffI-
needed to determine the effects of several of the cult relative to the general maturity level or experi-
factors cited in the fIrst paragraph of this section. In ential level of the group being studied.
addition, consideration should be given to identify- Research efforts may be directed toward the eval-
ing situations in which there is no time limit for a uation of alternative methods of presenting norms.
creative task to be completed, as in the instance of a In particular, attention could be given to how norms
mathematician or inventor who may be involved in may be affected by the sequence in which particular
the solution of a highly complex problem requiring tests within a battery are administered, by the level
considerable library research as well as time for of comprehensiveness of instructions for test taking
reflection and meditation. In such situations, a few including practice exercises, by the shortening or
individuals may have what is sometimes referred to lengthening of time limits especially for groups of
as a flash of brilliant insight that is diffIcult to ex- culturally disadvantaged, and by the personality
plain other than its being essentially intuitive; oth- characteristics of those administering the tests.
ers may reach the solution incrementally in a very It is probably evident from the text that shortcom-
systematic fashion that requires a great deal of trial ings in the application of psychometric procedures
and error and gradual elimination of alternative hy- to the study of creativity have existed. It may well
potheses or approaches associated with them. be that individuals who have contributed greatly to
the conceptualization of creativity have lacked the
requisite training in psychometrics and meth-
Concluding Statement odology to design well-controlled studies and to
apply appropriate statistical procedures. The au-
It is apparent that many issues enter into the accu- thors' impression is that the great surge of energy
rate assessment of creative behavior. As indicated devoted by researchers to the study of creativity
earlier, the establishment of the validity of mea- peaked in the mid-1960s. During the past 20 years,
sures intended to reflect creativity is the most cru- relatively few new ideas have been introduced. Per-
cial concern-a concern that rests on a sound con- haps this period from about 1965 to 1988 has been
ceptualization of the constructs of interest and on one of consolidation and reflection. What is very
the testing of hypotheses derived from a theory to much needed at this time is a renewal of interest on
which the constructs are central. It has been pointed the part of theoretically oriented psychologists who
out that sound principles of test construction that will combine their efforts with those of psycho-
will yield highly reliable measures is also necessary metricians, who are far better trained today than
if validity is to be achieved. The attainment of a they were two decades ago, in formulating new
high degree of validity and reliability will permit conceptual frameworks and in applying sophisti-
tests and scales intended to provide evidence of the cated multivariate procedures to extract the max-
manifestation of creativity to be useful in circum- imum information concerning the tenability of mul-
stances concerned with prediction, placement, di- tifold hypotheses that have been derived from
agnosis, prescription, and determination of compe- conceptually evolving frameworks regarding
tence levels in activities requiring creative beha- creativity.
viors.
In the presentation of any set of norms, care must References
be taken to indicate clearly the population for which American Psychological Association. (1974). Standards for ed-
the norms are applicable according to age, cultural ucational and psychological tests and manuals. Washington,
background, socioeconomic status, prior educa- DC: Author.
American Psychological Association. (1985). Standards for ed- Scientific creativity: Its recognition and development (pp.
ucational and psychological testing. Washington, DC: 101-118). New York: Wiley.
Author. Guilford, J. P. (1967). The nature of human intelligence. New
Anastasi, A. (1982). Psychological testing (5th ed.). New York: York: McGraw-Hill.
Macmillan. Guilford, J. P. (1968). Intelligence, creativity, and their educa-
Bentler, P. M. (1980). Multivariate analysis with latent vari- tional implications. San Diego, CA: Robert R. Knapp.
ables: Causal modeling. Annual Review of Psychology, 31, Guilford, J. P. (1970). Creativity: Retrospect and prospect.
419-456. Journal of Creative Behavior, 4, 149-168.
Berger, R. M., & Guilford, J. P. (1969). Plot titles. Beverly Guilford, J. P. (1971). Some misconceptions regarding mea-
Hills, CA: Sheridan Psychological Services. surement of creative talents. Journal of Creative Behavior, 5,
Bloom, B. S., Hastings, J. T., & Madaus, G. F. (1971). 77-87.
Handbook on formative and summative evaluation of student Guilford, J. P., & Fruchter, B. (1978). Fundamental statistics in
learning. New York: McGraw-Hill. psychology and education (6th ed.). New York: McGraw-
Campbell, D. T., & Fiske, D. W. (1959). Convergent and dis- Hill.
criminant validation by the multitrait-multimethod matrix. Guilford, J. P., & Hoepfner, R. (1971). The analysis ofintel-
Psychological Bulletin, 56, 81-105. ligence. New York: McGraw-Hill.
Campbell, J. P. (1976). Psychometric theory. In M. D. Dunnette Guilford, J. P., & Tenopyr, M. L. (1968). Implications of the
(Ed.), Handbook of industrial and organizational psychology structure-of-intellect model for high school and college stu-
(pp. 185-222). Chicago: Rand McNally. dents. In W. B. Michael (Ed.), Teachingfor creative endeav-
Christensen, P. R., Merrifield, P. R., & Guilford, J. P. (1958). or (pp. 25-45). Bloomington, IN: Indiana University Press.
Consequences. Beverly Hills, CA: Sheridan Psychological Gulliksen, H. (1950). Theory of mental tests. New York: Wiley.
Services. Harman, H. H. (1976). Modern factor analysis (3rd ed.). Chi-
Christensen, P. R., Guilford, J. P., Merrifield, P. R., & Wilson, cago: University of Chicago Press.
R. C. (1960). Alternative uses. Beverly Hills, CA: Sheridan Harris, M. L., & Harris, C. W. (1971). A factor analytic in-
Psychological Services. terpretation strategy. Educational and Psychological Mea-
Cattell, R. B., & Butcher, H. T. (1968). The prediction of surement, 31, 589-606.
achievement and creativity. Indianapolis: Bobbs-Merrill. Hocevar, D. (1979). Ideational fluency as a confounding factor
Clark, P., & Mirels, H. (1970). Fluency as a pervasive element in the measurement of originality. Journal of Educational
in the measurement of creativity. Journal ofEducational M ea- Psychology, 71,191-196.
surement, 7, 83-86. Hocevar, D. (1981). Measurement of creativity: Review and
Crocker, L., & Algina, J. (1985). Introduction to classical and critique. Journal of Personality Assessment, 45, 450-464.
modern test theory. New York: Holt, Rinehart & Winston. Hocevar, D., & Michael, W. B. (1979). The effects of scoring
Cronbach, L. J. (1951). Coefficient alpha and the internal struc- formulas on the discriminant validity of tests of divergent
ture of tests. Psychometrika, 16, 297-334. thinking. Educational and Psychological Measurement, 39,
Cronbach, L. J. (1971). Test validation. In R. L. Thorndike 917-921.
(Ed.), Educational measurement (2nd ed.) (pp. 443-507). Horn, J. L., & Knapp, J. R. (1973). On the subjective character
Washington, DC: American Council on Education. of the empirical base of Guilford's structure-of-intellect
Cronbach, L. J. (1980). Validity on parole: How can we go model. Psychological Bulletin, 80, 33-43.
straight? In W. B. Schrader (Ed.), Measuring achievement: Horn, J. L., & Knapp, J. R. (1974). Thirty wrongs do not make a
Progress over the decade (pp. 99-108). San Francisco: right: Reply to Guilford. Psychological Bulletin, 81, 502-
Jossey-Bass. 504.
Cronbach, L. J., & Meehl, P. E. (1955). Construct validity. Houghton Mifflin. (1931-1974). The Henmon-Nelson Tests of
Psychological Bulletin, 52, 281-302. Mental Ability. Boston: Author.
Dellas, M., & Gaier, E. L. (1970). Identification of creativity: Hoyt, C. J. (1941). Test reliability estimated by analysis of
The individual. Psychological Bulletin, 73, 55-73. variance. Psychometrika, 6, 153-160.
Ebel, R. L. (1951). Estimation of the reliability of ratings. Hudspith, S. (1985). The neurological correlates of creative
Psychometrika, 16, 407-424. thought: A comparison of the EEG activity of high and low
Ebel, R. L. (1979). Essentials of educational measurement (3rd creative subjects with an ergonomic presentation of the results
ed.). Englewood Cliffs, NJ: Prentice-Hall. for the lay person. Unpublished doctoral dissertation, Univer-
Fiske, D. W. (1987). Construct invalidity comes from method sity of Southern California, Los Angeles.
effects. Educational and Psychological Measurement, 47(2), JOreskog, K. G. (1969). A general approach to confirmatory
285-307. maximum likelihood factor analysis. Psychometrika, 34,
Getzels, J. W., & Jackson, P. W. (1962). Creativity and intel- 183-202.
ligence. New York: Wiley. JOreskog, K. G., & Sorbom, D. (1981). USREL V: Analysis of
Golann, S. E. (1963). Psychological study of creativity. linear structural relationships by method of maximum like-
Psychological Bulletin, 60, 548-565. lihood. Chicago: National Educational Resources. (Distribu-
Gorsuch, R. L. (1983). Factor analysis (2nd ed.). Hillsdale, NJ: tion by International Educational Services).
Erlbaum. Kazelskis, R. (1972). The convergent, divergent, and factorial
Gronlund, N. (1976). Measurement and evaluation in teaching validity of the Torrance Figural Test of Creativity. Southern
(3rd ed.). New York: Macmillan. Journal of Educational Research, 6, 123-129.
Guilford, J. P. (1954). Psychometric methods (2nd ed.). New Khattab, A-M., & Michael, W. B. (1986). The construct valid-
York: McGraw-Hill. ity of higher order structure-of-intellect factors reflecting se-
Guilford, J. P. (1963). Intellectual resources and their values as mantic and symbolic content abilities. Educational and Psy-
seen by scientists. In C. W. Taylor & F. Barron (Eds.), chological Measurement, 46, 1029-1035.
Khattab, A-M., and Michael, W. B., & Hocevar, D. (1982). torial validity of the Torrance Tests of Creative Thinking for a
The construct validity of higher order structure-of-intellect sample of III sixth-grade children. Educational and Psycho-
abilities in a battery of tests emphasizing the product of translogical Measurement, 34, 413-414.
formations: A confirmatory maximum likelihood factor analy- Roe, A. (1951). A psychological study of physical scientists.
sis. Educational and Psychological Measurement, 42, 1089- Genetic Psychology Monographs, 43(2), 121-235.
1105. Sax, G. (1980). Principles of educational and psychological
Kuder, G. F., & Richardson, M. W. (1937). The theory of the measurement and evaluation (2nd ed.). Belmont, CA:
estimation of test reliability. Psychometrika, 2, 151-160. Wadsworth.
Mace, D. E., Michael, W. B., & Hocevar, D. (1985). Validity Tenopyr, M. L. (1977). Content -construct confusion. Personnel
of higher-order ability constructs in structure-of-intellect tests Psychology, 30, 47-54.
all involving semantic content and operations of cognition or Thorndike, R. L. (1963). Some methodological issues in the
evaluation: A confirmatory maximum likelihood factor analy- study of creativity. In E. F. Gardner (Ed.), Proceedings of the
sis. Educational and Psychological Measurement, 45, 353- 1962 Invitational Conference on Testing Problems (pp. 40-
359. 54). Princeton, NJ: Educational Testing Service.
MacKinnon, D. W. (1962). The nature and nurture of creative Thorndike, R. L. (1982). Applied psychometrics. Boston:
talent. American Psychologist, 17, 484-495. Houghton Mifflin.
MacKinnon, D. W. (1965). Personality and the realization of Thorndike, R. L., & Hagen, E. P. (1977). Measurement and
creative potential. American Psychologist, 20, 227-281. evaluation in psychology and education (4th ed.). New York:
Marsh, H. W., & Hocevar, D. (1983). Confirmatory factor anal- Wiley.
ysis of multitrait-multimethod matrices. Journal of Educa- Torrance, E. P. (1959). Explorations in creative thinking in the
tional Measurement, 20, 231-248. early school years: VI. Highly intelligent and highly creative
Mednick, S. A., & Mednick, M. T. (1962). The associative children in a laboratory school. Research Mem. BER-59-7.
basis of the creative process. Psychological Review, 69, 220- Minneapolis: Bureau of Educational Research, University of
227. Minnesota.
Mednick, S. A., & Mednick, M. T. (1967). TheRemoteAssoci- Torrance, E. P. (1966). Torrance Tests of Creative Thinking:
ates Test (RAT). Boston: Houghton Mifflin. Norms-technical manual. Princeton, NJ: Personnel Press.
Meeker, M., & Meeker, R. (1975). Structure of1ntellect Learn- Torrance, E. P. (1974). Torrance Tests of Creative Thinking:
ing Abilities Test (SO/LAT). EI Segundo, CA: SOl Institute. Norms-technical manual. Lexington, MA: Ginn.
Mehrens, W. A., & Lehmann, I. J. (1984). Measurement and Treffinger, D. J., & Poggio, J. P. (1972). Needed research on
evaluation in education and psychology (3rd ed.). New York: the measurement of creativity . Journal of Creative Behavior,
Holt, Rinehart & Winston. 6,253-267.
Messick, S. (1975). The standard problem: Meaning and values Treffinger, D. J., Renzulli, J. S., & Feldhusen, J. F. (1971).
in measurement and evaluation. American Psychologist, 30, Problems in the assessment of creative thinking. Journal of
955-966. Creative Behavior, 5, 104-112.
Michael, W. B. (1977). Cognitive and affective components of Wolf, R. M. (1982). Validity of tests. In H. E. Mitzel, J. H.
creativity in mathematics and the physical sciences. In J. C. Best, & W. Rabinowitz (Eds.), Encyclopedia of educational
Stanley, W. C. George, & C. H. Solano (Eds.), The gifted and research (5th ed., Vol. 4, pp. 1991-1998). New York:
the creative (pp. 141-172). Baltimore: Johns Hopkins Uni- Macmillan.
versity Press. Yamamoto, K. (1965). Effects of restriction of range and test
Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New unreliability on correlation between measures of intelligence
York: McGraw-HilI. and creative thinking. British Journal ofEducational Psychol-
Parnes, S. J. (1964). Research on developing creative behavior. ogy, 35, 300-305.
In C. W. Taylor (Ed.), Widening horizons in creativity (pp. Yamamoto, K., & Frengel, B. (1966). An exploratory compo-
145-169). New York: Wiley. nent analysis of the Minnesota Tests of Creative Thinking.
Plass, H., Michael, J. J., & Michael, W. B. (1974). The fac- California Journal of Educational Research, 17, 220-229.
CHAPTER 3
A Taxonomy and Critique of

Measurements Used in the Study of
Creativity
Dennis Hocevar and Patricia Bachelor
Creativity measurement itself has been a creative In addition to the present review, readers in-
endeavor for both researchers and practitioners. terested in the measurement of creativity may want
When viewed as a group, the most salient charac- to refer to a number of previously published papers
teristic of creativity measurements is their diver- on the measurement of creativity and related issues
sity. The initial purpose of this review is to integrate (Crockenberg, 1972; Davis, 1971; Dellas & Gaier,
creativity measurements into a meaningful tax- 1970; Educational Testing Service, 1987; Kalt-
onomy and to illustrate the diversity of the available sounis, 1971, 1972; Mumford & Gustafson, 1988;
measurements by citing key examples of the many Petrosko, 1978; Treffinger & Poggio, 1972;
and varied ways in which creativity has been opera- Treffinger, Renzulli, & Feldhusen, 1971; Tryk,
tionalized. It also is hoped that the numerous exam- 1968). In contrast to most previous reviews, the
ples will give researchers a concise but thorough present review will cover a broad range of measure-
picture of the many options available when a mea- ments rather than a more narrowly defined category
sure of creativity is needed. The second goal of this of measurements (e.g., divergent thinking).
review is to use the taxonomy as a framework for
discussing the creativity construct in terms of sever-
al psychometric characteristics-namely, reliabili-
ty, discriminant validity, and nomological validity. A Taxonomy of Creativity
The third goal is to describe an analytic framework Measurements
in which measurement issues can be better
addressed. In the following pages, we classify more than
100 examples of creativity measurement into eight
categories: (1) tests of divergent thinking; (2) at-
Dennis Hocevar • Department of Educational Psychology, titude and interest inventories; (3) personality in-
University of Southern California, Los Angeles CA 90089- ventories; (4) biographical inventories; (5) ratings
0031. Patricia Bachelor • Department of Psychology,
California State University at Long Beach, Long Beach CA by teachers, peers, and supervisors; (6) judgments
90840. Permission to reproduce sections of this article has of products; (7) eminence; and (8) self-reported cre-
been granted by the Society for Personality Assessment, Inc. ative activities and achievements. The categories
53
are meant to be descriptive and nonevaluative, and, 1979a,b,c; Hocevar & Michael, 1979; Milgram,
in addition, they are not necessarily mutually 1983; Milgram & Arad, 1981; Milgram &
exclusive. Milgram, 1976a,b; Milgram, Milgram, Rosen-
bloom, & Rabkin, 1978; Runco, 1986a,b; Seddon,
1983; Zarnegar, Hocevar, & Michael, 1988).
Category 1: Tests of Divergent Thinking
Describing creativity in terms of divergent think-
Category 2: Attitude and Interest Inventories
ing is the most widely used approach to studying
creativity. On the basis of Guilford's (1956) Struc- Some investigators have suggested that creativity
ture-of-Intellect (SOl) model and over two decades can be identified in terms of interests and attitudes.
of factor-analytic research, Guilford and his col- This approach is based on the assumption that a
leagues identified various intellectual abilities. creative person will express attitudes and interests
Some of these abilities (e.g., fluency, flexibility, favoring creative activities. For example, on the
originality, redefinition, and elaboration) have Group Inventory for Finding Interests (GIFFI)
been collectively labelled divergent thinking. A (Davis & Rimm, 1982; Rimm & Davis, 1976,
wide variety of tests have been developed by 1980), subjects are asked to indicate their interest in
Guilford and his colleagues to measure divergent a wide variety of activities like the following:
thinking, such as Alternate Uses (Christensen,
Guilford, Merrifield, & Wilson, 1960), Plot Titles I have a good sense of humor.
(Berger .& Guilford, 1969) and Consequences I like to try new activities and projects.
(Christensen, Merrifield, & Guilford, 1958). Tests I like to invent things.
of divergent thinking are distinguished from tradi- I like to write stories.
tional intelligence tests in that they require a multi-
tude of responses rather than a single correct an- Similarly, on the Holland and Baird (1968) Pre-
swer. For example, in the Alternate Uses test, subconscious Activity Scale, individuals high on origi-
jects are asked to think of alternate uses for a variety nality agree with items like the following:
of common objects (e.g., a shoe or a pencil). Simi-
larly, in the Plot Titles test, subjects are asked to I often daydream about unsolved problems.
generate clever titles to two stories. I have to learn things in my own way rather than
Guilford (1968) and others (most notably, Tor- accepting ideas or relationships suggested in
rance, 1974; Wallach & Kogan, 1965; Wallach & textbooks, etc.
Wing, 1969) have suggested that the more creative If I had the necessary talent, I would enjoy being
individual should possess the types of abilities mea- a sculptor.
sured by tests of divergent thinking, and the I would like to be an inventor.
Guilford tradition has had considerable impact on
the study of creativity. Two of the most widely used Finally, Khatena and Torrance (1976) have de-
divergent-thinking test batteries, the Torrance Tests signed a personality inventory specifically for iden-
of Creative Thinking (Torrance, 1974) and the Wal- tifying creative adolescents. The Creative Percep-
lach-Kogan creativity tests (Wallach & Kogan, tion Inventory has two subscales, "Something
1965) may be seen as modifications and extensions about Myself," and "What Kind of Person Are
ofthe Guilford tests. Also, an important adaptation You?" that are designed to identify the extent to
of the Guilford tests at the elementary school level which a respondent has interests, thought patterns,
has been provided by Meeker and Meeker (1975), and personality characteristics that are thought to be
and this adaptation includes a divergent-thinking creative. Items on the "What Kind of Person are
component (Thompson & Andersson, 1983). You?" instrument call for the test-taker to select
Despite the considerable and valuable impact of characteristics in a forced-choice format. For exam-
divergent-thinking measures on the creativity liter- ple, the creative person will describe him- or herself
ature, they have not gone without criticism, and a as curious rather than self-confident, a self-starter
number of researchers have made suggestions as to rather than obedient, intuitive rather than remem-
how divergent-thinking tests might be revised to bering well, and altruistic rather than courteous.
better suit both practitioners and researchers (Evans Other attitude and interest inventories useful in
& Forbach, 1983; Fu, Kelso, & Moran, 1984; Har- studying creativity are Basadur and Finkbeiner's
rington, 1975; Hattie, 1977, 1980; Hocevar, (1985) measure of preference for ideation, the Cre-
CHAPTER 3 • A TAXONOMY AND CRITIQUE OF MEASUREMENTS 55
ative Behavior Disposition Scale (Taylor & Fish, represent the occurrence of events rather than feel-
1979), the Preference Inventory (Bull & Davis, ings and could be grouped into four categories: self-
1982), the Childhood Attitude Inventory for Prob- striving (e.g., displaying curiosity at an early age);
lem Solving (Covington, 1966), the Creative At- parental striving (e.g., father holding a graduate
titude Survey (Schaefer & Bridges, 1970), the cre- degree); social participation and social acceptance
ative interests scale of the Guilford-Zimmerman (e.g., active participation in the YMCA); and inde-
Interest Inventory (Guilford & Zimmerman, 1963), pendence training (e.g., setting one's own stan-
the Opinion, Attitude, and Interest Survey (Fricke, dards at an early age).
1965), the Allport-Vernon-Lindzey Study of Val- Two published biographical inventories are
ues (Heist, 1968), and a questionnaire developed by available: The Alpha Biographical Inventory by the
Holmes (1976). Institute for Behavioral Research in Creativity
(1968) and Schaefer's (1970) Biographical In-
ventory: Creativity. The Alpha Biographical In-
Category 3: Personality Inventories
ventory is the result of an extensive research pro-
Some investigators have characterized creativity gram carried out with National Aeronautics and
as a set of personality factors rather than cognitive Space Administration (NASA) scientists and en-
traits. Consequently, several well-known person- gineers (Taylor & Ellison, 1964), and, conse-
ality inventories have been scaled to identify quently, it is limited to measuring creativity in the
creativity. For example, in a number of studies, the scientific areas. It consists of 300 items covering a
Adjective Check List (Gough & Heilbrun, 1965) wide variety of areas, such as hobbies, interests,
has been suggested as a potential measure of childhood activities, self-ratings, and experiences,
creativity (Domino, 1970; Gough, 1979; Lacey & and it can be scored for both creativity and academ-
Erickson, 1974; Smith & Schaefer, 1969; Welsh, ic performance.
1977). The version that appears to have the most Schaefer's (1970) Biographical Inventory:
potential is Gough's (1979) Creative Personality Creativity is similar in nature but broader in ap-
Scale. Representative adjectives are: clever, indi- plicability. One hundred and sixty-five questions
vidualistic, insightful, original, self-confident, and are grouped into five sections: physical charac-
unconventional. teristics, family history, educational history, lei-
Additional personality inventories adapted to sure-time activities, and a miscellaneous category.
identify creativity are Heist and Yonge's (1968) The inventory has several dimensions measuring
Omnibus Personality Inventory (Heist, 1968), different fields of creative endeavor. Boys are
Gough's (1957) California Psychological Inventory ranked on a math-science dimension and an art-
(Helson, 1965), and Cattell and Eber's (1968) Six- writing dimension. Girls are ranked on a writing
teen Personality Factor Questionnaire (Cattell & dimension and an art dimension. Several validation
Butcher, 1968). studies have been conducted using this instrument
(Anastasi & Schaefer, 1969; Schaefer, 1969;
Schaefer & Anastasi, 1968).
Category 4: Biographical Inventories
The assumption that an individual's present be-
Category 5: Ratings by Teachers, Peers, and
havior is determined by past experiences forms the
Supervisors
basis for the use of biographical inventories in as-
sessing creative talent. A number of studies indicate Category 5 is the least homogeneous of our eight
that "tailor-made" biographical inventories can categories in that researchers vary considerably in
predict creativity in an industrial setting (Buel, the criteria they use when asking for ratings of
1965; Buel, Albright, & Glennon, 1966; McDer- creativity. For this reason, this method of measur-
mid, 1965; Michael & Colson, 1979; Owens, ing creativity is more consistent methodologically
Schumacher, & Clark, 1957). To illustrate, than conceptually.
Michael and Colson (1979) developed a lOO-item
questionnaire by examining the research literature Teacher Nominations. A large portion of the
and identifying biographical correlates of creative research on creativity takes place in educational set-
endeavor. This questionnaire was then used to mea- tings. Not surprisingly, teacher ratings are a com-
sure creative performance in a sample of electrical monly used criterion of creativity (e.g., Foster,
engineers in an industrial setting. Items generally 1971; Haddon & Lytton, 1971; Piers, Daniels, &
Quackenbush, 1960; Richards, Cline, & Needham, nick and Mednick's (1967) Remote Associations
1964). Test (Karlins, Schuerhoff, & Kaplan, 1969; Med-
Yamamoto's (1963) work is illustrative of the nick, 1963) and Schaefer's Biographical Inventory:
approach taken by several researchers. Yamamoto Creativity (Schaefer, 1969; Schaefer & Anastasi,
first established standards for teachers to use in their 1968) and Guilford and Torrance's tests of creative
ratings. These standards involved identifying the thinking (Houtz, Lewis, Shaning, & Denmark,
most and least creative thinkers in the class. Cre- 1983).
ative thinking was defined as fluency (lots of ideas),
flexibility (many different ideas), inventiveness
(inventing and developing ideas), originality Peer Nominations. While recognizing some
(unique ideas), and elaboration (detailed ideas). inherent limitations in using peer nominations as a
Harrington, Block, and Block (1987) used a dif- criterion of creativity, Torrance (1974) has sug-
ferent approach to obtain teacher nominations. gested that they may provide some useful data in the
They compiled a list of personality traits that were study of creativity. In order to eliminate the ambi-
cited in the research literature as characteristic of guity in peer nominations, Torrance (1962) in-
the creative personality. This list was then used as a structed young children to base their nominations
12-item checklist for teachers to use in making their on specific criteria, such as ideational fluency, un-
judgments. Some illustrative items follow: usual ideas, problem-solving ability, and in-
ventiveness. Yamamoto (1964a,b), Foster (1971),
Tends to be proud of own accomplishments and Reid, King, and Wickwire (1959) also have
Is resourceful in initiating activities used peer nominations as a criterion of creativity
Has unusual thought processes and have developed questionnaires that include
Is curious, exploring, eager for new experiences similar guidelines.
Surprisingly, peer ratings generally are used with
Along the same lines, Renzulli, Hartman, and Cal- children, although there are some exceptions. It is
lahan (1977) and Runco (1984) have developed rat- interesting to note that these exceptions stem from
ing scales for teachers which identify charac- research in business and industrial settings. In
teristics that theoretically relate to the creative working with research scientists, Taylor, Smith,
personality. and Ghiselin (1963) obtained creativity ratings,
Another common approach to teacher ratings is rankings, and nominations by peers in a research
to have teachers give one or two global ratings. For lab. Specher (1964) acquired peer descriptions on
example, in order to validate an inventory of cre- 12 characteristics taken to identify creative en-
ative interests, Rimm and Davis (1980) had teach- gineers. These characteristics included: skill in re-
ers rate students on a 1- to 5-point scale, where 5 porting results to others, persistence, liking for un-
was defined as "highly creative, has many and usual and challenging problems, analytic thinking,
original ideas related to art, music, class materials, ability to reach a decision on his own, fluency of
or out-of-class interests (p. 38)" and 1 was defined ideas, working energetically, knowledge of the
as "very low creativity level, has rarely expressed subject matter, friendly relations with co-workers,
any creative ideas in any verbal, musical or artistic foresight, development of original approaches to
form (p. 38)." At a secondary school level, Drev- problems, and development of valuable and worth-
dahl (1956) asked faculty members in an arts and while ideas. Finally, Keller and Holland (1978)
science department to rate their students on a 7- asked employees from three applied research and
point scale of creativity. The raters were asked to development organizations to nominate from one to
make two judgments: one based on their own defi- four co-workers who had contributed most to inno-
nition of creativity and one based on Drevdahl's vations in the organization. The number of innova-
definition. In an art school setting, Getzels and tions was used as an innovativeness index.
Csikszentmihalyi (1964) directed the art teachers to
rate their students on a 4-point originality scale,
where originality was defined as the "ability to Supervisor Ratings. Particularly in industrial
originate ideas and to draw on personal resources in settings, supervisor ratings have proven to be a
preparing assignments (p. 27)." useful method of choosing creative employees
Teacher ratings have served as criteria in the val- (Keller & Holland, 1978; Meer & Stein, 1955; Tay-
idation of a number of creativity tests, such as Med- lor, 1958). Supervisors, like teachers, have the ad-
vantage of being familiar with the work of many Each of these products was then rated, using guide-
individuals and of having the expertise to judge the lines generally based on three criteria: fluency,
work of these individuals. flexibility, and originality.
Buel (1965) asked research supervisors in a large More recently, Amabile's (1982) work is both
oil company to describe anonymously the most and important and illustrative of the creative product
the least creative researchers under their supervi- approach to measuring creativity. In the first of a
sion. On the basis of their descriptions, Buel devel- series of eight studies, she provided 100 pieces of
oped a list of creative behaviors which can be used paper in multiple shapes, sizes, and colors to pre-
as a descriptive checklist for helping supervisors adolescent girls and instructed them to make a col-
identify creative research personnel. Taylor and his lage that was "silly." Designs then were evaluated
colleagues at the Institute of Personality Assessment by art teachers on a 5-point scale, ranging from very
and Research (IPAR) also employed supervisor rat- uncreative to very creative. In another study,
ings in their research with scientists in a govern- Amabile extended her methods to literature by hav-
ment research lab (Taylor et al .. 1963). Immediate ing her subjects write a cinquain, a 5-line stanza of
supervisors were asked to rate scientists on a unrhymed poetry. These poems were then evalu-
number of traits including: productivity, drive, ated using scales similar to the one described above
mathematical ability, integrity, desire for facts, in- by two groups of judges-advanced poetry stu-
dependence, informative ability, flexibility, per- dents and published poets.
sistence, cooperation, and creation. Taylor also Numerous additional examples of research incor-
asked laboratory chiefs (higher level supervisors) to porating creative products are available in areas
rate scientists with the aid of several checklists. such as art (e.g., Brittain & Beittel, 1964; Bull &
Davis, 1982; Csikszentmihalyi & Getzels, 1970;
Jones, 1964; Sobel & Rothenberg, 1980), literature
(Borgstadt & Glover, 1980; Davis & Rimm, 1982;
Category 6: Judgments of Products Jones, 1964; Lynch & Kaufman, 1974; Malgady &
Barcher, 1977, 1979; Rookey, 1974; Wallen & Ste-
It is self-evident that creative people should pro- venson, 1960), and science (Harmon, 1963; Pelz,
duce creative products. Understandably, identify- 1963; Taylor etal., 1963). In addition, products are
ing creative people in terms of one or more products sometimes defined less concretely as ideas
is an approach that has encompassed virtually all (Graham, 1965; Ward & Cox, 1974).
areas of study and a varied sample of populations. The judges in product studies vary from experts
A little-known but ambitious example of this ap- to nonexperts, and the criteria vary from diverse
proach is a battery of tests developed by Foster definitions of creativity to social recognition (i.e.,
(1971). Foster's assessment devices involve a rewards or publication). Furthermore, subjective
broad range of activities typically associated with judgments usually are made on what products to
the secondary school. A brief description of these choose and in what situation they will be obtained
activities and their corresponding areas follow: and, consequently, the products generally represent
somewhat limited samples of behavior.
I. Sorting playing cards in sets of six that belong
together (mathematics)
2. Creating mathematical equalities with a given Category 7: Eminence
set of numbers and symbols (mathematics) An impractical but valuable approach to the
3. Working in a physical education class with study of creativity is to study eminent people. Per-
and without equipment (physical education) haps the most noteworthy research of this type is the
4. Playing charades with several pieces of work done at the IPAR (Barron, 1969; Hall &
equipment, such as masks, hats, tools, etc. MacKinnon, 1969; Helson, 1971; Helson &
(drama) Crutchfield, 1970; MacKinnon, 1962). The IPAR
5. Writing a story (literature) group initially asked a panel of professors of archi-
6. Making a model out of one or several mate- tecture to nominate the 40 most creative architects
rials, such as buttons, feathers, wire, glue, in the United States. On the basis of these nomina-
corks, cloth, etc. (model-making) tions, 64 architects were asked to visit Berkeley for
7. Listening to music and painting a picture of a weekend of intensive study at IPAR. Forty of
what the music brings to mind (music) these architects accepted. At IPAR, groups of cre-
ative writers and creative mathematicians also were able degree of face validity. Creative activities can
selected by a similar process. take place in a number of fields recognized by soci-
Another impressive programmatic analysis of ety as important. In studying talent for the National
eminence has been undertaken by Simonton Merit Scholarship Corporation, Holland and Nic-
(1976a,b,c, 1977a,b, 1979). By analyzing pub- hols (1964) described lists of achievements and ac-
lished historical and biographical records, Simon- tivities in science, art, literature, and music. The
ton has provided evidence on the effects of both achievements described by Holland are rare, de-
individual (e.g., age) and social (e.g., political in- mand commitment, and generally are publicly rec-
stability) variables on creative production in areas ognized through prizes and publications. Some il-
such as music, science, and others. lustrative items are: placed first, second, or third in
A number of older studies (e.g., Cattell, 1903; a regional or state science contest; exhibited or per-
Cox, 1926; Ellis', 1904; cited in Gilcrist, 1972) formed a work of art; had poems, stories, or articles
attempted to identify persons of eminence through a published in a public newspaper; invented a pa-
variety of methods. Cox (1926) used Cattell's tentable device; had minor roles or leads in plays
(1903) list of the 1,000 most eminent individuals in produced by a college or university. Research in-
history. Cattell's criterion of eminence was the corporating this type of checklist has been pub-
amount of space accorded to each individual in bio- lished in a number of studies (Holland & Astin,
graphical dictionaries. Ellis (1904) based his defini- 1962; Holland & Baird, 1968; Holland & Nichols,
tion of eminence on space devoted to individuals in 1964; Holland & Richards, 1965; Richards, Hol-
the Dictionary of National Biography, a British an- land, & Lutz, 1967a,b).
thology of eminent people. Cattell and Drevdahl Another checklist of creative activities and ac-
(1955) selected eminent researchers, teachers, and complishments has been provided by Hocevar
administrators in the social sciences and physical (1976; 1979d). Hocevar asked university students
sciences on the basis of committee selections and to name their three most creative accomplishments
membership in a professional society. Roe (1951 a) in each of six areas (art, mathematics and science,
selected a group of physical scientists by asking a literature, music, performing arts, and a mis-
committee of scientists to rate a total of 69 men on a cellaneous category). Another group of 50 college
3-point scale of excellence. Roe also chose a group students then rated these activities and accomplish-
of eminent biologists through a similar procedure ments on a scale ranging from most to least cre-
(Roe, 1951b). In Ireland, Barron (1969) asked the ative. These ratings were then used to create a 90-
staff of the Irish Management Institute, a 6,000- item inventory of creative activities and accom-
member management training group, to select indi- plishments that has six subscales: creativity in the
viduals who were unquestionably leaders in Irish fine arts, crafts, literature, music, performing arts,
economic life. and math-science.
The use of eminent people in the study of Other lists similar in content and emphasis to the
creativity raises the question of whether studying Holland list have been developed by Runco
creativity in the highly gifted is tantamount to (1986a), Torrance (1969), Skager, Schultz, and
studying creativity in "normals." Although most Klein (1965), and Wallach and Wing (1969). In
researchers have treated creativity as a normally general, the total creativity score is simply the
distributed trait, the argument that it is limited to a number of activities checked. Although these lists
very small segment of the population is a worth- have been used in educational and psychological
while consideration. research, there are no commercially available
checklists.
Category 8: Self-Reported Creative Activities
and Achievements Other Creativity Assessment Procedures
Perhaps the most easily defended way of identi- As stated earlier, these eight categories are meant
fying creative talent is in terms of self-reported cre- to integrate some common techniques for assessing
ative activities and achievements. Although there is creative talent. Because of the varied nature of re-
a problem in deciding which activities and achieve- search on creativity, there are other measures of
ments should be designated as creative, most of the creativity that are tailor-made for specific situations
lists that have been used in research have a reason- that do not easily lend themselves to categorization.
A partial listing follows: Remote Associates Test Reliability
(Mednick & Mednick, 1967); Starkweather
Two forms of reliability, internal consistency
Creativity Tests (Starkweather, 1971); Mosaic
and interjudge agreement, are particularly relevant
Construction Test (Hall, 1972); Barron-Welsh Art
to research on creativity, and each, therefore, will
Scale (Barron & Welsh, 1952); Welsh Figure Pref-
be discussed.
erence Test (Welsh, 1959); Lundsteen Test of Cre-
ative Problem Solving (Rickborn & Lundsteen,
1968); Onomatopoeia and Images (Khatena, 1978); Internal Consistency. Internal consistency is
Pennsylvania Assessment of Creative Tendency important whenever a group of items on a test or
(Rookey, 1971); Ingenuity Test (Flanagan, 1968); questionnaire are summed to obtain a composite
Sternberg's (1985) indices of people's implicit the- index. Such a practice is inherent to many of the
ories of creativity; the Rorschach VI. M scale previously described measurement approaches,
(Frank, 1979); the problem-solving and creativity namely, tests of divergent thinking, attitude and
dimension of Marsh's (in press) Self-Description interest inventories, personality inventories, bio-
Questionnaire III, a series of cognitive measures graphical inventories, and inventories of creative
developed by researchers at the Educational Testing activities and achievements. Although a detailed
Service (Frederiksen, Evans, & Ward, 1975); the analysis of the reported reliability coefficients in
measures of creativity fostering environments (Har- these studies is beyond the scope of this review, our
rington et aI., 1987); Your Style of Learning and subjective impression is that the internal consisten-
Thinking (Torrance, Reynolds, Riegel, & Ball, cy indices generally are satisfactory for research
1977; Reynolds, Kaltsounis, & Torrance, 1979); purposes in that most exceed .70. Internal con-
and the Lifetime Creativity Scales (Richards, Kin- sistency estimates are less relevant to supervisor
ney, Benet, & Merzel, 1988). ratings, teacher ratings, peer ratings, and product
In addition to the above, researchers also may ratings because often only a single item (or product)
want to examine measures of innovativeness, such is rated. Because of this practice, the extent to
as the Kirton Adaptation Inventory (Kirton, 1976), which these ratings would vary if another question
the Open Processing Scale (Leavitt & Walton, had been asked or another product had been tar-
1975); the Innovativeness Scale (Hurt, Joseph, & geted has not been determined fully.
Cook, 1977) (see Goldsmith, 1986, for an analysis
of these three scales). Finally, a number of intrigu- Interjudge Reliability. In comparison to inter-
ing measures of problem solving also could be nal consistency reliability, inteIjudge reliability is
useful to creativity researchers (see problems used of more concern to creativity researchers. A
by Milgram, 1983, and Houtz & Speedie, 1978, for number of the approaches identified earlier require
examples). people (i.e., experts, supervisors, peers, teachers)
The fact that the measures described in the pre- to make judgments about products, ideas, or other
vious two paragraphs do not fall neatly into any people. This technique presents a particular prob-
single category does not denigrate their potential lem because the researcher must decide who the
usefulness. Indeed, as a group, they generally stem judges should be and what the judges should be
from sophisticated and highly original theories and looking for. To illustrate, when judgments of art
offer much potential as new and interesting ways of products are made, there is some evidence that ex-
further investigating creativity. In the next section, perts and nonexperts disagree (Golann, 1963;
some general problems and criticisms of the eight Knapp & Wulff, 1963). Skager et al. (1965) have
major approaches to the measurement of creativity identified at least three different points of view re-
will be identified. garding art judgments. However, there also is evi-
dence that expertise does not have a significant in-
fluence on ratings (Amabile, 1982).
Discussion and Critique Related to the problem of inteIjudge reliability is
the problem of how the judges are asked to formu-
Our discussion and critique of the measures used late their decisions. Not surprisingly, different re-
to assess creativity will focus on three commonly searchers ask their judges to consider different cri-
discussed psychometric qualities: reliability, dis- teria. The list of criteria is virtually unending,
criminant validity, and nomological validity. including fluency, originality and other cognitive
characteristics, personality traits, self-expression, subjectively assessed, one must hope that judges
enthusiasm, productivity, and expertise. Some in- are able to distinguish creativity from other con-
vestigators use elaborate definitions of creativity; structs, such as intelligence, achievement, and
others use no definition at all. Some researchers competence. Unfortunately, a review of some ap-
train their judges; others believe that untrained propriate studies indicates otherwise. Holland
judges should be employed. Furthermore, there is (1959) had teachers, principals, and guidance coun-
no guarantee that judges will understand and be selors rate students on 12 traits, including origi-
guided by the sometimes complex definitions, and nality. Originality correlated .72 with speaking
there is little research that shows the instructions skills and .84 with writing skills. The other nine
even make a difference. To the contrary, when correlations that involved originality ranged from
Drevdahl (1956) compared defined and undefined .50 to .65. Wallen and Stevenson (1960) investigat-
conditions with teacher nominations, no significant ed creativity in fifth-grade writing and found that
differences in the judgments of their most creative the teacher's judgments correlated. 57 with IQ, .66
students were found. Similarly, Karlins et at. with school grades, and from .66 to .72 with three
(1969)reported a correlation of .97 between de- standardized achievement tests. In a study of re-
fined and undefined ratings of creativity. Finally, to search personnel, instructor ratings of creativeness
make matters even more complicated, with the ex- correlated .68, .72, and .75 with grades, an of-
ception of Amabile (1982), little is known about the ficership rating, and a logical reasoning rating, re-
extent to which extraneous factors (e.g., number of spectively (Mullins, 1964). Finally, in a study in-
dimensions judged, number of products judged, volving faculty ratings of creativity in architecture,
and judges training) affect judgments of creativity. Karlins et at. (1969) found that ratings of creativity
Despite the concerns raised in the previous para- correlated from .79 to .95 with ratings of depend-
graph, a review of the numerous interjudge reliabil- ability, adaptiveness, need to know, independence,
ity indices, which were reported in studies that in- and productivity.
corporate subjective judgments of people or Regarding art judgments, the overall results are
products, finds that interjudge reliabilities usually even more striking. Brittain and Beittel (1964)
exceed.70 and, consequently, are sufficient forre- found that creativity scores determined by judg-
search purposes. Practical constraints prevent our ments on three art performances were correlated. 89
reviewing all these studies, but the interested reader with judged aesthetic quality. Csikszentmihalyi and
is referred to Amabile's (1982) article, which repre- Getzels (1970) found that originality ratings of a
sents the best work in the area of interjudge reliabil- number of drawings correlated. 76 with a rating of
ity. Nevertheless, there are still a number of unre- technical skill and .90 with a rating of overall aes-
solved issues, and a number of improvements in thetic value. Similarly, in the Getzels and
rating procedures have been proposed (Amabile, Csikszentmihalyi (1964) study, mean ratings on
1982; Korb & Frankiewicz, 1976; Malgady & Bar- originality correlated with mean ratings on artistic
cher, 1979; Richards et ai., 1988; Runco, 1984). potential .72 for males and .77 for females. Finally,
Rossman and Gollob (1975) reported that among art
students, peer ratings of creativity correlated .84
Discriminant Validity with peer ratings of intelligence. Likewise, faculty
ratings of creativity correlated .59 with faculty rat-
A test of discriminant validity can be applied to ings of intelligence.
the creativity literature in two ways. First, one may
ask, "Is the creativity construct different from
Different Dimensions? Because judges have
other constructs?" Second, "Are the purported di- trouble discriminating creativity from other at-
mensions of creativity different from each other?" tributes, it follows that they will have even more
Answers to each of these questions will be consid- . trouble discriminating various dimensions of
ered in the sections that follow. creativity. In the study cited earlier, Foster (1971)
reported that the correlations between fluency, flex-
Different from Other Constructs? Creativity is ibility, and originality ratings were about .80 in
a scientific construct and thus it is reasonable to physical education, painting, model-making, writ-
expect that it be different from other scientific con- ing, and drama. Along the same lines, Yamamoto
structs. In particular, because creativity is often (1964b) found that when peer nominations were
used to measure fluency, flexibility, and inventive Nomological Validity

level, their intercorrelations ranged from .62 to .82. Nomological validity is achieved when variables
Finally, Hocevar (1979a,b) and Ho~evar and relate to each other in a logical fashion. I Each of the
Michael (1979) conducted several studies collec- previously defined approaches is a widely used and
tively demonstrating that the dimensio~s o~ d.iver- accepted technique for identifying creative tale.nt.
gent thinking tests have questionable dlscnmillant Because each method is purported to be measunng
validity. some aspect of creativity, it is reasonable to expect
that the methods should yield scores that are
correlated.
Additional Considerations. All the correla-
In a study of 166 scientists in a government labo-
tions reported in the previous two sections would be
ratory, Taylor et al. (1963) collected scor~s on 52
even higher if the data were free of measur~ment criteria of creativity, including 11 supervisor rat-
error (i.e., unreliability in the judgments). It IS ob- ings, six peer ratings, autobiographical informa-
vious that a considerable "halo" effect is present
tion, and numerous product variables. Although
when creativity is based on SUbjective judgments. scores based on the same method (e.g., the 11 su-
Judges seem only to be able to establi~h ~ome over- pervisor ratings) tended to be highly related, scores
all opinion that influences all of their Judgm~n~s. obtained from different methods were not. For ex-
Because of this, they fail to discriminate creativity
ample, supervisor ratings had a low relationship to
from other related constructs. The evidence indi-
peer ratings, and self-rati~g ~cores we~e unco?,e-
cates that asking judges to choose the most intel- lated with the more objective autobIOgraphical
ligent people or their favorite product is some~i~es
data. Furthermore, scores based on products, such
equivalent to obtaining judgments of creatlVlty.
as publications and reports, were not related to ei-
There are, of course, exceptions to the above
ther supervisor or peer ratings. When factor ana-
findings. Rossman and Gollob (1975) demon-
lyzed, this array of criteria yielded a total of 14
strated that subjects can make distinct judgments relatively independent categories.
regarding creativity and intelligence. Harrington,
Davis and Belcher (1971) compared four meth-
Block, and Block (1983) found only moderate cor-
ods of identifying creativity-two tests of diver-
relations in teacher ratings of creativity, general
gent thinking, a biographical inventory, and a c~te
intelligence, and cooperativeness and, conse-
rion composed of self-ratings on several creative
quently, concluded that their creativity criterion
activity questions. Except for the relationship of the
possessed substantial discriminant validity.
biographical inventory and the criterion, the inter-
Sternberg (1985) presented evidence that indi-
correlations of these tests were low and generally
viduals can discriminate creativity, intelligence,
nonsignificant for both males and females.
and wisdom. Amabile (1982) found that creativity
In their study of art students, Getzels and
could be distinguished from both aesthetic value
Csikszentmihalyi (1964) included 14 measures of
and technical competence in judgments of art quali-
creativity, including two teacher ratings, eight tests
ty. Richards et al. (1988) presented procedures for
of divergent thinking, two personality inventories
discriminating creativity from intelligence and ap-
(Cattell's 16 Personality Factors and the Study of
preciation of creativity and for distinguishing voca-
Values), grades, and IQ scores. The authors pre-
tional and avocational creativity. Finally, modifica-
dicted some overlap among these criteria because
tions in scoring techniques (Milgram, 1983;
they all have some claim as indices of creativity.
Milgram & Arad, 1981) and modifications in in-
Such was not the case. A factor analysis revealed
structions (Runco, 1986b; Runco & Albert, 1985)
only method factors rather than a general creativity
seem to separate better the dimensions of divergent
factor. In another study of art students, Ellison
thinking.
These exceptions notwithstanding, research~rs IWe chose the term nomological rather than construct validity
cannot assume that the creativity construct has diS- because construct validity is more all-encompassing; that is,
criminant validity, particularly when subjective construct validity usually subsumes all other types of validity.
judgments are involved. Empirical data as well as In addition we chose not to use the term convergent validity
because thi~ type of validity is generally applied when ?ne is
sophisticated statistical a.na~ysis are ~ee~ed. The studying different measures of a single construct. Our vIew ~s
topic of statistical analysIs IS dealt with III a later that the creativity instruments discussed herem measure a multI-
section. tude of constructs.
(1973) found low and negative relationships be- alternative measures of creativity have less than
tween judged creativity on a pastel drawing, the 10% variance in common. When creativity is iden-
Remote Associations Test, and the Barron-Welsh tified using multiple methods, individuals are
Art Scale. Along the same lines, Zimmer, Guip, ranked differently. High scorers on one method are
and Hocevar (1988) provided data that point to a not necessarily high scorers on another. This find-
similar lack of nomological validity in that no sig- ing fdises questions as to whether creativity should
nificant relationship was found between creative be viewed as a unitary disposition.
attitudes and interests and a two-judge evaluation of The inconsistent findings in relation to
art portfolios. nomological validity have another important im-
In the most ambitious analysis of the creative plication. Perhaps by using distinct techniques to
personality to date, Gough (1979) related seven dif- assess creativity, researchers are actually studying
ferent scales, all derived from the Adjective Check different phenomena. If this is the case, conclusions
List, to criterion ratings of creativity in seven male must be limited to studies incorporating the same
samples and in five female samples. For six of the method, and theory integration in the area of
seven scales, only 17 of 72 validity coefficients creativity becomes even more difficult.
were significant and only 1 of 72 exceeded .40. An explanation for the confusion is that many of
(Results were much more encouraging for the sev- the methods tagged with the creativity label really
enth scale, the Creative Personality Inventory, do not measure phenomena that society typically
which had 10 of 12 validity coefficients significant labels creative. In many cases, the methods are hy-
at the .05 level.) Finally, the correlations of diver- pothesized correlates ofreal-life creative behavior.
gent thinking with other measures of creativity Guilford and the cognitive group have hypothesized
often indicate no relationship (Andrews, 1975; Bar- that divergent thinking is somehow linked to cre-
ron, 1969; Beittel, 1964; Brittain & Beittel, 1964; ative behavior. The personality psychologists have
Fitzgerald & Hattie, 1983; Getzels & Csikszentmi- hypothesized that some personality traits are linked
halyi, 1964; Goolsby & Helwig, 1975; Gough, to creative behavior. And others have hypothesized
1976; Hadden & Lytton, 1971; Hocevar, 1980; Jor- that attitudes and interests or past experiences are
dan, 1975; Karlins etal., 1969; Kogan & Pankove, linked to creative behavior. Although there is some
1974; Popperova, 1972; Roweton, Farless, Do- evidence to support the validity of these various
nham, Wleklinski, & Spencer, 1975; Skager, hypotheses, the relationship of these measures to
Klein, & Schultz, 1967). real-life creativity is not one-to-one. Therefore,
There are exceptions to the trends reported in the caution should be taken in the uncritical use of cor-
last several paragraphs. In particular, many studies relates as criterion measures of creativity.
report positive relationships among divergent
thinking and various creativity criteria (Bartlett &
Davis, 1974; Bennett, 1973; Cropley, 1972; Dew-
ing, 1970; Halpin, Halpin, & Torrance, 1974; Har- Conclusions
rington, Block, & Block, 1983; Haven, 1965;
Hocevar, 1980; Jones, 1964; McCrae, 1987; It is apparent that reliability, discriminant valid-
Milgram & Milgram, 1976a; Rotter, Langland, & ity, and nomological validity cannot be taken for
Berger, 1971; Runco, 1984, 1986a; Torrance, granted in creativity research. Even though findings
1969; Wallach & Wing, 1969; Wallbrown & are mixed, enough problems have been identified in
Huelsman, 1975; Zegas, 1976). the extant literature to point to a need for greater
In addition to studies that involve divergent attention to measurement issues by those engaged
thinking, significant correlations between alter- in creativity research. It is noteworthy that through-
native approaches to measuring creativity have out this review we have relied on assessing discrim-
been reported in a variety of studies (e.g., Bull & inant and nomological validity through the subjec-
Davis, 1982; Davis & Rimm, 1982; Keller & Hol- tive interpretation of the size of correlations; that is,
land, 1978; Michael & Colson, 1979; Rimm & discriminant validity is indicated by low correla-
Davis, 1980; Taylor & Fish, 1979; Taylor, Sutton, tions and nomological validity is indicated by high
& Haworth, 1974). Nevertheless, it is important to correlations. There are problems with this ap-
recognize that in studies where significant positive proach. First, the correlations presented in most
correlations have been reported, the correlation is research reports are attenuated due to measurement
seldom higher than .30, suggesting that the two error. Second, there is no agreed upon guideline for
determining how Iowa correlation must be to dem- berg (1980) (e.g., poems, collages, and sketches)
onstrate discriminant validity or how high a correla- and the kind of activities listed in the Holland and
tion must be to demonstrate nomological validity. Nichols (1964) and Hocevar (1979d) inventories
Unfortunately, the literature itself necessitates this are recognized by society as being creative, and
crude approach to discriminant and nomological they are relatively rare. Most of the activities de-
validity. More sophisticated analytic approaches to mand a high degree of commitment, and they are
discriminant validity have been available for many not something just anyone can master.
years (i.e., exploratory factor analysis; multitrait- Although it can be argued that interest invento-
multimethod [MTMM] analysis); however, these ries, personality inventories, tests of divergent
approaches largely have been unused by most rethinking, and other measures of creativity have pre-
searchers in the area of creativity. It is hoped that dictive validity, past behavior is generally the best
more modem structural modeling, such as Joreskog predictor of future behavior. The efficacy of this
and Sorbom's (1986) LISREL analytic paradigm predictive strategy is evident in a number of areas
and Bentler's (1985) EQS paradigm, will not be (see Mischel, 1968, for a review). These findings
neglected similarly because these procedures offer appear to generalize to the area of creative behav-
much more precise and objective approaches to ior. MacKinnon (1968) reported that, in studies of
problems of discriminant and nomological validity. creative persons in a variety of fields, earlier ac-
Such procedures are best used when at least two complishments were consistently predictive oflater
traits are measured by two different methods (i.e., accomplishments.
the MTMM paradigm) (cf. Marsh & Hocevar, The most compelling evidence for predicting fu-
1983, 1988). However, they also are applicable to ture creative behavior with past creative behavior is
the more common single-method studies in which found in two related studies (Holland & Nichols,
the dimensionality of a test or questionnaire needs 1964; Richards et ai., 1967a). In a high-aptitude
to be determined or to studies in which the discrimi- sample, Holland and Nichols predicted talented,
nant validity of ratings on two or more traits needs nonacademic accomplishment in six areas: lead-
to be evaluated (cf. Zarnegar et at., 1988). Finally, ership, science, dramatic arts, literature, music,
these procedures also would be applicable to the and art. Predictors included interests, goals, ac-
exploration of possible higher-order factor models tivities, self-conceptions, aptitudes, and person-
(cf. Marsh & Hocevar, 1985) and to the analysis of ality traits. The results of this study indicated that
factorial invariance (cf. Hocevar & EI-Zahhar, past activities and achievements easily were the
1984; Marsh & Hocevar, 1985). best predictors of creative achievement in college.
Which approach to measuring creativity is best? The findings of this study have been replicated with
Tests of ability, interests, and personality are ap- a less select sample by Richards et ai. (1967a).
propriate when the researcher's goals are to explain In conclusion, our recommendation is that re-
something about creativity, but they are not accept- searchers interested in creativity focus on emi-
able as criteria of creativity. High scorers on these nence, creative products, or an inventory of cre-
tests are not necessarily creative people. If a re- ative accomplishments and achievements. Three
searcher is interested in the personality correlates of categories of creativity measurement-personality,
creativity or in the relationship of intelligence to interests, and divergent thinking-should not be
creativity, or in evaluating educational programs accepted uncritically as criteria of creativity.
designed to enhance creativity, why not go directly Rather, they should be viewed as correlates of
to the criteria that have face validity? This can best creativity that are possibly causally related but con-
be accomplished through studying eminent indi- ceptually distinct from the creativity construct. Fi-
viduals, evaluating creative products, or using an nally, the usefulness of ratings by others (teachers,
inventory of creative activities and accomplish- peers, supervisors) and of autobiographical in-
ments. ventories as criteria of creativity is yet to be fully
Because studies of eminent individuals are often determined.
impractical, assessing creativity through the analy- The possibility of using ratings by others de-
sis of creative products or through the administra- serves further consideration if the ratings focus on
tion of an inventory of creative activities and ac- real-life creativity instead of its correlates and if
complishments are the best of the currently satisfactory interjudge reliabilities, discriminant
available assessment strategies. The kind of tasks validities, and nomological validities can be ob-
analyzed by Amabile (1982) and Sobel and Rothen- tained. Although much of the literature on ratings
suggests that reliability and validity are problematic creative production were outcomes. We agree with
when people are asked to rate other people, there this approach.
are isolated examples in which ratings have ap- Does the previous paragraph imply that studies
peared to be promising. The work of Richards etal. that focus on divergent thinking or the creative per-
(1988) on the Lifetime Creativity Scale is a case in sonality can be discounted? The answer to this
point. They demonstrated that raters who indepen- question is "clearly no" for two reasons. First,
dently evaluate the protocols of clinical inter- there is overwhelming evidence that divergent
viewers could assess reliably lifetime vocational thinking and the creative personality are interesting
and avocational creativity in diverse samples of constructs in their own right, and, consequently,
subjects, and that these assessments appeared to deserve attention as distinct and important scientific
have content, discriminant, and construct validity. constructs. Second, there is at least some evidence
Besides ratings by others, it is possible that auto- that these two constructs are potential causes of
biographical inventories offer some potential as real-life creativity.
measures of real-life criteria of creativity. This is
because biographical inventories often include cre-
ative activities and accomplishments that are sim- Agenda for Future Research on
ilar to the lists previously recommended. On the Measurement Issues
negative side, biographical inventories also often
include life events that can only be classified as Rather than present a listing of "needed re-
possible correlates of creativity. Suffice it to say search," we will provide an expository example of
that autobiographical inventories are diverse, and an ideal hypothetical measurement study that is
this diversity makes their usefulness as a criterion of consistent with the conclusions just presented. Tra-
creativity difficult to determine. ditionally, measurement issues have been analyzed
In marked contrast to our recommendation that by computing reliability indices, inspecting attenu-
real-life criteria of creativity be included in studies ated correlations among measured variables, and
of creativity, prior studies often do not include such possibly conducting an exploratory factor analysis.
criteria. To illustrate, there are dozens of studies As discussed earlier in this review, this traditional
that relate correlates of creativity (e.g., divergent approach fails to take advantage of modem devel-
thinking) to other correlates (creative personality) opments in structural modeling. The intent of the
or to other variables that represent distinct con- following sections is to illustrate the use of struc-
structs (e. g., intelligence). Although these studies tural modeling in a nontechnical manner and to ex-
are valuable in that they help us understand the pound on the advantages of this approach. When
presumed correlates of creativity, they are limited applied to measurement models, structural model-
in that they are one step away from actual real-life ing is called confirmatory factor analysis. The sta-
creati vity . tistical framework for confirmatory factor analysis
Along the same lines, there are dozens of experi- presented herein is based mainly on the LISREL
mental studies that evaluate the effects of creativity- framework of Joreskog and Sorbom (Joreskog,
enhancing treatments without including a real-life 1969; Joreskog & Sorbom, 1986).
criteria of creativity. Again, these studies are Measurement research cannot be conducted
worthwhile in that.it is of intrinsic interest to know without a concrete conceptual framework, and fu-
whether the correlates of creativity (e. g., divergent ture research should reflect this fact. Thus, this il-
thinking) are modifiable. Whether changes in these lustration will follow one of many possible concep-
correlates can actually influence real-life creative tual models. The hypothetical model is shown in
behavior is another issue that can only be resolved Figure 1 (the parameter estimates should be ignored
when real-life criteria are included in the research for the moment). In line with conventional practice
design. (cf. Bentler, 1980), two types of variables are dif-
How can one best interpret the vast extant liter- ferentiated pictorially-measured or observed vari-
ature on creativity? A simple rule of thumb would ables and nonmeasured or latent variables. The
be to focus on only those studies that include a measured variables are shown in small squares and
measure of real-life creativity. Interestingly, this is represent hypothetical responses to the following
the strategy generally employed by Mumford and instruments:
Gustafson (1988) in their recent review of the
creativity syndrome. In the main, their review VFI = Wallach and Kogan's (1965) creativ-
focuses on studies in which eminence or real-life ity battery (verbal fluency only)
r.=lJl .80
,, ,
'.72
B
~
.13 .97
, ~B
, .71
IINT1 l't,
.81\
.91/
,
I INT2 1/J
Figure 1. Analytic model.
VF2 = Torrance's (1974) Tests of Creative JCl = Expert #1's judgments of the creativ-
Thinking (verbal fluency only) ity of a short story written by each
FFI = Wallach and Kogan's (1965) creativ- subject
ity battery (figural fluency only) JC2 = Expert #2's judgments of the creativ-
FF2 = Torrance's (1974) Tests of Creative ity of a short story written by each
Thinking (figural fluency only) subject
INTI = Group Inventory for Finding Interests JT 1 = Expert # 1' s judgments of the tech-
(Davis & Rimm, 1982; Rimm & nical quality of a short story written
Davis, 1976, 1980) (odd items only) by each subject
INT2 = Group Inventory for Finding Interests JT2 = Expert #2's judgments of the tech-
(Davis & Rimm, 1982; Rimm & nical quality of a short story written
Davis, 1976, 1980) (even items only) by each subject
The latent variables are designated by circles. They between latent factors (10 parameters) can be esti-
represent five different hypothetical constructs: mated also. Correlations between latent factors are
Verbal Fluency, Figural Fluency, Creative Interest, shown in Figure 1 as double-arrowed solid lines.
Writing: Creativity, and Writing: Technical Quali- As will be clear from the discussion below, the
ty. In line with the discussion in the previous sec- interpretation of a confirmatory factor analysis is
tion, the conceptual model specifies that fluency based on the factor loading and the factor intercor-
and interest constructs are considered as distinct relation parameter estimates previously mentioned.
constructs. The mathematical objective of a confirmatory fac-
One begins a confirmatory factor analysis by first tor analysis is to find estimates of the parameters
deriving a correlation or covariance matrix among that yield a predicted correlation matrix that most
measured variables. A hypothetical correlation ma- closely fits the observed correlation matrix. Pre-
trix is given in the Appendix. The next step is to dicted correlations are determined by path multipli-
determine which factor loadings will be estimated cation.
using theoretical considerations. Because both The parameter estimates generated by the confrr-
measured variables and latent factors are standard- matory factor analysis are shown in Figure 1. Sin-
ized, the factor loadings are correlations between gle-arrowed dotted paths going from a hypothetical
measured variables and latent factors. The factor construct to a measured variable are factor loadings
structure implied by Figure 1 is shown in Table 1. and they are correlations; double-arrowed paths
Two types of parameters are illustrated-factor going from one latent variable to another latent vari-
loadings which were estimated and factor loadings able are between factor correlations. These correla-
which were set to zero. The general rule is that tions are corrected for measurement error (i.e., dis-
target factor loadings are estimated and nontarget attenuated) and, consequently, they are more useful
factor loadings are set to zero; that is, measured than conventional attenuated correlations based on
variables are allowed only to load on the factors that raw data.
they are hypothesized to measure. For example, the The discussion and interpretation of our confrr-
two verbal fluency measured variables, VFl and matory factor-analytic results will focus on the
VF2, are constrained to load on the Verbal Fluency three issues that formed the basis for our previous
factor. As illustrated in Table 1, 10 parameters were critique of creativity measurement; namely, relia-
estimated in the factor-loading matrix, and 40 pability, discriminant validity, and nomological
rameters were constrained to zero. In relation to validity.
Figure 1, factor loadings are represented by paths
Reliability
that go from hypothetical constructs (e.g., Verbal
Fluency) to measured variables (e.g., VFl). These In an earlier section, two types of reliability were
paths are shown as dotted lines. identified as most relevant to the creativity re-
In a confirmatory factor analysis, correlations searcher-internal consistency and interjudge.
Table 1. Factor-Loading Matrix

Latent variables (factors)
Measured Verbal Figural Creative Writing: Writing:
variables fluency fluency interest creativity technical quality
VFl .80 0.0 0.0 0.0 0.0

VF2 .80 0.0 0.0 0.0 0.0
FFI 0.0 .74 0.0 0.0 0.0
FF2 0.0 .86 0.0 0.0 0.0
INTl 0.0 0.0 .81 0.0 0.0
INT2 0.0 0.0 .91 0.0 0.0
JCl 0.0 0.0 0.0 .72 0.0
JC2 0.0 0.0 0.0 .73 0.0
JTl 0.0 0.0 0.0 0.0 .71
JT2 0.0 0.0 0.0 0.0 .74
Internal Consistency. One approach to internal The total observed variance (S2) is the sum of the
consistency is to correlate separate halves of a mea- variances of the observed variables (1 + 1 because
sure. The split-half technique is well known, and we are analyzing a correlation matrix) plus two
other techniques (e.g., coefficient alpha, Kuder- times their observed correlation (2 x .74). In equa-
Richardson Formula 21) can be viewed as more tion form: S2 = 1 + 1 + 2 (.74) = 3.48. Substitut-
refined methods of computing split-half reli- ing into the formula given for reliability above, the
abilities. In the present example, the split-half tech- reliability of the measurement of the Creative In-
nique is illustrated for the Creative Interest con- terest factor is given as:
struct. However, it also could have been applied to
the Verbal Fluency and Figural Fluency constructs R I· b'I' = I _ Total error of measurement
in a separate analysis or through a higher order e!a 1 lty Total variance
factor analysis (cf. Marsh & Hocevar, 1988). In a = .34 + .17 = 85
confirmatory factor-analytic framework, a neces- 3.48 .
sary but not sufficient condition for reliability is that
the estimated correlations (i.e., factor loadings) be- A more strict test of internal consistency is the
tween measured variables and their respective con- examination of the extent to which different mea-
structs be significant. For the first half of the Cre- sures (e.g., the Wallach-Kogan test and the Tor-
ative Interest measure (INTI), the factor loading is rance test) of the same construct (e.g., Verbal Flu-
.81 and it is significant. (All the loadings in Figure 1 ency) give consistent results. As shown in Figure 1,
are significant, so the asterisks have been omitted such a research question can easily be incorporated
from the diagram to simplify the figure.) The factor into a confirmatory factor analysis. To illustrate,
loading can be squared and subtracted from unity to both the Wallach and Kogan and the Torrance mea-
yield an estimate of the percentage of error for any sures are constrained to load on Verbal Fluency
given measured variable. In this case, .81 squared latent variable. Both loadings are .80. These load-
is .66 and the error of measurement for INTI is ings easily exceed the .70 rule of thumb and the
therefore .34 (i.e., 1 - .66). actual reliability coefficient, using the formula
Because statistical significance can be obtained given above, is .78. These data point to the conclu-
even when measurements are less than satisfactory, sion that both the Wallach-Kogan and the Torrance
the degree of reliability also is important. With two verbal fluency subtests are measuring the same con-
measured variables, a sufficient reliability 2 is indi- struct in a satisfactory manner.
cated when factor loadings exceed. 70. The present
example meets this rule of thumb for all ten factor Interjudge Reliability. As previously dis-
loadings. A more precise indication of reliability is cussed, subjective judgments of creativity are fre-
obtained by computing the reliability coefficient. quently employed by researchers. Such subjectivity
The reliability coefficient is the percentage of requires that the extent to which independent judg-
"true" variation in a measurement, and it is given ments provide consistent results be assessed. In the
by the following formula: present example, this assessment is accomplished
by loading two independent judgments of creative
Reliability = I _ Total error of ~easurement writing (JCl and JC2) on the same latent factor
Total vanance
(Writing: Creativity). The procedure for computing
where the total error of measurement is the sum of and assessing interjudge reliability is identical to
the errors across the measured variables, and the the internal consistency calculations discussed ear-
total factor variance is the observed variance of a lier, and the interjudge reliability for the Writing:
factor composite that is created by adding the mea- Creativity indices (i.e., judgments) equals .69.
sured variables (in this case, the sum of the odd and The reader may have realized that, up to this
even halves of the creative interests measure). To point, there is little that confirmatory factor analysis
illustrate, the error for the first measured variable accomplishes that could not be accomplished with a
on the Creative Interest construct (INTI) is .34 (see
above). By similar calculation, the error for INT2 is equal to .70, the predicted correlation between measured vari-
1 - .9I2 or .17. Thus, the sum of the errors is .51 ables is .49 (.70 x .70). Substituting .49 into the split-half
formula, the reliability would be .66, which is usually regarded
(summing .34 and .17). as sufficient or close to sufficient for research purposes. (The
rule of thumb requires that one assumes that the predicted and
2The logic for the rule of thumb is as follows. If the loadings are observed correlations are roughly equivalent.)
more traditional approach to reliability (e.g., the which judges can discriminate creativity from other
split-half formula would yield almost identical reconstructs. For example, in reference to Figure 1,
sults). This is correct, and the major advantages of can judges discriminate creativity in a short story
confirmatory factor analysis fall into the areas of from the technical quality of the short story? Inspec-
discriminant and nomological validity. However, it tion of attenuated correlations offers only an im-
is important to point out that a confirmatory factor precise guess when answering questions about dis-
analysis requires that fundamental measurement is- criminant validity, but confirmatory factor analysis
sues be addressed. Also, in a confirmatory factor can answer such questions in a straightforward
analysis, measurement findings automatically be- fashion. To illustrate, Figure 1 shows that the disat-
come integrated into the evaluation and interpreta- tenuated correlation between Writing: Creativity
tion of discriminant and nomological validity. and Writing: Technical Quality is .97. The sheer
magnitude of this correlation clearly indicates that
judges cannot discriminate creativity from tech-
Discriminant Validity nical quality. (An even more precise test could be
In a previous discussion, two questions were accomplished in borderline cases by testing
identified as particularly relevant to the creativity whether the correlation differs from unity.)
literature. First, one may ask, "Are the dimensions Although it might be tempting to argue that the
of creativity different from each other?" Second, above questions could be answered using more tra-
one may ask' 'Is creativity different from other conditional correlational techniques, this is not the
structs?" The following discussion will demon- case. The inspection of uncorrected correlations is
strate how both questions can be examined in a limited in that such correlations are attenuated be-
confirmatory factor analysis. cause of measurement error and, consequently, can
provide the illusion of discriminant validity. The
Different Dimensions? According to most au- present example is a case in point in that the attenu-
thors, fluency has at least two dimensions: verbal ated correlations are only in the area of .50 (see the
and figural. Some authors have suggested that the Appendix); yet there is clear evidence against dis-
dimensions of divergent thinking do not have discriminant validity in the confirmatory factor analy-
criminant validity when operationalized em- sis. Simple correlational analyses are further lim-
pirically. Others have disagreed, and it is likely that ited in that discriminant validity requires the
a consensus will occur only when the issue is exam- consideration of patterns of correlation rather than a
ined in ways that are more precise than the subjec- single correlation. Such a process is unwieldy and
tive visual examination of attenuated correlations. difficult to interpret.
Confirmatory factor analysis is tailor made for such Traditional exploratory factor analysis has ad-
an issue, because it provides a statistical test of vantages over the visual inspection of attenuated
whether or not two constructs differ. correlations. Nevertheless, exploratory factor anal-
To illustrate using the present example, the disat- ysis is limited also in that the actual factors are
tenuated correlation between Verbal Fluency and determined using mathematical criteria rather than
Figural Fluency is .72. Whether Verbal Fluency theory, and the correlations between factors are de-
and Figural Fluency are actually different con- termined by arbitrary decisions that are set before
structs can be determined by testing whether the the analysis, e.g., the type of rotation and the pa-
dis attenuated correlation (.72) is different from uni- rameters (e.g., delta) that are set before the rota-
ty. This test is easily accomplished by testing tion.
whether the hypothesized model (Figure 1) better
reproduces the observed correlation matrix than a
Nomological Validity
model in which the Verbal/Figural correlation is set
at unity. In this case, the original model does better Nomological validity is achieved if a construct
reproduce the correlation matrix than the model in correlates with other constructs in a logical fashion.
which the Verbal/Figural correlation equals unity. For example, some logical relationships inherent in
Thus, one would conclude that fluency has distinct Figure 1 are:
Verbal and Figural dimensions.
1. Verbal Fluency should correlate with Writ-
Different Constructs? A reasonable and fre- ing: Creativity more highly than Figural Flu-
quent question in creativity research is the extent to ency because creative writing involves to a
greater extent the mental manipulation of fined using atheoretical criteria, and correlations
ideas rather than figures. are estimated using arbitrary criteria for rotation.
2. Verbal Fluency should correlate more highly
with Writing: Creativity than with Writing:
Technical Quality because Verbal Fluency Summary
has been identified as integral to creative writ-
ing by numerous authors. Available creatIvIty measures were classified
3. Creative Interest should correlate with Writ- into eight groups: tests of divergent thinking; at-
ing: Creativity; and this correlation should be titude and interest inventories; personality invento-
greater than the correlation between Creative ries; biographical inventories; ratings by teachers,
Interest and Writing: Technical Quality. By peers, and supervisors; judgments of products; emi-
definition, this prediction is self-evident. nence; and self-reported creative activities and ac-
complishments. These eight approaches to creativ-
The disattenuated correlations in a confirmatory ity measurement were collectively reviewed in
factor analysis can be used to check each of these terms of their reliability, discriminant validity, and
hypotheses in a precise manner. To illustrate, Hy- nomological validity.
pothesis 1 is supported because the correlation be- Both internal consistency reliabilities and inter-
tween Verbal Fluency and Writing: Creativity is judge reliabilities have been satisfactory for re-
.43, which is greater than the correlation between search purposes, although there remain many issues
Figural Fluency and Writing: Creativity (.13). Hy- to be resolved in the analysis on interjudge agree-
pothesis 2 is not supported because the correlation ment. A lack of discriminant validity continues to
between Verbal Fluency and Writing: Creativity be a problem, both in terms of how creativity differs
(.43) is slightly lower than the correlation between from other constructs and in terms of how the di-
Verbal Fluency and Writing: Technical Quality mensions of creativity differ from each other. Much
(.46). Finally, Hypothesis 3 is partially supported. evidence points to a lack of nomological validity if
Creative Interest does correlate with Writing: one considers the eight categories of measurement
Creativity (.42); however, this correlation is identi- as eight different ways of measuring some aspect of
cal to the correlation between Creative Interest and creativity; that is, measures that presumably relate
Writing: Technical Quality. More precise tests for to creativity frequently do not correlate with each
the difference between correlations in a confirmato- other. The obvious conclusion is that different re-
ry factor analysis are available (Joreskog & Sor- searchers are studying different phenomena by vir-
bom, 1986). Generally, in these tests, correlations tue of the measures that they select.
are constrained to equality, and a statistical test is Researchers in the area of creativity have not
conducted to examine whether the fit of the con- taken advantage of modem developments in confIr-
strained model shows a significant decrement. matory factor analysis when assessing reliability,
Taken together, the results in reference to discriminant validity, and nomological validity.
nomological validity are mixed and the reason for This chapter provides a concrete example of how
the mixed findings is clear: Judges fail to discrimi- confirmatory factor analysis might be applied to
nate creativity in writing from the technical quality resolve measurement issues, and the advantages of
of writing. the confirmatory factor analytic approach are
Although it is possible to analyze nomological discussed.
validity through visual inspection of attenuated cor- In reference to the issue of how creativity is best
relation coefficients and with exploratory factor measured, peer nominations, supervisor ratings,
analysis, these approaches are less advantageous and teacher nominations are often inadequate indi-
because low reliability can be mistaken for a lack of cators of creativity due to the rater's inability to
nomological validity. Further, nomological valid- discriminate creativity from other traits. In addi-
ity may necessitate the statistical comparison of tion, these methods do not guarantee that the raters
correlations. Even though such tests can be accom- will have the knowledge and intention to actually
plished using z transformations, it is more easily consider criteria that bear on real-life creativity in
accomplished in a confirmatory factor analysis. Fi- making their decisions. Divergent thinking, bio-
nally, though exploratory factor analysis can yield graphical characteristics, attitudes and interests,
factor correlations that are disattenuated, explorato- and personality characteristics are best described
ry factor analysis is limited in that factors are de- only as correlates of real-life creative behavior, and
they should not be taken as direct measures of ACKNOWLEDGMENTS. The authors acknowl-
creativity. Research that has the most bearing on edge the helpful comments of Susan Page Hocevar as
what society recognizes as actual creativity can best contributing to this manuscript. This chapter is a
be accomplished by studying eminent women and revised and updated version of D. Hocevar (1981),
men, analyzing creative products, and administer- Measurement of Creativity: Review and Critique.
ing inventories of creative activities and accomp- Journal ofPersonality Assessment, 45, pp. 450-64.
lishments.
Appendix
Correlation Matrix: Hypothetical Data
VFl VF2 FFI FF2 INTI INT2 JCl JC2 JTI
VF2 .64
FFI .45 .43
FF2 .48 .49 .64
INTI .07 .06 .02 .04
INT2 .12 .09 .13 .16 .74
JCl .25 .23 .11 .06 .27 .26
JC2 .26 .27 .10 .07 .27 .28 .53
JTI .25 .29 .12 .10 .26 .29 .50 .51
JT2 .28 .25 .09 .01 .28 .26 .53 .52 .53
Note: N = 200.
References review of psychology (Vol. 31, pp. 419-56). Palo Alto, CA:
Annual Reviews.
Bennet, S. N. (1973). Divergent thinking ability: A validation
Amabile, T. M. (1982). Social psychology of creativity: A con- study. British Journal of Educational Psychology, 43, 1-7.
sensual assessment technique. Journal of Personality and So· Berger, R. M., & Guilford, J. P. (1969). Plot Titles. Beverly
cial Psychology, 43,997-1013. Hills, CA: Sheridan Psychological Services.
Anastasi, A., & Schaefer, C. E. (1969). Biographical correlates Borgstadt, C., & Glover, J. A. (1980). Contrasting novel and
of artistic and literary creativity in adolescent girls. Journal of repetitive stimuli in creativity training. Psychological Re-
Applied Psychology, 53, 267-273. ports, 46, 652.
Andrews, F. (1975). Social and psychological factors which Brittain, W. L., & Beittel, K. R. (1964). A study of some tests of
influence the creative process. In I. A. Taylor & J. W. Getzels creativity in relationship to performances in the visual arts. In
(Eds.), Perspectives in creativity. Chicago: Adline. W. L. Brittain (Ed.), Creativity and art education.
Barron, F. (1969). Creative person and creative process. New Washington, DC: National Art Education Association.
York: Holt, Rinehart & Winston. Bue!, W. D. (1965). Biographical data and the identification of
Barron, F., & Welsh, G. (1952). Artistic perception as a factor in creative research personnel. Journal of Applied Psychology,
personality style: Its measurement by a figure preference test. 49, 318-321.
Journal of Psychology, 33, 199-203. Bue!, W. D., Albright, L. E., & Glennon, J. R. (1966). A note
Bartlett, M., & Davis, G. A. (1974). Do the Wallach and Kogan on the generality and cross-validity of personal history for
tests predict real creative behavior? Perceptual and Motor identifying creative research scientists. Journal of Applied
Skills, 39, 730. Psychology, 50, 217-219.
Basadur, M. , & Finkbeiner, C. T. (1985). Measuring preference Bull, K. S., & Davis, G. A. (1982). Inventory for appraising
for ideation in creative problem-solving training. Journal of adult creativity. Contemporary Educational Psychology, 7,
Applied Behavioral Science, 21, 37-49. 1-8.
Beittel, K. R. (1964). Creativity in the visual arts in higher Cattell, R., & Butcher, H. (1968). The prediction of achieve-
education: Criteria, predictors, experimentation and their in- ment and creativity. Indianapolis: Bobbs-Merrill.
teractions. In C. W. Taylor (Ed.), Widening horizons in Cattell, R., & Drevdahl, J. (1955). A comparison of the person-
creativity. New York: Wiley. ality profile (16 P.F.) of eminent researchers with that of
Bentler, P. (1985). Theory and implementation ofEQS: A Struc- eminent teachers and administrators and the general popula-
tural equations program. Los Angeles, CA: BMDP Statistical tion. British Journal of Educational Psychology, 46, 248-
Software, Inc. 261.
Bentler, P. M. (1980). Multivariate analysis with latent vari- Cattell, R., & Eber, H. (1968). Handbookfor the Sixteen Per-
ables. In M. R. Rosenweig & L. W. Porter (Eds.), Annual sonality Factor Questionnaire. Champaign, IL: IPAT.
Christensen, P. R., Merrifield, P. R., & Guilford, J. P. (1958). Fu, V. R., Kelso, G. B., & Moran, J. D., III (1984). The effects
Consequences. Beverly Hills, CA: Sheridan Psychological of stimulus dimension and mode of exploration on original
Services. thinking in preschool children. Educational and Psychologi-
Christensen, P. R., Guilford, J. P., Merrifield, P. R., & Wilson, cal Measurement, 44, 431-440.
R. C. (1960). Alternate Uses. Beverly Hills, CA: Sheridan Getzels, J. W., & Csikszentmihalyi, M. (1964). Creative think-
Psychological Services. ing in art students: An exploratory study. (Cooperative Re-
Covington, M. V. (1966). A childhood attitude inventory for search Project No. E-008). Office of Education, U.S. Depart-
problem solving. Journal of Educational Measurement, 3, ment of Health, Education and Welfare, University of
234. Chicago.
Crockenberg, S. (1972). Creativity tests: Boom or boon-doggie? Gilchrist, M. (1972). The psychology of creativity. Melbourne,
Review of Educational Research, 42, 27-45. Australia: Melbourne University Press.
Cropley, A. J. (1972). A five-year longitudinal study of the Golann, S. E. (1963). Psychological study of creativity.
validity of creativity tests. Developmental Psychology, 6, Psychological Bulletin, 60, 548-565.
119-124. Goldsmith, R. E. (1986). Convergent validity of four inno-
Csikszentmihalyi, M., & Getzels, J. W. (1970). Concern for vativeness scales. Educational and Psychological Measure-
discovery: An attitudinal component of creative production. ment, 46,81-87.
Journal of Personality, 38, 91-105. Goolsby, T., & Helwig, L. (1975). Concurrent validity of the
Davis, G. (1971). Instruments useful in studying creative behav- Torrance Tests of Creative Thinking and the Welsh Figural
ior and creative talent. Journal of Creative Behavior, 5, 162- Preference Test. Educational and Psychological Measure-
165. ment, 35, 507-508.
Davis, G., & Belcher, T. L. (1971). How shall creativity be Gough, H. (1957). The California Psychological Inventory.
measured? Torrance tests, RAT, Alpha Biographical and IQ. Palo Alto, CA: Consulting Psychologists Press.
Journal of Creative Behavior, 5, 153-161. Gough, H. (1976). Studying creativity by means of word asso-
Davis, G., & Rimm, S. (1982). (GIFFI) I and II: Instruments for ciation tests. Journal of Applied Psychology, 61, 348-353.
identifying creative potential in the junior and senior high Gough, H. G. (1979). A creative personality scale for the Adjec-
school. Journal of Creative Behavior, 16, 50-57. tive Check List. Journal of Personality and Social Psychol-
Dellas, M., & Gaier, E. L. (1970). Identification of creativity: ogy, 37, 1398-1405.
The individual. Psychological Bulletin, 73, 55-73. Gough, H. G., & Heilbrun, A. B. (1965). The Adjective Check
Dewing, K. (1970). The reliability and validity of selected tests List: Manual. Palo Alto, CA: Consulting Psychologists Press.
of creative thinking in a sample of seventh-grade West Aus- Graham, W. R. (1965). Creative and constructive idea men and
tralian children. British Journal of Educational Psychology, their participation in activities. The Journal of General Psy-
40,35-42. chology, 72, 383-391.
Domino, G. (1970). Identification of potentially creative per- Guilford, J. P. (1956). The structure of the intellect.
sons from the Adjective Check List. Journal of Consulting Psychological Bulletin, 53, 267-293.
and Clinical Psychology, 35, 48-51. Guilford, J. P. (1968). Intelligence, creativity, and their educa-
Drevdahl, J. E. (1956). Factors of importance for creativity. tional implications. San Diego, CA: Robert R. Knapp.
Journal of Clinical Psychology, 12, 21-26. Guilford, J. P., & Zimmerman, W. S. (1963). Guilford-Zimmer-
Educational Testing Service (1987). Annotated bibliography of man Interest Inventory. Beverly Hills, CA: Sheridan Psycho-
tests: Creativity and divergent thinking. Princeton, NJ: Edu- logical Services.
cational Testing Service. Haddon, F. A., & Lytton, H. (1971). Primary education and
Ellison, B. A. (1973). Creativity in black artists: A comparison divergent thinking abilities-four years on. British Journal of
of selected creativity measures using judged creativity as a Educational Psychology, 41, 136-147.
criterion. Journal of Non-White Concerns in Personnel and Hall, W., & MacKinnon, D. (1969). Personality inventory cor-
Guidance, 1, 150-157. relates of creativity among architects. Journal ofApplied Psy-
Evans, R. G., & Forbach, G. B. (1983). Facilitation of perfor- chology, 53, 322-326.
mance on a divergent measure of creativity: A closer look at Hall, W. B. (1972). A technique for assessing aesthetic pre-
instructions to "be creative." Applied Psychological Mea- dispositions: Mosaic Construction Tests. Journal of Creative
surement, 7, 181-187. Behavior, 6, 225-235.
Fitzgerald, D., & Hattie, J. A. (1983). An evaluation of the Halpin, G., Halpin, G., & Torrance, E. P. (1974). Relationships
"Your Style of Learning and Thinking" inventory. British between creative thinking abilities and a measure of the cre-
Journal of Educational Psychology, 53, 336-346. ative personality. Educational and Psychological Measure-
Flanagan, J. C. (1968). Ingenuity Test. Journal of Creative ment, 34, 75-82.
Behavior, 2, 215-216. Harmon, L. R. (1963). The development of a criterion of scien-
Foster, J. (1971). Creativity and the teacher. London: tific competence. In C. W. Taylor & F. Barron (Eds.),
Macmillan. Scientific creativity: Its recognition and development. New
Frank, G. (1979). On the validity of hypotheses derived from the York: Wiley.
Rorschach: VI. M and the intrapsychic life of individuals. Harrington, D. M. (1975). Effects of explicit instructions to "be
Perceptual and Motor Skills, 48, 1267-1277. creative" on the psychological meaning of divergent thinking
Frederiksen, N., Evans, F., & Ward, W. (1975). Development test scores. Journal of Personality, 43, 434-454.
of provisional criteria for the study of scientific creativity. Harrington, D. M., Block, J., & Block, J. H. (1983). Predicting
Gifted Child Quarterly, 19, 60-65. creativity in preadolescence' from divergent thinking in early
Fricke, B. (1965). Opinion, Attitude, and 1nterest Survey hand- childhood. Journal ofPersonality and Social Psychology, 45,
book: A guide to personality and interest measurement. Ann 609-623.
Arbor, MI: OAIS Testing Program. Harrington, D. M., Block, J. H., & Block, J. (1987). Testing
aspects of Carl Rogers's theory of creative environments: and extracurricular achievement in college. Journal ofEduca-
Child-rearing antecedents of creative potential in young ado- tional Psychology, 55, 55-65.
lescents. Journal of Personality and Social Psychology, 52, Holland, J. L., & Richards, J. (1965). Academic and non-
851-856. academic accomplishment: Correlated or uncorrelated?
Hattie, J. A. (1977). Conditions for administering creativity Journal of Educational Psychology, 56, 165-174.
tests. Psychological Bulletin, 84, 1249-1260. Holmes, D. (1976). A questionnaire measure of the creative
Hattie, J. A. (1980). Should creativity tests be administered personality. Journal of Creative Behavior, 10, 183.
under testlike conditions? An empirical study of three alter- Houtz, J. c., & Speedie, S. M. (1978). Processes underlying
native conditions. Journal of Educational Psychology, 72, diverggnt thinking and problem solving. Journal of Educa-
87-98. tional Psychology, 70, 848-854.
Haven, G. A. (1965). Creative thought, productivity, and the Houtz, J. c., Lewis, C. D., Shaning, D. J., & Denmark, R. J.
self-concept. Psychological Reports, 16, 750-752. (1983). Predictive validity of teacher ratings of creativity over
Heist, P. (Ed.). (1968). The creative college student: An unmet two years. Contemporary Educational Psychology, 8, 168-
challenge. San Francisco: Jossey-Bass. 173.
Heist, P., & Yonge, G. (1968). Manualfor the Omnibus Person- Hurt, H. T., Joseph, K., & Cook, C. D. (1977). Scales for the
ality Inventory-Form E. New York: The Psychological measurement of innovativeness. Human Communication Re-
Corporation. search, 4, 58-65.
Helson, R. (1965). Childhood interest clusters related to Institute for Behavioral Research in Creativity (1968). Alpha
creativity in women. Journal of Consulting Psychology, 29, Biographical Inventory. Greensboro, NC: Prediction Press.
352-361. Jones, C. A. (1964). Relationships between creative writing and
Helson, R. (1971). Women mathematicians and the creative creative drawing of sixth-grade children. In W. L. Brittain
personality. Journal of Consulting and Clinical Psychology, (Ed.), Creativity and art education. Washington, DC: Na-
36, 210-220. tional Art Education Association.
Helson, R., & Crutchfield, R. S. (1970). Mathematicians: The Jordan, L. A. (1975). Use of canonical analysis in Cropley's "A
creative researcher and the average PhD. Journal of Consult- five-year longitudinal study ofthe validity of creativity tests."
ing and Clinical Psychology, 34, 250-257. Developmental Psychology, 5, 117-126.
Hocevar, D. (1976). Dimensions of creativity. Psychological J6reskog, K. (1969). A general approach to confirmatory factor
Reports, 39, 869-870. analysis. Psychometrika, 34, 183-202.
Hocevar, D. (1979a). The unidimensional nature of creative J6reskog, K., & S6rbom, D. (1986). LISREL: Analysis of linear
thinking in fifth-grade children. Child Study Journal, 9, 273- structural relations by the method of maximum likelihood.
277. Chicago: International Educational Resources.
Hocevar, D. (1979b). Ideational fluency as a confounding factor Kaltsounis, B. (1971). Instruments useful in studying creative
in the measurement of originality. Journal of Educational behavior and creative talent. Journal of Creative Behavior, 6,
Psychology, 71, 191-196. 268-274.
Hocevar, D. (1979c). A comparison of statistical infrequency, Kaltsounis, B. (1972). Additional instruments useful in studying
subjective judgment, and random numbers as criteria in the creative behavior and creative talent. Journal of Creative Be-
measurement of originality. Journal of Personality Assess- havior, 6, 268-274.
ment, 43,297-299. Karlins, M., Schuerhoff, C., & Kaplan, M. (1969). Some fac-
Hocevar, D. (1979d). The development of the Creative Behavior tors related to architectural creativity in graduating architec-
Inventory. Los Angeles, CA: University of Southern Califor- ture students. Journal of General Psychology, 81, 203-215.
nia. (ERIC Documentation Reproduction Service No. ED 170 Keller, R. T., & Holland, W. E. (1978). A cross-validation
350). study of the Kirton Adaptation-Innovation inventory in three
Hocevar, D. (1980). Intelligence, divergent thinking and research and development organizations. Applied Psychologi-
creativity. Intelligence, 4, 25-40. cal Measurement, 2, 563-570.
Hocevar, D., & El-Zahhar, N. (1984). A paradigm for examin- Khatena, J. (1978). Identification and stimulation of creative
ing the psychometric characteristics of cross-cultural mea- imagination imagery. Journal of Creative Behavior, 12, 30-
surements. In H. M. van der Ploeg, R. Schwarzer, & C. D. 38.
Spielberger, Advances in test anxiety research (Vol. 4). New Khatena, J., & Torrance, E. P. (1976). Manual for Khatena-
York: Erlbaum. Torrance Creative Perception Inventory. Chicago: Stoelting.
Hocevar, D., & Michael, W. B. (1979). The effects of scoring Kirton, M. J. (1976). Adaptors and innovators: A description
formulas on the discriminant validity of tests of divergent and measure. Journal of Applied Psychology, 61, 622-29.
thinking. Educational and Psychological Measurement, 39, Knapp, R. H., & Wulff, A. (1963). Preference for abstract and
917-921. representational art. Journal of Social Psychology, 60, 255-
Holland, J. L. (1959). Some limitations of teacher ratings as 262.
predictors of creativity. Journal of Educational Psychology, Kogan, N., & Pankove, E. (1974). Long-term predictive valid-
50, 219-223. ity of divergent-thinking tests: Some negative evidence.
Holland, J. L., & Astin, A. W. (1962). The prediction of the Journal of Educational Psychology, 66, 802-810.
academic, artistic, scientific, and social achievement of un- Korb, R., & Frankiewicz, R. (1976). Strategy for a priori
dergraduates of superior scholastic aptitude. Journal of Edu- selection of judges in a product -centered approach to assess-
cational Psychology, 53, 132-143. ment of creativity. Perceptual and Motor Skills, 42, 107-115.
Holland, J. L., & Baird, L. L. (1968). The Preconscious Ac- Lacey, L., & Erickson, C. (1974). Psychology of scientist:
tivity Scale: The development and validation of an originality XXXI. Discriminability of creativity scale for the Adjective
measure. Journal of Creative Behavior, 2,217-225. Check List among scientists and engineers. Psychological Re-
Holland, J. L., & Nichols, R. (1964). Prediction of academic ports, 34, 755-758.
Leavitt, c., & Walton, J. R. (1975). Development of a scale for Mischel, W. (1968). Personality and assessment. New York:
innovativeness. In M. J. Schlinger (Ed.), Advances in con- Wiley.
sumer research (Vol. 2). Ann Arbor, MI: Association for Mullins, C. J. (1964). Current studies of the personnel research
Consumer Research. laboratory in creativity. In C. W. Taylor (Ed.), Widening
Lynch, M., & Kaufman, M. (1974). Creativeness: Its meaning horizons in creativity. New York: Wiley.
and measurement. Journal ofReading Behavior, 6, 375-394. Mumford, M. D., & Gustafson, S. B. (1988). Creativity syn-
MacKinnon, D. (1962). The nature and nurture of creative tal- drome: Integration, application and innovation. Psycholo-
ent. American Psychologist, 17, 484-495. gical Bulletin, 103, 27-43.
MacKinnon, D. (1968). Selecting students with creative poten- Owens, W. A., & Schumacher, C., & Clark, J. (1957). The
tial. In P. Heist (ed.), The creative college student: An unmet measurement of creativity in machine design. Journal of Ap-
challenge. San Francisco: Jossey-Bass. plied Psychology, 41, 297-302.
McCrae, R. R. (1987). Creativity, divergent thinking, and open- Pelz, D. C. (1963). Relationships between measures of scientific
ness to experience. Journal ofPersonality and Social Psychol- performance and other variables. In C. W. Taylor & F. Barron
ogy, 52, 1258-1265. (Eds.), Scientific creativity: Its recognition and development.
McDermid, C. D. (1965). Some correlates of creativity in en- New York: Wiley.
gineering personnel. Journal of Applied Psychology, 49, 14- Petrosko, J. (1978). Measuring creativity in elementary school:
19. The current state of the art. Journal of Creative Behavior, 12,
Malgady, R. G., & Barcher, P. R. (1977). Psychological scaling 109-119.
of essay creativity: Effects of productivity and novelty. Piers, E. V., Daniels, J. M., & Quackenbush, J. F. (1960). The
Journal of Educational Psychology, 69, 512-518. identification of creativity in adolescents. Journal of Educa-
Malgady, R. G., & Barcher, P. R. (1979). Some information- tional Psychology, 51, 346-351.
processing models of creative writing, Journal ofEducational Popperova, M. (1971, 1972). Some methodological problems
Psychology, 71,717-725. relating to tests of creative thinking. [Czskoslovenska
Marsh, H. (in press). Self-Description Questionnaire (SDQ): 111. Psycholgie, 15(4), 391-397.] Psychological Abstracts, 49,
San Antonio, Texas: Psychological Corporation. No. 5666.
Marsh, H., & Hocevar, D. (1983). Confirmatory factor analysis Reid, J., King, F., & Wickwire, P. (1959). Cognitive and other
of multitrait-multimethod matrices. Journal of Educational personality characteristics of creative children. Psychological
Measurement, 20, 231-248. Reports, 5, 729-737.
Marsh, H., & Hocevar, D. (1985). The application of confir- Renzulli, J., Hartman, R., & Callahan, C. (1977). Scale for
matory factor analysis to the study of self-concept: First and rating the behavioral characteristics of superior students. In
higher order factor structures and their invariance across age W. Barbe & J. Renzulli (Eds.), Psychology and education of
groups. Psychological Bulletin, 97, 562-582. the gifted (2nd ed.). New York: Irvington.
Marsh, H., & Hocevar, D. (1988). A new, more powerful meth- Reynolds, C. R., Kaltsounis, B., & Torrance, E. P. (1979). A
od of multitrait-multimethod analysis. Journal of Applied children's form of "Your Style of Learning and Thinking":
Psychology, 73, 107-117. Preliminary norms and technical data. Gifted Child Quarterly,
Mednick, M. T. (1963). Research creativity in psychology grad- 23, 757-767.
uate students. Journal of Consulting Psychology, 27, 265- Richards, J. M., Cline, V. B., & Needham, W. E. (1964).
266. Creativity tests and teacher and self judgments of originality.
Mednick, S. A., & Mednick, M. T. (1967). Examiner's manual: Journal of Experimental Education, 32, 281-285.
Remote Associates Test. Boston: Houghton Mifflin. Richards, J. M., Holland, J. L., & Lutz, S. W. (l967a). Predic-
Meeker, M., & Meeker, R. (1975). SOl Learning Abilities test. tion of student accomplishment in college. Journal of Educa-
El Segundo, CA: SOl Institute. tional Psychology, 58, 343-355.
Meer, B., & Stein, M. (1955). Measures of intelligence and Richards, J. M., Holland, J. L., & Lutz, S. W. (l967b). Assess-
creativity. The Journal of Psychology, 39, 117-126. ment of student accomplishment in college. The Journal of
Michael, W. B., & Colson, K. R. (1979). The development and College Student Personnel, 8, 360-365.
validation of a life experience inventory for the identification Richards, R., Kinney, D. K., Benet, M., & Merzel, A. P. C.
of creative electrical engineers. Educational and Psychologi- (1988). Assessment of everyday creativity: Characteristics of
cal Measurement, 39, 463-470. the lifetime creativity scales and validation with three large
Milgram, R. M. (1983). Validation of ideational fluency mea- samples. Journal of Personality and Social Psychology, 54,
sures of original thinking in children. Journal of Educational 476-485.
Psychology, 75, 619-624. Rickborn, I., & Lundsteen, S. (1968). The construction of and
Milgram, R. M., & Arad, R. (1981). Ideational fluency as a acquisition of reliability data for a test of qualitative levels in
predictor of original problem solving. Journal of Educational creative problem solving. California Journal of Educational
Psychology, 73, 568-572. Research, 19, 53-58.
Milgram, R. M., & Milgram, N. A. (l976a). Creative thinking Rimm, S., & Davis, G. (1976). GIFT: An instrument for the
and creative performance in Israeli students. Journal of Edu- identification of creativity. Journal of Creative Behavior, 10,
cational Psychology, 68, 255-259. 178-182.
Milgram, R. M., & Milgram, N. A. (I 976b). Group versus Rimm, S., & Davis, G. (1980). Five years of international re-
individual administration in the measurement of creative search with GIFT: An instrument for the identification of
thinking in gifted and nongifted children. Child Development, creativity. Journal of Creative Behavior, 14, 35-46.
47, 563-565. Roe, A. (l951a). A psychological study of physical scientists.
Milgram, R. M., Milgram, N. A., Rosenbloom, G., & Rabkin, Genetic Psychological Monographs, 43, 121-235.
L. (1978). Quantity and quality of creative thinking in chil- Roe, A. (1951b). A psychological study of eminent biologists.
dren and adolescents. Child Development, 49, 385-388. Psychological Monographs, 65 (No. 331).
Rookey, T. 1. (1971). The Pennsylvania assessment of creative prediction of academic and artistic achievement at a school of
tendency: Norms-technical manual. Harrisburg: Pennsylvania design. Journal of Educational Measurement, 4, 105-117.
Department of Education. Smith, 1., & Schaefer, C. (1969). Development of a creativity
Rookey, T. 1. (1974). Validation of a creativity test: The 100 scale for the Adjective Check List. Psychological Reports, 25,
students study. Journal of Creative Behavior, 8, 211-213. 87-92.
Rossman, B. B., & Gollob, H. F. (1975). Comparison of social Sobel, R. S., & Rothenberg, A. (1980). Artistic creation as
judgments of creativity and intelligence. Journal of Person- stimulated by superimposed versus separated visual im-
ality and Social Psychology, 31, 271-281. ages. Journal of Personality and Social Psychology, 39,
Rotter, D. M., Langland, L., & Berger, D. (1971). The validity 953-961.
of tests of creative thinking in seven-year-old children. Gifted Specher, T. B. (1964). Creativity and individual differences in
Child Quarterly, 15, 273-278. criteria. In C. W. Taylor (Ed.), Widening horizons in
Roweton, W. E., Farless, 1. E., Donham, R., Wleklinski, D. 1., creativity. New York: Wiley.
& Spencer, H. L. (1975). Indices of classroom creativity. Starkweather, E. K. (1971). Creativity research instruments de-
Child Study Journal, 5, 151-162. signed for use with preschool children. The Journal of Cre-
Runco, M. A. (1984). Teachers' judgments of creativity and ative Behavior, 5, 245-255.
social validation of divergent thinking tests. Perceptual and Sternberg, R. 1. (1985). Implicit theories of intelligence,
Motor Skills, 59,711-717. creativity, and wisdom. Journal of Personality and Social
Runco, M. A. (l986a). Divergent thinking and creative perfor- Psychology, 49, 607-627.
mance in gifted and nongifted children. Educational and P sy- Taylor, C. W. (1958). Some variables functioning in productivi-
chological Measurement, 46, 375-384. ty and creativity. In C. W. Taylor (Ed.), The second (1957)
Runco, M. A. (l986b). Maximal performance on divergent University of Utah research conference on the identification of
thinking tests by gifted, talented, and nongifted children. creative scientific talent. Salt Lake City: University of Utah
Psychology in the Schools, 23, 308-315. Press.
Runco, M. A.", & Albert, R. S. (1985). The reliability and valid- Taylor, C. W., Smith, W. R., & Ghiselin, B. (1963). The cre-
ity of ideational originality in the divergent thinking of aca- ative and other contributions of one sample of research scien-
demically gifted and nongifted children. Educational and Psy- tists. In C. W. Taylor & F. Barron (Eds.), Scientific creativity:
chological Measurement, 45, 483-501. Its recognition and development. New York: Wiley.
Schaefer, C. (1969). The prediction of creative achievement Taylor, I. A., & Fish, R. A. (1979). The Creative Behaviour
from a biographical inventory. Educational and Psychologi- Disposition scale: A Canadian validation. Canadian Journal
cal Measurement, 29, 431-437. of Behavioral Science, 11, 95-97.
Schaefer, C. (1970). Manual for the Biographical Inventory Taylor, I. A., Sutton, D., & Haworth, S. (1974). The measure-
Creativity (BIC). San Diego, CA: Educational and Industrial ment of creative transactualization: A scale to measure behav-
Testing Service. ioral dispositions to creativity. Journal of Creative Behavior,
Schaefer, C., & Anastasi, A. (1968). A biographical inventory 8, 114-115.
for identifying creativity in adolescent boys. Journal of Ap- Taylor, W., & Ellison, R. (1964). Predicting creative perfor-
plied Psychology, 52, 42-48. mances from mUltiple measures. In C. W. Taylor (Ed.),
Schaefer, C. E., & Bridges, C. I. (1970). Development of a Widening horizons in creativity. New York: Wiley.
creativity attitude survey for children. Perceptual and Motor Thompson, B., & Andersson, B. V. (1983). Construct validity
Skills, 31, 861-862. of the divergent production subtests from the Structure-of-
Seddon, G. M. (1983). The measurement and properties of di- Intellect Learning Abilities Test. Educational and Psycholog-
vergent thinking ability as a single compound entity. Journal ical Measurement, 43, 651-655.
of Educational Measurement, 20, 393-402. Torrance, E. P. (1962). Guiding creative talent. Englewood
Simonton, D. K. (l976a). Biographical determinants of Cliffs, Nl: Prentice-Hall.
achieved eminence: A multivariate approach to the Cox data. Torrance, E. P. (1969). Prediction of adult creative achievement
Journal of Personality and Social Psychology, 33, 218-226. arnong high school seniors. Gifted Child Quarterly, 13, 223-
Simonton, D. K. (l976b). The causal relation between war and 229.
scientific discovery: An exploratory cross-national analysis. Torrance, E. P. (1974). Torrance Tests of Creative Thinking:
Journal of Cross-Cultural Psychology, 7, 133-144. Norms-technical manual. Princeton, Nl: Personnel Press/
Simonton, D. K. (l976c). Philosophical eminence, beliefs, and Ginn.
zeitgeist: An individual-generational analysis. Journal ofPer- Torrance, E. P., Reynolds, C. R., Riegel, T., & Ball, O. E.
sonality and Social Psychology, 34, 630-640. (1977). "Your Style of Learning and Thinking" forms A and
Simonton, D. K. (l977a). Creative productivity, age and stress: B: Preliminary norms, abbreviated technical notes, scoring
A biographical time-series of 10 classical composers. Journal keys, and selected references. Gifted Child Quarterly, 21,
of Personality and Social Psychology, 35, 791-804. 563-573.
Simonton, D. K. (l977b). Eminence, creativity and geograph- Treffinger, D. 1., & Poggio, 1. P. (1972). Needed research on
ical marginality: A recursive structural equation model. the measurement of creativity. Journal of Creative Behavior,
Journal of Personality and Social Psychology, 35, 805-816. 6, 253-267.
Simonton, D. K. (1979). Multiple discovery and invention: Treffinger, D. 1., Renzulli, 1. S., & Feldhusen, 1. F. (1971).
Zeitgeist, genius, or chance? Journal of Personality and So- Problems in the assessment of creative thinking. Journal of
cial Psychology, 37, 1603-1616. Creative Behavior, 5, 104-112.
Skager, R. W., Schultz, C. B., & Klein, S. P. (1965). Quality Tryk, H. E. (1968). Assessment of creativity. InP. McReynolds
and quantity of accomplishments as measures of creativity. (Ed.), Advances in psychological assessment (Vol. 1). Palo
Journal of Educational Psychology, 56, 31-39. Alto, CA: Science and Behavior Books.
Skager, R. W., Klein, S. P., & Schultz, C. B. (1967). The Wallach, M. A., & Kogan, N. (1965). Modes of thinking in
young children: A study of the creativity-intelligence distinc- Yamamoto, K. (1963). Relationships between creative thinking
tion. New York: Holt, Rinehart & Winston. abilities of teachers and achievement of pupils. Journal of
Wallach, M. A., & Wing, C. (1969). The talented student: A Experimental Education, 32, 3-25.
validation of the creativity-intelligence distinction. New Yamamoto, K. (1964a). Creativity and sociometric choice
York: Holt, Rinehart & Winston. among adolescents. Journal of Social Psychology, 64, 249-
Wallbrown, F. H., & Huelsman, C. B. (1975). The validity of 261.
the Wallach-Kogan creativity operations for inner-city chil- Yamamoto, K. (1964b). Evaluation of some creativity measures
dren in two areas of visual art. Journal of Personality, 43, in a high school with peer nominations as criteria. The Journal
109-126. of Psychology, 58, 285-293.
Wallen, N. E., & Stevenson, G. M. (1960). Stability and corre- Zamegar, Z., Hocevar, D., & Michael, W. (1988). Components
lates of judged creativity in fifth grade writing. Journal of of original thinking in gifted children. Educational and Psy-
Educational Psychology, 51,273-276. chological Measurement, 48,5-16.
Ward, W. c., & Cox, P. W. (1974). A field study of nonverbal Zegas, J. (1976). A validation study of tests from the divergent
creativity. Journal of Personality, 42, 202-209. production plane of the Guilford Structure-of-Intellect model.
Welsh, G. (1959). Preliminary manual for the Welsh Figure Journal of Creative Behavior, 10, 170-177.
Preference Test .. Palo Alto, CA: Consulting Psychologists Zimmer, J. W., Guip, D., & Hocevar, D. (April 1988).
Press. Assessment of creativity-A longitudinal evaluation. Paper
Welsh, G. S. (1977). Personality correlates of intelligence and presented at the annual meeting of the American Educational
creativity in gifted adolescents. In 1. C. Stanley, W. C. Research Association, New Orleans, Louisiana.
George, & c. H. Solano (Eds.), The gifted and the creative: A
fifty-year perspective. Baltimore, MD: Johns Hopkins Uni-
versity Press.
CHAPTER 4
Individual Differences in Creativity

AN INTERACTIONIST PERSPECTIVE
Richard W. Woodman and Lyle F. Schoenfeldt
Creativity seems to be one of those concepts under- and social psychology perspectives on creativity.
stood by everyone in the world except behavioral The proposed interactionist model provides a
scientists. Although some segments of the public framework to explore in some detail a number of
might hypothesize other reasons for this state of factors thought to be of importance in explaining
affairs, we believe the reason for this seeming para- individual differences in creative behavior. Finally,
dox is as simple as the difference between the terms some conclusions and implications of our the-
concept and construct. As a concept used by layper- oretical perspective will be presented.
sons, creativity carries meaning in everyday speech
that, although somewhat imprecise, is nevertheless
widely shared; any surplus meaning is relatively Theoretical Explanations for
unimportant; and operationalization for measure- Differences in Creative Behavior
ment purposes is a nonissue. As a scientific con-
struct, however, creativity is held to a higher (or, at The potential sources for individual differences
least, different) standard, and the construct validity in creativity are legion. Any categorization scheme
issues surrounding the term can be frustrating in the is likely to be flawed; nevertheless, a potentially
extreme for researchers interested in investigating useful starting point is provided by an examination
the phenomena of creative behavior and creative of creativity from the perspective of (a) personality
persons. differences, (b) cognitive style or ability dif-
How best to understand creativity? Many disci- ferences, and (c) social psychology. Each of these
plines in the behavioral and social sciences have perspectives on creative behavior will be examined
provided perspectives that may be, in some mea- in tum.
sure, useful. Several of these perspectives will be
briefly reviewed here, pursuant to positing an in-
teractionist model of creative behavior that attempts Personality
to combine elements of the personality, cognitive, The study of personality characteristics associ-
ated with creative behavior has been a generally
Richard W. Woodman and Lyle F. Schoenfeldt • Depart- active area of research for some time, although such
ment of Management, Texas A&M University, College Station, research has waxed and waned in popularity over
TX 77843. the years (Helson & Mitchell, 1978). The research
77
and writing on personality factor-creative behavior In general, a fairly stable set of core cha;acteristics (e.g., high
relationships has tended to emerge in one of three valuation of esthetic qualities in experience, broad interests,
ways: (a) attempts by personality theorists to ex- attraction to complexity, high energy, independence of judge-
ment, autonomy, intuition, self-confidence, ability to resolve or
plain creativity in terms of comprehensive theories accommodate apparently opposite or conflicting traits in one's
of personality; (b) investigations regarding the per- self concept, and finally, a firm sense of self as "creative")
sonality and biographical characteristics of eminent continued to emerge as correlates of creative achievement and
creative individuals and/ or creative activity in a va- activity in many domains. (p. 453)
riety of fields; and (c) more narrowly focused work
that examines one or a few specific personality di- Despite such convergence in the research literature,
mensions for possible relationships to creative the generalizability of any specific constellation of
behavior. traits across fields of endeavor remains highly
First, creativity as a construct has seemed to pre- problematic.
sent a special challenge for personality theory, and Finally, a considerable body ofliterature has ac-
many well-known personality theorists have made cumulated concerning specific personality dimen-
attempts to incorporate an explanation for creative sions related to creative behavior. Examples of
behavior within their theoretical positions (Taylor, heavily-researched traits which are thought to be
1976; Woodman, 1981). The need for personality important factors in creativity include locus of con-
theory to explain creativity stems from its role as a trol(e.g., Bolen & Torrance, 1978), psychological
general theory of behavior. As such, a personality femininity and masculinity (e.g., Barron & Har-
theory that did not account for the creative act rington, 1981, p. 458), self-esteem or identity
would seem incomplete, at least in the eyes of some (e.g., Dellas, 1978), dogmatism (Faschingbauer,
critics. Of course, there is a great divergence across Moore, & Stone, 1978), and narcissism (Solomon,
theories with regard to explanations of individual 1985) among many others. Trait-specific research
differences in creativity. This divergence can be has both benefited from and contributed to the re-
traced, in part, to fundamental differences in per- search on creative persons discussed above, and, in
spective regarding the nature of human beings and addition, has close ties to the next major perspective
their behavior that exist within various streams of to be discussed-explanations of creativity based
psychological thought or research "traditions," on differences in cognitive styles or abilities.
that is, cognitive, humanistic, psychoanalytic, and
behavioristic (Woodman, 1981).
Cognitive Style/Ability
A second major area of research within the per-
sonality framework has been the study of charac- Whether differences in creativity are best
teristics of creative persons. Good examples of in- explained as a function of the personality or as a
vestigations in this area are provided by the writings function of differences in cognitive styles, prob-
of Barron (1969), Helson (1971), MacKinnon lem-solving approaches, or abilities continues to be
(1970), Roe (1953), and Simonton (1977, 1986). debated (e.g., Eysenck, 1983). Without dismissing
These types of studies have attempted to catalogue the importance of personality factors, a perusal of
personality correlates of creative productivity as the literature suggests that much of the current re-
well as biographical data that might be predictive of search focus seems to have shifted to explorations
later creative behavior. Researchers have investi- of cognitive ability-creative behavior rela-
gated similarities and differences in creativity tionships. Cognitive factors thought to have impor-
across broad fields of endeavor, such as art, liter- tant relationships to creativity include cognitive
ature, music, and science. In addition, many re- styles, such as field independence/dependence
searchers have focused on examining individual (e.g., Noppe & Gallagher, 1977; Spotts & Mack-
differences across more narrowly defined disci- ler, 1967), "creative" (e.g., "lateral," "refram-
plines, such as architecture, physics, and mathe- ing") thinking or problem-solving styles (Kershner
matics (e.g., Weiss, 1981). In the most general & Ledger, 1985; Noppe, 1985), cognitive com-
sense, this research on creative persons and their plexity (e.g., Quinn, 1980), divergent thinking or
creative products has delineated a set of "core char- "production" (Guilford, 1967), ability to link re-
acteristics" that are widely regarded as typifying mote "associations" among elements or ideas to
the "creative personality." For example, after re- achieve a creative solution (e.g., Mednick, 1962),
viewing 15 years of research on personality charac- ideational fluency (e.g., Basadur & Thompson,
teristics of creative individuals, Barron and Har- 1986; Carroll & Maxwell, 1979, pp. 615-616), and
rington (1981) concluded: imagery and verbal fluency (e.g., Suler &
CHAPTER 4 • INDIVIDUAL DIFFERENCES IN CREATIVITY 79
Aizziello, 1987). Many of these cognitive attributes on the importance of internal determinants of
or abilities are not sharply differentiated from each creativity, but much less emphasis on external de-
other, so additional work in modeling creative- terminants; a concentration on creative persons, but
thinking processes might be particularly instructive generally little appreciation for "creative situa-
(e.g., Cagle, 1985). One ofthe better-known mod- tions" or circumstances that might be conducive to
els in this regard is Guilford's (1967) "Structure- creative behavior. In contrast, the social psychol-
of-Intellect" (SOl) model. In Guilford's model, in- ogy of creativity seeks to understand and explain
telligence is defined as a collection of abilities or how particular social and environmental conditions
functions for processing information. Although might influence the creative behavior of individuals
many different mental functions relate to creativity, (Amabile, 1983b, p. 5).
the operation of "divergent production" is seen as Certainly, the important role of social interaction
being particularly critical for creative behavior. in creativity has long been recognized (cf., Hare,
This cognitive style or ability-perhaps identical to 1982). The groups in which an individual partici-
what other writers have called ideational fluency, pates-from the earliest family experiences to later
adaptive flexibility, or the ability to generate logical friendship and work groups-clearly have some
alternatives-provides a good example of a cog- impact on behavior, including creative behavior. A
nitive ability-creative behavior relationship that is reasonable number of studies have investigated the
empirically well supported. effects of particular social and physical environ-
An extreme position is staked out by Eysenck ments on creativity (e.g., Getzels & Jackson, 1961;
(1983) who argues that creativity and originality are Goyal, 1973; Klein, 1975; Torrance, 1965). Simi-
not aspects of operations of the mind, but instead larly, research on creativity and innovation in
are exclusively traits of personality and, as such, organizational settings has examined the effects of
are essentially noncognitive. A more common posi- various contingencies in the organizational environ-
tion seems to be to treat both personality and cog- ment, including salient work group characteristics
nitive factors as potentially important sources of that might foster or inhibit creative behavior (Staw,
individual differences in creativity (e.g., Arieti, 1984, pp. 655-657; Steiner, 1965; Woodman,
1976). Indeed, many examples of overlap between 1983). Much of the biographical research (alluded to
the personality and cognitive style perspectives on earlier) investigating the social and family back-
creativity can be found. For example, psycho- grounds of creative persons has informed both the
analytic theory places great emphasis on the critical personality and the social psychological perspec-
role of "primary process thinking" in creative be- tives on creativity. Despite this body of work, the
havior. Suler (1980) describes this role as follows: social psychology of creativity is probably the-
oretically less well developed than either the person-
The creative act can be conceptualized as a special form of ality or cognitive style perspectives.
interaction between primary and secondary process thinking in Possibly the most comprehensive social psycho-
which a novel idea or insight is generated by the loose, illogical,
and highly subjective ideation of primary process and is then logical explanations for creative behavior have
molded by secondary process into a context that is socially ap- been advanced by Amabile (1983ab). She has pro-
propriate and meaningful to others. (p. 144) posed, and investigated, a number of social and
environmental influences on creative behavior,
This is essentially a cognitive explanation of such as social facilitation, modeling, motivational
creativity, although most explanations of creative orientation, evaluation expectations, effects of ac-
behavior from the psychoanalytic tradition would tual evaluations, use of rewards for creative behav-
fit most clearly within the personality framework. ior, task constraints, and opportunities for behav-
In any event, the notion that creativity dif- ioral choices. A number of these social and
ferences can be at least partially explained by some environmental influences on creativity have been
cognitive aspects seems reasonably well estab- incorporated into the interactionist model of cre-
lished. Researchers are in general agreement that ative behavior presented next.
some cognitive operations are more likely to lead to
original ideas and to solutions to problems.
An Interactionist Model of Creative
Behavior
The Social Psychology of Creativity
In the overwhelming majority of psychological Imagine a bird in a cage. This situation contains
research on creativity, there has been a strong focus both the idea of "bird" and an easily definable
current reality within which the bird finds itself. In larly, cognitive explanations focus primarily on
this instance, the environmental press explains CS-O-B linkages. The social psychology perspec-
most, if not all, of the bird's behavior. Having active, which in many respects may be closest to the
counted for 100% of the variance with the situation, interactionist perspective (cf., Amabile, I983a),
is there anything left to say regarding the bird and its nevertheless might be described as showing the
behavior? Indeed, there is; in fact, perhaps every- most interest in explaining SI-O-B linkages, with
thing of importance remains to be explained. In somewhat lessened interest in CI-O-B or even A-
other words, a careful description of the situation SI-O-B and A-CI-O-B linkages. Going somewhat
and the bird's responses to its environment do not further, developmental psychologists, educational
begin to explain the behavior of all birds, all of the psychologists, and others interested in "gifted chil-
time. What other things might the bird do if the dren" pay a great deal of attention to A-O-B, A-SI-
situation changed? Why might the bird do these O-B, and A-CI-O-B linkages, as does the pre-
things? For starters, we could imagine (armed as we viously discussed research that examines the social
are with other information) that the bird may well and family background of creative individuals.
fly if the environmental constraint of "cage" is Much of the (limited) research on creativity and
removed. Reasoning in this fashion helps to reveal innovation in organizational settings could be de-
some of the advantages found in an interactionist scribed as being primarily concerned with the CI-0-
perspective on behavior. From an interactionist per- B chain. Many "stage" models of the creative pro-
spective, the behavior of an organism at any point in cess seem to focus strongly on O-B-C-O linkages
time is a complex interaction of the situation and and on reciprocal influences and changes in this
something else-this something else is the nature relationship over time. Finally, an operant-condi-
of the organism itself. Both situation and organism tioning explanation of creative behavior (e.g.,
and the interaction that unfolds over time must be Skinner, 1974) might be conceptualized as empha-
explained to fully understand the organism-in-its- sizing [A, CI, SI]-B-C linkages.
environment. Sometimes the contingencies of the A true interactionist explanation of creative be-
current situation account for or can be said to ex- havior contains all of these linkages as shown in
plain the greater part of current behavior; some- Figure 1. Thus, a potential advantage of the interac-
times the nature of the organism explains a great tionist model of creative behavior might be its abil-
deal; sometimes both plus their reciprocal influ- ity to integrate these diverse perspectives, each of
ences are necessary to even begin to understand which perhaps captures variables of some explana-
what is going on. From an interactionist position, tory power. Combining personality, cognitive, and
there is always something more to understanding social psychology explanations of individual dif-
behavior than just describing the observed behavior ferences in creative behavior could serve to im-
per se. This "something more" has to do with the prove our ability to understand creative persons,
essence of the organism and its behavioral processes, and products. A number of the variables
potentiality. contained in Figure 1 will be examined more close-
Figure 1 posits an interactional model of creative ly in the following section of this chapter.
behavior. The interactionist perspective has great
promise for explaining human behavior in complex
social settings, and interactional psychology pro- Factors Underlying Individual
vides a strong theoretical base from which to model Differences in Creativity
complex behavioral phenomena (Schneider, 1983;
Terborg, 1981). Figure 1 incorporates important Creativity is not a single, unitary characteristic,'
elements of the personality, cognitive, and social but instead can be thought of as an imprecise cate-
psychology explanations of creativity. gory of behavior (Barron & Harrington, 1981). To
Similar to a group of blind men describing an quote Snow (1986), "Creativity is not a light bulb
elephant, researchers from differing disciplines and in the mind, as most cartoons depict it. It is an
theoretical perspectives seem to emphasize differ- accomplishment born of intensive study, long re-
ent aspects of Figure 1 in their attempts to explain flection, persistence, and interest" (p. 1033). How-
creativity. For example, researchers and theorists ever defined, people differ in the extent to which
who focus primarily on personality difference ex- they exhibit creativity (Nicholls, 1972). The major
planations of creativity might be said to emphasize categories of variables delineated by the interac-
most heavily the P-O-B linkages of Figure 1. Simi- tionist model shown in Figure I-antecedent con-
Antecedent Person Situation Behavior Consequences
A Antecedent Conditions B Creative Behavior C Consequences
Examples:
past reinforcement history;
early socialization;
biographical variables -- sex,
family position, birth order
o "Organism" (person)
Gestalt of attitudes; values; intentions to behave; motivational orientations; and individual

differences.
CS = Cognitive Style/Abilities P = Personality DlmenslonslTralts

Examples: Examples:
Cognitive complexity Locus of control
Divergent thinking Dogmatism
VerbaVideational fluency Autonomy
Problem-Solving styles/approaches Self-esteem
Perceptual openness Narcissism
Field independence/dependence Intuition
CI Contextual Influences SI Social Influences
Examples: Examples:
Physical environment Social facilitation
Culture Evaluation expectalions
Group/organization ·climate· Rewards/punishments
Task and time constraints Role modeling
Figure 1. An interactionist model of creative behavior.
ditions, cogmtIve style/abilities, personality, ing), early socialization, and background character-
contextual and social influences-provide a istics.
framework to examine some of the specific factors Researchers have long been interested in the rela-
that account for or explain these differences in cre- tionship between relatively immutable birth charac-
ative behavior or the capability to produce creative teristics of individuals as explanatory factors in sub-
products. sequent creativity. For example, the question of
male versus female superiority in creative abilities
has been of research interest for several decades.
Antecedent Conditions
Studies show male superiority, female superiority,
Antecedent conditions refer to individual back- and no differences between the sexes, depending on
ground characteristics that result in differences in age and the procedure for measuring creativity
creativity. The list of antecedent conditions that af- (Gupta, 1981). Both Gupta (1981) and Richardson
fect creativity is potentially lengthy, and includes (1986) found similar results among Indian and Ja-
such things as past reinforcement history (or learn- maican adolescents, respectively. Sex differences
were not found to be significant in the case of scores center located in a housing project for welfare fami-
on verbal or nonverbal creativity. In the Indian lies (the disadvantaged group) with respect to fluen-
study, Gupta (1981) observed that boys showed su- cy and originality on verbal, kinetic, and integrative
periority on verbal fluency, verbal flexibility, and methods of expression. Results were consistent
verbal transformations. Girls scored significantly with previous studies in finding that lower so-
higher on nonverbal dimensions, such as origi- cioeconomic status subjects were more kinetically
nality, complexity, and productive designing abil- creative whereas middle-class subjects excelled
ity. The only difference between the Jamaican stu- verbally. In addition, the disadvantaged generated a
dents was a significantly higher score for girls on larger number of, but fewer original, solutions in
the test of verbal fluency (Richardson, 1986). In a studying creativity expressed through the inte-
further study involving 110 Indian ninth-grade sci- grative mode.
ence students, Raina (1980) found no difference The relationship between handedness and
between the sexes on scores of fluency, flexibility, creativity has received much attention (Katz, 1980;
and originality from the Torrance Tests of Creative Peterson & Lansky, 1974), mainly as a result of the
Thinking. Kershner and Ledger (1985) examined observation that many well-known creative ge-
the effects of sex on creativity among average and niuses have been left-handed. Recent examinations
gifted children and found that average and gifted of the question of handedness as an explanatory
girls scored significantly higher than boys on verbal construct in creativity (Hattie & Fitzgerald, 1983;
and figural fluency. Katz, 1980) have found little relationship. Hattie
It is possible that some of the sex-race dif- and Fitzgerald (1983) examined the relationship be-
ferences on creativity tests requiring interpretation tween handedness and some typical tests of
may be a function of examiner bias. Grim and Tor- creativity using a battery of tests administered to
rance (1977) assigned individuals trained at scoring 103 eighteen-year-olds. Handedness was assessed
and interpreting the Torrance Tests of Creative using the Edinburgh Handedness Scale and, as a
Thinking to one of four groups based on race result, 29 of the subjects were classified as left-
(black/white) and sex (female/male) combinations. handed, 22 integrated or of mixed-handedness, and
Each individual was given the same test record. To 52 right-handed. No differences were found with
quote from Grim and Torrance: respect to creativity.
Another research approach in relating antecedent
Analysis of the free descriptive materials revealed biases favor- conditions to creativity has been biographical inqui-
ing whites as more original, emotionally sensitive, intelligent, ry-historical information obtained on eminent
and quick. Boys were described more frequently than girls as
being aware of the environment and socially skillful.
creators. The original study was Galton's (1869)
The data from the Ideal Child Checklist depicted blacks as Hereditary Genius which became the prototype for
more negative, less independent in judgment, less quiet, more the "historiometric" approach. Other studies in-
talkative, more bored, more fearful, and less intuitive. Females clude The Early Mental Traits of Three Hundred
were described as more timid and shy, less willing to take risks, Geniuses by Cox (1926) which is a volume in Ter-
less self-initiating, less preference for complex tasks, less truth-
ful, more obedient and submissive, more willing to acceptjudg- man's monumental Genetic Studies of Genius.
ments of authorities, less self-confident, more neat and orderly, More recent works have included Cradles of Emi-
more visionary and idealistic, and less self-assertive. (p. 212) nence (Goertzel & Goertzel, 1962), and Three Hun-
dred Eminent Personalities (Goertzel, Goertzel, &
All results were significant, and should inject a note Goertzel, 1978), the anthology Genius and Emi-
of caution, at least with respect to results on tests of nence (Albert, 1983), and Genius, Creativity, and
creativity subject to interpretation in the scoring. Leadership (Simonton, 1984). In a 1986 study, Si-
Socioeconomic status has received attention as a monton subjected Goertzel's data to multivariate
variable associated with differences in creative ex- statistical analysis. The data included extensive
pression. Several studies have reported that chil- measurements on over 50 biographical charac-
dren from middle and lower socioeconomic levels teristics for 314 eminent individuals from over 40
have different strengths in creative expression countries. The individuals represented 20 distinct
(Deutsch, 1967; Gallager, 1975; Reissman, 1963; fields according to their chief contribution. The
Torrance, 1974, 1977). In a study by Haley (1984), measurements included general demographic char-
47 black children from a private day care center (the acteristics, family, education, and intellectual
advantaged) were compared to 42 black children traits. The significance, in terms of antecedent con-
attending a federally funded child development ditions on eminence, included the following:
Nonfiction authors were more prone to come from urban centers, analysis. The Structure-of-Intellect (SOl) model or-
military figures from small towns; women were more likely to ganized factors that were developed from intellec-
become nonfiction authors and performers, men to become ath- tual tests into a three-dimensional model involving
letes, politicians and artists (including composers and film
makers); those from atheistic or agnostic backgrounds have the content, operation, and product categories. The
best chance of going into the military, nonfiction writing, sci- content categories refer to the organization of infor-
ence or revolution; authors of either fiction or nonfiction tend to mation used in the human cognitive process, and
come from unhappy home environments, while better home con- specifically include item types involving figural
ditions produce scientists, religious leaders and philosophers,
labor leaders, editors and publishers and reformers .... [L]arge
(subsequently termed visual), symbolic, semantic,
families contain reformers and artists, small families poets and and behavioral properties. (Later research added
nonfiction authors; scientists, psychiatrists, politicians and re- auditory as a fifth content category; see Guilford,
formers have the most formal education, while athletes, labor 1975). The mental operations involved include
leaders, business persons, mystics or psychics, artists and per- evaluation, convergent production, divergent pro-
formers the least; artists and composers are most likely to have
had special schooling (e.g., in an art school or conservatory), duction, memory, and cognition. Finally, the prod-
business persons, explorers and editors or publishers to have ucts form an ascending scale: units, classes, rela-
gone to boarding school, while poets are the least likely to have tions, systems, transformations, and implications.
had any kind of special school experiences. (pp. 17-18) A total of 120 SOl categories (factors) result by
combining the four content categories with the five
In terms of the interactionist model, antecedent operations and six products.
conditions influence the cognitive and personality Guilford's (1977, 1984) research is not the only
characteristics of the individual and, in addition, taxonomy of cognitive abilities, and perhaps not
help define the situation existing for the individual even the approach most closely linked with the his-
at any given time. With this in mind, there can be tory of the factor analysis of abilities (Snow, 1986),
little doubt that antecedent conditions provide some but through the ability of divergent production, it is
explanation of differences in creativity. Further, it a research avenue associated with creativity. Thirty
is likely that different background characteristics different divergent abilities have been distinguished
relate to different types of creativity. On the other as defined by the five content and six product cate-
hand, it is not clear how much of the individual gories. Each divergent ability is characterized by its
variation in creativity is accounted for by anteced- unique combination of three features: its kind of
ent conditions. In all likelihood, much variation operation, kind of content, and kind of product.
remains after background characteristics are taken Further analysis by Guilford (1984) reported four
into consideration. second-order factors and a third-order factor. The
four second-order factors follow along the lines of
the contents, minus divergent-auditory dimensions.
Cognitive Factors
The six products (units, classes, relations, systems,
Most researchers regard cognitive factors as transformations, and implications) form the diver-
important in creativity. In addition to whatever else gent-visual dimensions. The same pattern is repli-
is involved, the development of creative products cated for divergent-symbolic, divergent-semantic,
seems to have an intellective component. The and divergent-behavioral factors. Finally, a third-
important (and controversial) relationships between order ability of divergent thinking is derived. The
general intelligence and creativity clearly fit within essence of Guilford's work demonstrates that diver-
the cognitive ability category. However, because gent production involves the cognitive processes of
intelligence-creativity relationships are the pri- fluency, flexibility, originality, and elaboration.
mary focus of another chapter in this volume, they Taken together, these are the cognitive components
will not be further explored here. Rather, we will of creative thinking.
examine several specific cognitive ability factors In a subsequent publication, Guilford (1983) dis-
considered to be important in explaining dif- cussed the role of the transformation abilities in
ferences in creative behavior. creativity. The 25 transformation abilities are
formed by the combination of five operations (cog-
Divergent Production. Much of the foundation nition, memory, divergent production, convergent
for current thought with respect to the role of intel- production, and evaluation) and five contents (visu-
lect in creativity comes from Guilford's (1967) real, auditory, symbolic, semantic, behavioral). One
search into the structure of intellect. Guilford's re- result of the research was a general cognitive style
search on creative thinking has been based on factor dimension, what Guilford (1980) called an intellec-
tual executive function or disposition that prompts a The impact of the research on cognitive style was
person to employ a certain kind of intellectual abil- summarized by Shouksmith (1970):
ity to look for transformations.
It is often found that a particular individual adopts a certain
Cognitive Style. A converging line of research strategy or group of strategies in his approach to a wide variety of
to cognitive aspects of problem solving is cognitive problem situations. We use the term "cognitive style" to refer to
this characteristic. Assessed over a wide range of situations, an
style. One of the most influential works on cognitive individual's cognitive style is the amalgam of strategies that he
style has been that of Witkin, Dyk, Paterson, Good- typically adopts in his attempts to solve problems with which he
enough, and Karp (1962). Following earlier work by is faced . . . . [C]ognitive style includes all those phenomena
Witkin, the group isolated two major approaches to a which have been variously referred to as "set," or "thinking
mode," or the like. (p. 149)
series of cognitive and perceptual tasks, field depen-
dence and field independence. Individuals with high
field independence have the ability to analyze the The significance for creativity is the realization that
environment, breaking up the total field and attend- the approach that is adopted to a task determines in
ing to relevant features while withholding attention part the level of success achieved in solving the
from the less salient aspects. Field-dependent indi- problem. Thus, the score derived from an ability
viduals have difficulty separating more important test may reflect a variety of factors, including ele-
from less important elements of a problem. Bloom- ments of cognitive style. In this regard, creativity
berg (1967) was one of the early investigators to may be considered a form of problem solving
realize the relationship between field indepen- (Guilford, 1964). Studies approaching creativity
dence/ dependence and creativity. from this perspective have often suggested that
The work of Witkin and his colleagues stimu- creativity is a form of thinking, one for which styl-
lated many other investigations. Several studies re- istic considerations are important (Kendler &
lated measures of cognitive style to those of various Kendler, 1962).
personality tests (Stuart, 1965; Wertheim & Med-
nick, 1958). Other investigators have examined the Other Cognitive Factors. Other cognitive rela-
relationship between various cognitive styles and tionships are worthy of note. Firestien and
creativity (Del Gaudio, 1976; Dellas & Gaier, Treffinger (1983) focused attention on another of
1970; Quinn, 1980). Others, as exemplified by the the SOl operations, convergent production. Most
work of French (1965), related measures of cog- research attention has been with respect to diver-
nitive style to performance on other conceptual gent production, the ability to generate imaginative
tasks. It was French who derived analogues to ideas, but, as noted by Firestien and Treffinger
Witkin's field dependence and field independence (1983, p. 32), "without convergence, no action can
in performance on cognitive tests. These dimen- take place, no decisions can be made." Firestien
sions were defined as a reasoned or systematic ap- and Treffinger outline important benefits to be
proach versus a less orderly scanning and achieved through the ability to converge and a
visualizing. model for combining divergent and convergent tal-
In a further study, Noppe (1985) investigated the ents to achieve an optimal group solution.
multivariate relationship between several measures With respect to other taxonomies and resulting
of cognitive style and creativity. In addition to field abilities that bear on creativity, Carroll (1985) re-
dependence/independence, Noppe included mea- analyzed previously collected data using modem
sures of fixity-mobility, the degree of individual methods. Three hierarchical levels resulted with
flexibility which permits cognitive functioning general intelligence the sole dimension at the pinna-
from both global as well as focused perspectives, cle. Seven second-order factors were specified, one
and organismic interpretations or stagelike aspects of which, idea production, is most closely associ-
of thought. Canonical analysis yielded a signifi- ated with creativity. The eight first-order factors
cant, positive correlation between the cognitive that are part of idea production include (1) asso-
style variables and the measures of creativity. The ciative fluency, (2) fluency of expression, (3) figur-
relationship was such that it accounted for approx- al fluency, (4) ideational fluency, (5) speech fluen-
imately 39% of the shared variance between the two cy, (6) word fluency, (7) practical ideational
sets of variables. The result was a clearer basis for a fluency, and (8) originality. Carroll's work does
cognitive foundation of creative thought, especially much to put the frequently cited structure of intel-
with respect to stylistic variables. lect in perspective as well as to bring to the fore
other functions that may relate to individual dif- in the psychoanalytic tradition (e.g., Freud, Jung,
ferences in creativity. Kubie) view creativity. as stemming from the pre-
Mathematical creativity of 476 ninth-grade Indi- conscious or the unconscious. Humanistic theories
an boys was studied by Tuli (1985). Each student (e.g., Murray, Maslow, Rogers) typically relate
was administered the Creative Ability in Mathemat- creativity to the individual's quest for self-actu-
ics Test, a subtest of the Differential Aptitudes alization. Behavioristic theories view creativity as
Tests, and the Mathematics Attitude Scale. Exam- novel or unusual behavior that is nevertheless
ination scores on a standardized achievement test learned, that is fundamentally no different than
were the criterion. A significant relationship was other behavior which can be explained in terms of
found between mathematical creativity and stimulus-response or contingencies of reinforce-
achievement in mathematics. Creativity was not ment.
significantly related to attitude toward mathema- Several investigators have examined the rela-
tics. tionship between locus of control (LOC) and
creativity. Internal LOC individuals believe that re-
Summary. According to the interactionist wards and reinforcements are primarily a result of
model, cognitive style and ability are part and par- their own actions. External LOC individuals be-
cel of the individual, and thus help define and draw lieve that outcomes are largely governed by chance,
from the contextual and social influences. There luck, fate, or other systems over which they have no
can be little doubt that cognitive factors playa ma- control. Yardley and Bolen (1980) explored the re-
jor role in the creative process, but just how this lationship of nonverbal creative abilities to LOC,
happens is not as clear as might be the case. In fact, sex, and race among 112 second-grade public
a frequent argument is whether or not various cog- school students in North Carolina. External females
nitive factors are distinct from creativity (Jackson & scored highest of all groups on fluency and flexibil-
Messick, 1965). For example, the work of ity, and were significantly higher than external
Guilford, which regards creativity as a subset of the males. The authors interpreted the results as sug-
SOl, treats cognition as, in essence, embodying gesting that creative males were internal while cre-
creativity, rendering the latter as superfluous. ative females tended to be external in orientation. In
A more enlightened approach would emphasize another study, Aggarwal and Verma (1977) com-
the process of cognitive functioning and would rec- pared high-creative and low-creative high school
ognize that strategies or styles of thinking are students from India on LOC. High-creative stu-
important to the production of novel and useful dents were significantly more internal than the low-
problem solutions. A number of cognitive styles creative students.
have been studied in relation to individual dif- Kumar (1981) compared interest patterns of
ferences in creativity, and this appears to be a prom- high- versus low-creative ninth-grade Indian stu-
ising line of investigation, but further research is dents. It was found that patterns associated with
needed. literary, scientific, and fine arts interests charac-
terized the high creatives. The low-creative stu-
dents tended to be interested in household and out-
Personality Factors
door activities. In a related study, Bachtold (1983)
As has been seen from the previous section, a investigated whether university students differed in
reasonable conclusion from existing research their potential for creativity according to their major
would be that creativity is related to various cog- field. It was found that art and economics students
nitive activities and abilities. Another focus of the excelled on two measures of creativity (figural elab-
research has been with respect to the contribution of oration, verbal flexibility/originality). Social sci-
noncognitive factors (personality, interests, at- ence students surpassed the other majors on verbal
titudes) toward individual differences in creativity. elaboration.
In a recent review, Woodman (1981) concluded The interactionist model treats personality much
that "a theory of personality ... which does not as cognitive style and ability, as an essential com-
account for the creative act seems incomplete" (pp. ponent of the individual. Yet the measures of per-
61-62). Woodman (1981) categorizes those theo- sonality are less impressive in terms of both reliabil-
ries that do attempt to explain creative behavior into ity and validity, hence the research evidence is less
one of three major streams: (a) psychoanalytic, (b) extensive than that for cognitive aspects of the indi-
humanistic, and (c) behavioristic. Theorists writing vidual. At the same time, biographical and experi-
mental research studies have suggested that dif- spect, treating imaginative ideas with respect, encouraging
ferences in creativity are a result of many factors and evaluating self-initiated learning, etc.) ...
and that among these qualities are patterns of per- 6. Activities which help children to value their own
ideas .... (pp. 171-172)
sonality traits.
The research evidence presented by Torrance
Contextual and Social Influences represents a major contribution to the un?erstand-
ing of social and contextual factors that mfluence
The contextual and social influences on creative
individual differences in creativity. Actually, the
behavior include such things as physical environ-
factors outlined by Torrance are those that have the
ment, culture, group or organizational climate,
potential to foster creativity, that is, to give those
time/task constraints, expectations, rewards/
individuals with the potential to be creative the
punishments, and role models. Taken together,
chance to develop those skills. The maximization of
these are the elements of the environment and social
opportunity increases individual differences in
setting in which the creative act takes place, and, as
creativity as those with creative talents improve and
such, have the potential to contribute to or detract
those with more limited capabilities remain rela-
from individual differences in creativity.
tively unchanged.
Although the contextual and social factors have In a major summary of the research literature on
been implicit in much of the research, they have not the social and environmental factors influencing
been an explicit topic in a large number of studies. creativity, Amabile (1983b) examined the topics of
The previously cited research on biographical char- evaluation, rewards, modeling, and training.
acteristics of eminent individuals is illustrative of Among the conclusions from the review of research
one way knowledge about contextual and social was the finding that the expectation of evaluation
factors can be deveIoped. Yet, as Simonton (1986) tends to undermine creativity. With respect to re-
noted, findings about eminent achievers raise the wards, there is considerable research support for
issue of how generalizations derived apply to others intrinsic motivation as facilitating creativity. An
with much more modest accomplishments. We can- extrinsic reward tends to be detrimental to creative
not automatically assume that the factors which dif-
performance to the extent that it can l~ad ot~ers to
ferentiate a renowned individual also contribute to view motivation as external. ConclUSIOns WIth re-
the creativity of those less acclaimed. spect to social facilitation are more tentative: Re-
In a research report, Torrance (1961) outlined sults of group versus individual work are mIxed.
specific factors that affect creative thinking in Exposure to models can impr.ove performa~ce, but
children: only if the modeled behavIOr bears a hIgh re-
semblance to the performance assessed. The
Of the factors in nature and society, . . . studies . . . thus far Amabile (1983b) review provides a definitive sum-
indicate that the following factors affect the development and/or
expression of creative thinking: mary of the literature on the social a~d ~n~ironme.n
1. Educational level tal processes that contribute to mdividual dIf-
2. Differential treatment of boys and girls ferences in creativity.
3. Premature attempts to eliminate fantasy In another study, Ekvall and Tangeberg-An-
4. Restrictions on manipulativeness and curiosity
5. Conditions resulting in fear and timidity, in both authority
dersson (1986) attempted to relate working climate
and peer relations to creativity in the field setting of a Swedish na-
6. Misplaced emphasis on certain verbal skills, especially on tional daily newspaper. A total of 12 working-cli-
mechanics mate dimensions were measured (including chal-
7. Overemphasis on prevention and on "success" lenge, support for ideas, richness in ide~s,
8. Lack of resources for working out ideas
... [Pjreliminary analyses indicate that the following experi- dynamism, and conflict) and were related to offIce
mentally manipulated variables affect significantly the produc- creativity. The authors concluded that freedom a~d
tion of ideas or functioning of the creative thinking abilities . . . autonomy, combined with a largely democratIc
I. The composition of the group in which the thinking takes work organization, contributed to a creative climate
place (homogeneous or heterogeneous) ...
2. Competition ...
not enjoyed elsewhere in the organization. To quote
3. The teaching of principles for thinking up ideas . . . these authors:
4. The nature of the "warm-up," instructions (quantity vs.
quality), and practice (unevaluated vs. evaluated The group shares certain basic values, altho.ugh its mem~ers do
practice) ... not necessarily have the same political, social or other kmds of
5. Rewarding creative thinking (treating questions with re- background. They approach the subjects discussed on the
[woman's] page in the same sort of way (humanistic, holistic, tions of current knowledge, existing disparities,
and basically psychological), which makes it easy to discuss the and needed knowledge.
factual content without wasting time and energy arguing about
fundamental values, on convincing each other or testing one
another's beliefs. Thus, the climate appars [sic] to be "soft" and
conflict-free. (p. 222) What Do We Know?
Creativity can best be thought of in terms of ac-
Norms that were reported include (a) everything is complishments, "achievements that are original
possible, nothing is fixed; (b) you create your own and make a meaningful contribution to culture"
reality; (c) pursue your own ideas; and (d) question (Nicholls, 1972, p. 717). People differ in the extent
all rules. Findings from the research summary by to which they can be said to be creative, and these
Amabile (1983b) are related. Amabile (1983b) con- differences can be thought of as a function of the
cluded that' 'work environments most conducive to interactive influence of antecedent conditions, cog-
the fulfillment of creative potential may include: a nitive factors, personality, and surrounding con-
high level of worker responsibility for initiating textual and social influences. The interactionist
new activities, a low level of interference from ad- model presented in Figure 1 is an attempt to capture
ministrative superiors, and a high stability of em- the gestalt of the relationships among these factors
ployment" (p. 184). This summary is in agreement and creative behavior.
with the Swedish findings. A plurality of research has focused on the role of
In summary, "research on creativity has led to a cognitive factors as the source of differences among
recognition of the fact that the kind of environment people with regard to creativity. Much of the re-
most likely to produce a well-adjusted person is not search has concerned the role of divergent thinking
the same as the kind of environment most likely to and, more recently, convergent thinking.
produce a creative person" (Tyler, 1974, p. 114). There are ample examples of the unrelatedness
This is to say, in terms of the interactionist model, between creativity and intelligence. Intellect is re-
antecedent conditions along with social and con- garded as a threshold characteristic, a necessary but
textual influences can have an important impact on not sufficient condition for creativity. As a result,
creativity. From the research evidence, it is clear more recent research has concerned the role of cog-
that individual differences are a function of the ex- nitive strategies or styles of thinking as factors in
tent to which the social and contextual factors nur- explaining differences in creativity. Creative indi-
ture the creative process. Further, given the factors viduals tend to approach problems with greater
shown to be important, many children and adults do amounts of intensity, reflection, and persistence
not operate in an environment that fosters creative than their colleagues of equal intelligence.
achievements. Concern for stylistic factors is, in effect, ac-
knowledgment of the joint role of cognitive and
personality (noncognitive) factors in creativity.
Summary and Conclusions: An There can be little doubt that noncognitive factors,
Interactionist Approach to Individual such as intensity, persistence, and interest, playa
Differences in Creativity role in creativity.
Finally, biographical studies have shown com-
The topic of individual differences in creativity, mon antecedent conditions among creative indi-
or, for that matter, individual differences in any viduals. At the same time, contextual and social
area of cognitive and noncognitive behavior, is in- factors can facilitate or retard creativity.
extricably bound to the issues of definition and as-
sessment. Further, individual differences are a
What Gaps Exist in Current Knowledge?
function of the criteria used to characterize creative
behavior. In short, the base on which individual The major gaps are with respect to the opera-
differences research is built can be no better than tional definition of creativity along with the mea-
thinking and research on issues of definition, as- surement of this concept. The vast majority of
sessment, cognitive, and noncognitive models. creativity research that now supports our under-
With this in mind, it is the purpose of this section to standing of individual differences in creativity is
review and pull together the evidence with respect based on a small number of measurement instru-
to individual differences in creativity through ques- ments. Further, the construct validity of many of
the instruments has not been adequately demon- techniques associated with construct validation
strated. have not been used in creativity research. For exam-
The construct of divergent thinking and the ple, the multitrait-multimethod matrix developed
closely related concepts of remote association and by Campbell and Fiske (1959) has become a staple
ideational fluency have dominated the measure- of construct-validation research, but has been sel-
ment of creativity. To quote Nicholls (1972), dom applied to research on creativity. The method
"though there is some theoretical support for the requires the demonstration of convergent and dis-
notion that ideational productivity is important for criminant validity. In other words, significant con-
creative achievement, the research evidence for the vergent agreement is required between measure-
predictive validity of divergent thinking tests is far ments of the same construct using different
from impressive" (p. 719). Most research using methods. At the same time, a validity coefficient
one or more measures of ideational productivity for a given variable must be greater than the correla-
assume that the resulting score or scores are syn- tion between the measurements of this variable and
onymous with creativity, that is, no further confir- the measurements of all other variables with any
mation is sought. other methods in order to demonstrate discriminant
Another research gap has to do with the subjects validity. Many of the measurement procedures used
used in many of the research studies. Most of those in creativity research, on which current knowledge
studied have been children or young adults, esis based, would not stand up to the requirements of
pecially college-age students. Questions exist as to the multitrait-multimethod matrix.
how the typical divergent thinking or novel uses There are other more recent techniques of con-
scores taken as evidence of creativity among these struct validation, such as meta-analysis, that may
groups translates into reshaping or extending have applicability to creativity research. The point
achievement in a field, which would be closer to the is that many of the techniques considered as basic to
definition of creativity among more mature indi- the establishment of construct validity of concepts
viduals. The concentration of research on easily have not been widely applied to the measurement of
available individuals may have inhibited our under- creativity.
standing of individual differences in this important In addition, an approach to explaining creative
area. behavior, such as that illustrated by the interac-
tionist model in Figure 1, could provide a useful
heuristic or framework for programmatic research
How Will We Develop Needed Knowledge?
on creativity. In other words, perhaps what is
Clearly there is room for innovation with respect needed at this point, in terms of theory develop-
to research on individual differences in creativity. ment, is not yet another new' 'theory" of creativity
In addition, further work along traditional lines is but rather an integrative framework that would
needed. serve to combine diverse perspectives on creativity
Construct and empirical validation represent ex- and thus build on current knowledge while, at the
amples of research along accepted lines that would same time, suggesting promising avenues for future
contribute to the understanding of individual dif- work. In short, we need a more complex system for
ferences in creativity. Empirical validation requires integrating what we now know rather than a wholly
a criterion of creativity against which measures are new perspective on understanding differences in
validated. MacKinnon's (1964, 1965) research pro- creative behavior. The interactionist perspective
vides an example. Creative architects were com- could possibly provide this more complex system.
pared to their less creative colleagues on a number
of psychological dimensions. None of the dimen-
sions were called creativity, as such, but instead Toward a Comprehensive View of Individual
were designed to provide a picture of how creative Differences in Creativity
indiv~duals differed from their normal colleagues. Creativity is a valued trait in a democratic, free-
Guilford's work on the structure of intellect, par- enterprise society. Those with the talent for
ticularly the effort on relationships associated with achievements that are original and make a mean-
the concepts of divergent and convergent thinking, ingful contribution are rewarded and/ or recognized
is exemplary of construct validation. There are for their accomplishments. Research efforts that are
other examples of factor-analytic research's mak- successful in guiding the selection and/ or develop-
ing contributions toward an understanding of ment of creative individuals will also have made a
creativity (Carroll, 1985). At the same time, other valued contribution. These are some of the reasons
that make individual differences in creativity of at different levels of socio-economic status. The Journal of
general interest and concern. Creative Behavior, 11, 150.
Albert, R. S. (Ed.). (1983). Genius and eminence. New York:
The principle contribution of the present review Pergamon.
is by way of the interactionist perspective adopted. Amabile, T. M. (1983a). Social psychology of creativity: A
Perhaps, understandably, researchers from differ- componential conceptualization. Journal of Personality and
ent orientations have approached the problem from Social Psychology, 45, 357-376.
different perspectives. It would not be surprising Amabile, T. M. (1983b). The social psychology of creativity.
New York: Springer-Verlag.
that a cognitive psychologist would rely on princi- Arieti, S. (1976). Creativity: The magic synthesis. New York:
ples of cognition as the basis of creativity, or that a Basic Books.
social psychologist would postulate social facilita- Bachtold, L. M. (1983). Differences in divergent thinking and
tion as critical to creative accomplishments. temperamental traits among university students. The Journal
of Creative Behavior, 17, 267.
The interactionist model suggests that creativity Barron, F. (1969). Creative person and creative process. New
is the complex product of a person's behavior in a York: Holt, Rinehart & Winston.
given situation. The situation can be characterized Barron, F., & Harrington, D. M. (1981). Creativity, intel-
in terms of the contextual and social influences that ligence, and personality. Annual Review of Psychology, 32,
either facilitate or inhibit creative accomplishment. 439-476.
Basadur, M., & Thompson, R. (1986). Usefulness of the idea-
The person is influenced by various antecedent con- tion principle of extended effort in real world professional and
ditions, and brings to bear both cognitive abilities managerial creative problem solving. Journal of Creative Be-
and noncognitive traits or predispositions. havior, 20, 23-34.
Perhaps of greater importance, the model, in Bloomberg, M. (1967). An inquiry into the relationship between
field independence and creativity. Journal of Psychology, 67,
conjunction with the research evidence, leads to an 127-140.
understanding of the sources of individual dif- Bolen, L. M., & Torrance, E. P. (1978). The influence on
ferences in creativity. For example, research seems creative thinking of locus of control, cooperation, and sex.
to suggest that cognitive factors are especially Journal of Clinical Psychology, 34, 903-907.
important. The specific abilities are those associ- Cagle, M. (1985). A general abstract-concrete model of creative
thinking. Journal of Creative Behavior, 19, 104-109.
ated with divergent production or ideational fluen- Campbell, D. T., & Fiske, D. W. (1959). Convergent and dis-
cy, along with convergent production. Creative ac- criminant validation by the multitrait-multimethod matrix.
complishment is more than cognitive ability; that Psychological Bulletin, 56, 281-302.
more is personality and that ability area that bridges Carroll, J. B. (1985, May). Domains of cognitive ability. Paper
presented at the meeting of the American Association for the
intellect and personality-cognitive style. Creative Advancement of Science, Los Angeles.
individuals seem to be more intense, more reflec- Carroll, J. B., & Maxwell, S. E. (1979). Individual differences
tive, and more persistent. In short, they approach in cognitive abilities. Annual Review of Psychology, 30, 603-
problems in a different manner from their less cre- 640.
ative colleagues. Cox, C. (1926). The early mental traits of three hundred ge-
niuses. Stanford, CA: Stanford University Press.
In sum, the interactionist model of creative be- Del Gaudio, A. C. (1976). Psychological differentiation and
havior proposed here provides a comprehensive mobility as related to creativity. Perceptual and Motor Skills,
view of individual differences in creativity that in- 43, 831-841.
corporates important elements of the personality, Dellas, M. (1978). Creative personality and identity.
Psychological Reports, 43, 1103-1110.
cognitive, and social psychology explanations of Dellas, M., & Gaier, E. (1970). Identification of creativity: The
creativity. Creative behavior is a complex person- individual. Psychological Bulletin, 23, 53-73.
situation interaction that is influenced by events of Deutsch, M. (1967). The disadvantaged child: Studies of the
the past as well as salient aspects of the current social environment and the learning process. New York:
situation. Within the person, both cognitive and Basic Books. .
Ekvall, G., & Tangeberg-Andersson, Y. (1986). Working
noncognitive aspects of the mind are related to cre- climate and creativity: A study of an innovative news-
ative behavior. The interactionist perspective pro- paper office. The Journal of Creative Behavior, 20, 215-
vides an integrating framework of sufficient com- 225.'
Eysenck, H. J. (1983). The roots of creativity: Cognitive ability
plexity to incorporate diverse streams of research or personality trait? Roeper Review, May, 10-12.
and to suggest exciting avenues for future program- Faschingbauer, T. R., Moore, C. D., & Stone, A. (1978). Cog-
matic research on creativity. nitive style, dogmatism, and creativity: Some implications
regarding cognitive development. Psychological Reports, 42,
795-804.
References Firestien, R. L, & Treffinger, D. J. (1983). Ownership and
converging: Essential ingredients of creative problem solving.
Aggarwal, Y. P., & Verma, L. K. (1977). Internal-external The Journal of Creative Behavior, 17, 32-38.
control of high creative and low creative high school students French, J. W. (1965). The relationship of problem-solving styles
to the factor composition oftests. Educational and Psycholog- MacKinnon, D. W. (1964). The creativity of architects. In C.
ical Measurement, 25, 9-28. W. Taylor (Ed.), Widening horizons in creativity. New York:
Gallager, J. J. (1975). Teaching the gifted child. Boston: Allyn Wiley.
& Bacon. MacKinnon, D. W. (1965). Personality and the realization of
Galton, F. (1869). Hereditary genius. London: Macmillan. creative potential. American Psychologist, 20, 273-281.
Getzels, J. W., & Jackson, P. W. (1961). Family environment MacKinnon, D. W. (1970). The personality correlates of
and cognitive style: A study of the sources of highly intelligent creativity: A study of American architects. In P. E. Vernon
and of highly creative adolescents. American Sociological (Ed.), Creativity. Harmondsworth, Middlesex, England: Pen-
Review, 26, 351-359. guin Books.
Goertzel, M. G., Goertzel, V., & Goertzel, T. G. (1978). Three Mednick, S. A. (1962). The associative basis of the creative
hundred eminent personalities. San Francisco: Jossey-Bass. process. Psychological Review, 69, 220-232.
Goertzel, V., & Goertzel, M. G. (1962). Cradles of eminence. Nicholls, J. G. (1972). Creativity in the person who will never
Boston: Little, Brown. produce anything new and useful: The concept of creativity as
Goyal, R. P. (1973). Creativity and school climate: An explora- a normally distributed trait. American Psychologist, 27, 717-
tory study. Journal of Psychological Researches, 17, 77-80. 727.
Grim, c., & Torrance, E. P. (1977). Race and sex bias in in- Noppe, L. D. (1985). The relationship of formal thought and
terpreting creativity test results of children. The Journal of cognitive styles to creativity. The Journal of Creative Behav-
Creative Behavior, 11, 212. ior, 19, 88-96.
Guilford, J. P. (1964). Creative thinking and problem solving. Noppe, L. D., & Gallagher, J. M. (1977). A cognitive style
Educational Digest, 29, 29-31. approach to creative thought. Journal of Personality Assess-
Guilford, J. P. (1967). The nature of human intelligence. New ment, 41,85-90.
York: McGraw-Hill. Peterson, J., & Lansky, L. (1974). Left-handedness among ar-
Guilford, J. P. (1975). Creativity: A quarter century of progress. chitects: Some facts and speCUlations. Perceptual and Motor
In I. A. Taylor & J. W. Getzels (Eds.), Perspectives in Skills, 38, 545-550.
creativity. Chicago: Aldine. Quinn, E. (1980). Creativity and cognitive complexity. Social
Guilford, J. P. (1977). Way beyond the IQ: Guide to improving Behavior and Personality, 8, 213-215.
intelligence and creativity. Buffalo, NY: Creative Education Raina, T. N. (1980). Sex differences in creativity in India: A
Foundation. second look. The Journal of Creative Behavior, 14, 218.
Guilford, J. P. (1980). Cognitive styles: What are they? Reissman, F. (1963). The culturally deprived child. New York:
Educational and Psychological Measurement, 40, 715-735. Harper & Row.
Guilford, J. P. (1983). Transformation abilities or functions. The Richardson, A. G. (1986). Sex differences in creativity among a
Journal of Creative Behavior, /7, 75-83. sample of Jamaican adolescents. The Journal of Creative Be-
Guilford, J. P. (1984). Varieties of divergent production. The havior, 20, 147.
Journal of Creative Behavior, 18, 1-10. Roe, A. (1953). The making of a scientist. New York: Dodd,
Gupta, A. K. (1981). Sex differences in creativity: Some fresh Mead.
evidence. The Journal of Creative Behavior, 15, 269. Schneider, B. (1983). Interactional psychology and organiza-
Haley, G. L. (1984). Creative response styles: The effects of tional behavior. Research in Organizational Behavior, 5, 1-
socioeconomic status and problem-solving training. The Jour- 31.
nal of Creative Behavior, 18, 25-40. Shouksmith, G. (1970). Intelligence, creativity and cognitive
Hare, A. P. (1982). Creativity in small groups. Beverly Hills, style. New York: Wiley.
CA: Sage. Simonton, D. K. (1977). Creative productivity, age, and stress:
Hattie, J., & Fitzgerald, D. (1983). Do left-handers tend to be A biographical time-series analysis of 10 classical composers.
more creative? The Journal of Creative Behavior, 17, 269. Journal of Personality and Social Psychology, 35,791-804.
Helson, R. (1971). Women mathematicians and the creative Simonton, D. K. (1984). Genius, creativity, and leadership.
personality. Journal of Consulting and Clinical Psychology, Cambridge, MA: Harvard University Press.
36, 210-220. Simonton, D. K. (1986). Biographical typicality, eminence and
Helson, R., & Mitchell, V. (1978). Personality. Annual Review achievement styles. The Journal of Creative Behavior, 20,
of Psychology, 29, 555-585. 14-22.
Jackson, P. W., & Messick, S. (1965). The person, the product Skinner, B. F. (1974). About behaviorism. New York: Alfred
and the response: Conceptual problems in the assessment of Knopf.
creativity. Journal of Personality, 33, 309-329. Snow, R. E. (1986). Individual differences in the design of
Katz, A. N. (1980). Do left-handers tend to be more creative? educational programs. American Psychologist, 41, 1029-
The Journal of Creative Behavior, 14, 271. 1039.
Kendler, H. H., & Kendler, T. S. (1962). Vertical and horizon- Solomon, R. (1985). Creativity and normal narcissism. Journal
tal processes in problem solving. Psychological Review, 69, of Creative Behavior, 19,47-55.
1-16. Spotts, J. V., & Mackler, B. (1967). Relationships of field-
Kershner, J. R., & Ledger, G. (1985). Effect of sex, intel- dependent and field-independent cognitive styles to creative
ligence, and style of thinking on creativity: A comparison of test performance. Perceptual and Motor Skills, 24, 239-268.
gifted and average children. Journal ofPersonality and Social Staw, B. M. (1984). Organizational behavior. Annual Review of
Psychology, 48, 1033-1040. Psychology, 35, 627-666.
Klein, P. W. (1975). Effects of open vs. structured teacher- Steiner, G. A. (Ed.). (1965). The creative organization.
student interaction on creativity of children with different lev- Chicago: University of Chicago Press.
els of anxiety. Psychology in the Schools, 12, 286-288. Stuart, I. R. (1965). Field dependency, authoritarianism and
Kumar, A. (1981). Interest patterns of high and low creatives. perception of the human figure. Journal ofSocial Psychology,
The Journal of Creative Behavior, 15, 270. 66, 209-214.
Suler, J. R. (1980). Primary process thinking and creativity. aptitude for achievement in and attitude towards mathematics
Psychological Bulletin, 88, 144-165. among boys. The Journal ofCreative Behavior, 19, 225-226.
Suler, J. R., & Aizziello, J. (1987). Imagery and verbal pro- Tyler, L. E. (1974). Individual differences: Abilities and moti-
cesses in creativity. Journal of Creative Behavior, 21, 1-6. vational directions. New York: Appleton-Century-Crofts.
Taylor, I. A. (1976). Psychological sources of creativity. Weiss, D. S. (1981). A multigroup study of personality patterns
Journal of Creative Behavior, 10, 193-202,218. in creativity. Perceptual and Motor Skills, 52, 735-746.
Terborg, J. R. (1981). Interactional psychology and research on Wertheim, J., & Mednick, S. A. (1958). The achievement
human behavior in organizations. Academy of Management motive and field independence. Journal of Consulting Psy-
Review, 6, 569-576. chology, 22, 30-38.
Torrance, E. P. (1961). Factors affecting creative thinking in Witkin, H. A., Dyk, R. B., Paterson, H. F., Goodenough, D.
children: An interim research report. Merrill-Palmer Quar- R., & Karp, S. A. (1962). Psychological differentiation. New
terly of Behavior and Development, 7, 171-180. York: Wiley.
Torrance, E. P. (1965). Rewarding creative behavior: Experi- Woodman, R. W. (1981). Creativity as a construct in personality
ments in classroom creativity. Englewood Cliffs, NJ: Pren- theory. The Journal of Creative Behavior, 15,43-66.
tice-Hall. Woodman, R. W. (1983). A proposed process model of organi-
Torrance, E. P. (1974). Differences are not deficits. Teachers zational innovation. Southwest Division Academy ofManage-
College Record, 75, 472-487. ment Proceedings, 189-193.
Torrance, E. P. (1977). Discovery and nurturance of giftedness Yardley, C. R., & Bolen, L. M. (1980). Relationship oflocus of
in the culturally different. Reston, VA: The Council for Ex- control to figural creativity in second-grade students. The
ceptional Children. Journal of Creative Behavior, 14, 276-277.
Tuli, M. R. (1985). Mathematical creativity: Its relationship to
CHAPTER 5
The Nature-Nurture Problem in

Creativity
P. E. Vernon
Very little discussion of the nature-nurture topic Genetic Theory

has appeared in the literature, mainly because of the
complexities of the problem and the difficulties of The common notion of hereditary or genetic ver-
collecting objective evidence. A useful attempt was sus environmental factors is much oversimplified.
made by Scheinfeld in his book, Heredity in Hu- The genes do indeed provide for the transmission of
mans (1972), but this was a popular rather than a hereditary qualities, but they do not determine an
scientific treatment. A brief account by Zigler and individual's height, or intelligence, or creativity.
Farber (1985) can also be recommended. In the They are predispositions, whose effects develop
absence of hard data from tests, sociological sur- differently in different environments; that is, they
veys, or controlled experiments, this chapter has to interact with environmental conditions or experi-
rely very largely on observational and interview ences and produce not a fixed effect but a certain
studies, historical or biographic information, the "range of reaction." Nature and nurture are not
opinions of critics or colleagues about creative indi- opposed factors but are complementary to each
viduals, and discussions of theories, all of which other. Sometimes also the genes control or modify
are liable to ambiguity and subjectivity. the environment, as when a highly intelligent child
The traditional view of creativity and genius as- shapes his own environment by choosing books to
cribes these to creative abilities or traits within the read and other intellectual activities.
individual; that is, they are a kind of gift that is Nevertheless, it is possible to estimate quan-
inborn. But even the pioneer writers, such as Gal- titatively the relative strength of genetic and en-
ton, Terman, Burt, and Cox, realized that genius is vironmental differences in the growth of such
affected by environment and upbringing as well as qualities as intelligence; for example, by comparing
the genes; also that such personality factors as per- the resemblance in intelligence quotients between
sistence, drive, and commitment are at least as various relatives (who share certain numbers of
important as intellectual or cognitive abilities. genes in common), or between foster children and
their adopting parents. BU,t, in practice, there are so
many sources of error and unreliability in collecting
and testing suitable groups of subjects that we still
P. E. Vernon • Late of the Department of Psychology, Uni- find wide disagreements between different writers
versity of Calgary, Calgary, Alberta, Canada T3A 2E3. in their estimates of genetic and other effects (e. g. ,
93
Eysenck & Kamin, 1971). Doubtless the same lack of creative abilities. He points out the present con-
of agreement will occur in attempting to estimate fusion between Galton's "eminence," Terman's
genetic variance in measures of creativity. The sec- "genius," general intelligence, and the factorists'
ond half of this chapter surveys what evidence we g, or multiple-ability factors. He describes six do-
have, and what answers it provides (see also Mistry mains of creativity, which he designates as "intel-
& Rogoff, 1985). ligences," namely, (1) verbal-linguistic, (2) musi-
cal, (3) logic-mathematical, (4) visual-spatial, (5)
bodily-kinesthetic, and (6) social-personal. These
Definitions domains are not causative entities, though they may
involve some genetic differences. They were ar-
Let us fIrst defIne the nature of creativity and rived at more from observational studies of child
other related variables, which might differ in their development and from real-life behavior, rather
genetic-environmental components. Although nu- than from paper-and-pencil tests, such as the instru-
merous different defInitions of creativity have been ments that Thurstone, Guilford, and others have
proposed (I. A. Taylor, 1959), there would be a used. Each of these domains has a separate system
considerable consensus on the following general of signals and, although relatively autonomous,
statement. Creativity means a person's capacity to they often work together in human problem solving
produce new or original ideas, insights, restructur- and invention. Gardner provides considerable
ings, inventions, or artistic objects, which are ac- cross-cultural, neuropsychological, and other kinds
cepted by experts as being of scientifIc, aesthetic, of evidence to back his categorization.
social, or technological value. In addition to novel- There are several other terms that are linked to
ty as our major criterion, we must incorporate in our creativity and that are often used interchangeably;
defInition the acceptability or appropriateness of but it is better to distinguish them.
the creative product, even though this valuation Talents are skills that differ in the different sci-
may change with the passage of time. ences and in the arts and that do appear typically in
Naturally we would expect differences between the performing arts. They almost certainly involve
different types or areas of creativity. But the sci- genetic components as well as learning and may
ences-physical or biological, mathematical, psy- occur, too, in idiots savant (e.g., in numerical or
chological and social, and also engineering-do musical ability).
show a good deal in common. Likewise there are Genius is virtually identical with very high cre-
some resemblances, and also some differences ative abilities, but one cannot specify how high.
between scientifIc and artistic creativity, as, for Because it also has other interpretations (e.g., a
example, in literary, philosophical, musical, visual kind of inner spirit), it is better to avoid use of the
or decorative artistic, and sculptural creativity. Ar- term. Grinder (1985) gives a useful historical sur-
chitecture seems to partake of both types (MacKin- vey of conceptions of genius from Plato onward.
non, 1962). Scientific creativity always involves Prodigy refers to a child with skills or talents
some addition to our previous knowledge, either an much superior to what is normal for his or her age.
improved theory or a new object or procedure, But the prodigy does not necessarily become a high-
whereas artistic creation may give some new repre- ly creative artist or scientist until he or she produces
sentation of life (e.g., a painting or poem) or feel- new work of accepted value; and this seldom occurs
ing, but not usually a progression from previous much before age 20. It has been said that all ge-
representations. Another possible type of creativity niuses were prodigies (Agassi, 1985), but many
might be termed social or spiritual (e.g., Mahatma exceptions come to mind (e.g., Darwin and
Gandhi). But in my view, the performing arts (play- Einstein).
ing vs. composing music, acting, or dancing) do not Gifted applies to any child or adult of much above
merit the label creative, or only rather rarely , as, for average ability; and most psychologists would
example, Pavlova's dancing or Glenn Gould's agree that it is partly genetic in origin. It usually
highly interpretive piano music. refers to those high in general intelligence or all-
Gardner's recent book, Frames of Mind (1983), round ability. Thus, a child who is specially able in
has attracted wide attention because it provides an mathematics only would be talented rather than
alternative yet logical approach to the classifIcation gifted.
CHAPTER 5 • THE NATURE-NURTURE PROBLEM 95
Distribution of Creativity abIes. However, chi-squared or other non-

parametric techniques could be used for comparing
Because some other qualifications are necessary, productivity indices with, say, socio-economic sta-
I would prefer to restrict the term creative to a very tus (SES), age, or other relevant variables.
small proportion of scientists or artists whose pro-
ductions are highly valued, say, the most eminent
2% or so of the total populations of Western or
developed cultures. Although this conception is Sociology of Creativity
supported by Nicholls (1972), Mansfield and Busse
(1981), and others, it runs counter to the popular Most psychologists might be willing to give up
views of Guilford, Torrance, and their followers, the notion of creativity as a kind of ability or trait
who believe that creative potential is present in all possessed by the artist or scientist (whether inher-
members of the population and that it can be treated ited or acquired) if they were better acquainted with
as a normally distributed variable. For example, the the work of such sociologists as Merton (1968),
householder who plans an attractive garden, or his Cole and Cole (1973), Brannigan (1981), Black-
wife who designs a new dress, would be continuous well (1969), and Barber and Hirsch (1962). These
with, say, Leonardo da Vinci. These authors also authors stress the part played by the social climate
provide tests which require simple tasks, such as or Zeitgeist in the production of creative ideas. Si-
Alternate Uses for a Brick or Torrance's Picture monton (1977), among others, tabulated the pro-
Completion, on which almost all children from kin- ductivity of large numbers of musical composers,
dergarten up can score something. philosophers, and writers in different countries and
Now one would agree that there is some con- at different dates, and put forward hypotheses re-
tinuity between, say, a brilliant research chemist garding the social conditions that might be expected
and his assistants who contribute more by teaching to affect their work, for example, occurrence of
but who add a few novel ideas to his research. One wars, political stability, civil disturbances, or the
cannot prove that the highly or genuinely creative existence of many other persons working in the
scientist or artist differs qualitatively from those same field of creativity. Each of these variables is
working at lower levels in the same fields (though carefully defined and measured by objective indices
this problem might be tackled by factor analysis). and then intercorrelated. Some of these hypotheses
But the distribution of this productivity is nowhere are confirmed by multiple correlations or path co-
near normal, and approaches more closely to the efficients between the predictor variables and pro-
exponential or so-called J-curve. This distribution ductivity. But often the correlations are too small to
usually occurs when scores are affected by a large be significant (see also Narroll, Benjamin, Fohl,
number of influences that interact with one another Fried, Hildreth, and Schaeffer, 1971).
multiplicatively, and not additively. It applies, for Frequently, it has been noted in science that two
example, to incomes, numbers of criminal acts, and or more authors report almost identical discoveries
populations of cities; and it explains why the top 5% or theories at about the same date (e. g. , Darwin and
or so of all scientists produce over 50% of the total Wallace) . Further, quite a large part in the ap-
publications in science. Price (1962) noted that pearance of a novel discovery is played by chance
such distributions must be S-shaped rather than J- circumstances or observations (e.g., penicillin);
shaped, because growth must slow down at some though, at the same time, there must be sufficiently
time, otherwise, for example, the total membership able scientists to recognize and work out the im-
of the American Psychological Association would plications of the chance observation. F. C. Bartlett
one day exceed the total American adult (1958) pointed out that the appearance of some new
population. technique (e.g., the electron microscope) triggers
These facts are important, because they imply off a whole series of new advances. Nowadays,
that measures of creativity cannot be analyzed or also, advances in science are more likely to result
treated statistically by conventional parametric from groups of scientists who are working together
techniques; and factor or principal component anal- on some project rather than from the researches of
ysis would not be suitable for trying to analyze the individual geniuses; (this does not seem to occur in
chief underlying components of creativity vari- the arts).
It follows then that ascribing creative production The earliest cognitive analyses were those of von
to a creative mind is largely tautologous, because Helmholtz (1903), Poincare (1924), Hadamard
we have no evidence that a person is creative or not, (1945), and Wallas (1926), who distinguished four
other than by observing his or her productions and main stages in creative activity. These stages were
evaluating them (see also Agassi, 1985; Albert, oflimited value, as it was obvious that there was no
1975; Mistry & Rogoff, 1985). A topic that has rigid pattern of processing applicable to the arts as
been frequently discussed is that of the "unrecog- well as to the sciences. Naturally, also, there are
nized" or "potential" genius (see, e.g., Thomas variations between different creators and different
Gray's "Elegy Written in a Country Church- products. The most interesting aspect of the stages
Yard"). We would have to say that such a person is is their recourse to unconscious mental processes,
mythical, because he is not a genius until he has both in the incubation or cerebration stage and in the
produced work of high creative quality. Thus, sudden flash of insight or inspiration. Ghiselin
Gregor Mendel should not be so called until the (1952) published a collection of descriptions by
value of his work was recognized long after his writers, musicians, and scientists of their own
death. However, this is a tricky issue, for adoles- thoughts, methods, and feelings. More recent are
cents and young adults may produce work that Eiduson's (1962) and Rosner and Abt's (1970)
strikes their teachers as promising, although they books, with contributions from living scientists and
probably require greater maturity, knowledge, and other creative persons. Some of the authors recog-
skills before winning widespread recognition. Len- nize the importance of these subconscious pro-
neberg (1980) described the difficulties young com- cesses, whereas others entirely reject them. Some
posers often face in getting their early works played point out that in any kind of thinking we retrieve
or published. He suggested that this is exaggerated large numbers of associations from long-term mem-
and that many innovative composers (e.g., Franz ory, which might be called preconscious. But they
Schubert and Robert Schumann) were indeed rec- deny that the mind goes on working on an unsolved
ognized by their contemporaries, and their works problem when they are engaged in other activities.
were published quite soon after completion. This is the view favored in Perkins's (1981) detailed
Finally, there are obvious cultural factors in analysis of cognitive processes in creative thinking.
creativity that are shown by the many differences in He also allows that such thinking includes the ap-
contents, styles, and skills, which characterize the pearance of sudden leaps to new conclusions,
artistic productions of different societies. These dif- which are often accompanied by strong feelings of
ferences are more marked the more'geographically rightness or correctness.
remote one country is from another, but such cultur- In general, scientists rely less on inspirational
al differences are much less common in the sci- flashes or sudden insights and claim that their cre-
ences. It is possible, however, that some of the ative discoveries depend far more on accumulating
artistic differences may have a genetic or racial wide knowledge of the field, on elaboration and
basis. For example, it is quite likely that the North working out implications, and on evaluating new
American Eskimo inherit certain genes for three- ideas. Artists are more likely to admit that their
dimensional spatial ability, which help them to sur- creative work is accompanied by strong emotions.
vive life in an Arctic environment and to produce Yet they, too, may make many drafts or modifica-
their unique stone carvings. tions before they are satisfied. Compare, for exam-
ple, Wolfgang Amadeus Mozart's rapid production
of fully fledged compositions with Ludwig van
Cognitive and Motivational Processes Beethoven's painful striving. Richard Wagner was
occupied with The Ring for 30 years; Edgar Allen
Studies of creativity are often classified under Poe's "The Raven" was not completed until 20
three headings: the product, the process, and the years; and Charles Darwin's The Origin of Species
personality. So far we have been mainly concerned took 23 years before publication.
with the first of these headings, and less attention Few psychologists or psychiatrists follow
will be given to the second, since it is covered in Freud's initial view of art as a sublimation of uncon-
several other chapters. Also, there is little evidence scious conflicts and sex drives. But many subse-
to suggest that some creative processes are more quent writers favor psychodynamic theories that in-
likely to involve genetic factors than others. volve Freud's primary process thinking as well as
secondary process or rational thinking (e.g., Arieti, Although this work was not based on adult creative
1976; Biber, 1984; Kubie, 1958; Storr, 1972; Sul- thinking, most of it probably could be so applied.
er, 1980). This is interesting because primary and Doubtless there are great variations in the de-
secondary process activities probably emanate from tailed processing for different areas (e.g., mathe-
different areas or levels of the brain. Secondary matical, biological, musical, literary); hence, we
process activity or conscious thinking would ob- have not reached a stage of being able to specify an
viously depend to a much greater extent on learned overall catalogue of creative operations. Some writ-
skills. ers deny the existence of qualitative differences be-
Sternberg and Davidson (1985) gave a useful sur- tween creative processing and logical reasoning or
vey of theories of cognitive development among the problem solving in general (Campbell, 1960; New-
highly talented and pointed out the weaknesses in ell & Simon, 1972; Perkins, 1981; Rabinowitz &
(1) S-R theories, (2) Piagetian psychology, and (3) Glaser, 1985). Blackwell (1969) argued that there
psychometric and factorial approaches. They had a is no type of mental operation specific to creative
preference for cognitive psychology, which studies work. Such work ranges all the way from the rou-
the kinds and levels of information processing and tine to the revolutionary. However, in my view,
aims to specify the main components in which high- creative thinking and production should be differ-
ly able children or adults are much in advance of entiated. They involve a greater degree of free-
normal thinkers. For example, these individuals are flowing imagination and recourse to subconscious
characterized by the careful monitoring of their own activities, such as incubation and inspiration. Al-
thought processes and the correcting of their tech- though these activities appear more frequently with
niques when required. They not only have more writers and musical and visual artists, they can be
knowledge but are better able to apply it. They have recognized in a good deal of scientific discovery.
automatized many of their elementary skills, thus The roles of neurotic and affective processes are
freeing their attention for focusing on novel tasks still obscure, but they, too, are evident among re-
and seeing new concepts and relations. They also search scientists (Kubie, 1958). And, as pointed out
show exceptional drive and dedication in tackling in the introduction, creativity in the sciences and in
new problems in their own field. Sternberg also the arts requires outstanding motivation and person-
concurred with Gardner (1983), Feldman (1982), ality qualities besides cognitive skills.
and Gruber (1974) in believing that an individual
case-study approach for tracing the methods of
highly creative thinkers is more fruitful than trying
to measure the main background factors or current The Relation of General Intelligence
abilities. Gruber's (1974) study of Charles Dar- and Creativity
win's life and work is a good example.
A further survey by Rogers (1986) summarized The analysis thus far suggests that creative pro-
106 recent contributions to the characteristic prob- duction depends to some extent on general intellec-
lem-solving styles or strategies of highly gifted tual ability, and more so in scientists than in artists.
children, which include: But, insofar as it has been possible to measure both
types of variables, we would expect the correlations
1. Preference for less structured or conventional to be moderate rather than high. Fortunately, there
materials is a certain amount of hard evidence.
2. Taking responsibility for their own thinking In their genetic studies of genius, Terman (1925)
(i.e., internal locus of control) and Cox (1926) clearly believed that adult genius
3. Independence of thought, as described by was directly connected to high intelligence in child-
Witkin, rather than high verbal intelligence hood. As is well known, they made estimates of the
4. Quick recognition of the nature of the prob- likely childhood (17-year) IQs of 300 eminent per-
lem, and the appropriate steps for solving it sons from history and found an overall average fig-
5. Selecting relevant previous information and ure of about 135. But Cox provided good ground for
experience for application to the problem regarding this as an underestimate, attributable to
6. Monitoring their thinking systematically the paucity of information about many cases. She
7. Exceptional insights in arriving at correct considered the correct figure to be nearer to 140-
solutions 155. There were big variations, with some of her
cases ranging up to 190 and a very few as low as creatIvIty also has some genetic components;
100. Some 16% ofthe cases were estimated as fall- though there are obvious weaknesses and difficul-
ing below an IQ of 120. When classified by type of ties in such an argument.
achievement, philosophers were most able with
mean IQs of 147. Writers (140) and scientists (135)
were very high, artists were lower at 122, and sol-
diers were down to 115. Other Cognitive Indices of Creativity:
There are some more direct test results of living Divergent Thinking
scientists. Gibson and Light (1967) gave the
Wechsler Adult Intelligence Scale (WAIS) to 131 In MacKinnon's (1962) series of researches on
university scientists. Not many of these indi- eminent architects, writers, and other professional
viduals, perhaps, were in the genius class, but they groups, the more creative individuals did not usu-
would all have made many creative contributions to ally show better college grades than the less cre-
research in their own areas. The mean IQ was ative. Likewise, Hudson (1966) found that the Fel-
126.5. Chemists and mathematicians were highest lows of the Royal Society (who can be regarded as
at 130, agricultural scientists lower at 121. 7. the most eminent living research scientists in Great
Notice, therefore, that quite a large proportion must Britain) had frequently obtained only mediocre BA
have scored below 130. Little work has been done or BS degrees at the university. This fact poses
on more specialized talents or ability factors among serious problems for the staff members of graduate
those in different areas. Roe (1952), however, con- programs, who try to select the most promising
structed and applied tests of verbal, spatial, and students for doctoral work. The shortage of good
mathematical abilities to her groups of eminent sci- scientists in the Western world, the costs of training
entists. Theoretical physicists were much higher them, and the high drop-out rate in science and
than experimental ones in verbal ability, with biolo- engineering courses are well known. Frequent use
gists and social scientists intermediate. All phys- is made by academic selection committees of the
icists were high on spatial ability, biologists and Miller Analogies, the Terman Concept Mastery, or
anthropologists rather low. Biologists and psychol- other difficult tests. But these instruments show
ogists were high on the mathematical test; (no math- quite low, if any, correlations with the quality of
ematicians were tested). graduate theses. MacKinnon (1962) did obtain very
Terman's (1925) major contributions were based high Concept Mastery scores in a sample of creative
on very high IQ children (135 upward), and he writers, though not among architects or scientists.
followed them through to adulthood. Although The most consistently predictive test in this area
most of them achieved successful educational and seems to be the Barron-Welsh Art Judgment test,
occupational careers, very few showed such out- which is based on preference for complex, asym-
standing abilities as to merit the term genius. A metrical designs over simpler, symmetrical pat-
large volume of subsequent psychometric studies terns. Barron (1969) found considerable dif-
has confirmed that the exceptionally creative tend ferences in mean scores between groups rated as
to show high intelligence scores and vice versa; but high or low creative:
there are many who are relatively high in one and
lower on the other. MacKinnon (1962) and many Artists 39
other writers have subscribed to the "threshold hy- Highly creative architects 37
Research scientists 30Yz
pothesis," namely, that there is a high correlation Architects generally 26
of intelligence with indices of creativity up to the Low-creative scientists 19
average level. But when a certain threshold is Normal college students 18
reached, say, IQ 120, further gains in intelligence
do not bring about much further rise in creative Welsh (1975) gave this test along with Concept
abilities. However, direct evidence for this hypoth- Mastery to groups of able high school students and
esis is often equivocal. found interesting personality differences between
Despite continuing controversy regarding genet- high or low scorers on both measures. But these
ic factors in intelligence, it is widely agreed that the have not so far been related directly to creativity.
heritability variance is somewhere around .50. In- Guilford (1967), Torrance (1965), Wallach and
sofar as there is a relation between creativity and Kogan (1965), and others have devised a large vari-
general intelligence, this would suggest that ety oftests of "divergent thinking," which contrast
with the conventional "convergent" tests, and are lach's study, that accomplishments in the per-
claimed to measure original thinking. However, forming arts (music and drama) were least well
these tests are based on such trivial tasks as Alter- predicted.
nate Uses, Similarities, and Verbal Fluency, which
could not be expected to be of any relevance to real-
life or high-level creativity (see the earlier section
on "Distribution of Creativity"). In MacKinnon's Evidence for Genetic Factors in
study of architects, the divergent thinking tests Creativity: Rarity of Genius
clearly had no predictive value for creativity. Some
more promising differences between children with Six main types of evidence regarding the nature-
high and low divergent scores have been reported nurture problem may be adduced. None of these is
by Getzels and Jackson (1962), Torrance (1965), conclusive, but some are more favorable to genetic,
and Wallach and Kogan (1965). But these studies others to environmental, determinants.
have not been adequately replicated, and a more Many persons who are generally accepted as ge-
critical analysis of Wallach's results by Cronbach niuses differ so remarkably from the norms as to
(1968) showed that they add very little to predic- seem inexplicable in terms of favorable environ-
tions based on ordinary general verbal intelligence ment. Clearly this applies to Mozart, who was not
measures. Hudson (1966) provided a brilliant dis- only performing music brilliantly by the age of 6 but
cussion of why divergent thinking should not be was also composing. However, his most precocious
confused with creativity (see also Vernon, Adam- compositions can hardly be said to have shown
son, & Vernon, 1977). R. B. Cattell (1971) has much originality until he reached about the age of
made considerable use of ideational fluency tests 10. Now Mozart did grow up in a highly musical
from the 1930s on and now includes them as part of environment, indeed, under excessive pressure
a general retrieval capacity, gr' in his list of mental from his father, but his superiority was far greater
powers. But he has also found them to relate to his than that found in other musicians who had equally
second-order personality factor, Exvia, which is stimulating environments. Regardless of all other
not a characteristic of highly creative scientists. kinds of evidence, Mozart by himself provides
However, several batteries of tests have been proof that genetic differences in creativity do exist.
tried out on scientific researchers and other high- Another compelling example is Leonardo da
creative groups and have been found to yield multi- Vinci. Even though his artistic talents doubtless
ple correlations of around .50 (Drevdahl, 1956; owed much to a favorable environment and good
Shapiro, 1968; Vernon, 1972). Most of the tests training, his almost equally outstanding scientific
were cognitive, that is, similar to divergent think- genius was almost wholly idiosyncratic, that is,
ing, but usually at a more complex level. None has self-initiated and self-developed. Madame Marie
achieved wide usage, probably partly because of Curie and Sophie Germain (1776-1811) are very
the time needed to give and score them, or because exceptional since so few women have specialized in
subjects' performance is too much affected by what physics or mathematics. The latter was almost
they think the tests measure. Mednick and Med- wholly self-taught, and had no parental encourage-
nick's (1967) Remote Associations Test is another ment. An English girl, Ruth Lawrence, is currently
example that has proved disappointing. attracting attention, because at age 13 she has sur-
The most striking evidence that such tests may be passed all other mathematics students at Oxford
of some value at adult level has been provided by University. She did receive, however, much help
Wallach and Wing (1969). They gave a divergent from her parents and schools. A considerable
thinking battery to some 500 college freshmen, and number of Stanley's (~tanley, Keating, & Fox,
inquired by questionnaire into their participation in 1974) mathematical students (mostly males) were
creative extracurricular activities, such as writing, five years or more ahead of their age-peers, and
art, and scientific projects. Those who had partici- some were graduated from college at age 13.
pated substantially obtained higher divergent Youthful prodigies appear to be much more rare in
scores, though the degree of correlation was gener- the literary and visual arts. However, Gardner's
ally small. Bartlett and Davis (1974) confirmed this (1980) Nadia, an autistic girl who produced highly
finding with correlations in the .30s, but Kogan and talented drawings at about age 5, was barely able to
Pankove (1974) reported on the weak reliability of communicate orally with other children or adults.
the creativity measures. It was noteworthy, in Wal- Thus, one might hesitate to ascribe her talent to
some unusual genes, when she was so obviously males and females, which might presumably have
psychotic and/or brain-damaged. Mention also something to do with sex differences in creativity
should be made of idiots savant who show excep- (see below). Unfortunately, the importance oflater-
tional facility, particularly in mathematics or mu- alization has been much exaggerated, and many
sic. They could hardly be called creative, and their writers believe that all verbal activities are left-
talent is often attributed to intensive practice and brained, and all spatial and nonverbal activities
concentration in a very narrow field. But it seems right-brained. On the basis of quite inadequate evi-
reasonable to infer that they, too, must possess ex- dence, it has been claimed that all creative activities
ceptional genes. are right-brained, so it is desirable for schools to try
to train the right as well as the left (Gardner, 1978;
Springer & Deutsch, 1981; Vernon, 1984). Orn-
stein's book, The Psychology of Consciousness
Neurological Factors (1977), has also had a considerable influence,
though he is more concerned with the contrast be-
If particular types of creativity are associated tween rational, logical, or Western-type thought,
with particular areas of the brain, this would pro- and intuitive, creative Eastern-type meditation. He
vide strong evidence of genetic causation. We do claims that electroencephalograph (EEG) alpha
know quite a lot about the language functions con- waves and rapid eye movements in sleep are related
nected with Broca's and Wernicke's areas, though to lateralization.
no evidence is known to me that exceptionally large The grounds for crediting the right brain with
growth in these areas is correlated with excep- such a wide range offunctions, including creativity
tionally high verbal abilities. Likewise vision is de- in general, are certainly weak; and it is much more
pendent on the occipital cortex and auditory sen- probable that most, if not all, thinking and creative
sitivity with the temporal lobe. Gardner's (1983) activities involve many parts and both sides of the
book does lay claims to cortical localization of his brain. However, some relations to music and to
six areas of talent, though he also emphasizes the drawing are fairly well established. Musical stimuli
importance of upbringing and environment in the seem to be better recognized and recalled with the
growth of such functions. We should further recog- left ear (i.e., the right brain). But among trained
nize that environmental stimulation can bring about musicians, the analytic perception of music as a
neurological growth, that is, that the causation is series of relations between pitched tones is more a
not necessarily one way (Vernon, 1979). Most of left-brain process. In visual art, Edwards (1979) has
the above evidence relates to sensation and percep- shown that drawing ability can be improved by re-
tion and therefore mayor may not have much ap- ducing the influence of our verbal concepts of what
plication to creative thinking. However, Arieti objects, people, and trees look like. We all have
(1976) discussed the role of different cortical areas stereotyped ideas of, for example, a chair, and we
in creativity, stressing the temporal, parietal, and draw it, albeit inaccurately, to accord with these
prefrontal rather than the sensory areas. preconceptions. But if a chair is shown upside-
Lateralization offunction is, of course, an aspect down and we draw not the object but the spaces
of localization. Most of our knowledge of later- between the legs, seat, and back, our left brain is
alization derives from brain-damaged persons who not interested in empty spaces, and the result is
show very different symptoms according to much more successful artistically. Conceivably,
whether the damage occurred to the left or right then, creative visual artists do have a different style
hemisphere. Also, many inferences have been of perceiving and reproducing visual stimuli than
made from the abilities or defects of com- the average nontalented person.
missurotomized patients. But impaired functioning Another reasonable hypothesis is that the right
observed when some cortical area is injured, or brain is mainly responsible for grasping the overall
excised, does not necessarily mean that that area structure and contours of an object or scene to be
controls the same function in normal persons. For- drawn, while the left is more involved in recalling
tunately, there is other evidence from experiments and delineating details. Further, a case might be
with normals that left or right brain stimulation pro- made for the involvement of subcortical centers in
duces different effects, both in the visual and au- the emotional aspects of creative thinking, though
ditory fields. There are also differences in later- no direct evidence has been provided.
alization of visual and spatial functions between Chronological age is another physiological vari-
able of some significance according to Lehman's mediocre adult careers. On the whole, the most
(1953) finding that the peak of creative production positive evidence is that collected by sociologists
in physicists tends to occur in the 20s or 30s, among using Citation Indices, total numbers of publica-
biologists and social scientists more in the 40s, and tions, or other statistics on academic success, which
much more spread out in writers and philosophers. show substantial correlations and reliability over 1
His conclusions are strongly criticized by Dennis to 5 years or so. But, unfortunately, these do not
(1958). Highly original work occurs in most arts seem to have been explored over longer periods.
and sciences even up to 70 or later (e.g. ,Sophocles, Even if it did tum out that there is high consisten-
Giusppe Verdi, Pablo Picasso, Goethe, Sigmund cy over, say, 30 years, it would be dubious to argue
Freud, Bertrand Russell). Thus, although there may that stability in creative production is proof of ge-
be some mental vigor or flexibility characteristic netic determination. It might be attributed, at least
that tends to decline with age, we have not been able in part, to stability in environmental circumstances
to specify it. and stimulation. For example, persons brought up
in continuously favorable artistic or scientific en-
vironments, with access to good universities and
employment opportunities (say, as university
Consistency, or Predictability, from teachers) might be more consistently creative than
Early Childhood to Adulthood others reared under more unstable conditions.
On the analogy Of human physical charac-

teristics, such as height or red hair, which are Sex or Gender Differences
known to be very largely genetically determined
and are remarkably stable throughout most of the Considering the virtual unanimity of American
life period, one might expect mental charac- social and differential psychologists on the cultural
teristics, such as intelligence or creativity, to show origins of differences between males and females in
considerable stability from early to later years, if abilities, interests, attitudes, and personality, it
they, too, depend on genetic factors. But there are may seem odd to cite sex differences in creativity as
many difficulties. Often the genes do not manifest giving strong evidence of genetic factors. In sec-
their full effects until puberty. Certainly, measures ondary schools and colleges (in Western culture),
of intelligence show big fluctuations from infancy far fewer girls than boys opt for science and mathe-
until later childhood, though there are fairly high matics courses (see Cole & Cole, 1971, 1973;
correlations between 12-year and early adult IQs, Dresselhaus, 1984; Fox, Brody, & Tobin, 1980;
and even more consistency from then on until old Maccoby & Jacklin, 1974). Very few women ob-
age. Even physical attributes are to some extent tain Ph.D.s in science. Cole quotes only 3% in the
modifiable by environmental conditions conducive physical sciences in the 1970s; 8% in chemistry,
to good or poor health; and mental attributes are 15% in biology, but 18% in sociology and 24% in
likely to be much more liable. Hence, Gruber psychology. He claims a considerable growth in
(1974) goes so far as to say that no link has been their numbers during the 1970s, but it is difficult to
demonstrated between giftedness in childhood and disentangle this assertion when there are increasing
creative genius in adulthood. Indeed, we can hardly numbers of women entering graduate work in all
expect high creativity to manifest itself much before areas. Sometimes, too, the total number of female
age 20, because creative scientists and artists have science students increases, but the number of males
to acquire a great deal of knowledge and skills to be increases even more, so that the female proportion
creative with in the adolescent and early adult peri- actually drops. The discrepancy is greatest in high-
ods. Also, at least in the arts, it may be that the level employment, with extremely few women
emotional maturity and drive necessary to creative reaching full professorships in the natural sciences,
production do not develop until late adolescence though somewhat greater in the social sciences. The
(with rare exceptions). Hence, although several in- numbers of science publications by women also lag
vestigations, such as Cox's (1926), Roe's (1952), greatly behind those of men. Also women are more
and Eiduson's (1962), show most creative adults to often employed as teachers than as research
have manifested considerable talents during child- workers.
hood, there are probably many more equally tal- What is more surprising is that the numbers of
ented who either fade out or achieve only quite highly creative women in most of the arts are about
as low as in the sciences. This is true in music and scientific staff in universities. Their jobs are mostly
the visual arts, including sculpture and architecture. lower grade, and their rate of promotion is below
However, there are many well-known female writ- that of the male staff. They publish only about half
ers (though few poets). Probably there are more as many books or articles, so that their Citation
women than men in decorative and applied arts. Indices tend to be low. They have more difficulty in
The situation is very different, too, among perform- getting research grants or other rewards. But Cole's
ers. In music and drama, there is probably little analysis shows that, when their scientific qualifica-
difference in the totals of men and women, but in tions are comparable to those of the male staff (e. g. ,
ballet dancing, women are clearly in the majority. whether they have Ph.D.s), their status and awards
There can be no doubt that, in Western cultures, are much more similar. Although there is certainly
there are different pressures on boys or men and discrimination generally, it is less marked in uni-
girls or women to engage in different kinds of in- versities than is usually believed.
terests, educational courses, and careers. From In their well-known book on sex differences,
about the age of 1 year, boys are given different toys Maccoby and Jacklin (1974) showed how many of
from girls, for example, bricks or cars as against the widely accepted beliefs about the abilities and
dolls, cuddly animals, and books. Boys are encour- personalities of the sexes are traditional stereotypes
aged to engage in physical activities and are ex- that are not confirmed by controlled investigations;
pected to be more aggressive, girls are expected to they also stressed the amount of overlapping when
be quieter, more conformist, and to express them- real differences do occur. Frequently, too, such dif-
selves more in verbal than in motor actions. The ferences as are found are rather specific to particular
educational systems of Western society have ac- tests. For example, on many tests of verbal fluency
cepted the same stereotyping, and often it is the and comprehension, females tend to score higher,
influence of the teachers that deters girls from tak- but not on vocabulary or verbal information. They
ing advanced mathematics and science courses and are usually lower, again, on Witkin-type tests of
concentrating more in the arts, humanities, or do- field independence, but not necessarily on all tests
mestic training. Girls themselves accept these dif- of spatial abilities. In both these areas, the dif-
ferences, and a majority of them aspire to marriage ferences could be explained as due to social expec-
or to a limited range of jobs, such as teachers, secre- tations and stereotypes as by genetic factors. How-
taries, hairdressers, or nurses "which they can read- ever, there are obvious biological differences in
ily abandon for marriage. But Fox et ai, (1980) physical musculature and hormones that produce
noted that the proportions of mathematically or sci- greater strength and suitability for fighting in
entifically inclined girls vary considerably from one males. Hence, it is highly probable that males are,
high school to another, presumably because some by and large, more aggressive and dominant,
schools have higher quality staff, or different tradi- females more submissive and concerned with child
tions. Single-sex schools for girls tend to produce nurture and domestic activities. Such differences
more scientists than mixed sex, because their stu- are by no means restricted to Western culture but,
dents are less affected by fear of competition with with few exceptions, apply throughout all human
boys. Most girls in ordinary mixed schools show societies. Likewise they occur in many nonhuman
some degree of "fear of success." That is, they mammalian species, where males do most of the
prefer not to get higher marks or to appear to be hunting and fighting, and are less aggressive toward
better achievers than boys, because this makes them females than to males of their own species.
less attractive as potential marital partners. Thus, it is possible to make a case for certain
There are, of course, large numbers of girls who biological sex differences, particularly those that
reject these stereotypes and aim to become scien- are responsible for aggression, exploration, and ini-
tists or eventually to succeed in business careers. tiative in males, and nurturance in females; and
Some of them marry and bring up children as well. these traits might well underlie spatial, mechanical,
But these rebels against convention are a minority, and technological talents, even predisposing to sci-
and they may undergo considerable emotional entific inclinations in males, and verbal and domes-
stress as a result. Helson (1967) and Barron (1969) tic inclinations in females. It is hardly possible to
give good descriptions of the difficulties of women prove this, when social-environmental influences
mathematicians in pursuing their careers. are so strong. But it is entirely implausible that
Many writers, such as Cole (1979), have crit- human society should approve of females becoming
icized the apparent discrimination against female highly talented performers of music, dance, and
drama, and even allowing them to become creative Among the detailed pedigree tables that have
writers, while, at the same time, disapproving of been published, probably the most famous is that of
their becoming musical composers or painters. To the Bach family of musicians. From Veit Bach
me, this is the crux of the argument for attributing (died 1619) to Johann Christian Bach (died 1782),
sex differences in creativity at least, in part, to ge- there is some information on 60 related males,
netic factors. spanning seven generations. Of these, 53 were well
known as musicians, cantors, and/ or organists,
many of whom were eminent during their lifetimes.
Family Resemblance But it is noteworthy that only one of Johann Sebas-
tian Bach's sons is still regarded as an important
The statement that creativity or outstanding tal- creative musician, namely, Carl Philipp Emanuel.
ent tends to "run" in families is frequently heard A few others produced compositions that are still
but is almost impossible to justify, because such played, but none of these sons have lasting reputa-
resemblance between relatives can be attributed ei- tions comparable to their father's. Moreover, no
ther to common genes, or to environmental influ- other pedigree is anywhere near as striking, though
ences, or to both, and we have no method to decide several father-son resemblances are known, for ex-
which. To my knowledge, no one has studied or ample, Alessandro and Domenico Scarlatti. The
compared the creativity of mono- and dizygotic pedigrees are even more rare in the other arts and in
twins or has used any other controlled techniques the sciences. Often one member of a pair is consid-
for providing more reliable evidence. However, ered much more creative than the other, or they
many cases of identical twins who were reared differ in their particular branch of science. For ex-
apart, but who entered very similar occupational ample, Robert Thorndike's area of specialization is
careers, have been reported. largely different from that of his father, Edward.
According to Galton (1869), just half of his his- Surely, if there is a powerful genetic component in
torically eminent persons had fathers or other close scientific research or other forms of creativity, one
relations who were similarly gifted. Although this would expect it to last over several generations. Yet
finding has not been replicated, Terman's (1925) we must also remember that only half a parent's
gifted children came largely from upper-class fami- genes are passed on to his offspring, one quarter
lies. But this fact was not universal. As many as from the grandparent to the grandchild, and so on.
19% of Terman's group had unskilled or laboring- Hence, even with a naive genetic theory, the under-
class parents, and many of the geniuses were of lying genes are rapidly diluted.
lower-class origins, neither the father nor mother How far could parent-child resemblance be ac-
showing any superior abilities (so far as their biog- counted for by environmental influence? In the
raphies indicated). For example, Immanuel Kant, home of a highly musical parent, one would expect
Charles Dickens, Michael Faraday and John Bun- the offspring to be exposed to a wide range of musi-
yan all seem to have been raised in very poor cir- cal experience, from infancy through adolescence.
cumstances, but the available data are often too This would be less true in the sciences, though an
patchy to allow quantitative measures of familial eminent physicist, for example, would very likely
resemblance. look out for any signs of scientific interests or skills
Many instances of resemblance among siblings in his children, even, say, in biology, and would
could be cited, where two or more have shown usually give strong reinforcement by discussing and
outstanding creative talent for example the Bronte approving the child's activities, by providing scien-
sisters. Often siblings are eminent in different tific toys, books, and relevant TV documentaries.
areas, for example, Aldous and Julian Huxley, one Moreover, many of the father's friends would also
a writer, the other a biologist. Their grandfather, be scientists, who would give further encourage-
Thomas Huxley, was also an eminent scientist. In ment. Yet, in other homes, it often occurs that ado-
the area of music, Scheinfeld (1972) claimed that lescents rebel against parental pressures and set out
70% of the children, both of whose parents are to follow very different careers. Sometimes such
musical, are likely to be equally musical; with 1 offspring grow out of this stage and later revert to an
musical parent, 60% of the children are the same, occupation quite like the father's, but we have no
and when neither parent is musical, only 15% of the convincing data for indicating how frequently these
offspring are. This does not seem to have been patterns of resemblance and difference occur, nor
confirmed. why.
Upbringing and Environment Terman (1955) followed up all of his gifted chil-
dren who had become scientists, contrasting them
Although our evidence for genetic determinants with those in business, law, or the humanities. The
in creativity is rather weak, it should be possible to main difference, when first tested at about 11 years,
get better proof of the effects of home and school was that the former already had considerable scien-
environment, and differences in upbringing. These tific interests and activities, and were lower in so-
factors appear to stimulate, or inhibit, creativity in cial participation than the nonscientists.
general, or to affect certain types or aspects of cre- The work of MacKinnon (1962) and his col-
ative production. Several investigations have been leagues tends to agree with and expand that of Roe.
carried out using case studies, interviews, or bio- Creative architects and others usually show more
graphical data obtained from highly creative scien- radical, nonconformist attitudes, introverted and
tists or artists. Other studies have used personality independent and often neurotic traits, though, as
tests or biographical questionnaires, which have the Barron (1969) argues, these are controlled by well-
advantages that they can be given to large groups developed ego-strength. Taylor (1962) described a
and scored to yield quantitative measures. Besides study of typical American male adults, as con-
Cox's (1926) work, Roe's (1952) surveys of several trasted with creative individuals. They were gener-
groups of scientists and Bloom and Sosniak's ally conventional, extroverted, unimaginative, and
(1981) study of talent development are outstanding. concerned with home and sports rather than cultural
Taylor and Ellison's (1964), Cattell and Butcher's or intellectual interests.
(1968), and Chambers' (1964) investigations are In the early 1960s, Taylor and Ellison (1964)
characteristic of the second, more quantitative ap- realized that neither academic achievement, intel-
proach. MacKinnon's (1962), Barron's (1969), and ligence, nor other ability tests were of use in picking
Eiduson's (1962) studies combine both types of evi- out industrial or research scientists, and tried in-
dence. A useful overall survey of psychosocial de- stead to construct predictive biographical invento-
velopment of highly gifted children is provided by ries. They included large numbers of items dealing
Janos and Robinson (1985). with childhood and adolescent interests, favorable
Although Cox drew attention to the importance factors in home background, and current leisure
of good home conditions and of the personality activities. These inventories were given to adult
qualities of her geniuses, she also emphasized the scientists who had been rated high or low in
great variations in upbringing and the wide person- creativity by their supervisors; and the best differ-
ality differences. Roe likewise claimed that, in her entiating items were retained. When given to other
groups of scientists, every personality was unique. rated groups, the scores correlated consistently at
Yet she arrived at a rather frequent basic pattern of around .50 with the criterion. Several such invento-
childhood growth. A great many of her subjects ries were prepared, but it appeared that items were
showed a high degree of independence and soli- somewhat unreliable over time; hence, an elaborate
tariness; their relations with their parents were process of reconstruction and validation might be
rather impersonal, though they were strongly en- needed after a few years. Several other such instru-
couraged in achievement and in intellectual in- ments have been produced, for example, by Buel
terests. An unusual proportion of them were first (1965) and by Schaefer and Anastasi (1968). Also
born or only children, and many tended to show Nichols and Holland (1963) found that similar
poor health, which reduced their contacts with other questionnaires given to college freshmen yielded
age-peers. Many, again, had been strongly influ- small but significant correlations with their self-
enced by another relative, adult friend, or an inspir- rated accomplishments in scientific, musical, dra-
ing teacher. As adults they were rather detached and matic, and visual arts a year later. They further
were much more strongly committed to their work found predictive items in questionnaires given to
than to social or sexual activities. Roe's study is parents, dealing with their own interests and social
exceptional in showing a very marked difference in attitudes.
the upbringing and personalities of 42 physicists However, the above methods do not provide us
and biologists on the one hand, and 22 social scien- with a kind of recipe for the "good" home, which
tists on the other. The latter had had more stormy produces creative adult offspring, nor a basis for
childhoods and conflicts with their parents, but differentiating those who are likely to be outstand-
were more gregarious as adults and more unstable ing in any particular art or science. The question-
emotionally; some 40% of them had been divorced naires tend to support environmental explanations
as against 10% of the natural scientists. of creativity but not to prove them. Indeed, some-
times they contradict our expectations. Ammons colleges attended and the geographic areas which
and Ammons (1962) presented several case studies produced most scientists (see also Brannigan, 1981;
of eminent creatives (e.g., John Stuart Mill and and Coler, 1963). Clark (1957) surveyed the back-
Mozart) who as children were grossly over- grounds and careers of American psychologists,
pressured by their fathers. Apparently, the kind of and Moulin (1955) analyzed the countries of origin
conditions regarded by child psychologists as es- and other characteristics of Nobel prizewinners
sential to mental health were entirely lacking. Some from 1901 to 1950. Other writers (e.g., Datta,
geniuses, also, seem to have been stimulated rather 1967; Roe, 1952) found that religious background
than impeded by opposition, for example, Galileo is influential, the greatest number of eminent scien-
and Freud. However, the romantic notion that all tists coming from Jewish or liberal Protestant
great artists must have experienced suffering and homes, the fewest from fundamental Protestant or
conquered their stresses and hardships has no sound Catholic homes.
backing. It has often been noted that "golden ages" occur
Bloom and Sosniak (1981) carried out a detailed in particular countries at particular periods, when a
inquiry into the development of outstanding talent large number of creative individuals in one or more
among 25 men in each of 6 areas-2 scientific, 2 of the arts or in the sciences reach an outstanding
artistic, and 2 athletic. They emphasized the enor- level of creativity. Among the most notable were
mous amount of support and planning given by the the dramatists, philosophers, and sculptors in clas-
homes from age 12, or even earlier. The boys com- sical Greece around 400-300 B.C., and the paint-
monly spent more time on practice and coaching in ers, architects, and writers of the fourteenth- and
their talent areas than on ordinary schoolwork; and fifteenth-century Renaissance period in northern
this training was more individualized and more in- Italy. Certainly, such occurrences have environ-
tensive than they received at school. However, this mental effects, though we have little or no knowl-
applied more to the athletes and artists than to the edge as to why or how they occur.
scientists.
There is some evidence of a greater degree of
authoritarianism than of permissiveness in the intel- Trainability
lectuallives of future gifted scientists (Kennedy &
Willcut, 1963). Parloff, Datta, Kleman, and Hand- If the influence of genetic factors on the develop-
low (1968) gave a personality inventory to some ment of creativity is of minor or no importance, it
900 science talent scholars, who had produced cre- should be possible to teach or train many aspects of
ative scientific projects, and found that they scored creative thinking in much the same ways as educat-
highly on a personality factor designated as Disci- ing to read and to do arithmetic. Currently, there is a
plined Effectiveness. The subjects were not usually strong trend in this direction. Many writers, such as
characterized by the nonconformist or rebellious Torrance (1965), de Bono (1970), and Parnes
traits found among MacKinnon's architects or (1963), believe that, from the earliest school years,
among some of Roe's scientists. Hudson (1966) the curriculum should include extensive training in
found English high school boys and college stu- a variety of divergent thinking tasks, and that these
dents, who were taking science, to be mainly con- should improve the all-around capacity to show
vergent thinkers, whereas those taking literary imaginative, flexible thinking, leading up to cre-
courses were more often divergent thinkers. Storr ative problem solving. On the other hand, other
(1972) suggested that many highly creative indi- psychologists, such as Mansfield, Busse, and Kre-
viduals can be classified under pathological head- pelka (1978), Cattell and Butcher (1968), and
ings, such as schizophrenic, manic depressive, or MacKinnon (1968), are highly critical of such an
obsessional, but provided no empirical evidence. approach. They point out that the work of creative
Some of the earliest studies of environmental ef- scientists and artists is totally unlike Osborn's
fects were James McKeen Cattell's (1906) demo- brainstorming, de Bono's lateral thinking, or the
graphic surveys of the backgrounds and careers of numerous schemes and programs now being pub-
American scientists. He showed that the production lished for the enhancement of creative thinking.
of eminent scientists was greatest in those Ameri- From the earliest days of educational psychology,
can states with good educational systems, many the difficulties of bringing about transfer of training
libraries and laboratories, and a high degree of tol- have been stressed. As G. M. Stratton (1922) put it:
erance and liberality toward scientific research. "what you do to the mind by means of education
Knapp and Goodrich (1952) studied the types of knows its place; it never spreads. You train what
you train." In other words, coaching and practice Stanley discovered far fewer mathematically gifted
may considerably improve ability in some particu- girls, he was able to pull many of them up much
lar activity, but they have limited or scarcely any nearer the boys' standards by designing summer
transfer to other situations or tasks which might courses or other special classes.
appear to be similar, or which have the same name. Many books have doubtless been written on
Torrance (1972) published a survey of 142 fol- school or college teaching, which aim to show how
low-up studies of divergent thinking training and to stimulate creative students and not merely im-
claimed that three-quarters of these gave positive plant conventional knowledge and skills. Brand-
evidence of improvement. But, in most of these wein's (1955) book, on the recruitment and teach-
cases, the criterion of creative thinking was much ing of promising science students at high school,
the same as the practice exercises, namely, scores may be an old one, but it brings together a wealth of
on divergent thinking tasks. For example, scores on data on stimulating interest in, and commitment to,
the Alternate Uses test may be affected by the train- science, the need for very high academic and intel·
ing, but they cannot be taken to demonstrate gener- lectual standards among the students, the plannin~
alization to real-life problem-solving or artistic pro- and individualization of courses, and the quality 01
ductivity. Also few, if any, investigators appear to the staff.
have carried out long-term follow-up studies, in At the college level, Heist (1968) deplored the
which older students or adults, who have received loss of potentially creative scientists because of
so-called creativity training as children, were found their frustration with the rigidity and spoonfeeding
to be superior as scientists or artists to those who of undergraduate science courses. Many requisite
received no such childhood training. Obviously, it graduate courses can also be more stifling than
would be very difficult to ensure that those without stimulating; and similar criticisms apply to much
childhood experiences were equivalent to those course work in the fine arts. Chambers (1973) stud-
with such experience in general intelligence, quali- ied the characteristics of ineffective psychology and
ty of schooling, and many other possible influ- chemistry teachers by means of class observations
ences. Yet, until such evidence is forthcoming, we and student ratings. "Facilitating," as contrasted
are not entitled to claim that creative ability is with "inhibiting," teachers encourage class dis-
trainable. cussions and seldom read from notes. They are en-
Nevertheless, Stein's (1975) book, Stimulating thusiastically committed to their subjects and are
Creativity, surveyed a number of relatively well- readily available to students for consultations.
controlled experimental studies that make a better Their courses/ are also intellectually demanding.
case for some positive short-term effects. Fre- Just as in home upbringing, permissive teaching
deriksen (1984) demonstrated improvements in can do more harm than good; and research students
quite a range of complex problem-solving skills, have to experience hard work, with frequent frus-
which are more carefully defined and measured trations, in order to achieve the grounding neces-
than creativity in general. sary for creative research. At the same time, the
Rather more is known of the effects of organiza- more unconventional and rebellious students, who
tion of schools and the planning of suitable courses may be a nuisance to their instructors and super-
from such work as Stanley's and his colleagues on visors, are likely to require more sympathetic han-
mathematically precocious adolescents (Stanley et dling and encouragement. But these problems are
al., 1974; Daurio, 1979). Large numbers of mathe- more matters of discussion and argument than of
matically gifted students were making very poor controlled research.
progress because of the slow pace at which school
mathematics was taught, unless they happened to
get help at home or obtained books for self-teach- Summary and Conclusions
ing. But when placed in special programs designed
to let them progress at their own rates, they gained 1. The difficulties of defining and measuring
three or more years in grade standing and often creativity are so great that no firm conclusions can
obtained university degrees much earlier than usu- be reached on the relative importance of genetic and
al. Unfortunately, there has been little attempt to environmental factors; the latter including so-
follow them up several more years in order to see ciological variables as well as individual dif-
how many, as a result of this boost, have become ferences in upbringing. Complex interactions occur
research mathematicians or scientists. Although among all these factors in the growth of creativity.
However, a considerable amount of psychological, brain growth. The popular view that creativity de-
sociological, biographical, neurological, and other pends much more on the right than on the left side of
types of evidence is available. the brain is too speculative to be acceptable. Like-
2. Creativity is taken to mean the production of wise, changes in productivity with brain growth and
novel ideas, theories, and objects, either in the sci- later decay are of minor importance.
ences or in the arts, which are accepted by compe- 7. Giftedness observed during childhood and ad-
tent experts as original and valuable. Such produc- olescence gives rather poor indications of creative
tivity is quite rare and, though a matter of degree, productivity in adulthood. Although there is clearly
tends to be concentrated among relatively small some consistency and continuity in adults, fluctua-
numbers of scientists and artists. Creative effort is tions also occur that might be attributable more to
not the same as an ability or trait that is presumed to changes in environmental circumstances or person-
underlie unconventional behaviors or to responses ality. Thus, this line of evidence neither confirms
to "divergent thinking" tests, which do approxi- nor disproves genetic theories.
mate to a normal distribution throughout the 8. The numbers of creative women scientists and
population. artists tend to be much lower than those for men.
3. In scientific discovery, the term environment Most writers attribute this difference to sex-role
includes not merely home, school, and cultural in- stereotypes and differential pressures on the sexes
fluences, but also the present state of knowledge that is exerted by parents, peers, schools, and soci-
and instrumentation, together with chance stimuli ety in general. However, there are obvious biolog-
that suggest new ideas, favorable conditions of ical sex differences favoring more aggressive and
work, and cooperation or competition with other dominant behavior in males and more compliant
scientists. Thus, it should not be regarded as a kind and nurturant behavior in females. These traits ap-
of possession of the individual scientist, whether it ply to most known human societies and to many
be innate or acquired. mammalian species. Moreover, purely social
4. Descriptive and experimental approaches to causation could not readily explain the greater
the mental processes involved in scientific and ar- female achievement in particular areas, for exam-
tistic creativity are being fruitfully studied by cog- ple, in writing and in the decorative and the per-
nitive psychologists and information theorists. But forming arts as against painting and musical
the degree to which creative thinking differs from composition.
other complex thinking is not yet decided. A great 9. Considerable resemblances are often observed
deal of observational and introspective evidence in- between offspring, parents, sibs, and other relatives
dicates the role of the subconscious processes in in creative production. But there are also frequent
inspiration and in creativity, though-particularly differences in achievement or in areas of specializa-
in the sciences-rational processes of acquiring tion. Although genetic factors cannot be ruled out,
and analyzing knowledge and working out implica- such similarities are probably better explained by
tions of new ideas tend to predominate. factors of upbringing and environment.
5. Creative production correlates moderately 10. When highly creative scientists or artists are
with general intelligence. Scientists tend to show compared with less creative or normal individuals,
the highest IQs, but a great many exceptions occur. a great many differences in personality, home back-
Thus, neither intelligence tests nor other cognitive ground, and upbringing are found. Demographic
tests, such as Terman's Concept Mastery or differences are also associated with productivity.
Guilford's divergent thinking, are valid predictors But because there is much overlapping, no particu-
of future scientific or artistic creativity. This result lar personality or environmental factors differenti-
should be expected in the light of the conclusions in ate in all cases. Thus, we cannot manipulate or
numbers 1 to 3 above. control the development of creativity. Many out-
6. The strongest evidence for genetic factors is standing creative individuals have succeeded de-
the appearance of persons of outstanding genius, spite apparently unfavorable home, school, or other
both in the arts and the sciences, beyond what could conditions.
reasonably be attributed to family upbringing and Differences between artists and scientists, or be-
environment. There is also some evidence linking tween those working in different areas, have also
aspects of creativity with particular areas of the been observed, but again the results show many
brain, though we cannot rule out the alternative inconsistencies.
theory that creative activity itself tends to stimulate 11. Several attempts to train or develop creative
talent by school activities and exercises are less The other line of recommended exploration is an
effective than claimed. Specific skills can usually extension of longitudinal studies similar to those of
be trained by instruction and practice, but these Roe and of MacKinnon. A large group of students
seldom generalize or transfer to broader activities, identified as highly gifted at, say, 12 years might be
such as creative production. No long-term follow- followed through until they reached about 25 years,
up studies can be quoted to show that such training in order to find the home upbringing and school
produces more high-creative scientists or artists. progress characteristics of those who end up as
However, progress is being made by cognitive psy- highly creative, and those who fade out and become
chologists in influencing the development of com- mediocre adults. Fluctuations in interests and per-
plex problem-solving skills. Better planning and sonality during adolescence would be particularly
organization of the curriculum in high schools and revealing. Less time-consuming studies might start
colleges have been shown to affect achievement with college freshman and attempt to analyze the
(e.g., in mathematics). But so much depends on the college courses and avocations of those who, per-
motivation of the students and the quality of their haps 6 years later, became really promising scien-
teachers that it is difficult to specify general tists or artists. Again, there seem to have been no
principles. follow-up studies of those designated as creative at
age 25 to elicit the conditions under which they
achieve further successes or declines over the next
Future Developments 10 or more years.
Quite short-term studies that would be of value
Readers of a chapter such as this might expect to could be based on work in progress. They should
be provided with a list of the most promising ave- focus more on the motivational and emotional as-
nues for future research. But given the complexities pects of creativity than on the cognitive features.
of collecting any hard evidence, it is likely that These studies, too, should cover several scientific
every author would recommend a different list. We and/or artistic areas.
are still in the stage of carrying out fishing expedi- Given a large enough sample of highly creative
tions rather than recommending specific hypoth- young adults (e.g., through public advertisements),
eses and methods. However, two general consid- it would be informative to study the backgrounds,
erations can be suggested. First, that the processes present activities, and personalities of pairs of
to be studied should be those that directly involve mono- and dizygotic twins, and a control group of
creative activities in one or more of the arts or sci- siblings. The calculation of heritability indices
ences. There is no good reason for concentrating, as would be of minor importance; but analyses of re-
at present, on such simple tasks as those used in semblances and differences by case-study methods
divergent thinking. could be quite fruitful.
Second, the samples of subjects to be studied
should be mainly adults of, say, 25 years or over,
who can be assessed as genuine creatives (e.g., by References
their recent teachers). Adequate control samples
who lack such creativity are also needed. As in Agassi, J. (1985). The myth of the young genius. Interchange,
Terman's follow-up of his gifted group, such cre- 16. 51-60.
ative individuals are likely to be interested in psy- Albert, R. S. (1975). Towatd a behavioral definition of genius.
American Psychologist. 30, 140-151.
chological investigations of their work and to coop- Ammons, C. H., & Ammons, R. B. (1962). How to prevent
erate in interviews or questionnaire surveys. There genius: McCurdy revisited. Proceedings of Montana Acade-
has been con~iderable progress in our knowledge of my of Science, 21, 145-152.
creativity over the past 30 years or so. Thus, given Arieti, S. (1976). Creativity: The magic synthesis. New York:
adequate funding, the prospects for further ad- Basic Books.
Batber, B., & Hirsch, W. (Eds.). (1962). The sociology of sci-
vances are favorable. ence. New York: Free Press.
The major approach for experimentalists and Barron, F. (1969). Creative person and creative process. New
theorists should be the extension of cognitive psy- York: Holt, Rinehart & Winston.
chological studies, like those of Sternberg and of Bartlett, F. C. (1958). Thinking: An experimental and social
study. London: Allen & Unwin.
Frederiksen, which should cover genuine scientific Bartlett, M. M., & Davis, G. A. (1974). Do the Wallach and
research and discovery, or the actual production of Kogan tests predict real creative behavior? Perceptual and
musical, visual, or literary works. Motor Skills. 39. 730.
Biber, B. (1984). Early education and psychological develop- Eysenck, H. J., & Kamin, L. (1981). The intelligencecontrover-
ment. New Haven: Yale University Press. sy. New York: Wiley.
Blackwell, R. J. (1969). Discovery in the physical sciences. Feldman, D. H. (Ed.). (1982). Developmental approaches to
Notre Dame: Notre Dame University Press. giftedness and creativity. San Francisco: Jossey-Bass.
Bloom, B. S., & Sosniak, L. A. (1981). Talent development vs. Fox, L. H., Brody, L., & Tobin, D. (Eds.). (1980). Women and
schooling. Educational Leadership, 39(2), 86-94. the mathematical mystique. Baltimore: Johns Hopkins Uni-
Brandwein, P. F. (1955). The gifted student asfuture scientist. versity Press.
New York: Harcourt, Brace. Frederiksen, N. (1984). Implications of cognitive theory for
Brannigan, A. (1981). The social basis of scientific discoveries. instruction in problem solving. Review of Educational Re-
New York: Cambridge University Press. search, 54, 363-407.
Buel, W. D. (1965). Biographical data and the identification of Galton, F. (1869). Hereditary genius. London: Macmillan.
creative research personnel. Journal of Applied Psychology, Gardner, H. (1978). What we know (and don't know) about the
49, 318-321. two halves of the brain. Journal of Aesthetic Education, 12,
Campbell, D. T. (1960). Blind variation and student retention: 113-119.
Creative thought as in other knowledge processes. Gardner, H. (1980). Artful scribbles: The significance of chil-
Psychological Review, 67, 380-400. dren's drawings. New York: Basic Books.
Cattell, J. McK. (1906). A statistical study of American men of Gardner, H. (1983). Frames of mind: The theory of multiple
science. III. The distribution of American men of science. intelligences. New York: Basic Books.
Science, 24, 732-742. Getzels,1. W., & Jackson, P. W. (1962). Creativity and intel-
Cattell, R. B. (1971). Abilities: Their structure, growth and ligence: Explorations with gifted children. New York: Wiley.
action. Boston: Houghton Mifflin. Ghiselin, B. (1952). The creative process: A symposium.
Cattell, R. B., & Butcher, H. J. (1968). The prediction of Berkeley: University of California Press.
achievement and creativity. New York: Bobbs-Merrill. Gibson, J., & Light, F. (1967). Intelligence among university
Chambers, J. A. (1964). Relating personality and biographical students. Nature, 213, 441-442.
factors to scientific creativity. Psychological Monographs, Grinder, R. E. (1985). The gifted in our midst: By their divine
78, No. 584. deeds, neuroses, and mental test scores we have known them.
Chambers, J. A. (1973). College teachers: Their effect on In F. D. Horowitz & M. O'Brien (Eds.), The gifted and tal-
creativity of students. Journal ofEducational Psychology, 65, ented: Developmental perspectives. Washington, DC: Ameri-
325-334. can Psychological Association.
Clark, K. E. (1957). America's psychologists: A survey of a Gruber, H. E. (1974). Darwin on man: A psychological study of
growing profession. Washington, DC: American Psychologi- scientific creativity. New York: Dutton.
cal Association. Guilford, J. P. (1967). The nature of human intelligence. New
Cole, J. R. (1979). Fair science: Women in the scientific commu- York: McGraw-Hill.
nity. New York: Free Press. Hadamard, J. (1945). An essay on the psychology of inven-
Cole, J. R., & Cole, S. (1971). Measuring the quality of so- tion in the mathematicalfield. Princeton: Princeton University
ciological research: Problems in the use of the Science Cita- Press.
tion Index. American Sociologist, 6, 23-29. Heist, P. (Ed.). (1968). The creative college student: An unmet
Cole, J. R., & Cole, S. (1973). Social stratification in science. challenge. San Francisco: Jossey-Bass.
Chicago: University of Chicago Press. Helmholtz, H. von (1903). Vortrage und Reden (Fiinfte
Coler, M. A. (Ed.). (1963). Essays on creativity in the sciences. Auflage). Braunschweig: F. Vierweg.
New York: Universities Press. Helson, R. (1967). Sex differences in creative style. Journal of
Cox, C. M. (1926). The early mental traits of three hundred Personality, 35, 214-233.
geniuses. Stanford, CA: Stanford University Press. Hudson, L. (1966). Contrary imaginations. London: Methuen.
Cronbach, L. J. (1968). Intelligence? Creativity? A par- Janos, P. M., & Robinson, N. M. (1985). Psychosocial develop-
simonious reinterpretation of the Wallach-Kogan data. ment in intellectually gifted children. In F. D. Horowitz & M.
American Educational Research Journal, 5, 491-511. O'Brien (Eds.), The gifted and talented. Washington, DC:
Datta, L. (1967). Family religious background and early scien- American Psychological Association.
tific creativity. American Sociological Review, 32, 626-635. Kennedy, W., & Willcutt, H. (1963). Youth-parent relations of
Daurio, S. P. (1979). Educational enrichment versus accelera- mathematically gifted adolescents. Journal of Clinical Psy-
tion: A review of the literature. In W. C. George, S. J. Cohn, chology, 19, 400-402.
& J. D. Stanley (Eds.), Educating the gifted. Baltimore: Johns Knapp, R. H., & Goodrich. H. B. (1952). Origin of American
Hopkins University Press. scientists. Chicago: University of Chicago Press.
De Bono, E. (1970). Lateral thinking. New York: Harper & Kogan, N., & Pankove, E. (1974). Long-term predictive valid-
Row. ity of divergent-thinking tests: Some negative evidence.
Dennis, W. (1958). The age decrement in outstanding scientific Journal of Educational Psychology, 66, 802-810.
contributions. American Psychologist, 13, 457-460. Kubie, L. S. (1958). Neurotic distortion of the creative process.
Dresselhaus, M. S. (1984). A physicist's perspective on the New York: Noonday Press.
complementary roles of the physical and behavioral sciences. Lehman, H. C. (1953). Age and achievement. Princeton: Prince-
American Psychologist, 39, 333-340. ton University Press.
Drevdahl, J. E. (1956). Factors of importance for creativity. Lenneberg, H. (1980). The myth of the unappreciated (musical)
Journal of Clinical Psychology, 12, 21-26. genius. Musical Quarterly, 66, 219-231.
Edwards, B. (1979). Drawing on the right side brain. Los An- Maccoby, E. E., & Jacklin, C. N. (1974). The psychology of sex
geles: Tarcher. differences. Stanford, CA: Stanford University Press.
Eiduson, B. T. (1962). Scientists: Their psychological world. MacKinnon, D. W. (1962). The personality correlates of
New York: Basic Books. creativity: A study of American architects. Proceedings of the
XIV Congress of Applied Psychology (Vol. 2). Copenhagen: Shapiro, R. J. (1968). The identification of creative research
Munksgaard. scientists. Psychologica Africana, Monograph Supplements,
MacKinnon, D. W. (1968). Educating for creativity: A modern No.4.
myth. In P. Heist (Ed.), The creative college student. San Simonton, D. K. (1977). Eminence, creativity, and geograph-
Francisco: Jossey-Bass. ical marginality: A recursive structural equation model.
Mansfield, R. S., & Busse, T. V. (1981). The psychology of Journal of Personality and Social Psychology, 35, 805-816.
creativity and discovery: Scientists and their work. Chicago: Springer, S. P., & Deutsch, D. (1981). Left brain: Right brain.
Nelson-Hall. San Francisco: Freeman.
Mansfield, R. S., Busse, T. V., & Krepelka, E. J. (1978). The Stanley, J. C., Keating, D. P., & Fox, L. H. (Eds.). (1974).
effectiveness of creativity training. Review of Educational Mathematical talent: Discovery, description and develop-
Research, 48, 517-536. ment. Baltimore: Johns Hopkins University Press.
Mednick, S. A., & Mednick, M. E. (1967). Remote Associa- Stein, M. I. (1974). Stimulating creativity (Vol. I). New York:
tions Test. Boston: Houghton Mifflin. Academic Press.
Merton, R. K. (1968). Social theory and social structure. New Stein, M. I. (1975). Stimulating creativity (Vol. 2). New York:
York: Free Press. Academic Press.
Mistry, 1., & Rogoff, B. (1985). A cultural perspective on the Sternberg, R. J., & Davidson, J. E. (1985). Cognitive develop-
development of talent. In F. D. Horowitz & M. O'Brien ment in the gifted and talented. In F. D. Horowitz & M.
(Eds.), The gifted and talented. Washington, DC: American O'Brien (Eds.), The Gifted and Talented. Washington, DC:
Psychological Association. American Psychological Association.
Moulin, L. (1955). The Nobel prizes for the sciences from 1901 Storr, A. (1972). The dynamics of creation. London: Secker &
to 1950. British Journal of Sociology, 6, 246-263. Warburg.
Narroll, R., Benjamin, E. c., Fohl, F. K., Fried, M. J., Hil- Stratton, G. M. (1922). Developing Mental Power. New York:
dreth, R. E., & Schaeffer; J. M. (1971). Creativity: A cross- Houghton-Mifflin.
historical pilot survey. Journal ofCross-Cultural Psychology, Suler, J. R. (1980). Primary process thinking and creativity.
2, 181-188. Psychological Bulletin, 88, 144-165.
Newell, A., & Simon, H. A. (1972). Human problem-solving. Taylor, C. W. (1962). Who are the exceptionally creative?
Englewood Cliffs, NJ: Prentice-Hall. Exceptional Children, 28,421-431.
Nicholls, J. G. (1972). Creativity in the person who will never Taylor, C. W., & Ellison, R. L. (1964). Predicting creative
produce anything original and useful: The concept of performance from multiple measures. In C. W. Taylor (Ed.),
creativity as a normally distributed trait. American Psychol- Widening horizons in creativity. New York: Wiley.
ogist, 27, 717-727. Taylor, I. A. (1959). The nature of the creative process. In P.
Nichols, R. c., & Holland, J. L. (1963). Prediction of the fust Smith (Ed.), Creativity: An examination of the creative pro-
year college performance of high aptitude students. cess. New York: Hastings House.
Psychological Monographs, 77, No. 570. Terman, L. M. (1925). Genetic studies of genius: Vol.J. Mental
Ornstein, R. E. (1977). The psychology of consciousness. New and physical traits of a thousand gifted children. Stanford,
York: Harcourt Brace Jovanovich. CA: Stanford University Press.
Parloff, M. B., Datta, L., Kleman, M., & Handlow, J. H. Terman, L. M. (1955). Are scientists different? Scientific Ameri-
(1968). Personality characteristics which differentiate cre- can, 192, 25-29.
ative male adolescents and adults. Journal of Personality, 36, Torrance, E. P. (1965). Rewarding creative behavior.
528-552. Englewood Cliffs, NJ: Prentice-Hall.
Parnes, S. J. (1963). Education and creativity. Teachers College Torrance, E. P. (1972). Can we teach children to think cre-
Record, 64, 331-339. atively? Journal of Creative Behavior, 6, 114-143.
Perkins, D. N. (1981). The mind's best work. Cambridge: Har- Vernon, P. E. (1972). The validity of divergent thinking tests.
vard University Press. Alberta Journal of Educational Research, 18, 244-258.
Poincare, H. (1924). Thefoundations of science. London: Sci- Vernon, P. E. (1979). Intelligence: Heredity and environment.
ence Press. San Francisco: Freeman.
Price, D. J. (1962). The exponential curve of science. In B. Vernon, P. E. (1984). Intelligence, cognitive styles, and brain
Barber & W. Hirsch (Eds.), The sociology of science. New lateralization.lnternational Journal of Psychology, 19, 435-
York: Free Press. 455.
Rabinowitz, M., & Glaser, R. (1985). Cognitive structure and Vernon, P. E., Adamson, G., & Vernon, D. F. (1977). The
process in highly competent performance. In F. D. Horowitz psychology and education of gifted children. London:
& M. O'Brien (Eds.), The gifted and talented. Washington, Methuen.
DC: American Psychological Association. Wallach, M. A., & Kogan, N. (1965). Modes of thinking in
Roe, A. (1952). The making of a scientist. New York: Dodd, young children. New York: Holt, Rinehart & Winston.
Mead. Wallach, M. A., & Wing, C. W. (1969). The talented student.
Rogers, K. B. (1986). Do the gifted think and learn differently? New York: Holt, Rinehart & Winston.
A review of recent research and its implications for instruc- Wallas, G. (1926). The art of thought. London: Jonathan Cape.
tion. Journal for the Education of the Gifted, 10, 17-39. Welsh, G. S. (1975). Creativity and intelligence: A personality
Rosner, S., & Abt, L. E. (Eds.). (1970). The creative experi- approach. Chapel Hill, NC: Institute for Research in Social
ence. New York: Grossman. Science.
Schaefer, C. A., & Anastasi, A. (1968). A biographical invento- Zigler, E., & Farber, E. A. (1985). Commonalities between the
ry for identifying creativity in adolescent boys. Journal of intellectual extremes: Giftedness and mental retardation. In F.
Applied Psychology, 52, 42-48. D. Horowitz & M. O'Brien (Eds.), The Gifted and Talented.
Scheinfeld, A. (1972). Heredity in humans. Philadelphia: Washington, DC: American Psychological Association.
Lippincott.
CHAPTER 6
Creativity and Intelligence

Patricia A. Haensly and Cecil R. Reynolds
The Relationship between Creativity support a view of the two phenomena as integrated
and Intelligence in optimal mental performance.
A frequently recurring theme in the consideration
Consideration of the relationship between creativity of the dual phenomena of creativity and intelligence
and intelligence during this past half-century has has been that creativity is not independent of the
occupied the attention of psychologists with varied general factor of intelligence (Yamamoto, 1965).
perspectives (e.g., Cattell, 1963a; Coler, 1963; Intelligence appears to be a necessary but not suffi-
Guilford, 1950, 1959, 1968, 1981; MacKinnon, cient condition for creativity (Rossiman & Hom,
1962; Roe, 1951, 1963; Terman, 1954, 1955; Tor- 1971); that is, although intelligence appears to al-
rance, 1960, 1967). Scientists working in a variety low the development of creativity, it does not en-
of other disciplines from genetics to engineering sure that creative expression always will be forth-
(see Simonton, 1984, 1985; Taylor & Barron, coming (Schubert, 1973). However, many, if not
1963) also have been intrigued by the contribution most, students of the phenomena hold the view that
of intelligence to creative discovery and invention. intelligent thinking must also include some degree
And philosophers from Aristotle and Plato to Im- of creative thinking. The most prevalent view,
manuel Kant, Brand Blanshard, and Jacques Mari- then, has been that creativity is a distinct category
tain (citing here only those of Western civilization) of mental functioning that has limited overlap with
(Rothenberg & Hausman, 1976) have pondered the intelligence, both in the processes used and in the
origin of creativity and its relationship to rational characteristics of individuals who exhibit them.
thought. Professionals and lay people alike are fas- Because the terms creativity and intelligence
cinated by the topic, and, even more so, by the carry strong valences in the images they convey, it
consequences of applying creative effort and/or in- is important that our examination not be hampered
telligent action, and by the implications for their by conceptual biases stemming from images that
nurturance and educability. This chapter will pro- may not be accurate. As deBono (1971) suggests,
vide a brief overview of past perspectives regarding we must break "out of the old, self-perpetuating
the nature and extent of a relationship between patterns [and generate] new ways of looking at
creativity and intelligence and will propose and things" (deBono, 1971, p. 11).
Transitions in Contemporary Perspective

Patricia A. Haensly and Cecil R. Reynolds • Department
of Educational Psychology, Texas A&M University, College The description of the relationship between
Station, TX 77843. creativity and intelligence has, throughout the his-
111
tory of psychology, most often been approached studies of the processes followed by "creative"
from a position of defensiveness; that is, intel- artists and scientists conducted by himself and
ligence appears to be a "given," a characteristic Howard Gruber, stated that such individuals focus
organisms exhibit in varying degrees permitting their abilities in particularly adept ways. Gardner
them to adapt to their environment. Creativity, on asserted that these individuals integrate the spe-
the other hand, often is viewed as an appendage, an cialized activities of the left and right hemispheres
ancillary characteristic permitting humans to pur- of the brain in particularly effective ways (Gardner,
sue roads not usually traveled. Creative endeavor 1982, pp. 352-353). In addition, Gardner and
has even been viewed as an alternate pathway itself, Gruber's findings verify Feldman and Benjamin's
one not necessarily basic to survival or adaptation. (1986) ideas regarding the effect of genetic, famil-
Under these circumstances, creativity has been ex- ial, motivational, and cultural factors in the out-
plained as something external to models of intel- standing contributions of the creative genius, and,
ligence, something out of the realm of rational more importantly, on their coincidence. The bio-
thought. We propose that such an approach has graphical observations suggested that only when
been incorrect and counterproductive in that it lim- these variables come together in an appropriate time
its our understanding of how individuals function and place does creative genius seem to occur.
mentally. In fact, most of the research of the past 30 The effect of these additional variables on cre-
years or so on creativity has been inspired by the ative productivity is seen in other biographical anal-
interests Torrance developed from his military yses as well. Simonton (1979) found that anteced-
work in survival training, in which creativity was ents to creative production among Cox's 301
seen as integral to survival. eminent geniuses included not only intelligence but
Contemporary perspective or a "new think" also family background, role models, and formal
(deBono, 1971) regarding the relationship has be- education. Judging by the paucity of creative ge-
gun with consideration of the process of develop- nius, the correct combination of these factors ap-
ment of creative effort. However, in this develop- pears to be very difficult to assemble. Simonton
ment, the path by which an individual becomes even suggested that education at the graduate level
creative cannot be described without first describ- may lessen one's chances of achieving eminence.
ing the nature of its context, that is, the environment Simonton described achieved eminence as, in part,
or medium upon which the creative response acts a curvilinear or inverted U function of the amount of
(Csikszentmihalyi & Robinson, 1986). The nature formal education a person has. Based on his data,
of the response required to obtain the medium for its he suggested that "excessive amounts [of intel-
expression, or the experiences required to develop ligence] may inhibit creative development by en-
the skills necessary to express the creative re- forcing an overcommitment to traditional perspec-
sponses, also must be described (e.g., Bloom, tives" (Simonton, 1979, p. 81).
1985; Feldman & Benjamin, 1986). Consider, for Even when intelligence is given primary focus, a
example, the elements of intelligence that had to be similar view is perceived. Estes (1985), who de-
present for the mentally retarded artist Alonzo scribed intelligence as adaptive behavior, stated
Clemons (a 26-year-old man reported to have the that creativity or creative responses are often not
mental capacities of a 6-year-old) to become an adaptive for a particular setting or in a limited time
accomplished sculptor (Harper, 1983). Once, when frame. This may be seen, for example, in some
counselors at a state institution where Clemons academic settings, limited levels of scientific en-
grew up took away his modeling clay, Clemons deavor, traditional societal groups, in conventional
sneaked onto the roof and scraped up tar with his business, and other such settings. Estes suggested
fingernails from which he sculpted tiny black that the more intelligent individuals are, the more
horses. Such a response required an analysis of likely they will block their creative output in order
what could be done with soft tar and the goal-ori- to produce an adaptive response to the particular
ented behavior to acquire it without being caught. setting. A problem exists in that, in the long run,
Without "intelligence" within a very limiting con- this kind of blocking of creativity is likely to be
text, could his creative expression have taken maladaptive for the individual and for society.
place? The matter of choice in how and when an intel-
Analyses from a neurological perspective also ligent individual will allow his or her own creative
add insight to the relationship between intelligence expression may be seen in Krippner's systems ap-
and creativity. Gardner (1982), in describing the proach to creativity (Krippner, 1983). Krippner
CHAPTER 6 • CREATIVITY AND INTELLIGENCE 113
suggested that different avenues for creative effort though the difference between a dichotomous and
occur through expressed preferences for certain the proposed unitary or integrated view may be sub-
types of creative activity; that is, individuals exhibit tle, the implications of the latter for application,
preferences for the kinds of information they seek nurturance, and educability are extensive.
about their world, the mode through which they To initiate this analysis of past thinking about the
seek information, the way they process that infor- relationship between creativity and intelligence and
mation, and the way they make decisions regarding to develop a foundation for the integrated view,
their processed information. Creative effort may some basic questions should be addressed. What
arise in each of these preferences and will be exhib- specifically does creativity have to do with intel-
ited accordingly. Thus, Krippner classified indi- ligence? Can an individual ever be creative without
viduals who perceive information from their having intelligence? Can an individual demonstrate
senses, who are practical and oriented to what is a high degree of intelligence that has no creativity
feasible now, as analytic scientists; he suggested associated with it? Creativity and intelligence-can
that E. Paul Torrance is an example of this type and they be isolated from each other in verbal context,
that one of Torrance's greatest creative contribu- or in any other context? Despite careful factor anal-
tions has been the set of principles by which the ysis, is it possible to verify that the clusters of traits
environment that would stimulate creativity in the that emerge are completely independent of each
learning process could be created. On the other other? Would, for example, one cluster actually
hand, the individual who absorbs information occur if other clusters had not preceded or accom-
through his or her imagination and reaches deci- panied its emergence?
sions based on values unique to that individual- More particularly, can creativity or intelligence
the conceptual humanist-is exemplified in Rollo be assessed as if one exists without the other? Is this
May who wrote The Courage to Create (1975). any more possible than the culture-free assessment
Intelligence, or adaptation to the perceived de- of intelligence? Indeed, have the psychometric in-
mands of one's environment, as well as preference struments, so carefully designed to measure the de-
for specific models of functioning may, in this way, gree to which individuals possess creativity or intel-
elicit creative responses. ligence, been confounded by the presence of the
Although Jackson and Messick, as early as 1965 other? This dilemma seems similar to that of verify-
(Kagan, 1967), stated that efforts to distinguish em- ing a wholly unique role for either heredity or the
pirically between the two phenomena lack signifi- environment in human behavior. A resolution of
cant informative power, contemporary efforts in this latter' 'problem" was proposed 30 years ago by
this direction have continued. These efforts have Anastasi (1958), when she suggested that the intri-
focused on demonstrating, for example, that tests cate interweaving of the two components of heredi-
designed to elicit production of unusual responses ty and environment had rendered the problem point-
bring into play distinctly different mental processes less, that the more vital consideration had become
than tests designed to assess intelligence. In fact, the determination of how each contributes to the
the results of such efforts are somewhat arbitrary, final product, and what contexts would be most
since performance on the tests is, at best, a limited suitable for promoting maximal expression of each.
indicator of either phenomenon in practical situa- Similarly, we might more profitably view mental
tions (Sternberg & Wagner, 1986). functioning as a complex interweaving of intel-
ligent and creative types of response, according to
specific task demands and parameters, and depen-
A Focus on Integration
dent upon individual predilection or capacity for a
Little attention has been given to the idea that particular response type.
creativity and intelligence may be viewed more in- Further, any consideration of mental functioning
formatively as two different facets of a singular must include awareness of the effect of context.
function originating from mental ability. In such a This environmental component, upon which genet-
view, each facet has a different type of endpoint and ically controlled ability may develop and act, not
each manifests a different way of attending to infor- only permits the expression of ability, shapes and
mation, retrieving and reorganizing it, and applying defines it, but also can limit it (Haensly, Reynolds,
the resulting reorganization. Yet, together, both & Nash, 1986). Yet, as canalization theory (Fish-
contribute to an integrated, fully adaptive response bein, 1976) suggests, when adverse environments
to a specific task or immediate environment. AI- occur, the human's genetic plan tries to restore po-
tential maximization. So too might mental ability, mental processing, according to the demands of the
modified by context, attempt to manifest itself ac- task addressed.
cording to the individual's potential for different This chapter focuses on how and why creativity
types of expression. The type and degree of re- and intelligence should be viewed as a unitary phe-
sponse also may be shaped by the individual's nomenon in which optimal mental functioning oc-
understanding of task demands and his or her men- curs in an integrated, complementary fashion. The
tal efficiency, which is defined by Sternberg (1984) remainder of the chapter will be devoted to (a) a
as the internal world of the individual. Additional brief examination of several early approaches to the
response shaping occurs through experiences that assessment of creativity in order to describe the
have developed or trained readiness for the task (the theoretical and empirical inclusion of intelligence
external world of the individual), and the recep- in each, and (b) an examination of several ap-
tiveness of the available audience (context), as de- proaches to assessment of intelligence in order to
scribed in Sternberg's componential theory of intel- describe the theoretical and empirical inclusion of
ligence (Sternberg, 1984). creativity in each. The intent in both of these exam-
Attempting to isolate creativity from intelligence inations is to introduce and elaborate on the comple-
seems no more informative from an information- mentary nature of creativity and intelligence, rather
processing perspective of intelligence than from a than to review comprehensively all available theo-
developmental perspective, nor, for that matter, ries. We then address a particular problem regard-
through analyzing the differential contributions of ing the perceived relationship between creativity
brain hemispheres and cortical specializations. and intelligence that stems from the restricted range
When the individual confronts new information, is encountered in the assessment data of many studies.
it perceived and encoded intelligently or creatively? This is followed by a discussion of the synergistic
Are the neurological patterns that arise intelligently effect between creativity and intelligence. The
planned and stored, or do they associate flexibly? chapter concludes with a brief presentation of a
And, depending upon the prior storage of patterns, perspective of mental processing that views intel-
do they organize predictably, or do they become ligence as an intrinsic component of creative behav-
established in novel arrangements? When retrieved ior in which creative expression is the ultimate ex-
from short- or long-term memory, what determines tension of intelligence. It also notes some implica-
whether resulting information qualifies as an intel- tions of the integrated viewpoint.
ligent response or a creative response? Does this
change when one considers the particular indi-
vidual, or does it depend upon the aggregate of Intelligence as a Component in Theories
individuals who are privy to the response? Or, must
it await the criterion of acceptance over time? How
and Assessment of Creativity
long must the novel or highly original response be
The Association Hierarchy
available before others view it as an intelligent re-
sponse to the task (Haensly & Roberts, 1983; Among studies of creativity initiated with a focus
Haensly et ai., 1986)? on the creative process itself is that of Mednick
Critical examinations of traditional views of in- (1962). Mednick defined creativity as "the forming
telligence and reconceptualizations of the processes of associative elements into new combinations
involved (e.g., Case, 1978; Gardner, 1983; Pas- which either meet specified requirements or are in
cuale-Leone, 1970; Piaget, 1972; Sternberg, 1985; some way useful" (p. 220). Based on this defini-
Sternberg & Wagner, 1986) demand a reconsideration, Mednick established two immediate criteria
tion of the phenomenon of creativity as well. Unfor- for assessing creativeness: the degree of mutual re-
tunately, there is no eluding the problem of termi- moteness of the elements of a new combination,
nology. To attempt to replace terms in current use and the combination's usefulness. For example, a
will only decrease our ability to talk about mental propulsion device, chair, and parachute were once
functioning. We can, however, develop paradigms quite remotely associated elements, but their com-
for description, explanation, and prediction in- bination into an ejection seat has permitted indi-
clusive of both creativity and intelligence. Such viduals to safely depart precrash, in-flight vehicles.
paradigms would demonstrate that optimal mental Mednick also postulated that the probability and
functioning occurs when there is appropriate align- speed of creative solutions would be affected by the
ment and balance between the various modes of condition of the organism. One of the most salient
of those conditions is the repertoire of elements necessarily mean results of a creative nature will be
available to an individual. In other words, indi- produced; the latter would depend upon the pres-
viduals who have larger repertoires resulting from ence of specific motivational and temperamental
quality of preparation and experience in a content traits. Thus, Guilford addressed the genetic aspect
area, experience in related content areas, or even in in intellect (ability) and personality (temperament)
parallel systems will have a greater availability of yet did not exclude the contribution of environment
elements. This state of the repertoire of available (motivational factors) to the development of such
elements might well be equated with an individual's traits.
level of intelligence, if quantity of information and Early in his studies of creative process, Guilford
diversity of experiences with that information can (1963) attempted to demonstrate how factors of in-
be said to reflect intelligence. If this holds, then the telligence fit into the operations used in creative
more intelligent the individual, the higher the prob- thinking. He maintained that typical problem-solv-
ability that a creative and useful association will ing models (e.g., Dewey, 1933) essentially de-
arise quickly. scribe the same phenomenon as typical creative
Other conditions that may affect the association production paradigms do (e.g., Wallas, 1926,
of elements into new combinations include cog- 1945), with comparable, parallel steps or stages.
nitive style, the way in which an individual origi- Dewey's first stage of recognition of the problem
nally organized the associative elements, and pre- was described by Guilford as the unique intellectual
vious or current disposition to the problem at hand ability of sensitivity to problems. Further, he point-
or to methods of approaching problems. Organiza- ed out that Dewey's stage of solution-suggesting is
tion of information refers to encoding for efficient comparable to Wallas's illumination or generation
retrieval, which may include both visual and verbal of ideas; the difference is that, in the latter case, the
representations. Again, the more intelligent an indi- step occurs in a highly compressed timeframe, per-
vidual is, the more likely that the encoding will haps with unusual efficiency in mental processing.
make available to the individual a wider and more Guilford described the generation of ideas, whether
complex hierarchy of associations in a variety of in Wallas's illumination stage or in Dewey's solu-
representations from which a creative solution tion-suggesting stage as retrieving information to
could be selected. be used in new forms or with new connections. This
Mednick's theory led to a simple assessment de- process invokes the phenomenon of transfer, which
vice, the Remote Associates Test (RAT), which goes beyond just recalling associations to selec-
was designed to elicit creative combinations and tively reassociating pieces of information within the
thus measure an individual's creative potential. memory storage and with newly encountered
Mednick focused on the semantic component of problems.
associations, and the RAT depends on a semantic In Guilford's description of intellectual opera-
mode of response. The contribution of intelligence tions, the degree of continuity in producing associa-
to the establishment of a rich repertoire from which tions (in which speed might be an assumed charac-
creative combinations can be generated can be seen teristic) is termed fluency (see Guilford, 1968, p.
clearly. 125). Fluency is applied to multiple types of con-
tent-figural, semantic, symbolic-and to the
complexity of associations-units, relations, sys-
Guilford's Divergent Production Model
tems. Within Guilford's divergent production oper-
Guilford, too, approached the phenomenon of ation, generation of ideas may exhibit greater or
creativity from a process orientation, using the op- lesser degrees of diversity in their basic categoriza-
erations component of his structure-of-the-intellect tion, which Guilford termed flexibility. Flexibility
(SOl) model to examine creativity (Guilford, in idea production describes the ability to redefine
1968). The relevant operation from within his SOl problem parameters, a condition necessary for re-
model that is known as "divergent production" or ducing functional fixedness, and, in so doing,
"divergent thinking" (Guilford, 1959) frequently providing a richer array of possibilities for problem
has been thought of as analogous to creativity. Yet solution. Thus, fluency and flexibility represent
Guilford asserted that "creativity and creative pro- measures of the quantity and quality of divergent
ductivity extend well beyond the domain of intel- production taking place. It can be assumed that flu-
ligence" (Guilford, 1950, p. 445). He further sug- ency depends, at least in part, on the quantity of
gested that having the requisite abilities does not ideas available in memory, and flexibility, on the
variety of ideas. Both fluency and flexibility also repertoire of ideas from which to retrieve, compare,
depend not only upon the efficiency of the mental combine, and transfer. In either case, in assess-
search but also on the manner of encoding-the ment, the more specific the divergent production
variety of stimulus types (verbal, visual, tactile) test (i.e., semantic, figural, and symbolic content,
and the organizational efficiency of the encoding and the units, relations, and systems products), the
(the categorization system). less likely will be a tight relationship with a broader
Through incorporating the divergent production based, general intelligence test.
process as only one of the five basic operations in It may be that theories generated to explain the
intellectual activity, Guilford's theory inextricably creative and intelligence processes have not been
binds intelligence and creativity. Despite this the- matched with an appropriate system of verification ,
oretical amalgam, Guilford's search for correla- considering that reliance for verification has been
tional data to demonstrate a relationship between on the interpretation of analyses from statistical
creativity and intelligence using traditional IQ tests procedures that confound rather than unravel the
resulted in data that do not necessarily support a theories. Nevertheless, in Guilford's theory, intel-
direct relationship (Guilford, 1968). Using a bat- ligence appears not only to contribute to creative
tery of divergent production tests and a group intel- ability, but creativity is actually one form of the
ligence test, the California Test of Mental Maturity process of intelligence.
(CTMM), Guilford and Hoepfner found that figural
divergent production factors correlated lowest with
CTMM IQ and with verbal comprehension. Sec-
Torrance's Scientific View of Creative Process
ond, they found that symbolic and semantic diver-
gent production factors correlated higher with both "I defined creativity as the process of becoming
of these latter criteria. For example, the mean of sensitive to problems, deficiencies, gaps in knowl-
correlations between the figural divergent produc- edge, missing elements, disharmonies, and so on;
tion tests and the verbal comprehension test was identifying the difficult; searching for solutions,
.10, and with the CTMM IQ, .22; the mean of making guesses, or formulating hypotheses and
correlations between the symbolic and semantic possibly modifying them and retesting them; and
tests and verbal comprehension was .26 and .29, finally communicating the results" (Torrance,
respectively, and with CTMM IQ, .40 and .37, 1967, pp. 73-74). Thus, Guilford's (1963) elab-
respectively (Guilford, 1968, pp. 131-133). As oration of the unique intellectual ability of sen-
Guilford pointed out, because the intelligence test sitivity to problems as Dewey's first stage of prob-
reflects great complexity and thus has more sources lem recognition has also become the cornerstone of
of variance, using a limited type of divergent pro- Torrance's definition. In generating this definition
duction test can greatly affect the resulting correla- for creativity, Torrance set the stage for an objec-
tion. Further, Guilford stated that scatterplots de- tive observation of an elusive process. It also pro-
veloped from their data (see Guilford, 1968, pp. vided a foundation from which to study the personal
132-134) vividly demonstrate that low-IQ indi- and environmental factors that facilitate or inhibit
viduals do not show high divergent production creative process. In reflecting on his definition,
scores, whereas high-IQ individuals exhibit a wide Torrance commented that some find it too loosely
range of divergent production scores. (This issue constrained, while others find it too narrowly ori-
will be discussed further in the section "Problems ented to scientific process to the exclusion of artistic
of Restricted Range in Studies of Creativity and endeavors. However, the definition was an attempt
Intelligence.") From the theoretical foundations of to find an area of focus in creativity that would be
Guilford's SOl model, it could be expected that the productive in helping scientists and lay people alike
low-IQ individual (assessed on whatever intel- better understand the phenomenon.
ligence measure) would lack the available quantity Toward that end, Torrance's definition suggests,
and quality of stored ideas to be either fluent or though does not specify, a relationship between
flexible. If efficiency of mental processing reflects creativity and intelligence. It fits closely Spear-
intelligence, as previously assumed, retrieval of man's ideas (1930) about mental creativity as the
ideas and complex transference also would be lack- power to transfer relations and generate new corre-
ing in low-IQ individuals. Likewise, it could be lates and Newell, Shaw, and Simon's(l962) defini-
expected that, with increased intelligence, diver- tion of creative problem solving. In each case, the
gent production would access an increasingly wider focus is on process, that is, the use of particular
mental abilities applied to some specific content in "manifestation of talent in a field of endeavor" to
order to accomplish an end not previously in place. describe the process applied as well as the resulting
Torrance's methods of assessment of creative po- product; "constellation of supportive intellectual
tential, especially the figural and verbal forms of abilities" describes the individual's internal "en-
Torrance's Tests of Creative Thinking (1966), em- vironment" from which the process emanated. In
phasized (as in Guilford's theory described earlier) the former, we infer creativity based upon judgment
the ability to generate many new ideas (fluency) of observed products (i.e., manifestation of talent).
that are unusual (originality) and represent a variety In the latter, supposedly, we must infer creativity
of categories (flexibility), as well as the ability to based on observation of multiple behaviors (i.e.,
embellish the ideas (elaboration). This approach abilities, personality variables, problem-solving
would seem to reflect Mednick's association hier- traits), across multiple situations. Thus, when Aus-
archy, and Guilford's divergent production strand ubel describes creativity as a "particularized sub-
of intelligence, more than a "sensitivity to prob- stantive capacity" versus creative abilities that are
lems and gaps in knowledge." However, Tor- "supportive intellectual-personality functions,"
rance's philosophical perspective suggests that we are faced with a recurrent dilemma of obser-
ability represented in the listed behaviors will be vability and measurement. How can capacity be
applied with increasing frequency or transferred to separated from function, the capability to act from
any of the phases of problem solving. Torrance's the act by which we may infer the capability? How-
approach to creativity and its implications for the ever, Ausubel suggested that a truly creative prod-
transference of behaviors to multiple situations thus uct only arises when the general constellation of
reflects an intrinsic inclusion of intelligence in supportive abilities is of such breadth and depth as
creativity. to be manifested in a particularly rare and outstand-
ing way. Is creativity's manifestation a question
then of degree of breadth and depth of abilities to
Creative Capacity, Constellation of Traits, and
exhibit function (i.e., process)? Ausubel provided
Their Application
an answer to this (one already attributed to Anastasi
The crux of the problem of defining creativity & Schaefer, 1971, and to Eisner, 1965), when he
may lie in our seeming inability to separate the said" Assessments of creative potentiality can only
individual from the process, and the process from be based on expert judgments of actual work prod-
its application; that is, do we define creativity as the ucts, suitably tempered by considerations of age
"things" one does, or as the ability to do them and experience" [italics added] (Ausubel, 1978, p.
given the right conditions and circumstances, an 181). We have thus returned to a focus on
ability that cannot be determined until the behavior application of the process by which a judgment of
is exhibited? Ausubel (1978), in fact, attributed the ability or capacity of the individual to deliver the
confusion in our conceptualization of creativity to process can be made. Apparently, not only must we
the failure to separate creativity as a trait (or cluster return to process, but we must also consider the
of traits) from the individual who possesses these application of the process relative to what it might
traits to a singularly high degree. This may be com- be, "tempered by ... age and experience."
pared to the difficulty in separating intelligence as a Ausubel (1978) saw confounding in yet another
broad set of characteristic behaviors from the per- situation, resulting from discrepancies between
son who possesses these behaviors. But let us delay cognitive trait data and assessment of ability data.
a discussion of this difficulty to address more im- He pointed out that cognitive traits associated with
mediately the process/ application entanglement. creativity (i.e., supportive of creativity) have been
Ausubel (1978) maintained that semantic, if not shown to correlate just as highly with intelligence as
theoretical, confusion results from a failure to dis- they do among each other, citing studies such as
tinguish "between creativity as a highly particu- Anastasi and Schaefer (1971), Crockenberg
larized and substantive capacity . . . a manifesta- (1972), and Getzels and Jackson (1962). Yet diver-
tion of talent in a particular field of endeavor [and] gent thinking scores tend to correlate only moder-
as a general constellation of supportive intellectual ately with measures of intelligence. The identified
abilities, personality variables, and problem-solv- cognitive traits might then contribute but not be the
ing traits" (p. 180). Ausubel's argument ex- sole support for creative expression, and/or they
emplifies the lack of separation of individual from might contribute but not be sole support for intel-
process and process from application. He used ligent behavior. Although Ausubel (1978) postu-
lated that a minimal degree of intelligence is neces- and knowledge-acquisition processes. "Meta-
sary for individual creative actualization, he balked processes are used to plan, monitor, and evaluate
at the idea that a relationship with creativity exists one's problem solving ... knowledge-acquisition
throughout the continuum of intelligence. processes are used to figure out how to solve the
Using traits and personality variables to define problems in the first place" (Sternberg, 1986, p.
and examine creativity has been a part of the ap- 145). Thus, if creativity is equated with qualitative
proach of others as well. In their 1965 study, Wal- problem solving, then both the metaprocesses and
lach and Kogan identified a number of personality knowledge-acquisition processes represent creative
variables that appeared to characterize the indi- behavior at some level of response. Sternberg sug-
viduals in their study (151 middle-class fifth gra- gested that conventional intelligence tests measure
ders) when they were grouped according to their the metaprocesses directly (planning, monitoring,
intelligence and creativity. The continuum of intel- and evaluating one's strategies in solving posed
ligence in this sample represented a limited range problems) and knowledge-acquisition processes in-
and especially did not represent the lower end of the directly through measuring past effectiveness of ac-
continuum. The type of characteristics derived in- quisition of knowledge. It should be noted that there
cluded such things as high-creative-high-intel- are inadequacies in the tests (the problems may not
ligent individuals exhibiting a minimal anxiety be as complex nor as practical as those faced in real
level, "as if they were bursting through the typical life). Further, the tests cannot measure the effec-
behavioral molds that the society has constructed" tiveness of the knowledge-acquisition processes of
(Wallach & Kogan, 1972, p. 213). High-intel- a child who has not had the opportunity to apply
ligent-Iow-creative individuals exhibited un- them in his or her environment on the type of mate-
willingness to take chances, fear of errors, and re- rial typical of intelligence tests. Synthetic abilities
liance on conventional responses and were not (as defined by Sternberg, 1986), then, are not as-
disruptive but were aloof from peers. Both of these sessed in conventional tests, and available
groups shared strong capacities for concentration creativity tests do not measure creativity at the level
on academic work and long attention span. Low- needed for significant intellectual contributions.
intelligent-high-creative individuals were least However, Sternberg suggested that they probably
able to concentrate and maintain attention, were do as good a job of predicting creativity in the ar-
socially isolated, behaved disruptively, and tificial setting as intelligence tests do in the artificial
possessed the lowest self-esteem. And finally, low- setting of academia. We must therefore look for
intelligent-low-creative individuals, who func- prediction of future creative behaviors to past man-
tioned better in the classroom than their highly ifestations of such behavior, that is, to products or
creative counterparts in intelligence, showed the observable accomplishments. With this conclu-
weakest aesthetic sensitivity. The personality traits sion, are we really making a leap in understanding
associated with creativity thus appear to reflect a or only digging a deeper hole?
significant interaction with intelligence level, or at
least with the way in which individuals of different
A Synthesis: Intelligence as an Element of
intelligence levels are able to function.
Creative Process
More recently, Sternberg (1986) also addressed
interactive effects on performance between the Varied views on the nature of creativity have
traits of creativity and those of intelligence. In his been presented and the contribution to each of intel-
frequently described student typology, he specified ligence as a differentJacet of mental functioning,
that Student Alice does not have synthetic abilities rather than a different phenomenon, has been dis-
and described Student Barbara as having been rec- cussed. Although a relationship between creativity
ommended as a tremendously creative woman with and intelligence is evident, and commonalities ap-
the ability to generate and follow through on cre- parent, differences in perspective hinder agreement
ati ve ideas producing important research. Thus, we on the relationship as one of "universal mutual
have a temporary definition for creativity, that is, origin," or even mutual purpose. For example, in
synthetic ability, and a measure of it, that is, com- Mednick's view, assessment involves creative re-
pleted work based on new ideas or approaches. sponses that are essentially semantic in nature, an
Elaborating his triarchic componential theory , almost sure link with intelligence as we have known
Sternberg described the three basic processes of in- it. However, through focusing on the semantic
telligent behavior-metaprocesses, performance, component of associations and depending on the
semantic mode of response, other forms of stored creativity tests with academic achievement, com-
information that might be more typically associated parable to those found with intelligence test scores.
with creative endeavor, such as the figural, images, However, Ausubel (1978, p. 184) asserted that it is
or even kinesthetic patterns, may be ignored. From "quite unlikely that creativity should be related to
the standpoint of creative talent and development, academic achievement, inasmuch as mastery of a
another problem surfaces. The test ignores the indi- given subject-matter discipline does not in any way
vidual's disposition to the specific problems, that presuppose conspicuous capacity for making origi-
is, his or her particular experiences with and interest nal or creative contributions to that discipline."
in the problems presented for response. Individual Additional confounding occurs when the ex-
differences in prior conditions available to the prob- pression of a specific process appears to be product-
lem solver and familiarity or experience with the oriented yet lacks the essential integrity of a prod-
particular domain of terms included in the test also uct; for example, assessment of creative potential
are not taken into account. Assessment depends on based on the total number (fluency of responses)
the individual's ability to generate uncommon asso- and uniqueness (originality of responses) of rele-
ciations based on a current store of knowledge and vant associations (simulated product) in an
experience, even though this measure of fluency in artificially defined situation (Wallach & Kogan,
divergent production may not be generalizable to 1965).
situations in the individual's occupation or practical Contemporary theories of intelligence suggest
living conditions. similar problems of definition and measurement, as
A statistical relationship, or lack of it, between well as difficulties in extracting the process from its
measures of creativity and measures of intel- application or endproduct and the individual from
ligence, or between these measures and other char- the process. An even more informative parallel with
acteristics of the individuals studied, frequently has these contemporary theories (already alluded to in
been viewed as definitive evidence. Wallach and the discussion of Mednick's association hierarchy)
Kogan posed the thesis that such proof shows dif- is the idea that context, age, and experience always
ferences in cognitive capacity between creative in- must be taken into account when attempting to as-
dividuals and intelligent individuals. The evidence sess the presence of intelligence through applica-
takes a strange tum when the results focus on social tion of the process; or when predicting the capacity
and motivational differences (as discussed in the of the individual to function in an intelligent way
previous section) rather than on the actual cognitive based on fragmented expressions of intelligence.
processes, the social and motivational differences In fact, according to Sternberg and Davidson's
that have long been associated with failure to suc- theory of insight and problem solving (Sternberg &
ceed in relation to peers. Capacity to perform might Davidson, 1983), or to deBono's theory of lateral
be inferred but the ability of these findings to dis- thinking (deBono, 1971), creative problem solution
tinguish a difference in capacity for creativity ver- entails selective combination of seemingly isolated
sus capacity for intelligence does not seem logically pieces of information for creative associations, or
sound. seemingly illogical combinations, associations
Comparisons also have been based on (a) the that, once examined, create breakthroughs in think-
presence of specific, intact characteristics or clus- ing about a problem. Further, these theories suggest
ters of traits and personality variables, used to infer that selective comparison of new problem situations
capacity in individuals, (b) the expression of both to old bits of information or associations can be-
creative and intelligent process by individuals, and come the extraordinary insights of creative problem
(c) products, as in academic achievement, resulting solving. Whether Sternberg and Davidson consid-
from the application of the processes. Unfortunate- ered these processes characteristic of creative be-
ly, comparisons of intelligence with creativity often havior or of intelligent behavior is not clear, but, in
have been made that confound person, process, and either case, their studies indicated that a rela-
product. For example, a process variable for tionship between intelligence (assessed through
creativity, such as divergent thinking, may be cor- group intelligence tests) and the various insight
related with a product variable for intelligence, abilities is a positive one. Thus, the greater the
such as academic achievement. Several investiga- intelligence, the greater the likelihood that indi-
tors (Getzels & Jackson, 1959, 1962; Torrance, viduals will employ spontaneously the selective
1960; Yamamoto, 1964a,b,c) have found correla- comparison and combination (insights) necessary
tions between Guilford-type and Torrance-type to solve insight problems.
There seems no doubt that intelligence is an ele- Learning Theory

ment of the creative process. But is creativity as
Classical learning theory focuses exclusively on
important an element of intelligence as it has been
process, and intelligence becomes "learning pr~
defined by various theoreticians? We continue with
cess functioning at its optimal level under appropn-
our examination of these phenomena, or the inte-
ate situational conditions and satisfying some exter-
grated phenomenon, as the case may be.
nal value-type criteria not inherent in the behavior
itself" (Maloney & Ward, 1976, p. 176). Learning
theory claims to have the least use for i~div~d~al
characteristics, ignoring hereditary capaCIty m Its
Creativity as a Component in Theories
consideration of what behaviors the individual
and Assessment of Intelligence emits and the circumstances under which those be-
haviors occur.
Intelligence as a hypothetical construct has been From this perspective, Thorndike (1926) focused
aptly described by Conger (1957) as a "scientific on associations or connections as the "accumula-
fiction" invented to help explain and predict behav- tion" of intelligence, in which greater intelligence
ior. Psychologists from Cattell to Wechsler to the occurs when a larger number of connections have
present day have focused in various ways on three been made. If creative production may be consid-
themes or components of this scientific fiction: ca- ered the result of accessing the more remote asso-
pacity to act purposefully, that is, to learn or change ciations in one's mental storage (Mednick, 1962),
one's behavior based on experiences; ability to then we can begin to see a tenuous notion of
think rationally and abstractly; and the ability to creativity and intelligence as having a common ori-
respond effectively to new situations in one's en- gin. One can further see the role a wide variety of
vironment. Much earlier, however, Thorndike prior learning experiences might play in establi~h
(1921) viewed this hypothetical construct in a more ing the vast and diverse hierarchy from WhICh
concrete way as facility in manipulating objects, in unique associations could be generated.
the use of symbols, and in interacting with other The influence of classical learning theory in the
human beings. Remnants of these three themes may consideration of creativity may also be seen in the
be found in each of the current views of intel- role of behavior modification in expanding origi-
ligence. Through these themes the groundwork was nality in responses and creative production in prob-
laid for a translation of theoretical construct into lem solving. Studies by Brigham, Graubard, and
observable and measurable behaviors. Notwith- Stans (1972), Goetz and Baer (1973), Glover and
standing this simplified thematic context and sim- Gary (1976), Glover and Sautter (1977), Haensly
plicity of measurement possibilities, four distin~tly (1979), Locurto (1974), Maltzman (1960), Nash
different theoretical approaches to understandmg (1975), Wallach and Kogan (1965), and othe~s
and measurement of intelligence have arisen (Mal- have focused on various ways to successfully modI-
oney & Ward, 1976). Each places differential em- fy stimuli for eliciting creative respon~es an? in-
phasis on the three possible themes and each has creasing frequency of responses and theIr quahty or
generated different methods for measurement. The originality.
four general approaches can be classified as learn-
ing, neurological-biological, psychometric, and
Neurological-Biological Theory
developmental. We briefly describe, in tum, each
general approach to intelligence, focusing on how The neurological-biological approach to intel-
the creative process is addressed in each. The pur- ligence seeks to understand behavior and the pa-
pose of this examination of a sample of available rameters that influence or determine behavior
theories is to demonstrate that perspectives on intel- through looking for relationships within the under-
ligence typically have included creative process, lying neurological system, its anatomy and phys-
whether or not it was labelled as such. (Some of iology (Maloney & Ward, 1976). Attempts to ver-
these approaches, such as the developmental and ify earlier versions of this theory focused on trying
psychometric, are addressed at length in other chap- to locate specific areas of the brain devoted to mem-
ters of this handbook. They are included here in ory, reasoning, and judgment. In this approach,
order to develop a point regarding the relationship behavioral correlates of neuroanatomical and neu-
of creativity with intelligence.) rophysiological functioning are assessed.
Perhaps the most successful of these efforts was the pyramid and specific factors breaking out from
that of Broca who established the presence of a the general are models that depict a common origin
speech center in the posterior portion of the left for general mental functioning. These theories were
frontal lobe of the brain. However, principles of devised to explain observed patterns of differences
mass action and equipotentiality formulated by among individuals, particularly groups of indi-
Lashley (Maloney & Ward, 1976, p. 180), contra- viduals. They have been the product of two specific
dicted the efforts of those trying to localize specific mathematical and statistical organizations of data,
functions in specific areas ofthe brain. This contro- correlations, and their derivative procedure, factor
versy continues to be addressed in research on analysis. Factor analysis by itself cannot provide
hemispheric lateralization of the brain for abilities answers about theoretical conceptions of intel-
ranging from spatial to language expression and ligence. Interpretations must be developed for the
reading comprehension and from musical reception observed relationships among specific abilities de-
to mathematical abstraction (Gardner, 1983). This termined through correlations of performance on
category of theory has been distinguished by different types of tasks. These interpretations be-
Hebb's (1972) ideas on intelligence A and B and by come quite critical as the statistical procedure be-
Cattell's (1963b) theory of fluid and crystallized comes more detailed and complex in factor analy-
intelligence, referring, respectively, to innate bio- sis. Frequently, several interpretations appear
logical capacity and to experience with the environ- equally viable in explaining the data. Further, the
ment. The biological contribution in these latter origin of the data (i.e., the subject population from
theories refers to the intactness and efficiency of the which the data are gathered) and its homogeneity or
nervous system. heterogeneity (sample selectivity) may greatly alter
Recent theories in this category have focused on the results, as does the choice of variables for inclu-
the integrative functions within the brain and of the sion in the analysis.
brain and nervous system, on the biochemical as- Although two of the most frequently applied psy-
pects within the brain and at the nerve endings, as chometric assessments of intelligence-the Binet
well as on the measurable excitations of specific and Wechsler scales-were constructed to sample a
areas of the brain in response to specific types of variety of intellectual abilities and are thus some-
tasks. Though perhaps difficult, it is not impossible what conceptually related to the multiple factor hy-
to conceive of creativity and intelligence as differ- pothesis (Maloney & Ward, 1976), their reliance on
entiated either in chemical constituency or in area of a single index would imply an assessment of the
cerebral localization. "g" or general factor. Nevertheless, the single
A nebulous reference to creative production quantitative index does not fit well with the factor
might be suggested for Hebb's intelligence A, with approach, and the tests are a theoretically complex
the assumption that biological capacity is related to mix of such factors. Psychometric theory, which is
problem-solving abilities, and for Cattell's fluid in- basically qualitative in explanation (based on ob-
telligence, with the assumption that untutored abil- served patterns of performance), does not seem to
ity would represent the raw material of creative as be represented well by these basically quantitative
well as intelligent responses. In this latter case, tests that result in an intelligence quotient. Current
experiences with the environment could enhance psychometric conceptualization continues to lean
creative responses only if the experiences did not toward multiple factors, the number lying some-
overwhelmingly reinforce mundane responses. where between the unacceptable extremes of too
few (Spearman) and too cumbersome (Guilford).
Interestingly enough, Gardner (1983), who pro-
Psychometric Theories
fesses extreme opposition to the psychometric
Psychometric theories view intelligence as a pat- "IQ," has promoted a neurobiological theory that
tern of particularly adaptive abilities. Individuals distinctly focuses on factors similar to those sug-
differ in the specificity and extent of their abilities. gested by Thurstone.
With few exceptions these theories have arisen Now, given these hedges, does creativity have a
from data gathered on individual differences in abil- place in psychometric theories of intelligence?
ities. Spearman's two-factor theory (1927), Thur- Since creativity and creative production are not ex-
stone's multiple-factor theory (1931), and hier- clusive to any content or domain and conceivably
archical models with a general factor at the apex of can encompass all manner of performance, it would
seem inappropriate to assign it one specific ability Developmental Theory

slot in one of the factor theories. Vernon's model
(1950) postulated two major group factors (verbal- Intelligence in the developmental perspective is
numerical-educational and practical-mechanical- conceptualized as action-dynamic, epigenetic,
spatial-physical), one of which may be more and ontogenetic-represented by mental organiza-
function-oriented and the other content-oriented; tional structures that change with maturation and
creative production might find its niche in such a experience. The third of the three all-purpose
function-oriented factor. Within IQ tests them- themes for the intelligence construct, the ability to
selves, little if any allowance has been made for respond effectively to new situations in one's en-
divergent types of response; in fact, correctness in vironment, seems to dominate this perspective. Ad-
response is the mode. Yet the abilities assessed aptation of the individual to the environment is the
form the foundation from which creative endeavor goal of intellectual activity, as it is the goal of all
must arise. biological organisms.
Although Guilford's theory falls into the psycho- Epistemological orientation, or emphasis on the
metric category because it purported to measure genesis and growth of knowledge in the individual,
individual differences in ability through perfor- is reflected in, for example, Piaget's concept of
mance, it is a reversal of typical approaches in that a schema and his principles of assimilation, accom-
theory of intelligence preceded the gathering of data modation, and equilibration. Piaget emphasized
for its verification. Most relevant to this discussion, that these internal organizational processes are
it assigned creative function as one of the five types available to all, and the focus is on universality
of mental operations, specifically, divergent think- rather than individuality. Intelligence is not only
ing or divergent production (as described in a pre- quantitatively different, at different ages, but also
vious section of this chapter). This operation can qualitatively different, and the age-related stages
take place with any of the four types of content, at are sequential and hierarchical; that is, organiza-
any of the six increasingly complex levels of prod- tional processes or mental structures available to the
ucts. However, a problem arises in an exclusive individual are different at infancy than at adoles-
assignment of creativity to the divergent production cence. They are different in type of stimulus that
operation. Unless cognition is present, no content is activates response and different in type or mode of
available for memory. Unless memory is available, response generated, with each new stage interfac-
divergent production would be limited to responses ing with the structures or knowledge of the previous
in cases immediately at hand, and, without evalua- stage and, in this way, adding to the complexity of
tion, divergent production cannot be usefully di- understanding.
rected to form a creative contribution. Thus, we see Maloney and Ward (1976) suggested that these
that an intricate integration of operations must take psychological structures have a neurological ana-
place for divergent production to be realized as logue. This possibility lends credence to Piaget's
creativity. Such integration seems to reflect position that stage development is limited by neu-
Sternberg's "synthetic ability" (Sternberg, 1986) rological development; thus, maturation would ex-
-application of metaprocesses to knowledge-ac- ert an effect on capacity and performance. Accord-
quisition and performance processes. As in other ing to Piagetian perspective, individuals may differ
assessments stemming from Guilford's theory, we in the rate at which they accomplish the concep-
must conclude that it is the pattern of integration tualization characteristic of a particular stage and,
that provides diversity among responses and among likewise, not all individuals will necessarily attain
individuals. Even though various levels of the dif- the highest stages. A critical element in the rate
ferent types of operations might contribute to effec- difference is the environment, with the individual
tive functioning, some minimal levels of each dependent upon the appropriateness and richness of
would be imperative. Creative genius would occur environmental experiences for optimization of
as each of the operations found its optimal balance knowledge growth and stage attainment. However,
for a particular content, operating on one of the six even with the richest environmental experiences,
possible levels of products, in response to a particu- individuals would be limited in stage attainment by
lar environment with its specific demands, and cri- age-related neurological development. In practice,
teria or qualifications. a few individuals with extraordinary mental capaci-
ty transcend these ordinary maturational limita- imum advantage of the individual's readiness and
tions. Neo-Piagetian perspective (e.g., Case, 1978; inclination to produce the creative contribution.
Pasquale-Leone, 1970) proposed that some of such Thus, outstanding creative contributions, generated
limitations have to do with the mental capacity to resolve problems, will occur when specific abili-
available for processing increasing quantities of in- ties have coalesced in a relevant context, and are
formation. This perspective opens up the possibility emitted with sufficient quality, intensity, and dura-
that strategy awareness, training, or task simplification (Haensly et at., 1986).
tion might permit responses not thought possible
under strict Piagetian perspective.
A Synthesis: Integration of Creative Process
From a developmental perspective, then, multi-
with Intelligence
ple creative possibilities exist in the schemata each
individual constructs from his or her particularly As stated earlier, the purpose of an examination
unique interaction with the environment. Carrying of basic approaches to intelligence was to demon-
this further, the concept of generative processing strate that perspectives of intelligence do include or
(Wittrock, 1974) proposes that, at any point in assume creative process as integral to intelligence,
one's epistemological development, construction even when creative process is not so labeled.
of knowledge is based on quantity and quality of In summary, classical learning theory appears
one's prior, specifically individual experience. not to distinguish between an intelligent and a cre-
Therefore, no two individuals can have exactly the ative response, viewing responses in general as ob-
same construction and/or understanding of any servable and measurable behaviors, elicited and re-
concept, nor are they likely to have exactly the same inforced by attendant stimuli. On the other hand,
view of events, persons, places, and things. neurological-biological theory may be seen as the
Uniqueness becomes available to everyone. The most elusive of the four described approaches, yet
quality of original ideas will greatly depend upon the most promising for future investigation.
the appropriateness or "fit" with the understanding Through microanalyses of physiological and ana-
of others in the population. In turn, this may depend tomical correlates of behavior, the possibilities of
upon the extent to which an individual has devel- differentiating the biochemical state in creative pro-
oped related schemata, based on reliable experi- cess from that in intelligent process, especially with
ences. The application of unique perspectives will relationship to antecedent conditions for the re-
depend upon the individual's problem-solving sponses, is well within the realm of consideration.
strategies (which will also evolve via the process However, as of now, localization of response to
described above) and on the opportunities for ap- specific content or domains of knowledge appears
plying them-being in the opportune place at the to have been the most productive application of this
opportune time. approach.
The possibility of a contrib,ution of creativity to Despite bias regarding intelligence testing and its
intelligence from the developmental perspective assumed exclusion of creative potential, psycho-
becomes much more evident upon examination of metric theory provides an avenue for explaining the
Csikszentmihalyi and Robinson's (1986) elabora- complementary nature of creativity and intel-
tion of the time lines in development of talent. ligence. However, it is developmental theory that
These authors superimposed Erikson's life-span appears most clearly suited to explanation of the
transitions on Piaget's stages of cognitive develop- mutual origin of creativity and intelligence in men-
ment along with a time line for progression in a tal function, the interrelatedness of creative and in-
particular domain and in a particular field of en- telligent responses within that function, and the rea-
deavor. The latter two time lines may depend upon sonableness of universal goal-oriented adaptation.
maturational factors, but even more importantly on This adaptation is accomplished through appropri-
the presence of appropriate training experiences at ate application, timing, and intermingling of intel-
the most opportune intersects with cognitive and ligent responses (planning, executing, and monitor-
personality development. Further, the evolvement ing) and creative responses (extending beyond
of an acknowledged creative contribution will de- previously derived solutions to novel and useful
pend upon its intersection with an appropriate and ones).
accepting audience at a time that will take max- Before proceeding to further elaboration of the
nature of an integrated relationship between crea- sample of 272, 52% were among the top 20% in
tivity and intelligence as elements of a unitary phe- both creativity and intelligence as measured with
nomenon, we discuss a recurring barrier to such a the Lorge Thorndike Verbal Battery, a phe-
view, a barrier resulting from interpretation of re- nomenon Torrance attributed to the distribution of
stricted-range studies. talent in this school or to the nature of the particular
measure of intelligence used.
For comparison, however, the highly creative
group (who did not rank in the upper 20% in intel-
Problems of Restricted Range in Studies
ligence) were contrasted on achievement measures
of Creativity and Intelligence with the highly intelligent group (who did not rank
in the upper 20% on creative thinking). Using the
Data accumulated from studies of the rela- Iowa Basic Skills Battery as criterion for achieve-
tionship between assessed creativity and intel- ment, the highly intelligent group consistently (but
ligence have lent support to the idea that creativity slightly) outscored the highly creative group,
does represent a function distinctly different from though the difference was not statistically signifi-
intelligence. In these studies, correlations between cant in most samples. Torrance considered this evi-
creativity and intelligence have often differed sig- dence supportive of the Getzels and Jackson 1958
nificantly between low or moderate levels of intel- data and suggested that it also validated the Tests of
ligence and high levels of intelligence. Some of Creati ve Thinking as measures of factors other than
these studies examined a restricted range of intel- intelligence. He alluded to the possibility that the
lectual ability. Others focused on criteria of mental abilities sampled in measures of IQ may be
creativity that limited the options for creative re- more useful in certain kinds of achievement, where-
sponse for particular subjects with specific back- as the creative-thinking abilities may be more
grounds and thus may have wrongfully identified useful in other kinds of achievement. Differences in
them as noncreative. Using a sample of relevant the mean of intelligence between the highly creative
studies, problems arising from interpretation of this and the highly intelligent group, a statistically sig-
type of limited-range data are discussed in the fol- nificant difference in each sample, were further
lowing section. Interpretations of restricted-range suggested as supporting a distinction between cre-
data based on the idea of conceptual differences ative ability and intelligence. However, a dif-
between creativity and intelligence will be de- ference in mean of intelligence between these two
scribed. The discussion will center, however, on groups had to exist by virtue of the way in which the
alternative interpretations of statistical differences groups were comprised, with the most highly intel-
at low- and high-IQ levels. ligent creatives eliminated from the comparison
group just as the most highly creative intelligent
individuals were eliminated from the intelligence
Limited Range in Intellectual Ability comparison group. The results appear to support
In a study designed to partially replicate with more appropriately the contention that a high de-
elementary, high school, and graduate school stu- gree of intelligence is more likely to be associated
dents the Getzels and Jackson 1958 study, Torrance with a high degree of creativity, whereas at the
(1962) found that a significant portion of students same time, creative responses may be associated
identified as most highly creative (upper 20%) on with intelligent responses at any level of either
his battery of creative-thinking tasks (the Min- mode.
nesota Tests of Creative Thinking) also were identi- Using performance on the Remote Associates
fied as most highly intelligent (upper 20%) on a Test (RAT) (Mednick, 1962) as measure of
variety of individual and group intelligence mea- creativity and the Scholastic Aptitude Tests of Ver-
sures . Yet the overall correlations between bal Ability (SAT-V), Mednick and Andrews (1967)
creativity scores and intelligence were less than found intelligence moderately related to creative
moderate, the highest, .32, in a sample where intel- ability in college populations, r = .43. Among
ligence was assessed with the Otis Quick-Scoring 1,211 University of Michigan freshman, there was
Test of Intelligence. Among five samples of ele- a slight tendency for the relationship to increase as
mentary school children, this overlapping highly the level of intelligence increased (from .09 to .19
creative-highly intelligent group comprised from over the six increments of SAT-V scores from mini-
25% to 41 % of the total sample; in a high school mal to maximal). Mednick and Andrews (1967)
found no support for the suggestion that creativity concerned" [italics added] (Guilford, 1967, p. 9).
and intelligence are more closely related at lower However, when considering a wide range ofIQ, for
levels of intelligence, and that they are relatively example, from 62 to 150, the relationship is sub-
independent processes among the brightest stu- stantial in the lower and middle area, but less strong
dents. In this same data sample, the relationship at the upper end. Although Guilford suggested that
between the SAT-Math scores and the RAT re- IQ sets an upper limit on creative potential, other
mained negligible across the same range, varying explanations should be considered. First, we may
from .03 to .08, with an overall correlation of .20, not have sophisticated or precise enough instru-
as might be expected when using a verbal measure ments to differentiate levels of creativity at its upper
of creativity to correlate with an intelligence test range. Second, limiting description to a linear
focused on mathematics. model also may limit our understanding of the rela-
Guilford and Christensen (1973) used a divergent tionship. It may be that as the information base
production (DP) battery of tests to demonstrate that (intelligence) expands and an ever widening variety
a triangular scatter plot of the relationship between of possible responses becomes available, the type
IQ and creative potential would be found more of intelligence tests and creativity tests must be-
often than an elliptical scatter plot. These authors come more diversified as well. Guilford's emphasis
pointed out that the IQ as usually assessed probably on the importance of "transfer recall" (Guilford,
represents most strongly semantic or verbal infor- 1967), or transformation in the understanding of
mation. They hypothesized that, consequently, se- insight within creative thinking, points to multidi-
mantic divergent production tests would be more mensionality to explain the existing relationship. It
likely candidates for demonstrating the triangular very well may be necessary to use a three-dimen-
type of relationship, with an elliptical relationship sional model to demonstrate the corelationship of
when visual-figural tests were used. As hypoth- creativity with intelligence, especially to accom-
esized, their data supported triangular scatter plots modate the upper end of the continuum.
for semantic-DP tests and the more usual elliptical
form for visual-figural DP tests. The data were ob-
Multiple Criteria for Assessing Creativity
tained in a study of six groups of 40 to 95 children in
each case, grades 4, 5, and 6, groups from which Owens (1972) introduced a different type of con-
children with a Stanford-Binet IQ of 130 (Califor- cern into the creativity-intelligence question with
nia "gifted") or higher had been removed. The regard to restricted range. In contrast to sampling
resulting triangular plots showed a continuous grad- only a portion of the distribution of intelligence,
ual shift in divergent production from low to high Owens referred to the limited range of types of cre-
IQ with no breaks at any IQ level. Guilford and ative response or application included in studies
Christensen (1973) interpreted these findings to that focus on a specific field, such as psychology,
demonstrate that "the higher the IQ, the more like- architecture, mathematics. Thus, in Owens's study
ly we are to find at least some individuals with high of research and development engineers, the type of
creati ve potential" (p. 251). They further stated creative response that may most frequently be ex-
that their evidence does not support the threshold pressed in relationship to the creativity criterion for
hypothesis regarding absence of relationship be- engineers would be invention. Invention does not
tween IQ and creative potential above a critical IQ really represent the range of creative responses that
level, such as 120. might exist across other fields or occupations. As a
Guilford (1967) again emphasized the lack of matter of fact, in his study, Owens found that, in
correlation between creativity and intelligence at machine design, a specific, highly structured test
the upper end of the creativity range as he described with multiple definers for the product was a better
the results of studies by Roe (1952), MacKinnon, predictor of creativity than a general ability test.
Barron and associates (MacKinnon, 1960), and Whether the measures used to assess individual
others. Scientists, writers, architects and mathe- creativity actually possess criterion-related (con-
maticians with an observable record of creative ac- struct) validity long has been a controversial issue
complishment do not appear to exhibit a verifiable and is addressed at length in another chapter of this
statistical corelationship between creativity and in- volume. The following remarks specifically apply
telligence- "the relation of creative potential and to the topic of restricted range. Dellas and Gaier
creative production to the traditional IQ has been (1970) addressed this issue citing MacKinnon's
found close to zero where groups of superior IQ are (1961) study of architects, in which the Guilford
tests did not correlate highly with expert-peer judg- creative responses while others with an equal ca-
ment of the architects' creativity; Gough's (1961) pacity for intelligence limit their responses to the
data on the creativity of peer-rated research scien- conventional and usual. What should be added to
tists exhibited low and even negligible correlations this fostering base is an environmental respon-
with the Guilford tests of Unusual Uses, Conse- siveness to the particular direction of symptomatic
quences, and Gestalt Transformations; Bittel's behavior, without which the symptom cannot be
(1964) findings showed little relationship between recognized as creative even when the producer
measures of divergent thinking and performance in deems it is (Haensly & Roberts, 1983). Thus, ac-
art; and Skager, Klein, and Schultz's (1967) study cording to Dellas and Gaier (1970), "the roots of
of artistic achievement at a school of design yielded creativity do not seem to lie in convergent or diver-
inconsistent results. However, Barron (1963) found gent thinking, but rather ... in the personality and
contradictory results using the Guilford tests with motivational aspects of character" (p. 68). Yet the
Air Force officers and Elliott (1964) with public organic entity that can produce the creative re-
relations personnel. As Dellas and Gaier (1970) sponse remains the individual with at least a
pointed out, the criteria used for individual modicum of intelligence. Nevertheless, these au-
creativity nominations and judgments are a critical thors stopped short of suggesting that either the
factor in whether or not the assessment measures personality characteristics or motivational factors
can be expected to relate with the identifiers used in are the determining factors for identifying who will
the practical applications of a particular field or or will not be creative.
occupation. The semantic nature of many of the In a retrospective view of creativity investiga-
divergent production tests may provide an optimal tion, Taylor (1975) listed a number of definitions
medium for the creativity of individuals who are and criteria of creativity, including his own, which
engaged in advertising or public relations, but a suggest the basic concept that new ideas are in-
much less appropriate medium for the expression of volved and that what is viewed as creative at one
artistic, inventive, or discovery-based talent. time and in one culture may not be so viewed at
Della and Gaier's (1970) position goes beyond another time and culture. Wade (1968), who specu-
this suggestion of inappropriately constructed mea- lated on the role of environment in fostering the
sures using an irrelevant medium. They empha- child's creativity, helps us focus on the possibility
sized that production of creative ideas simply may that variance in quantity and quality of creative re-
be, to use a medical metaphor, the "symptom" that sponses may be a function of factors that permit the
exhibits when inherent potential meets with an ap- child to go beyond basic, intelligent responses. Fac-
propriate intervening environment. The intervening tors, such as psychological safety and freedom,
environment, or intermediary factors, from which which provide a climate where external evaluation
creativity symptoms might arise would be the mutu- is not restrictive, may foster imagination, in-
ally reinforcing clusters of personality charac- ventiveness, and fantasy. Although this type of en-
teristics (such as intuitiveness, unconventionality, vironment should have little effect on development
openness to feelings, independence, empathic ca- of the processes of efficient acquisition, effective
pability) and motivation variables (such as attrac- storage, and successful retrieval of information
tion to unconventional types of achievement, self- needed for intelligent responses, it is likely to have
actualizing need, history of reinforcement for risk- a profound effect on the production of unusual com-
taking endeavors). Reapplying the medical meta- binations, innovative ideas, or highly elaborated
phor, the individual may experience pain and de- "creative" responses. Wade's (1968) data on 105
scribe it, even suggesting its presumed source, but tenth-grade students from upper-middle-class
unless the diagnosing physician also views this pain schools resulted in a correlation of .37 between
as a recognizable symptom of condition X, the indi- creativity (using three Guilford tests of divergent
vidual's pain remains a personal expression of an thinking) and intelligence (using the Lorge-Thorn-
unrecognized condition. Although creative symp- dike Verbal test). As Wade pointed out, of greater
toms cannot evolve from a vacuum, but from an importance is the finding that "in no case did any
"effective use of intelligence" (Dellas & Gaier, subject scoring within the upper twenty percent on
1970) that prepares the information and/or skills the creativity tests score below the mean on the
from which the creative response emanates, neither verbal intelligence measure" [italics added]
will they evolve without the fostering factors. Thus, (Wade, 1968, p. 99). The mean intelligence score
some intelligent individuals go beyond to produce for the total group was 111.5, but the mean for the
highly creative subjects was 119. Additionally, the It would seem that MacKinnon's ideas directly
combination of intelligence and creativity was focused on the idea of a well-developed intelligence
found to explain almost twice as much of the as the optimal medium from which creative re-
achievement variance (grade point average) as in- sponse can arise, whether that creative response
telligence alone, a strong indication of "cognitive involves a skillful and sensitive use of an artistic
overlap" and the additivity of creativity in the intel- medium to generate an original and worthy master-
ligent response (Wade, 1968). piece or whether it involves discovering a unique
Even though MacKinnon (1962) supported the solution for containing nuclear radiation. Restricted
intelligence threshold concept and denied that his range and the resulting correlations, which have led
sample of architects represented a restricted range to an assumption of a threshold of intelligence for
of intelligence, he admitted that this sample did not the appearance of creativity and no relationship at
truly represent the entire range. MacKinnon's higher levels of intelligence, may be more of a
(1962) definition of creativeness included novelty, problem of accurate statistical description than of
problem solving or accomplishment of a goal, and the functional relationship between creativity and
development of the response to its fullest. Addi- intelligence.
tionally, his descriptors of the creative person in- Most of the studies to date that show either a
cluded "a clear preference for the complex and threshold effect in the intelligence-creativity rela-
asymmetrical" (MacKinnon, 1962, p. 488) and tionship, or that the two are independent at the' 'up-
disposition "to admit complexity and even disorder per levels," have failed to control for regression
into their perceptions without being made anxious effects on one of these variables or have fallen into
by the resulting chaos ... disordered multi- the fallacy of conclusions that can result from the
plicity" (p. 489). restriction of range phenomenon. It appears for now
From an information-processing paradigm, this that intelligence and creativity are related across the
preference for complex stimuli can be seen as set- full range of both variables, but the precise nature of
ting the stage for a wider scan of stored information the relationship remains unclear, largely for reasons
and for holding open the possibilities of multiple described by Brown in Chapter 1 in this volume.
associations between bits of information evolving
in novel ways. Thus, Sternberg's selective com-
bination and selective comparison that make up in-
sight (Sternberg & Davidson, 1983) operate upon The Synergistic Effect between
the richest of substrates-the more complex the set Creativity and Intelligence
of perceptions made available, the greater the prob-
ability that a novel and appropriate response can be "What is needed for the understanding of the
generated to accomplish a particular goal. The in- relationship between creativity and intelligence is
tensity and persistence with which this open-ended not only data at the correlational level, but concep-
search and combination and selection of ideas pro- tual reorganization as well" (Golann, 1963, p.
ceed will affect the quantity and quality of creative 560).
responses. We can see then that the personality Jackson and Messick (1967) pointed out that per-
characteristics of openness, acceptance of ambigu- formance on tests does not provide the definitive
ity, and flexibility make possible special access to foundation for that conceptual reorganization, even
the same store of information available to a variety when the conventional tests for intelligence have
of intelligent individuals. The difference in re- appeared to require different abilities than tests de-
sponse possibilities (conventional vs. creative) can signed to measure creative ability. They described
occur because of differences in personality charac- the population, so to speak, of responses that might
teristics and in the motivational attitudes. MacKin- be considered to satisfy requirements for cor-
non's reservations about the relationship of rectness in an intelligent response, and those that
creativity and intelligence could then be handled by could be considered to satisfy requirements for
viewing the process of derivation of creative re- "goodness," in a creative response. Although
sponses from a specialized information-processing originality or unusualness of response might result
perspective. From such a perspective, the focus from a request for uses of a tin can (e. g., compress-
would be on insight as the creative variant of a basic ing the tin can to use as convenient currency), the
sequence that exhibits intelligent encoding, manip- goodness of such a response results from the fact
ulation, and retrieval of stored information. that compression for reuse has value for the conser-
vation of resources. Thus, the concept of whether an unusual idea for Arctic transportation
usefulness, appropriateness, or value, introduced also could be considered appropriate for inclusion
by Mednick (1962), becomes superimposed as a in the overlap of pools between intelligent and
requirement for creativeness, to be applied to that creative.
original response; the usefulness would not appar- "Correctness" and "goodness of fit" are com-
ently have to be immediate, but useful in that the plex criteria, and judgment of test performance is
response leads to other responses of value. These much more complicated than test constructors may
populations of responses, the so-called intelligent have realized. In his model of intelligence, Guilford
and the so-called creative, become pools that have (1959) included a dimension of transformations, a
more or less of an overlap, as in a Venn diagram. particular dimension that allows for or considers
The intelligent responses outside of the area of leaps such as described above as an analyzable
overlap lack the quality of goodness, are correct but mode, different from the mode that addresses mate-
mundane, suffice but do not move the solution for- rial and problems within their expected realm.
ward to greater possibilities. The creative responses Transformations, then, well may create new
that lie outside of the overlap area are original and realms, realms that now can elicit their own expec-
unusual, but, in so far as they lack justification, also tations for response. Jackson and Messick (1967)
lack ability to move the solution forward to pos- discussed the personal impact of the products in
sibilities of value. Yet a marvelously synergistic each of these cases, describing their impact as sur-
effect takes place when the two pools overlap. As prise to the criterion of unusualness, satisfaction to
described in Webster's dictionary (1980, p. 1174), appropriateness, and stimulation (of new ideas) to
synergism occurs as the result of "the cooperative transformation.
action of discrete agencies such that the total effect Jackson and Messick (1967) added one further
is greater than the sum of the effects taken criterion to the intelligence/creativity collection
independently. " that, perhaps in their judgment and in ours, is the
Jackson and Messick (1967) reflected that these most critical of all in considering these overlapping
overlapping criteria may apply in the aesthetic pools of responses characterized as having cor-
realm as well as in the realm of thinking and logic; rectness and those having goodness. This fourth
for example, it can be assumed that there is a wide criterion of condensation refers to the bringing to-
gap between the paint smattered canvas of the child gether of a set of ideas in such a simple way as to
and the abstract canvas of a Picasso, even though produce a sense of "why didn't I ever think of that
both have applied color to background in seemingly before-it is so clearly fortuitous," yet of such
ill-structured representations or "spatial group- complexity that different viewers will not only un-
ings." The child's painting, though it might be understand its simplicity in different ways, but also
usual, may lack the direction, substance, and cor- will be able to expand and generate entire new sets
rectness (sophistication?) of form present in the of understandings. Thus, Jackson and Messick
Picasso; the "goodness" criterion can then only be (1967) described the recipient's response to the per-
bestowed in comparison with the works of others formance or product as one of pondering or
with similar degrees of background and practice. "savoring. "
Appropriateness of the response still may be ques- This fourth criterion of condensation, as does
tioned, and then only those individuals who have that of transformation, requires judgment by indi-
similar mental sets, or are in tune with the thinking viduals who have a sufficient foundation of infor-
of the artist, may be able to make that particular mation, have personally experienced transforming
judgment-individuals with quite highly devel- of ideas or understanding and appreciation of such
oped intelligence and/ or creativity in such realms as transformations, and have the capacity to com-
engineering, medicine, or literature may fail to see prehend both the simplicity and complexity of the
meaning in Picasso's abstract shapes or to be moved product, that is, of the response or performance. As
to aesthetic appreciation of them. Yet those same Gowan has suggested, sometimes such individuals
individuals may see those' 'missing" qualities in an may not be available until a future time. In his
innovatively designed bridge span or in the symme- words, he would donate copies of his books to li-
try of an anatomical arrangement. Returning from braries so that' 'when the future caught up with my
the aesthetic to the practical, we might see a similar thinking, it would be there" (Haensly & Roberts,
need for judgment by an individual with compara- 1983, p. 11). Actually, we may have few indi-
ble background, interests, and mind set to determine viduals who are prepared to make such judgments
on the individual intelligence and/or creativity of ble within a rational framework. In fact, we have
many test performance responses. Such judgment avoided, for the most part, the humanistic and psy-
presumes individuals do and will always exist who choanalytic perspectives of creativity which view it
have equal or greater capacity than the test as the result of internal self-actualizing forces, or as
respondent. a function of the subconscious shaped by elements
Perhaps the most enlightening suggestion of of personality outside of the realm of individual
Jackson and Messick's model (1967) is the idea that control. We submit that, in so doing, we may be
test performance or product formation is the ele- overlooking important qualities of creativity, ones
ment upon which assessment of intelligence and/ or that could contribute to its extraordinariness,
creativity must be applied, rather than upon the uniqueness, and even mysteriousness.
individual. This idea, however, is so obvious that Attempting to determine how much intelligence
critics must quickly respond "But of course, that is associated with how much creativity would ap-
was always the intention of intelligence tests, of pear to be an exercise in small views of the universe
creativity tests, to assess responses and thus cate- of mental capability. Creativity and intelligence are
gorize the individual as intelligent, as creative, or as
an interrelated process and we propose that the in-
both." Nevertheless, the entire body of research on telligent response, at whatever level the individual
the characteristics of creative individuals is predi-is capable of, produces the ordinary, while the cre-
cated on the idea of defining clusters of traits thatative response lifts the ordinary into the realm of
permit us to label an individual as intelligent or asextraordinary. This interpretation may help us un-
creative. From this model, however, the personal derstand better Guilford's (1981) data, which indi-
characteristics, such as openmindedness, reflec- cated that a nonlinear relationship existed between
tiveness, spontaneity, playfulness, inconsistency, verbal IQ and creativity test scores, with a higher
intuitiveness, sensitivity, and flexibility, can be correlation below 120 IQ and a lower one above 120
seen to facilitate the responses that combine un- IQ. As Guilford emphasized in his earlier work
usualness versus commonality with appropri- (1968) on complex problem-solving tasks, each in-
ateness versus lack of fit, transformation versus re- dividual must produce his or her own strategy to
taining a traditional or accepted form, and capitalize on personal strengths and work around
condensation versus ambiguous complexity or in- weaknesses. Individuals whose IQ exceeds 120
correct simplicity. These are the response charac- may be able to depend more often on convergent
teristics that differentiate whether the response falls
process to produce acceptable or correct responses
in the correctness pool, the unique idea pool, or the to difficult and complex problems, whereas success
overlapping pool of ideas that transcend simple cor- for individuals whose IQ is below 120 may more
rectness and novelty. directly depend upon the extent to which they can
access divergent production. It would be of interest
to repeat studies, such as Guilford's, using educa-
tion as an independent variable, partialling intel-
Integration: A Model for Creativity and ligence from the relationship, and then repeating
Intelligence the work partialling education from the IQ-
creativity relationship.
Wallach (1985) stated that "creativity may be The annals of scientific invention, of musical and
best understood as what constitutes the work done artistic accomplishment, of biological discovery, of
at the cutting edge of a given field by those who mathematical logic or of governmental and political
have mastered it" (p. 117). Identification of such greatness are filled with descriptions of the creative
individuals remains a legitimate goal. In his Walter process as it evolved in singularly extraordinary
Bingham Lecture, MacKinnon (1962) referred to contributions. As deBono (1971) stated, some peo-
the loss to individuals and to society when indi- ple have better developed capacities for generating
viduals do not recognize their own possibilities for new ideas than others. New ideas do not always
creative response and when environments fail to come to those who have worked the hardest or
nourish or respond positively to creative expres- learned the most about their domain. Yet some indi-
sion. viduals go beyond deep thinking to bring into exis-
In our examination of the relationship between tence new ideas, new insights, and new break-
creativity and intelligence, we have made an as- throughs at "the cutting edge" of their fields.
sumption that creativity is a phenomenon analyza- These are the creative contributions. Not all of these
creative acts are of the same importance or sophis- deBono, E. (1971). New think. New York: Avon Books.
tication, nor are all recognized by a societal "medi- Dellas, M., & Gaier, E. L. (1970). Identification of creativity:
um" that is perhaps unreceptive to the particular The individual. Psychological Bulletin, 73 (1), 55-73.
Dewey, J. (1933). How we think. New York: Heath.
creative application or insufficiently prepared to Eisner, E. W. (1965). Children's creativity in art: A study of
understand it. Yet each creative act may be the types. American Educational Research Journal, 2, 125-136.
ultimate expression of intelligence, in which all of Elliott, J. M. (1964). Measuring creative abilities in public rela-
the cognition and comprehension that individuals tions and in advertising work. In C. W. Taylor (Ed.),
Widening horizons in creativity. New York: Wiley.
have developed at that point in their time (age) and Estes, W. K. (1985). Learning, memory, and intelligence. In R.
situation (context) with their degree of training (ex- J. Sternberg (Ed.), Handbook of human intelligence (pp. 170-
perience) have been brought to bear upon a particu- 224). Cambridge, England: Cambridge University Press.
lar idea or problem. We propose that creativity is Feldman, D. H., with Benjamin, A. C. (1986). Giftedness as a
developmentalist sees it. In R. J. Sternberg & J. E. Davidson
not another "breed" of mental processing, but is
(Eds.), Conceptions of giftedness (pp. 285-305). Cambridge,
the ultimate expression of that finely honed system England: Cambridge University Press.
of thinking we know of as intelligence. Fishbein, H. D. (1976). Evolution, development, and children's
learning. Pacific Palisades, CA: Goodyear.
Gardner, H. (1982). Art, mind, and brain. A cognitive approach
to creativity. New York: Basic Books.
References Gardner, H. (1983). Frames of mind. New York: Basic Books.
Getzels, J. W., & Jackson, P. W. (1959). The highly intelligent
Anastasi, A. (1958). Heredity, environment, and the question and the highly creative adolescent: A summary of some re-
"How?" Psychological Review, 65, 197-208. search findings. In C. W. Taylor (Ed.), The Third University
Anastasi, A., & Schaefer, C. E. (1971). Notes on the concepts of of Utah Research Conference on the Identification ofCreative
creativity and intelligence. Journal of Creative Behavior, 5, Scientific Talent. Salt Lake City: University of Utah Press.
113-116. Getzels, J. W., & Jackson, P. W. (1962). Creativity and intel-
Ausubel, D. P. (1978). The nature and measurement of ligence: Explorations with gifted students. New York: Wiley.
creativity. Psychologia, 21, 179-191. Glover, J., & Gary, A. L. (1976). Procedures to increase some
Barron, F. (1963). Creativity and psychological health: Origins aspects of creativity. Journal ofApplied Behavior Analysis, 9,
of personality and creative freedom. Princeton, NJ: Van 79-84.
Nostrand. Glover, J., & Sautter, F. (1977). Procedures for increasing four
Bittel, K. R. (1964). Creativity in the visual arts in higher educa- behaviorally defined components of creativity within formal
tion. In C. W. Taylor (Ed.), Widening horizons in creativity. written assignments among high school students. School Ap-
New York: Wiley. plications of Learning Theory, 9, 3-22.
Bloom, B. (1985). Developing talent in young children. New Goetz, E. M., & Baer, D. M. (1973). Social control of form
York: Ballantyne. diversity and the emergence of new forms in children's block-
Brigham, T. A., Graubard, P. S., & Stans, A. (1972). Analysis bUilding. Journal ofAppliedBehavioralAnalysis, 6, 209-217.
of the effects of sequential reinforcement contingencies on Golann, S. E. (1963). Psychological study of creativity.
aspects of composition. Journal of Applied Behavior Analy- Psychological Bulletin, 60, 560.
sis, 5, 421-429. Gough, H. (1961). Techniques for identifying the creative re-
Case, R. (1978). Piaget and beyond: Toward a developmentally search scientist. In D. W. MacKinnon (Ed.), The creative
based theory and technology of instruction. In R. Glaser person. Berkeley: Institute of Personality Assessment and Re-
(Ed.), Advances in instructional psychology (Vol. 1, pp. 167- search, University of California.
228). Hillsdale, NJ: Erlbaum. Guilford, J. P. (1950). Creativity. American Psychologist, 5,
Cattell, R. B. (1963a). The personality and motivation of the 444-454.
researcher from measurements of contemporaries and from Guilford, J. P. (1959). Three faces of intellect. American Psy-
biography. In C. W. Taylor & F. Barron (Eds.), Scientific chologist, 14, 469-479.
creativity: Its recognition and development (pp. 119-131). Guilford, J. P. (1963). An informational theory of creative think-
New York: Wiley. ing. Paper presented at the Convention of the Western Psycho-
Cattell, R. B. (1963b). Theory of fluid and crystallized intel- logical Association, Santa Monica, CA. (Available in USAF
ligence: A critical experiment. Journal of Educational Psy- Instructors' Journal, I, 28-33.)
chology, 54, 1-22. Guilford, J. P. (1967). Creativity: Yesterday, today, and tomor-
Coler, M. A. (1963). Essays on creativity in the sciences. New row. Journal of Creative Behavior, I, 3-14.
York: New York University Press. Guilford, J. P. (1968). Intelligence, creativity and their educa-
Conger, J. J. (June, 1957). The meaning and measurement of tional implications. San Diego, CA: Robert Knapp.
intelligence. Rocky Mountain Medical Journal, 54, 10-12. Guilford, J. P. (1981). Potentiality for creativity. In J. C.
Crockenberg, S. B. (1972). Creativity tests: A boon or boondog- Gowan, J. Khatena, & E. P. Torance (Eds.), Creativity: Its
gle for education? Review of Educational Research, 42, 27- educational implications (2nd ed., pp. 1-5). Dubuque, IA:
45. Kendall/Hunt.
Csikszentmihalyi, M., & Robinson, R. E. (1986). Culture, time, Guilford, J. P., & Christensen, P. R. (1973). The one-way rela-
and the development of talent. In R. J. Sternberg & J. E. tion between creative potential and IQ. Journal of Creative
Davidson (Eds.), Conceptions of giftedness (pp. 264-284). Behavior, 7, 247-252.
Cambridge, England: Cambridge University Press. Haensly, P. A. (1979). Behavior modification for problem solv-
ing: Originality and creative production expanded. Roe, A. (1952). The making of a scientist. New York: Dodd,
Unpublished manuscript. Texas A & M University, Gifted & Mead.
Talented Institute, College Station, TX. Roe, A. (1963). Psychological approaches to creativity in sci-
Haensly, P. A., & Roberts, N. M. (1983). The professional ence. In M. A. Coler (Ed.), Essays on creativity in the sci-
productive process and its implications for gifted studies. ences (pp. 153-182). New York: New York University Press.
Gifted Child Quarterly, 27, 9-12. Rossiman, T., & Horn, J. (1971). Cognitive, motivational and
Haensly, P. A., Reynolds, C. R., &Nash, W. R. (1986). Gifted- temperamental indicants of creativity and intelligence.
ness: coalescence, context, conflict, and commitment. In R. J. Journal of Educational Measurement, 9, 265-286.
Sternberg & J. E. Davidson (Eds.), Conceptions of giftedness Rothenberg, A., & Hausman, C. R. (1976). The creativity ques-
(pp. 128-148). Cambridge, England: Cambridge University tion. Durham, NC: Duke University Press.
Press. Schubert, A. (1973). Intelligence as necessary but not sufficient
Harper, T. (1983, October 4). Celebrated sculptor can't even for creativity. Journal of Genetic Psychology, 122, 45-47.
read (Associated Press). Boulder, CO: Bryan-College Station Simonton, D. K. (1979). The eminent genius in history: The
Eagle, page Ie. critical role of creative development. In J. C. Gowan, J.
Hebb, D. O. (1972). Textbook of psychology (3rd ed.). Phila- Khatena, & E. P. Torrance (Eds.), Educating the ablest (2nd
delphia: W. B. Saunders. ed., pp. 79-87). New York: F. E. Peacock.
Jackson, P. W., & Messick, S. (1967). The person, the product, Simonton, D. K. (1984). Genius, creativity, and leadership:
and the response: Conceptual problems in the assessment of Historiometric inquiries. Cambridge: Harvard University
creativity. In J. Kagan (Ed.), Creativity and learning (pp. 1- Pr~s.
19). Boston: Houghton Mifflin. Simonton, D. K. (1985, December). Genius, creativity and lead-
Kagan, J. (Ed.). (1967). Creativity and learning. Boston: ership. EEE Potentials, pp. 31-32.
Houghton Mifflin. Skager, R. W., Klein, S. P., & Schultz, C. B. (1967). The
Krippner, S. (1983). A systems approach to creativity based on prediction of academic and artistic achievement in a school of
Jungian typology. Gifted Child Quarterly, 27, 86-89. design. Journal of Educational Measurement, 4, 105-117.
Locurto, e. M. (1974). Verbal operant conditioning and self- Spearman, C. (1927). The abilities of man: Their nature and
reinforcement of originality. Dissertation Abstracts Interna- measurement. New York: Macmillan.
tional, 35, 1945-B. Spearman, e. (1930). Creative mind. London: Nisbet.
MacKinnon, D. W. (1960). The highly effective individual. Sternberg, R. J. (1984). Toward a triarchic theory of human
Teachers College Record, 61, 367-378. intelligence. The Behavioral and Brain Sciences, 7, 269-315.
MacKinnon, D. W. (1961). Creativity in architects. In D. W. Sternberg, R. J. (1985). Beyond 10: A triarchic theory of human
MacKinnon (Ed.), The creative person. Berkeley: Institute of intelligence. Cambridge, England: Cambridge University
Personality Assessment and Research, University of Press.
California. Sternberg, R. J. (1986). Identifying the gifted through IQ: Why a
MacKinnon, D. W. (1962). The nature and nurture of creative little bit of knowledge is a dangerous thing. Roeper Review, 8,
talent. American Psychologist, 17, 484-495. 143-147.
Maloney, M. P., & Ward, M. P. (1976). Psychological assess- Sternberg, R. J., & Davidson, J. E. (1983). Insight in the gifted.
ment: A conceptual approach. New York: Oxford University Educational Psychologist, 18, 51-57.
Press. Sternberg, R. J., & Wagner, R. K. (1986). Practical intel-
Maltzman, I. (1960). On the training of originality. ligence: Nature and origins of competence in the every-
Psychological Review, 67, 229-242. day world. Cambridge, England: Cambridge University
May, R. (1975). The courage to create. New York: Norton. Press.
Mednick, S. A. (1962). The associative basis of the creative Taylor, C. W., & Barron, F. (Eds.). (1963). Scientific
process. Psychological Review, 69, 220-227, 232. creativity: Its recognition and development. New York:
Mednick, S. A., & Andrews, F. M. (1967). Creative thinking Wiley.
and level of intelligence. Journal of Creative Behavior, I, Taylor, I. (1975). A retrospective view of creativity investiga-
428-431. tion. In I. A. Taylor & 1. W. Getzels (Eds.), Perspectives in
Nash, W. R. (1975). The effects of warm-up activities on small creativity (pp. 1-36). Chicago: Aldine Publishing.
group divergent problem solving with young children. Terman, L. M. (1954). Scientists and nonscientists in a group of
Journal of Psychology, 89, 237-241. 800 gifted men. Psychological Monographs: General andAp-
Newell, A., Shaw, J. e., & Simon, H. A. (1962). The processes plied, 68 (7, Whole No. 378).
of creative thinking. In H. E. Gruber, G. Terrell, & M. Terman, L. M. (1955). Are scientists different? Scientific Ameri-
Wertheimer (Eds.), Contemporary approaches to creative can, 192(1), 25-29.
thinking (pp. 63-119). New York: Atherton Press. Thorndike, E. L. (1921). Intelligence and its measurement.
Owens, W. A. (1972). Intellective, non-intellective, and en- Journal of Educational Psychology, 12, 124-127.
vironmental correlates of mechanical ingenuity. In e. Taylor Thorndike, E. L. (1926). Measurement of intelligence. New
(Ed.), Climate for creativity (pp. 253-268). New York: Per- York: Teachers' College, Columbia University.
gamon Press. Thurstone, L. L. (1931). Multiple factor analysis. Psychological
Pascuale-Leone,1. (1970). A mathematical model for transition Review, 38, 406-427.
in Piaget's developmental stages. Acta Psychologica, 32, Torrance, E. P. (1960). Eight partial replications of the Getzels-
301-345. Jackson study. Research Memorandum BER-60-15.
Piaget, J. (1972). Intellectual evolution from adolescence to Minneapolis: Bureau of Educational Research, University of
adulthood. Human Development, 15,1-12. Minnesota.
Roe, A. (1951). A psychological study of physical scientists. Torrance, E. P. (1962). Guiding creative talent. Englewood
Genetic Psychology Monographs, 43 (2), 121-235. Cliffs, NJ: Prentice-Hall.
Torrance, E, p, (1966), Torrance Tests of Creative Thinking: Wallas, G. (1926, 1945). The art of thought. New York: Har-
Norms-technical manual. Princeton, NJ: Personnel Press. court, Brace.
Torrance, E. P. (1967). Scientific views of creativity and factors Webster's New Collegiate Dictionary (1980). Springfield, MA:
affecting its growth. In J. Kagan (Ed.), Creativity and learn- G. & c. Merriam.
ing (pp. 73-91). Boston: Houghton Mifflin. Wittrock, M. C. (1974). Learning as a generative process.
Vernon, P. E. (1950). The structure of human abilities. London: Educational Psychologist, 11, 87-95.
Methuen. Yamamoto, K. A. (1964a). Role of creative thinking and intel-
Wade, S. (1968). Differences between intelligence and ligence in high school achievement. Psychological Reports,
creativity: Some speculation on the role of environment. 14, 783-789.
Journal of Creative Behavior, 2, 97-101. Yamamoto, K. A. (1964b). Threshold of intelligence in academ-
Wallach, M. A. (1985). Creativity testing and giftedness. In F. ic achievement of highly creative students. Journal ofExperi-
D. Horowitz & M. O'Brien (Eds.), The gifted and talented: mental Education, 32, 401-405.
Developmental perspectives (pp. 99-124). Washington, DC: Yamamoto, K. A. (l964c). A further analysis of the role of
American Psychological Association. creative thinking in high-school achievement. Journal of Psy-
Wallach, M. A., & Kogan, N. (1965). Modes of thinking in chology, 58, 277-283.
young children. New York: Holt, Rinehart & Winston. Yamamoto, K. A. (1965). Validation of tests of creative think-
Wallach, M. A., & Kogan, N. (1972). Creativity and intel- ing: A review of some studies. Exceptional Children, 31,
ligence in children. In A. Rothenberg & c. R. Hausman 281-290.
(Eds.), The creativity question (pp. 208-217). Durham, NC:
Duke University Press.
PART II
Cognitive Models of
Creativity
Part II presents a collection of chapters in which the authors have developed models of creative thought
based on contemporary cognitive psychology. Hayes begins with an overall perspective, "Cognitive
Processes in Creativity." Flowers and Garbin then focus on the role of perceptual processes in their
chapter, "Creativity and Perception." Next, Stein examines the role of memory in creative thought in
"Memory and Creativity." Armbruster examines the potential roles in creativity of metacognition in her
chapter on "Metacognition in Creativity." Moshman and Lukin take a cognitive-developmental perspec-
tive in their chapter, "The Creative Construction of Rationality: A Paradox?" This second part of the
volume then closes with a look at the development of adult thought and creativity in the chapter by Benack,
Basseches, and Swan, "Dialectical Thinking and Adult Creativity."
CHAPTER 7
Cognitive Processes in Creativity

John R. Hayes
"Creative" is a word with many uses. Sometimes it Despite the vagueries of such judgments, there
is used to describe the potential of persons to pro- appears to be a core of three evaluations that under-
duce creative works whether or not they have pro- lie the identification of a creative act: (1) The act
duced any work as yet. Sometimes it is used to must be seen as original or novel; (2) the act must be
describe everyday behaviors as, for example, when seen as valuable or interesting; and (3) the act must
a nursery school curriculum is said to encourage reflect well on the mind of the creator. All three of
creative activities, such as drawing or storytelling. these criteria appear to be essential if an act is to be
In this chapter, I will restrict the meaning of the considered creative. No matter how well executed a
term in two ways. First, I will be concerned solely work may be, it will not be considered creative
with creative productivity, that is, with creativity unless it incorporates substantial new ideas not easi-
expressed in the actual production of creative works ly derived from earlier work. Thus, even the best
and not with the unexpressed potential for produc- copies of paintings are not judged creative, not, at
ing such works. Second, I will be concerned only least, if the source is known. And no matter how
with creative acts at the highest level, that is, with original an act is, it will not be considered creative
the best and most valued works of artists, scientists, unless it also is judged to be valuable. A composer
and scholars. may arrange notes in a novel and unexpected way,
Society defines creative acts through a complex but the work will not be considered creative unless
process of social judgment. It relies most heavily on it also is judged to have musical value. Finally, an
the opinions of relevant experts in making such act will not be judged creative unless it reflects the
judgments-music critics, art historians, scholars, intelligence of the person who is the creator. If a
and scientists who are presumed to know the field. work is produced entirely accidentally, then it is not
But even expert judgments are highly subjective judged to be creative. This does not mean that
and are frequently influenced by irrelevant factors. chance cannot playa role in genuinely creative acts.
For example, judgments may be influenced by the Austin (1978) made an interesting distinction
experts' current focus of attention-Gregor Men- among four kinds of chance events. Chance I is just
del had to wait decades before the appropriate ex- blind luck; it could happen to anyone and does not
perts recognized that his work was important-and depend on any special ability of the person it hap-
by the reputation of the creator-it is difficult for an pens to. In Chance IT, luck depends on the person's
unknown writer to get a publisher's attention. curiosity or persistence in exploration. The fact that
a curious person attends more, say, to the habits of
John R. Hayes • Department of Psychology, Carnegie- beetles makes him or her more likely to discover
Mellon University, Pittsburgh, PA 15238. something interesting about beetles than a person
135
136 PART II • COGNITIVE MODELS OF CREATIVITY
who regards beetles simply as something to be dingly productive. Although the first group out-
squashed. In Chance III, luck depends on the per- published the second group at a rate of eight to one,
son's having extensive knowledge of the field that there was no difference in IQ between them. In a
is not shared by most people. Thus, the discovery of similar study, MacKinnon (1968) compared scien-
radium by Pierre and Marie Curie depended on their tists, mathematicians, and architects who had made
recognizing that a certain mineral was more radio- distinguished contributions to their respective fields
active than it ought to be on the basis of the known with a matched group who had not made dis-
elements it contained. Clearly, only a very knowl- tinguished contributions. There was no difference
edgeable person could make such a discovery. This between the two groups in either IQ or in school
is the sort of chance that Louis Pasteur was referring grades.
to when he said that "chance favors only the pre- How can it be that creative scientists and architects
pared mind. " Finally, in Chance IV, luck depends have higher than average IQs and yet IQ does not
on the person's particular and, perhaps, unique in- predict which one of two professionals will be the
tellectual style or pattern of interests. Acts that in- more creative? At least two alternative theories seem
volve chance events of the last three types do reflect plausible. I will call the first alternative the' 'thresh-
credit on the mind of the actor and are thus poten- old theory. " According to this theory, a person's IQ
tially creative. must be above some threshold value, say, 120, if that
In the remainder of this chapter, I will discuss person is to be successful in creative activities. IQ
data that bear on two major questions: "What are differences above the threshold level, however,
the characteristics of creative people?" and' 'What make no difference in creativity. The reason that
cognitive processes are involved in creative acts?" there is no correlation between IQ and creativity
Finally, I will present a theoretical framework to among professionals is that schooling weeds out
account for these data. professionals with IQs that are less than 120.
I have proposed a second alternative theory
which I call the "certification theory" (Hayes,
1978). According to the certification theory, there
Characteristics of Creative People is no intrinsic relation between creativity and IQ.
Being creatively productive, however, depends on
Do Creative People Have High IQs? Yes and No attaining a position in which one can display
It is often assumed that creativity is closely related creativity-such as a college professor, an indus-
to IQ. Indeed, both Roe (1953), who did studies of trial chemist, or an architect. Being considered for
eminent physicists, biologists, and social scientists, these positions typically requires a college or also a
and MacKinnon (1968), who also studied dis- graduate degree. Because academic performance is
tinguished research scientists, mathematicians, and correlated with IQ, it may be that one's opportunity
architects, found that these creative individuals had to be creative depends on IQ simply because of the
IQs ranging from 120 to 177 -well above the gener- degree requirement. Thus, creative people may not
al average. However, these higher-than-average IQs need high IQs to be creative but they may need them
cannot be taken as an explanation of the observed to be certified to get jobs where they can put their
creativity and, indeed, may be unrelated to it. creativity to work. This second alternative is worth
Several studies have indicated that highly cre- considering, because, if it is correct, or even partly
ative individuals in a particular field do not have correct, our society inappropriately may be dis-
IQs higher than the IQs of matched individuals in couraging a large portion of the creative individuals
their field who are not judged to be creative. For in the population.
example, Harmon (1963) rated 504 physical and
biological scientists for research productivity and
Other Cognitive and Personality Traits
found no relation between creativity and either IQ
or school grades. Bloom (1963) studied two sample A large number of studies have been conducted
groups of chemists and mathematicians. One sam- to identify the cognitive and personality traits that
ple consisted of individuals who were judged out- characterize creative people. Surprisingly, studies
standingly productive by their colleagues. The of cognitive traits generally have yielded disap-
other group consisted of scientists who were pointing results, with perhaps the most disappoint-
matched in age, education, and experience to the ing being the results on divergent thinking. Diver-
first sample, but who were not judged to be outstan- gent thinking is widely believed to be an important
CHAPTER 7 • COGNITIVE PROCESSES IN CREATIVITY 137
part of the creative process (Guilford, 1967), and opportunity to do really creative research and to
measures of divergent thinking constitute a major choose problems of interest to them" (p. 6).
component in the most popular creativity tests MacKinnon (1961) found that creative architects
(e.g., the Torrance Tests of Creative Thinking). also strongly preferred independent thought and ac-
However, Mansfield and Busse (1981) reviewed tion rather than conformity. Ypma (1968) found
the studies of divergent thinking in scientific that creative scientists were more likely than other
thought and concluded that essentially there is no scientists to say that they would like to have "a
evidence relating divergent thinking to creative per- good deal of responsibility" in their jobs. Further,
formance in science. These same researchers also Ypma found that creative scientists were much
reviewed studies of 16 other cognitive tests and more likely than others to answer yes to the ques-
concluded that none "has consistently shown high tion, "Did you ever build an apparatus or device of
correlations with measures of real-life creativity" your own design on your own initiative and not as
(p. 50). Researchers have been more successful, part of any required school assignment during your
however, in identifying personality traits in creative later school years?" (In this instance, "later school
people. I will therefore review evidence concerning years" refers to high school and college.) This last
four traits that appear to differentiate more creative result is interesting in the light of the success that
from less creative people: devotion to work, inde- the Westinghouse Science Talent Search has had in
pendence, drive for originality, and flexibility. identifying outstandingly creative scientists. This
talent search has selected 40 high school students
Devotion to Work. One of the most consistent each year since 1942 on the basis of self-initiated
observations about creative people is that they work projects rather than written tests or grades. The pro-
very hard. Roe (1953) studied a group of top rank- jects are then evaluated for excellence by two scien-
ing physicists and biologists and described them in tists in the project's field. In the group of 1,520
this fashion: students selected between 1942 and 1979, there are
five Nobel prize winners, five winners of Mac-
There is only one thing that seems to characterize the total group, Arthur Fellowships, and two winners of the Fields
and that is absorption in their work, over long years, and fre-
Medal in Mathematics. This remarkable perfor-
quently to the exclusion of everything else. This was also true of
the biologists. This one thing alone is probably not of itself mance suggests that the tendency to initiate inde-
sufficient to account for the success enjoyed by these men, but it pendent action is, indeed, an important trait of the
appears to be a sine qua non. (p. 233-234) creative person and that it may be exhibited quite
early in the person's career.
Chambers (1964) and Ypma (1968) also reported
that creative people work harder than others. Harris The Drive for Originality. Because creative
(1972) reported that University of California pro- acts are by definition original, it would not be sur-
fessors spend an average of 60 hours weekly on prising if creative people showed a special drive to
teaching and research. Herbert Simon, who re- be original. In fact, that is just what the research has
ceived the 1978 Nobel Laureate in economics, shown. MacKinnon (1963) described the typical
spent about 100 hours per week for years doing the creative architect in his study as "satisfied only
work for which he eventually won the Nobel Prize with solutions which are original and meet his own
(personal communication, May 15, 1986). high standards of architectural excellence" (p.
276). Ypma (1968) found that when they are asked
Independence. Researchers consistently have about their major motivations, the more creative
found that creative people have a strong drive for scientists were likely to answer, "to come up with
independence of thought and action. In particular, something new." Barron (1963) and Bergum
they seem to want very strongly to make their own (1975) have made similar observations.
decisions about what they do. Chambers (1964)
found that the creative scientist "is not the type of Flexibility. Helson and Crutchfield (1970) ad-
person who waits for someone else to tell him what ministered the California Psychological Inventory
to do, but rather thinks things through and then to 105 mathematicians who had been rated for
takes action on his own with little regard to conven- creativity by other mathematicians. The more cre-
tion or current 'fashion'" (p. 14). He also found ative mathematicians scored significantly higher on
that "when seeking a position, ... the over- the flexibility scale than did the less creative mathe-
whelming choice for the creative scientists is the maticians. In an extensive review of research on
creativity in engineers, Rouse (1986) also found 2. Look for linear relations among variables.
that flexibility was strongly correlated with creative 3. If two variables increase together, consider
performance. Creative engineers tended to mix al- their ratio.
gorithmic and associative thinking and to represent 4. If one variable increases while another de-
knowledge both visually and symbolically. creases, consider their product.
When provided with appropriate data, this program

What Cognitive Processes Are Involved successfully induced Boyle's law, Kepler's third
in Creative Acts? law, Galileo's law, and Ohm's law.
Lenat (1976) had demonstrated earlier that a
In this section, I will present an analysis of cre- well-specified set of heuristics, incorporated in his
ative acts in terms of familiar cognitive processes, program AM (for Automated Mathematician),
that is, in terms of processes involved in everyday could make interesting discoveries in mathematics.
thought and action. Before doing so, though, I For example, AM discovered de Morgan's laws,
should note that there are (at least) two points of the unique factorization of numbers into primes,
view which hold that such an analysis is impossible. and Goldbach's conjecture.
The first of these is that creative acts are, in princi- Of course, these results do not mean that human
ple, unanalyzable and the second, that creative acts creative processes can be accounted for entirely in
involve special processes not involved in other terms of such search heuristics. If a person did make
kinds of thought. a discovery by applying search heuristics to data, it
would still be interesting to ask what motivated the
person to examine those data. However, the results
Are Creative Processes Unanalyzable? do demonstrate the plausibility of accounting for an
Popper (1959) asserted quite forcefully that the important part of the creative process through com-
process of scientific discovery is indeed unanalyza- monsense search heuristics.
ble. In his book, The Logic of Scientific Discovery,
Popper claimed that Is There a Special Creative Process?
The initial stage, the act of conceiving or inventing a theory, At present, the special process view appears to
seems to me neither to call for logical analysis nor to be suscepti- have achieved "straw man" status in the scientific
ble of it. ... My view of the matter, for what it is worth, is that literature on creativity and is much more frequently
there is no such thing as a logical method of having new ideas, or attacked than defended. Further, there are no live
a logical reconstruction of this process. My view may be ex-
pressed by saying that every discovery contains "an irrational candidates for' 'special creative process" that have
element," or "a creative intuition," in Bergson's sense. (pp. substantial empirical backing. Although we should
31-32) not rule out the possibility that such special pro-
cesses may be discovered someday, we should con-
In their book, Scientific Discovery: An Account tinue to exercise a healthy skepticism toward candi-
of the Creative Processes, Langley, Simon, dates proposed in the popular press (e. g., "lateral
Bradshaw, and Zytkow (1987) presented a position thinking" and "right-brain thinking"). Parsimony
directly challenging Popper's view. These authors appears to be serving us well in this area.
argued that it is indeed possible to account for scien-
tific discovery in terms of well-specified heuristic
The "Nothing-Special" Position
procedures. In particular, they hold that discoveries
are achieved when the scientist applies sensible This position, due primarily to Simon and his co-
heuristic procedures in drawing inferences from workers (Newell, Shaw, & Simon, 1964; Simon,
data. They argued quite convincingly for the ade- 1966), holds that creative acts are a variety of prob-
quacy of this view by incorporating such heuristics lem solving and involve only those processes that
in computer programs and by showing that these also are involved in everyday problem-solving ac-
programs can induce well-known scientific laws tivities. According to this view, creative acts are
from data. For example, one program, BACON 1, problem-solving acts of a special sort. First, they
incorporates the following search heuristics: are problem-solving acts that meet criteria such as
those above-that is, they are seen as novel and
1. Look for variables (or combinations of vari- valuable and they reflect the cognitive abilities of
ables) with constant value. the problem solver. Second, they typically involve
ill-defined problems-that is, problems that cannot at least five different recordings were currently
be solved unless the problem solver makes deci- available.) From these data, I calculated when in
sions or adds information of his or her own. Ill- the composer's career, that is, how many years after
defined problems occur frequently in practical set- the onset of serious interest, each work was com-
tings. For example, in architectural practice, the posed. Out of more than 500 works, only three were
client typically specifies a few of the properties of a composed before year 10 of the composers' careers,
building to be designed, but the architect must sup- and these three works were composed in years 8 and
ply many more features before the design problem 9. Averaged over the group, the pattern of career
can be solved. productivity involved an initial lO-year period of
To describe creative activities as problem solving silence, a rapid increase in productivity from year
need not, but to many does, suggest that creation 10 to year 25, a period of stable productivity from
happens only when the creative person is in some year 25 to about year 45, and then a period of gradu-
sort of trouble. To an extent, this is true. Necessity al decline.
is the mother of invention-at least, of some In the same paper, I reported a parallel study of
invention. But there are other sorts of situations that 131 painters, using biographical data to determine
lead to creation. Creators are not always digging when each artist became seriously involved in
themselves out of trouble. In many cases, it is rea- painting. I defined the notable works ofthese paint-
sonable to think of them as taking advantage of ers as those that were reproduced in any of 11 gener-
opportunities, of recognizing the possibility of im- al histories of art. The pattern of career productivity
proving what is currently a satisfactory situation. for these painters was similar to that observed in the
Whether an individual is exploring an opportunity composers. There was an initial period of non-
or resolving a difficulty, the important point is that creativity lasting about six years. This was followed
he or she is setting goals and initiating activities to by a rapid increase in productivity over the next six
accomplish those goals. years, then a period of stable productivity until
Having reviewed these alternative points of about 35 years into their careers, and lastly a period
view, I will now return to the analysis of creative of declining productivity.
acts in terms of familiar cognitive processes. I will Wishbow (1988) conducted a biographical
discuss various cognitive processes for which there study, similar to those just described, of 66 eminent
are either data or a plausible inference to suggest poets. She defined a notable poem as one that was
that they are especially important in creative acts. included in the Norton Anthology of Poetry. She
found that none of her 66 poets wrote a notable
Preparation poem earlier than five years into their careers, and
55 of the 66 poets produced none earlier than 10
There is very wide agreement among researchers years into their careers.
that preparation is one of the most important condi- The early periods of inactivity observed in all
tions of creativity (Mansfield & Busse, 1981; Wal- three of these studies suggests that a long period of
las, 1926; Ypma, 1968). By preparation, we refer preparation is essential for creative productivity,
to the effort of the creative person, often carried out even for the most talented of our composers, paint-
over long periods oftime, to acquire knowledge and ers, and poets.
skills relevant to the creative act. Recently, I In conducting this research, both Wishbow and I
(Hayes, 1985) provided strong evidence that even encountered considerable skepticism expressed by
the most talented composers and painters (e.g., experts in music, art, and literature that such inves-
Mozart and Van Gogh) required years of prepara- tigations could produce any consistent result. Their
tion before they began to produce the work for skepticism was based on the following very reason-
which they are famous. I surveyed all the com- able argument:
posers mentioned in The Lives of the Great Com-
posers (Schonberg, 1970) for whom there were suf- 1. These studies included individuals of very di-
ficient biographical data to determine when they verse aesthetic orientations (e. g., Richard
first became seriously interested in music (e.g., Wagner and Erik Satie) who were attempting
when they began piano lessons in earnest). Seven- to do very different things.
ty-six composers were included in the study. Next, 2. These studies included individuals from four
I identified the notable works of these composers different centuries (the seventeenth through
and the dates on which they were composed. (For the twentieth) who produced their works in
this study, I defined a notable work as one for which very different social contexts.
3. Therefore, there is no reason to expect that painting, Luncheon on the Grass, was historically
there would be consistency in the conditions significant because it was" a visual manifesto" of a
favoring creative performance across such di- new set of goals-goals that emphasized the impor-
verse times and groups. tance of visual effects on the canvas in contrast to
social or literary' 'meanings" a painting might con-
There is nothing logically wrong with this argu- vey. He said, "here begins an attitude that was later
ment. It might be that differences in social context summed up in the doctrine of Art for Art's Sake"
and aesthetic goals would dominate all other condi- (p. 607).
tions of creative productivity. As it turns out, they Of course, goal setting is not always difficult.
do not. Creators appear to require a long period of There are many situations in which the goals are
preparation despite differences in time and aesthetic obvious even though the means for achieving them
objectives. are not. Everyone knows that curing cancer and
What is this period of preparation used for? Si- reducing auto accidents are valuable goals. What
mon and Chase (1973) observed that chess players distinguishes the creative people in the examples
require about 10 years of preparations before they described above is that they recognized an oppor-
reach the level of grand master. They suggested tunity or a problem when other people did not.
that, during this time, the serious player learns a What might be responsible for differences in peo-
vast store of chess patterns through hundreds of pies' abilities to find problems or to recognize op-
hours devoted to study and play. They estimated portunities? Because we know very little about such
that a player needs to know roughly 50,000 chess processes, any account admittedly must be spec-
patterns in order to play at the grand-master level. ulative. Here are some hypotheses.
One can easily imagine that composers, painters, 1. Extensive knowledge of a field should give
and poets need a comparable period of time to ac- one increased ability to recognize both oppor-
quire sufficient knowledge and skills to perform in tunities and problems by analogy to previous expe-
their fields at world class levels. rience. For example, if a chess situation resembles
one the player has been in before, it could signal an
opportunity if the previous outcome was favorable,
Goal Setting and a problem if it was not.
Goal setting often appears to be the most critical 2. A unique pattern of knowledge outside of a
element in a creative act. According to Einstein and field, which is acquired perhaps through hobbies or
Infeld (1938), through switching professions, could provide a per-
son with analogies that are not generally available
Galileo formulated the problem of determining the velocity of to others in the field. Such analogies could suggest
light, but did not solve it. The formulation of a problem is often unsuspected possibilities or problems in the field.
more essential than its solution, which may be merely a matter of
mathematical or experimental skill. To raise new questions, new Consistent with this view, Gordon (1961) recom-
possibilities, to regard old problems from a new angle, requires mended that problem-solving teams in industry
creative imagination and marks real advance in science. (p. 92) should include people from very diverse fields.
3. Strong evaluation skills may lead a person to
Pavlov's discovery of the conditioned reflex is recognize problems in a line of research that others
another case in point. As part of a study of digestive fail to recognize and, as a result, to initiate new
processes, Pavlov was investigating the salivary re- studies that others would not have thought of. Eval-
flex in dogs. Dogs salivate automatically when food uation skills in the social sciences seem to depend
is placed in their mouths. At first, the experiment heavily on the sorts of critical-thinking skills de-
went well but, after a while, the dogs began to scribed in the rival hypotheses of Huck and Sandler
salivate before the food was placed in their mouths. (1979). Perhaps some aspects of creative perfor-
This development seriously complicated the study mance could be improved through training in these
Pavlov was trying to carry out. However, rather skills.
than seeing it as an annoyance to be eliminated, he These hypotheses could be viewed as examples
saw it as an interesting phenomenon to be investi- of the operation of Austin's (1978) Chance III and
gated. Against the advice of his colleagues, Pavlov Chance IV.
abandoned his original objective and set a new goal
which led to his historic work on the conditioned
reflex. Representation
Janson (1983) claimed that Edouard Manet's Because tasks that allow scope for creativity typ-
ically are ill defined, a person doing such a task is novices may lie in the way they define the task to be
forced to make many choices in building a represen- performed. Hayes, Flower, Schriver, Stratman,
tation of the task. For example, an architect may be and Carey (1987) found that novice writers repre-
given the task of designing a shop together with sented the task of revision as a sentence-level task;
specifications of the location, size, type of mer- that is, they attended to each sentence separately,
chandise to be displayed, and clientele. To repre- fixing the grammatical and lexical problems it con-
sent the design problem in sufficient detail so that it tained, and concerned themselves rarely or not at all
can be solved, the architect must make a great many with global problems, such as transitions, co-
decisions. For example, he may decide that the herence, and the effectiveness of the whole text.
shop should have a certain kind of access, should be The experts, in contrast, were primarily concerned
"transparent," and should have "levels" (see with global problems, although they fixed the local
Hayes, 1978, pp. 206-210). Ill-defined problems problems as well. The experts did a far better job of
offer a great deal of latitude in the way they can be revision than did the novices, and it seems clear in
represented or defined. this case that their better performance depended on
The way in which a person represents a task can their having defined a better task for themselves.
have a critical impact on how hard the task is to do One cannot really expect to do a good job of revi-
or even whether it can be done at all. Kotovsky, sion with a task definition that ignores a very impor-
Hayes, and Simon (1985) showed that a problem tant class of problems. Carey and Flower (Chapter
represented in one way may be 16 times as hard to 17 in this volume) provide an excellent discussion
solve as the same problem represented in a different of how expert-novice differences in task definition
way. The 16-to-l range almost certainly underesti- influence expository writing.
mates the full range over which changes in repre- Although these expert-novice studies cannot be
sentation can change problem difficulty. Thus, taken as proof, they do make it seem plausible that
choosing to represent a problem visually rather than creative people may differ from less creative people
verbally, or choosing to represent the problem by in some degree because they define better tasks for
one metaphor rather than another, could make a themselves.
sufficient difference in problem difficulty that one
scholar may be able to solve the problem and an-
other may not be able to. In some cases then, the Searching for Solutions
creative person-the one who solved the problem Many approaches to improving creative think-
when others could not-may be the person who ing, such as brainstorming (Osborn, 1948) and syn-
chose the best representation of the problem. etics (Gordon, 1961), focus on the fostering of
Kotovsky, Hayes, and Simon (1985) compared divergent thinking, that is, on generating many al-
different representations of the same problem. Even ternative solutions to the same problem. These
though the problem-solvers' representations of the techniques appear to be useful for some kinds of
problem were different in the sense that a problem group problem solving. However, as was noted
element might be represented as a position in one above, divergent-thinking skills appear to be unre-
case and as a size in another, the underlying prob- lated to the sort of creative productivity this chapter
lem was always the same. It is rare, though, for two is concerned with.
people, acting independently, to define an ill-de- It is interesting to contrast the emphasis in the
fined problem in the same way. If two architects creativity literature on the importance of generating
were commissioned to design the same house, they many solution paths with the emphasis in the cog-
would almost certainly interpret that commission in nitive science literature (see Newell & Simon,
different ways, placing different emphases on the 1972) on the importance of heuristic search, that is,
various design requirements. Each architect would with narrowing many solution paths down to a few.
define his or her own design task. It is tempting to Perhaps high-level creative activities are more like-
speculate that creative people define "better" or ly to demand heuristic searches than divergent
"more interesting" tasks for themselves than do thinking.
less creative people. In an early, but still influential, discussion of
Although there are no studies comparing task creativity, Wallas (1926) claimed that incubation is
definition in creative and noncreative people, there one of the characteristic stages of the creative pro-
are some task-definition studies comparing experts cess. By incubation, he meant a stage in which the
and novices. These studies show that a very impor- problem solver has stopped attending to the prob-
tant part of the difference between experts and lem but during which progress, in any case, is being
made toward the solution. Researchers have at- talked of "the endless discipline of writing and
tempted to demonstrate the reality of this phe- rewriting and rewriting" (Murray, 1978, p. 85).
nomenon experimentally. Typically, experimental Revision, of course, is not confined to writing. It
and control subjects are given a complex problem to happens in the development of scientific theory, in
solve. The control subjects are allowed to work painting, and in musical composition. For example,
continuously on the problem until they solve it. The in a letter, Tchaikovsky said,
experimental subjects are interrupted in their solu-
tion efforts and are asked to attend to another task Yesterday, when I wrote you about my method of composing, I
did not enter sufficiently into that phase of the work which relat-
for a period of time before they are allowed to return
ed to the working out of the sketch. This phase is of primary
to the problem and solve it. If the experimental importance. What has been set down in a moment of ardour must
subjects required less total time working on the now be critically examined, improved, extended, or condensed.
problem to solve it than the control subjects, this (quoted in Vernon, 1970, p. 59)
would be taken as evidence of incubation. Although
a number of early investigators failed to obtain If revision is an important part of creative activity, it
positive results with this method, Cook (1934, is reasonable to expect that creative people may be
1937), and Ericksen (1942), more recent experi- better at revision than others. Although evidence on
menters, have obtained positive results (Fulgosi & this issue is scant at best, the question is interesting
Guilford, 1968; Murray & Denny, 1969; Silviera, enough to pursue. There are at least three possible
1971). factors which might make creative people superior
The success of such experiments, however, can- revisors:
not be taken as definite proof that incubation has 1. Creative people may have higher standards
occurred. As Ericsson and Simon (1984) pointed for performance than others. Although this is a
out, the problem is that it is very difficult to estab- very plausible assertion, its validity has been tested
lish that the experimental subjects obeyed (or, in- only in the area of standards for creativity. As was
deed, could obey) instructions not to attend to the noted above, creative people aspire more than oth-
problem during the incubation period. ers to be creative. The impact this might have on
Even if incubation is real, it does not follow that performance is illustrated in a study carried out by
it is characteristic of the creative process. In my Magone (personal communication). Magone col-
reanalysis (Hayes, 1978) of the data on which Wal- lected think-aloud protocols of people who were
las based his conclusions (the testimony of creative taking a creativity test in which they were asked to
individuals), I found many instances in which cre- complete a drawing in as many different ways as
ative acts proceeded from beginning to end without they could. She found that people who scored high
any pause that would allow for incubation. Al- on the test were much more likely than those who
though Wallas's claims for incubation are interest- scored low to reject ideas as "trite" or "boring."
ing, it appears that there is little empirical evidence Even though a creativity test probably does not pre-
to support them. dict real creativity, the study does illustrate the
point that high standards for creativity can shape
performance.
Revision
2. Creative people may be more sensitive than
In performing skilled activities, people often stop others in perceiving that standards have not been
to evaluate what they have produced and to improve met. There are no studies comparing creative peo-
on any shortcomings they may find. This revision ple with others in this skill. However, Hayes et al.
process appears to be especially important in cre- (1987) have found that expert writers were far more
ative activities because of the very high standards sensitive detectors of text problems than novices.
involved. Donald M. Murray, a Pulitzer prize-win- 3. Experts may be more flexible than others in
ning essayist, spoke eloquently about the impor- considering change. Results of the personality sur-
tance of revision: "Rewriting is the difference be- veys, which were cited above, suggest that creative
tween the dilettante and the artist, the amateur and people are, in fact, more flexible than others. Flexi-
the professional, the unpublished and the pub- bility could increase a person's chances of per-
lished." William Glass testified, "I work not by forming creatively in a number of ways: A more
writing but rewriting." Dylan Thomas stated, "Al- flexible person might be more likely than others to
most any poem is fifty to a hundred revisions-and drop everything to pursue a hot new lead as Pavlov
that's after it's well along." Archibald MacLeish did in the example presented earlier. A flexible per-
son might be more likely than others to sacrifice less 4. Taking effective action to revise short-
important goals in order to accomplish more impor- comings
tant ones. And, lastly, a more flexible person might
be more likely than others to change problem repre- Clearly, both cognitive and motivational factors
sentation if progress toward solution is are involved in creative performance. However, the
unsatisfactory . failure of cognitive ability measures, such as IQ, to
predict creative performance leads me to propose
that creative performance has its origin not in innate
cognitive abilities but rather in the motivation of the
Summary creative person. Over a period of time, this moti-
vation has cognitive consequences, such as the ac-
In this chapter, I explored two major questions. quisition oflarge bodies of knowledge, that contrib-
In answer to the question, "What are creative peo- ute in a critical way to creative performance; but the
ple like?" fairly good empirical evidence supports origin is in motivation, not cognition.
the following conclusions: The creative person may be thought of as a vec-
tor special in strength and in direction. Motiva-
1. Creative people work very hard. tion of great strength is necessary because creative
2. Creative people are more disposed to setting people face daunting tasks. They must work for
their own agenda and to taking independent many years, perhaps for a decade or more, before
action than are others. they can begin to accomplish their creative goals.
3. Creative people strive for originality. They may have to reject easily available rewards in
4. Creative people show more flexibility than order to pursue their fields. One of my students
others. said, "I must like art a lot to be willing to go to
5. Creative people do not have higher IQs or get school for four years in order to be out of work. "
better school grades than others when we con- Sometimes they may have to face active opposition
trol for age and education. In fact, no cog- as Pavlov did.
nitive abilities have been identified that relia- The direction of motivation is as critical as its
bly distinguish between creative and noncrea- strength. Success in many areas of life requires
tive people. strong motivation and hard work. In many practical
situations, the hard work must be directed to satis-
The surprising thing about these findings is that fying the demands of a boss, some set of standards,
all the variables that discriminate between creative or the interests of the public. Creative people, how-
and noncreative people are motivational. No cog- ever, are motivated to be in charge of their own
nitive abilities have been discovered that discrimi- actions, and, through those actions, to do some-
nate between these two groups. thing that has not been done before, perhaps has not
In exploring the question, "What cognitive fac- even been thought of before.
tors are involved in creative acts?" I have un- The nature of their motivation may lead creative
covered convincing evidence on two points: people to take different paths than others take. For
example, creative people may choose fields, such
1. Years of preparation are essential for creative as the arts or sciences, in which they believe they
productivity in many fields. can exercise their interest in creative activities,
2. Goal setting is the critical element in many rather than sales or medical practice, in which cre-
creative acts. ative activities may not be appreciated.
Motivational differences can result in important
In addition, plausible arguments can be made for differences in cognitive factors. If a person is will-
the importance of the following procedures in cre- ing to work longer and harder than others, he or she
ative acts: can acquire a larger body of information than oth-
ers. In solving a problem, this extra information
1. Choosing good problem representations might be used directly to make an essential in-
2. Defining good problems in ill-defined prob- ference or might provide an analogy that would
lem situations suggest a solution path. Willingness to work hard
3. Accurately evaluating the shortcomings of could also lead persons to define harder and better
one's own work problems for themselves and in general to set higher
standards for themselves. Higher standards could Harris, S. E. (1972). A statistical portrait of higher education.
then lead them to be more critical of shortcomings New York: McGraw-Hill.
Hayes, J. R. (1978). Cognitive psychology: Thinking and creat-
in their own work. ing. Homewood, IL: Dorsey Press.
The motivation to be independent would pre- Hayes, J. R. (1985). Three problems in teaching problem solv-
dispose persons to set their own goals and moti- ing skills. In S. Chipman, J. W. Segal, & R. Glaser (Eds.),
vation to be creative would lead them to reject goals Thinking and learning skills (Vol. 2, pp. 391-406). Hillsdale,
that were "trite" or "boring." New Jersey: Erlbaum.
Hayes, J. R., Flower, L. S., Schriver, K. A., Stratman, J., &
Finally, motivation to be flexible could make it Carey, L. (1987). In S. Rosenberg (Ed.), Cognitive processes
easier to change direction completely when a new in revision. Cambridge, England: Cambridge University
opportunity presents itself, to sacrifice minor objec- Press.
tives to accomplish major ones, and to change rep- Helson, R., & Crutchfield, R. S. (1970). Mathematicians: The
creative researcher and the average Ph.D. Journal ofConsult-
resentations when progress is unsatisfactory. The
ing and Clinical Psychology, 34, 250-257.
primary thrust of the position I am presenting here is Huck, S. W., & Sandler, H. M. (1979). Rival hypotheses: Alter-
that differences in creativity have their origin in nate interpretations of data based conclusions. New York:
differences in motivation. These differences in mo- Harper & Row.
tivation then cause cognition differences, and these Janson, H. W. (1977). History of Art. (2nd ed.). Englewood
Cliffs, N.J.: Prentice-Hall, and New York: Abrams.
motivational and cognitive differences account Kotovsky, K., Hayes, J. R., & Simon, H. A. (1985). Why are
jointly for the observed differences between cre- some problems hard? Evidence from tower of Hanoi.
ative and noncreative individuals. Cognitive Psychology, 17, 248-294.
Langley, P., Simon, H. A., Bradshaw, G. L., &Zytkow, J. M.
(1987). Scientific Discovery. Cambridge, Massachusetts:
MIT Press.
Lenat, D. (1976, July). AM: An artificial intelligence approach
References to discovery in mathematics as heuristic search, SAIL
AIM-286. Stanford CA: Stanford University, Artificial Intel-
Austin, J. H. (1978). Chase, chance, and creativity. New York: ligence Laboratory.
Columbia University Press. MacKinnon, D. W. (1961). Creativity in architects. In The Cre-
Barron, F. (1963). The disposition toward originality. In C. W. ative Person. Proceedings of a conference presented at the
Taylor & F. Barron (Eds.), Scientific creativity: Its recogni- University of California Alumni Center, Lake Tahoe, Ca.
tion and development (pp. 139-152) New York: Wiley. MacKinnon, D. W. (1963). Creativity and images of the self. In
Bergum, B. O. (1975). Self-perceptions of creativity among R. W. White (Ed.), The study of Lives. New York: Atherton
academic inventors and non-inventors. Perceptual and Motor Press.
Skills, 40, 78. MacKinnon, D. W. (1968). Selecting students with creative
Bloom, B. S. (1963). Report on creativity research by the exam- potential. In P. Heist (Ed.), The creative college student: An
iner's office at the University of Chicago. In C. W. Taylor & unmet challenge (pp. 101-118). San Francisco: Jossey-Bass.
F. Barron (Eds.), Scientific creativity: Its recognition and Mansfield, R. S., & Busse, T. V. (1981). The psychology of
development (pp. 251-264) New York: Wiley. creativity and discovery. Chicago: Nelson-Hall.
Chambers, J. A. (1964). Relating personality and biographical Murray, D. M. (1978). Internal revision: A process of discovery.
factors to scientific creativity. Psychological Monographs, 78 In C. R. Cooper & L. Odell (Eds.), Research on Composing:
(7, Whole No. 584). Points of Departure (pp. 85-104). Urbana, IL: National
Cook, T. W. (1934). Massed and distributed practice in puzzle Council of Teachers of English.
solving. Psychological Review, 41, 330-335. Murray, H. G., & Denny, 1. P. (1969). Interaction of ability
Cook, T. W. (1937). Distribution of practice and size of maze level and interpolated activity (opportunity for incubation) in
pattern. British Journal of Psychology, 27, 303-312. human problem solving. Psychological Reports, 24, 271-
Einstein, A., & Infeld, L. (1938). The evolution ofphysics. New 276.
York: Simon & Schuster. Newell, A., & Simon, H. A. (1972). Human problem solving.
Ericksen, S. C. (1942). Variability of attack in massed and Englewood Cliffs, NJ: Prentice-Hall.
spaced practice. Journal of Experimental Psychology, 31, Newell, A., Shaw, J. c., & Simon, H. A. (1964). The process of
339-345. creative thinking. In H. Gruber, G. Terrell, & M. Wertheimer
Ericsson, K. A., & Simon, H. A. (1984). Protocol analysis: (Eds.), Contemporary approaches to creative thinking. New
Verbal reports as data. Cambridge: MIT Press. York: Atherton Press.
Fulgosi, A., & Guilford, J. P. (1968). Short-term incubation in Osborn, A. (1948). Your creative power. New York: Scribner.
divergent production. American Journal of Psychology, 7, Popper, K. R. (1959). Thelogic of scientific discovery. London:
1016-1023. Hutchinson.
Gordon, W. J. (1961). Synectics. New York: Harper & Row. Roe, A. (1951). A psychological study of physical scientists.
Guilford, J. P. (1967). The nature of human intelligence. New Genetic Psychology Monographs, 43, 121-235.
York: McGraw-Hill. Roe, A. (1953). The making of a scientist. New York: Dodd
Harmon, L. R. (1963). The development of a criterion of scien- Mead.
tific competence. In C. W. Taylor & F. Barron (Eds.), Rouse, W. B. (1986). A note on the nature of creativity in
Scientific creativity: Its recognition and development (p. 44- engineering: Implications for supporting system design.
52). New York: Wiley. Information Processing and Management, 22, 279-285.
Schonberg, H, C. (1970). The lives of the great composers. New Vernon, P. E. (1970). Creativity. Harmondsworth, England:
York: W. W. Norton. Penguin Books.
Silviera, J. A. (1971). Incubation: The effects of interruption Wallas, G. (1926). The art of thought. New York: Harcourt,
timing and length on problem solution and quality of problem Brace.
processing (Doctoral dissertation, University of Oregon, Wishbow, N. (1988). Creativity in poets. Unpublished
1972). Dissertation Abstracts International, 32, 5500B. doctoral dissertation, Carnegie-Mellon University, Pitts-
Simon, H. A. (1966). Scientific discovery and the psychology of burgh, PA.
problem solving. In R. G. Colodny (Ed.), Mind and cosmos: Ypma, E. G. (1968). Predictions of the industrial creativity of
Essays in contemporary science and philosophy (Vol. III). research scientists from biographical information (Doctoral
Pittsburgh: University of Pittsburgh Press. dissertation, Purdue University, 1970). Dissertation Ab-
Simon, H. A., & Chase, W. (1973). Skill in chess. American stracts International, 30, 5731B-5732B.
Scientist, 61, 394-403.
CHAPTER 8
Creativity and Perception

John H. Flowers and Calvin P. Garbin
Informal thought about the nature of mental opera- occasionally can be found in indexes of mono-
tions important to creative human behavior sug- graphs specifically dealing with the topic of
gests that perceptual processes are of considerable creativity, most of these references refer to research
importance. The ability to "see relationships related to specific theories about individual dif-
among elements" is an attribution commonly made ferences in perceptual styles or processing modes,
toward authors of major scientific discoveries or of as opposed to broader contemporary issues of per-
noteworthy artistic achievements. For example, ceptual processing. Clearly, most researchers in the
Shepard (1978, 1981) documented self-reports field of perception have not touched upon the topic
from several creative scientists and authors that of creativity, and relatively few researchers in
strongly emphasize the role of visual imagery and creativity have chosen to integrate their work with
the manipulation of visual codes in the creative perceptual issues.
process.
Given the anecdotal and self-report evidence for
a relationship between creative behavior and as-
pects of perceptual processing, it initially may seem Why Have Perceptual Psychologists
surprising that there is a notable void in either re- Had Little to Say about Creativity?
search or theoretical articles specifically focused on
these issues. In preparing this chapter, for example, One reason that may have inhibited psychol-
we noted that, during the last six volumes of the ogists who were studying aspects of perception
Journal of Creative Behavior, there was only one from becoming involved in research on creativity is
title that included the word perception, and that the fact that the term itself is not viewed as a scien-
paper (Goodman & Marquart, 1978) was limited to tifically "tight" concept of the variety preferred by
a one-page abstract. In addition, we noted that relatively "operationalistic" behavioral scientists.
among seven current textbooks in perception that As has been noted in earlier research (e.g., Stein,
presently reside on our bookshelves, none contain 1956; Taylor, 1960), as well as the contributors to
the term creativity in their indexes, nor is the term this volume, providing an easily agreed upon opera-
creative ability addressed at any point in the texts. tional definition of creativity that can be related to
Although references to the term perception specific aspects of observable behavior or specific
information-processing operations is problematic.
John H. Flowers and Calvin P. Garbin • Department of Although there is general agreement among re-
Psychology, University of Nebraska, Lincoln, NE 68588. searchers in human thinking that a key component
147
of creativity is the process of generating novel men- zational processes (see, e.g., Hochberg, 1981, for a
tal representations, assessment of creativity by so- review), most theories approach perception as an
ciety is generally done on the basis of product; that information reduction process whereby noisy, vari-
is, does a particular solution, invention, discovery, able, and impoverished patterns of environmental
or artistic contribution meet the joint criteria of nov- energy become resolved into stable and consistent
elty and worthwhileness? Obviously, a wide variety internal representations optimal for human perfor-
of mental operations and processes could contribute mance. Terms such as perceptual organization and
to the characteristics of a product that elicit those perceptual constancy reflect emphasis on the infor-
subjective evaluations. mation reduction processes. The latter term, in par-
The lack of existing literature explicitly relating ticular, emphasizes the role of perception in provid-
issues in perception to the study of human creativity ing an individual with similar or identical mental
may also be attributable, in part, to fundamental representations of events that may have widely
differences in the "level of approach" among be- varying physical representations in the environ-
havioral scientists who study fundamental charac- ment.
teristics of mental and behavioral processes com- If a major function of "efficient" perceptual pro-
mon across individuals, as opposed to those having cessing is to provide perceptual constancy, as well
primary interest in differences among individuals. as to encourage different observers to obtain similar
Both historically and at present, the majority of or identical representations from common environ-
research efforts directed at understanding percep- mental stimuli, then it does indeed seem that this
tual processes has been directed at theories, mod- organizational aspect of perception works against
els, and descriptions of behavior that apply to per- the generation of novel representations. One might
ception in general, as opposed to individuals. The therefore predict that some measures of perceptual
very term creativity, on the other hand, denotes an performance that tap perceptual organizational pro-
attribute that individuals presumably possess (or at cesses would be negatively related to measures of
least exhibit) in different amounts, hence, its study creative ability. A potential consequence of such a
generally assumes an individual differences relationship would be a· positive association be-
approach. tween perceptual deficiencies or handicaps and cre-
In our view, however, neither the fuzziness at- ative ability. Case study instances supportive of a
tributed to the definition of creativity nor its degree relationship between perceptual handicaps and cre-
of association with ideographic rather than ative behavior can, in fact, be found in discussions
nomothetic approaches is as responsible for the lack of artistic and musical achievement. For instance,
of an existing body of literature relating creativity some of the innovative stylistic changes in Beetho-
and perception as is the fact that perception tradi- ven's later works are commonly attributed to the
tionally has been studied as a process of organizing increasing severity of his hearing impairment.
information within the nervous system, whereas However, there are many reasons to believe that
studies of mental processes associated with creative some aspects of perceptual performance should be
behavior usually imply the generation of novel rep- positively correlated with creative ability-or at
resentations of information within the nervous sys- least with the output of creative products. Many
tem. At first inspection, these two classes of mental products of creative activity are not simple spon-
activity seem to have little to do with each other- taneous generations, but result from effortful pro-
or worse yet, they may even be viewed as incom- duction, interspersed with the evaluation of feed-
patible operations. back. Thus, handicapping the senses at a peripheral
level (e.g., blindness or deafness) would certainly
disrupt the ability to evaluate visual and auditory
productions as they are being created. Despite what
has been said about Beethoven's deafness, it seems
Is Perceptual Organization absurd to predict that there would be a general nega-
Incompatible with Creative Thought? tive relationship between basic measures of audito-
ry sensitivity and frequency discrimination ability
Although individual theories of perception may and creative output of music among composers, or a
differ substantially in their emphasis upon the im- negative relationship between visual acuity mea-
portance of stimulus structure versus mental organi- sures and creative output of painters.
CHAPTER 8 • CREATIVITY AND PERCEPTION 149
The Importance of Executively B
Controlled Processing
v
In our view, however, there are other aspects of
perception for which performance measures ought
to relate positively to creativity-in particular, pro-
cesses under the conscious control of the perceiver.
These include the ability to control various aspects
of selective attention, to control figural organiza-
tion when ambiguous sensory data are presented, to
perform manipulations of internal visual and/ or au-
ditory representations of perceptual information,
and the ability to equate perceptual experiences ob-
tained from different sense modalities (cross-modal
abilities). Although the earlier stages of perceptual
----rV
M!III!IlI!It
C
D
processing leading to perceptual organization (e.g.,
'/
sensory transduction, feature extraction, figural
synthesis) typically are viewed as involuntary pro-
cesses not under executive control, it is clear that
perceptual organization and pattern recognition are Figure 1. The Necker Cube-a familiar example of how con-
influenced jointly by both involuntary or automatic trolled visual attention can change perceptual organization. In
and executively controlled processing. Example A, where the cube is essentially "transparent," two
At the level of perceptual organization, consider distinct organizations of depth can be obtained with moderate
the examples of the Necker cube and of the revers- ease. In Example B, the shading and obscuring of the internal
contours provide additional context making one organization
ible or ambiguous figures commonly found in most dominant, although the second organization can still be obtained
introductory psychology textbooks as well as in with effort. In Example C, shading of other parts of the figure
textbooks on perception (some examples are given makes the opposite depth organization dominant. In Example D,
in Figure 1). In each of these examples, there are where the internal contours are totally removed, the alternative
depth organization cannot be obtained at all (by typical subjects),
contextual cues that cause a particular organization even with considerable mental effort. Note also that, after view-
of the object to be perceived upon initial inspection. ing cube C, it is easier to view cube A (and even cube B) in the
In most versions of these objects, observers have an same organization as is dominant for cube C.
ability to reverse the organization through volun-
tary changes in visual attention, although the time
and effort required for such reversals, as well as Thus, we see two distinct categories of percep-
which of the possible organizations is most likely to tual processes for which individual variation among
be initially seen, can be greatly influenced by modi- different observers might well relate to creativity.
fying the contextual cues supporting either of the The first category contains the largely involuntary
organizations. With some modifications, reversal processes of perceptual organization. These are in-
becomes virtually impossible, and thus only one formation-reducing processes that promote stability
interpretation is obtained. Executively controlled and organization of percepts, and thus normally
processes can, within limits, significantly modify tend to act against the formation of novel represen-
figural synthesis and organization, allowing a sin- tations of information. The second category of per-
gle physical stimulus to have multiple perceptual ceptual processes, those under executive control of
representations. Such processes may thus playa the observer, serve to modify and control the ac-
role in the generation of novel mental representations of the involuntary processes and also may
tions of information, which, in tum, could form the serve to encourage the generation of novel repre-
basis for creative products. Indeed, it is this kind of sentations of information. Extensive data exist on
processing that appears to contribute to high-perfor- how each of these classes of processes affects per-
mance levels on tests, such as Guilford's (1967) formance as a function of stimulus and task param-
divergent production test, that are specifically eters. Unfortunately, the data base on individual
aimed at measuring individual differences related to differences among either of these categories of pro-
creativity. cesses is limited. Thus, it is difficult to relate di-
rectly specific empirical findings from existing per- of creative activity to the same perceptual mecha-
ceptual studies to either the prediction of creative nisms that are normally coupled to the analysis of
behavior in individuals or to modification of the sensory input, and that are responsible for many
environment to encourage creativity. We feel this aspects of perceptual organization. According to
indicates a need for additional basic research in Shepard, these highly evolved and specialized per-
individual differences per se, as well as research ceptual mechanisms have the ability to operate
that directly relates these individual differences to upon data other than that obtained from normal sen-
creative ability. sory input, when decoupled from the sensory en-
vironment, as in such altered consciousness states
as dreams. It is conceivable that the ability to use
such spontaneously generated mental constructions
A Configura! Conception of Creativity in creative thought could be relatively independent
of an individual's power of executively controlled
The particular topics (and perceptual tasks) that mental operations, but positively correlated with the
we will be discussing are selected largely because strength of perceptual organizational processes.
they focus on the joint role of involuntarily and This three-factor view of creativity, directly
executively controlled processes in perception. We based upon perceptual mechanisms, suggests there-
are guided by a broad conception of creativity that fore, a highly configural, nonlinear relationship be-
assumes that there are essentially three factors that tween creative ability in the population and specific
influence an individual's creative potential. One in- individual differences in mental processes. By
volves the relative "looseness" of involuntary or- looking-from an individual differences perspec-
ganizational processes. An individual for whom the tive-at various perceptual tasks that seem likely to
involuntary processes operate somewhat less deter- tap differentially involuntary organizational pro-
ministically (but perhaps less efficiently) is more cesses and executively controlled processes, re-
likely to represent environmental data mentally (as searchers may be in a better position to form a more
well as data recalled from memory) in novel ways. detailed model relating specific representational
The potential relationship between schizophrenia and transformational processes to creative behav-
and creativity (Keefe & Magaro, 1980; Prentky, ior. Additionally, such research might guide devel-
1979) would seem to illustrate this factor. opment of programs or products for the enhance-
The second factor involves the power of execu- ment of creative ability.
tively controlled processes, such as spatial selective
attention, manipulation of mental images, and con-
trolled cross-modal representation. Presumably, an
individual having superior executive control of Impairment of Sensory Processes
these processes is able to produce novel representa-
tions of information through effortful construction One of the most highly noticeable individual dif-
and modification of mental representations. This ference variables in perception is that of impairment
concept of "creativity through controlled mental of one or more of the primary senses-at least inso-
effort" is very different from the concept of far as distinguishing normal individuals (including
creativity attributable to "loose organizational pro- prosthetically correctable persons) from those who
cesses." For example, schizophrenics and indi- have apparent handicaps. Case study analyses of
viduals with schizotypic patterns of cognitive ac- creative output from visually or auditorily handi-
tivity are notably weak on performance measures capped individuals is a topic that deserves intensive
that presumably tap executively controlled mental study on its own; our limited mention of the topic
processes. here is merely to recognize that the broad categories
The third factor, we feel, is of importance, par- of differences in perceptual and cognitive processes
ticularly to creative thought that involves sudden that characterize adaptation to severe defects in ei-
insight, that involves processes not under executive ther hearing or vision obviously can affect the cre-
control nor driven by sensory data, but that pro- ative process, albeit in complex ways. It is a well-
duces seemingly spontaneous mental representa- known principle of developmental psychobiology
tions, often involving visual imagery. Based that prolonged deprivation of sensory experience,
largely upon self-reports of creative scientists and from birth or early in life, can result in permanent
authors, Shepard (1978, 1981) attributed this aspect changes in neural structures that often prevent full
recovery of function, even if the handicap is later sensory stimulation upon creative thought pro-
repaired. Furthermore, neural structures may adapt cesses. In particular, the fact that common re-
functions qualitatively different from those as- sources are used in executive control of mental rep-
sumed by corresponding structures in sensorily in- resentations and in the processing of corresponding
tact individuals. In human behavior, the conse- forms of sensory data implies that perception within
quences of sensory handicaps may affect aspects of a particular modality may interfere with thought
memory and cognition that depend upon the use and processes sharing a common form of representa-
manipulation of auditory or visual "codes," in ad- tion. For instance, the generation and the manipula-
dition to the ability to gain information from the tion of visual images are inhibited when a task re-
environment. Recognition of this fact has led requires processing visual input, and manipulations
searchers and educators to develop specialized edu- of linguistic representations may be inhibited by
cational programs for hearing-impaired individuals tasks requiring the processing of speech. The inter-
that are specifically adapted to differences in modes ference between perceptually generated and inter-
of information processing among such individuals. nally generated codes that share a common
In addition to producing performance differences modality easily can be demonstrated by using a task
per se, differences in memory codes, attention, and developed by Brooks (1968). In this task, subjects
other control processes attributable to early sensory are required to imagine a moving dot traveling
handicaps obviously have the potential for produc- around a mental image of a block drawn letter, such
ing products that are novel when compared with as a capital F (see Figure 2), and to respond orally
norms of the general population of nonhandicapped yes or no based upon whether each comer con-
individuals. stitutes an external angle (requiring a "right tum"
We see two major problems, however, in extrap- on the part of the dot) or an internal angle (requiring
olating general conclusions about creativity from a "leftturn"). Time to complete the circuit with the
the study of the perceptually handicapped. The first oral response mode is generally much less than that
problem is the extreme degree of heterogeneity in required for the same task when subjects must
types and degree of impairment among such indi- check off Ys versus Ns in a visually presented an-
viduals; not to mention potential differences in swer sheet. However, when subjects are presented
compensatory processes among individuals who with a linguistic search task, requiring them to indi-
share relatively similar physical impairments and cate whether successive words in memorized sen-
developmental histories. The second problem is tences are nouns versus other parts of speech, the
that the ease of retrieving examples of perceptually visually guided check-response mode produces
handicapped creative individuals from memory
makes the topic ripe for attributions of correlations
between the handicap and the creativity that may
not be warranted (e.g., Kahneman & Tversky,
1973). One may be able to learn a great deal about
the creative processes of a given handicapped indi-
vidual through careful case study, protocol analy-
ses, and other techniques, and such research poten-
tially could be of tremendous value to the
development of special therapeutic and educational
approaches. As a simple individual difference vari-
able to be applied for understanding the charac-
teristics of creativity in the general population, the
handicapped-nonhandicapped dichotomy does Figure 2. An illustration of the sort of "imaginary stimuli" used
not, on the other hand, appear to be very useful. by Brooks (1968) to study the effect of verbal or visually guided
response processes on the ability to search visually coded infor-
mation in active memory. Subjects were required to mentally
construct an image of a letter, such as this capital F, and to
Environmental Constraints on Sensory indicate whether each comer passed by the imaginary asterisk
Input consisted of an external or an internal angle. Subjects could
perform better on this task if an oral yes-no response was used
than if a visually guided manual response system was employed,
It may be more useful to consider the effects, indicating that manipulation of visual codes shares resources
within unimpaired individuals, of different levels of with visual perception.
considerably faster searches than does the oral yes- edge base that is generated from such a research
no mode. Modality specific interference between effort might prove very useful in the development
imagery and perceptual detection tasks also has of both text-related prcducts and picture-related
been demonstrated by Segall and Fusella (1970), software aids used in computer-assisted design.
who asked subjects to construct mental images, More generally, however, such knowledge might
~uch as the appearance of a tree (visual) or the ring- prove useful in the development of techniques for
mg of a telephone (auditory), and showed selective fostering increased output of creative activity in a
deficits in the ability to detect weak auditory or much wider range of settings, through the teaching
visual signals. of specific strategies for minimizing perceptual in-
Because documented self-reports of mental terference with image-based mental operations and
events associated with creative thought often in- other forms of mental code transformations.
clude extensive use of mental imagery (Shepard,
197~, 1981), it seems plausible that creating an
envIronment that minimizes potentially interfering
sen~ory input might be useful in facilitating manip- Altered and Transitional States-
ulatIOns of mental image processing and, hence,
Decoupling Perceptual Mechanisms
contribute to creative thought. Additional research
in how regulation of stimuli (visual and linguistic from Sensory Input
feedback from a display screen on a word pro-
cessor) can affect the speed and quality of creative Our discussion above has concerned the overlap
output is one area of applied perceptual research in resources between mechanisms normally tied to
that has current value for product development. An the involuntary processes of analyzing and organiz-
increasing number of products are being released ing sensory stimulation and those of executively
into the personal and professional software markets controlled construction and manipulation of visual
that are designed to aid the initial stages of manu- and auditory images. A related issue concerns the
script planning and organization. These so-called potential role of the perceptual mechanisms nor-
idea processors are aimed specifically at the facili- mally driven by sensory input during unaltered
tation of creative output for a variety of applications states of observer consciousness, when those mech-
(Kellogg, 1986). With the increasing use of small anisms are decoupled from sensory input. Dream
computers in a variety of settings, one might expect states, and perhaps some drug-induced states, rep-
the developments of similar products to continue. resent examples of such a decoupling. Shepard
One attribute of at least some of the existing hard- (1978, 1981) argued that transitions from such
ware is that users are prevented from viewing pre- states represent a fertile ground for the development
viously entered text-a design feature that seems to of creative ideas, because the perceptual mecha-
assume that the availability of visual feedback, or at nisms automatically linked to organizing the senso-
least the existence of printed text that stimulates the ry world (which are normally transparent to our
visual system of the user, is detrimental to conscious experience) run "on their own," occa-
performance. sionally constructing novel and useful percepts and
As Kellogg (1986) pointed out, evaluative re- images from fragments of internal neural noise and
search on these products and on prototypes for fu- loosely guided consultations with memory. Ac-
ture related software aids is woefully lacking. The cording to Shepard (1981), contact with the lin-
same can be said for research that is not tied to a guistic system allows the abstract images and rela-
specific product but that provides general informa- tionships to be translated into communicable form.
tion about the production as a function of the pres- Shepard clearly viewed the mechanisms of percep-
ence of perceptual stimuli of, for example, text, tual organization that involve spatial relationships
figural design, and metaphorical or analogical rela- in particular as a powerful source of general knowl-
tionships. Based upon the implications of such edge about relationships that can be analogically
studies as those of Brooks (1968), it may be the case applied to invention and problem solving:
that the presence of auditory or visual stimuli, while
an individual is attempting to "be creative," can
The creative productions of a brain presumably stem from what-
have certain interfering stimulus-related costs for ever intuitive wisdom, whatever deep organizing principles have
mental representation and transformation, and, ?een built into that brain as a result of the immense evolutionary
hence, inhibit creative performance. The knowl- Journey that has issued in the fonnation of that brain. If the
arguments sketched out in this chapter have any merit, the most transitional states actually constitutes a major
basic and powerful innate intuitions and principles underlying source of creative productions. Mental construc-
verbal and nonverbal thought, alike, may well be those govern- tions occurring during an altered or transitional
ing the relations, projections, symmetries, and transformations
of objects in space. (Shepard, 1981, p. 339) state can be useful only insofar as they are remem-
bered, and insofar as they can be evaluated for ap-
plication and worthwhileness. It may be that the
Thus, Shepard described a very direct rela- degree to which decoupled automatic perceptual
tionship between perceptual processes and creative processes contribute to creative output has far more
thought, making the claim that (1) implicit knowl- to do with facility in higher level cognitive pro-
edge of visual relationships among objects, and cesses, such as memory storage, retrieval, search,
rules for transforming those relationships, may con- and comparison, than in individual differences in
stitute the fundamental mental operations inherent perceptual organization processes per se.
in much of creative thought, and (2) that decoupling Consideration of the role of stimulus-decoupled
certain normally involuntary processes from their perceptual organization processes as a fundamental
data source may allow those same mechanisms to source of creative thought, as Shepard proposed,
operate as a primary generator of creative thought. leads to an interesting conjecture about the link be-
It seems apparent from Shepard's descriptions of tween efficiency of perceptual organization and
self-report data that this form of creative thought is creativity-particularly with respect to how gener-
not of the controlled variety (as discussed in the al cognitive defects, such as those associated with
previous section) but is the product of involuntary schizophrenia, might relate to creative thought.
mental operations that lead to spontaneous insight. One common view of why individuals who possess
Thus, the effects of decoupling of normal sensory some of the traits of schizophrenic thought might be
input during alternative states of consciousness viewed as creative is that deficiencies in the normal
should be viewed as distinct from restricting senso- involuntary perceptual organization processes lead
ry input in a normal-waking-state individual in to an increased likelihood of an atypical representa-
order to prevent interference with controlled manip- tion of a perceptual event (see, e.g., Keefe &
ulation of perceptual codes. Magaro, 1980). In other words, it is the anomalous
In popular religious and scientific circles, vast organization of sensory input, coupled with suffi-
amounts have been written about alternative states ciently good higher order processes to evaluate the
of consciousness that involve such decoupling of potential worthwhileness of a mental construction
sensory input, although, in our view, it is some- (or, alternatively, the opportunity for a peer to
times difficult to determine into which of those cir- notice the worthwhileness of a product even if the
cles a given piece of research/literature should be cognitively impaired individual cannot perform the
categorized. It is also apparent that, throughout evaluation), that lead to creative output. However,
much of the history of civilization, some human creativity that is attributable to looseness in percep-
beings knowingly have exploited alternative states tual organization in the presence of stimuli is very
(including dreams) as a deliberate strategy for fos- different from creativity that is attributable to per-
tering creative behavior-not to mention the use of ceptual organization processes decoupled from nor-
alternative consciousness states as a causal mal sensory inputs. An increased frequency oftran-
construct for particular creative acts. Like research sitions from hypnopompic or hallucinatory states
relating either handicaps or psychopathology to (as might be reasonably expected to occur in asso-
creativity, selection biases and the availability ciation with certain psychotic disorders), combined
heuristic can play havoc with attempting to relate with unimpaired (or even superior) mechanisms of
existing literature on alternate states and creativity perceptual organization, thus represents a potential
either to individual differences in creative potential alternative route for contributing to creative thought
or to research into conditions that foster creativity. by some individuals who possess dispositions to-
It may, however, be profitable to study the extent ward cognitive disorders. Moreover, the relative
to which individual differences in dream recall weakness or looseness in organizational processes
might relate to instances of insightful creative and the ability to exploit involuntary organizational
thought, as well as to individual differences in at- processes decoupled from sensory input are sug-
tention and perceptual organization. Such data gested as distinct individual difference variables,
could provide for a better empirical grasp on the both of which might relate to creativity in the gener-
degree to which information processing in dream or al popUlation.
Assessing Individual Variation in used by Pomerantz, Sager, and Stover (1977). The
Perceptual Organization Ability task required of the subject was to detect the "odd-
ball" quadrant as rapidly as possible, and choice
Our discussion thus far suggests that obtaining a reaction time (RT) was the dependent variable. In
better understanding of patterns of individual varia- Example A, the baseline task involves choosing the
tion in perceptual organization tendencies might be oddly oriented single parenthesis, located in this
highly useful in examining the role of perceptual example in the upper left. This display requires a
organization processes in creative thought. Unfor- quite effortful scrutiny as the mean RT obtained by
tunately, with the exception of various measures of Pomerantz et al. was 2,400 msec. However, adding
embedded figures performance, Gestalt grouping the four extra context elements in the display leads
processes and other related "automatic" processes to a perceptual grouping phenomenon that makes
of perceptual organization have not been subjected the oddball element stand out, reducing the RT by
to systematic individual difference measurements. nearly a second (for a group average obtained with
Performance measures on tasks, such as those in- college student subjects). Now, consider the dis-
volving embedded figures, often do not provide a play shown in Example B. Here, different context
convenient way of separating performance at- elements are added (rotated parentheses) for which
tributable to loose perceptual organization (failure the result of organizational processes produces ob-
to group), from performance attributable to strong, jects perceptually less distinct than the single paren-
executively controlled focused attention (ability to theses, resulting in an average increase in RT of 550
break apart). It appears to us, however, that one msec. Thus, in these examples, we have two highly
should be able to separate the contributions of auto- similar tasks, one for which Gestalt organization
matic organization and efficient focused attention, helps, one for which it hurts.
by selecting a combination of structurally similar Presumably, individuals with weak organiza-
tasks, including some in which perceptual organi- tional processes would fail to gain as much benefit
zation is helpful to performance, and others for from the helpful context, in comparison with more
which perceptual organization is harmful. typical subjects, while performing tasks such as that
As an example of such an approach, consider the of Example A. On the other hand, such subjects
example of the stimulus displays shown in Figure 3, also might be less susceptible to interference from
context stimuli that disrupt performance of typical
subjects. In contrast, subjects possessing nominal
automatic grouping processes, but exceptionally
[J C] u
EXAMPLE A;
powerful executively controlled focal attentional
processes, might exhibit far less interference from
the harmful context, while still maintaining benefits
from the helpful context.
) ) ) ) » » A related set of classification or sorting tasks that
have the potential for separately assessing loose-
5T I MUll ALONE CONTEXT ELEMENTS STIMULI & CONTEXT ness of involuntary perceptual organization and the
RI .. 2'±OO mSBC RI ... 1 '±SO
strength of voluntary visual attention is the one used
mSBC
EXAMPLE B: by Pomerantz and Schwaitzberg (1975). In this ex-
[J [J
- - (d
periment, there were three different basic sorting
tasks required of a subject, and the stimulus re-
sponse mappings for each of these is shown in Fig-
ure 4. In Condition A (the control condition), sub-
) ) ) - ) - jects sorted stimuli on the basis of orientation of the
leftmost parenthesis; the orientation of the right pa-
ST I MUL I ALONE
RT .. 2'iOO msec
CONTEXT ELEMENTS STI MUll & CONTEXT
RT .. 2950 msec
renthesis was held constant. In Condition B, which
required subjects to divide attention and attempt to
Figure 3. Examples of stimuli used in an "oddball quadrant" ignore the irrelevant rightmost parenthesis, the re-
detection task employed by Pomerantz, Sager, and Stover sponse assignment was determined by the left pa-
(1977). In Example A, the addition of the context elements
causes Gestalt grouping processes that make the unique quadrant
renthesis, and the right parenthesis varied in a man-
more distinctive, whereas in Example B, the rotated context ner orthogonal to the left parenthesis. In this
elements obscure the distinction (see text for details). condition, classification times were considerably
CONDITION A (control) tual attributes (e.g., "the fat ones vs. the skinny
ones"). However, at wider separations, in which
« )(
involuntary grouping processes break down, sub-
jects had to apply effortful divided attention pro-
cesses, either to try to perceptually group the paren-
thesis pairs or to separately evaluate each
left response right response parenthesis and apply a classification rule. A typ-
ical pattern of results for these three tasks is shown
in Figure 5, in which sorting time is plotted as a
function of separation of the parentheses. Again,
CONDITION B (focused attention) these data are based upon the means of subjects'
performances that were obtained by Pomerantz and
Schwaitzberg (1975).
« () )( » One might expect, however, that individual dif-

ferences in the potency of involuntary organiza-
tional processes and individual differences in the
executively controlled ability to break up percep-
left response right response
tual configurations (as well as divide attention, in
this case) would produce systematic differences in
the functions from those shown in Figure 5. An
appropriate application of psychometric scaling
CONDITION C (divided attention) procedures to differences among such patterns, in
addition to similar analyses of patterns in such tasks
as the "oddball detection" examples in Figure 3,
() )( « »
could lead to separate scales for strength of percep-
tual organization tendencies and power of voluntary
selective attention.
To our knowledge, there is no existing research
left response right response
on patterns of individual differences in the costs and
Figure 4. Examples of stimuli used by Pomerantz and Schwaitz- benefits of the configural effects that might be ob-
berg (1975) in a task illustrating how Gestalt grouping affects tained by systematic administration of these or sim-
selective aIid divided attention in a visual classification task (see ilartasks to large numbers of individuals , in order to
text for details). obtain an index of relative organizational power
among different individuals that might be of use for
determining its relationship to creative behavior.
slower than those of the control condition, provided However, given that individual differences in more
the spatial separation between the parenthesis pairs traditional measures of embedded figures tasks ex-
was close. As the spatial separation between the ist (even though the voluntary selective component
elements was increased beyond a single types pace has not been factored out), and given the recent
in other stimulus sequences, the interference from evidence of the very unique and actually superior
the irrelevant right parenthesis diminished. performance of schizophrenics (compared with
Condition C required that subjects split the four normals) on tasks for which avoidance of grouping
possible pairings of parentheses into two groups, contributes to performance (Place & Gilmore,
such that the response assignment was determined 1980; Wells & Leventhal, 1984), we believe that
by the combination of both parentheses. Unlike ei- such individual differences exist, are potentially
ther Condition A or Condition B, this task required measurable, and probably do bear a relationship
an evaluation of both the leftmost and the rightmost with other important cognitive attributes including
parentheses. At close separations, such grouping creativity. Given that the presumed organizational
seemed to occur involuntarily, as subjects reported looseness of schizophrenics is often related to cre-
perceiving the parenthesis pairs as single objects. In ative thought in individuals who perhaps have a
this case, classification times were actually shorter lesser degree of the deficiency (e.g., Keefe &
than for Condition B, because the objects assigned Magaro, 1980), the data on the psychiatric popula-
to each category appear to share common perceptions are of considerable interest and are encourag-
19 Divided Attention ,.
/
/
/
18
,.
1/1 /
"'C /
c: /
0
u 17
Q)
1/1
/
/
c: /
16 Selective Attention /
/
'-
Q) A-._............. "
E -~---
I- , ............
...----....
15 /' '&...
01
c:
....I-
._ ......
14
0
Vl
13 Control
12
0.25 0.50 1.0 2.0 4.0 8.0

Near Far
Proximity
(in approximate degrees of visual angle)
Figure 5. A graphical representation of results obtained by Pomerantz and Schwaitzberg (1975) for Tasks A, B, and C of Figure 4.
Sorting times for decks of 32 cards displaying these stimuli are plotted as a function of the physical separation between the pairs of
parentheses. Stimulus sets involving wide spatial separations between the two stimuli greatly slowed sorting in the divided attention
task (Task C of Figure 4), while allowing subjects to avoid interference from perceptual grouping in Task B. Presumably the form of
such functions might differ for subjects differing in the potency of their involuntary perceptual organization processes as well as the
power of their controlled visual attention. Reprinted by permission of LEA Associates.
ing for clarifying the "creativity-madness" asso- mining which of two clock times (given as digital
ciation that has been around for centuries. expressions) would produce the smallest angle be-
tween the hands of a standard analog clock. Paivio
found that, in general, reaction time was a monoto-
nic function of differences in the analog hand an-
Individual Variation in Intramodal gles, providing strong evidence that the task was
Visual Code Transformation mediated by image comparison. In addition, how-
ever, subjects were divided into "high-imagery"
In contrast to our lack of knowledge about indi- and "low-imagery" groups, on the basis of a medi-
vidual variation in perceptual organizing ability, an split of performance on a composite of
psychometric tests that directly tap the ability of Guilford's Block Visualization Test, the Minnesota
individuals to compare transformed visual objects Paper Form Board, and the Thurstone Space Rela-
(e.g., rotated block drawings, folded and unfolded tions Test. High-imagery subjects produced RTs
block figures) have been available for decades that averaged about 1 second faster at each level of
(e.g., see Guilford, 1967). Individual variation in angular difference than low-imagery SUbjects. In
abilities to perform such operations is clearly mea- contrast, a similar split of the subject population on
surable, and it is apparent that performance on such the basis of a verbal fluency measure (Paivio, 1971)
psychometric tests correlates with performance of produced no differences between groups. This
mental tasks or problems that appear to be mediated study as well as numerous others (e.g., see Paivio,
by use of visual analog codes. 1971, 1978 for a review) provide relatively con-
One specific example is a study by Paivio (1978), vincing evidence that individual differences in abil-
who measured subjects' reaction times for deter- ity to perform top-down manipulation and trans for-
mation of visuospatial codes can be predicted most of the anecdotal reports of such spontaneous
effectively with existing psychometric instruments. insight, summarized by Shepard and others, have
Given the apparent importance of fluent visual come from individuals who were likely to score
coding in anecdotal reports of creative cognition, it very high on measures of mental abilities and who
seems intuitive that psychometric measures of such were (or still are) known to have put in large
visual abilities should relate positively to measures amounts of controlled mental effort on problems
of creative behavior, particularly for those indi- related to their creative achievements. The selec-
viduals whose creative work is arrived at primarily tion, rejininement, and use of images spon-
through the effortful, voluntary route. One must taneously generated during transitional or altered
keep in mind, however, that simple possession of a states likely may depend upon effortful executive
mental ability does not necessarily imply that abil- processes, suggesting that spontaneous insight
ity will be used to produce novel and worthwhile should not be viewed as a totally involuntary occur-
products. Creative behavior requires an evaluative rence. Moreover, mental activity that substitutes
component for recognizing when a particular novel for the sensory signals that normally drive mecha-
representation is of value. Given these considera- nisms of perception, insofar as they are influenced
tions, and our configural view of the creative pro- by memory activities, may be highly structured by
cess, one should not expect extremely high linear previous effortful mental code manipulations. In
correlations between measures of visual coding flu- short, spontaneous insight may not be so spon-
ency and creative behavior per se. taneous and is conceivably quite closely related to
One issue of importance in determining how spe- mental transformation abilities of both intramodal
cific information-processing characteristics relate and intermodal varieties.
to creativity is whether it is appropriate to view
individual differences in performance of specific
classes of perceptual tasks as measures of relatively An Example of the Interplay of
narrowly defined information-processing abilities,
or whether correlations among specific task perfor- Involuntary and Executively Driven
mances suggest the existence of a general ability Creative Processes: Synesthesia and
"factor." Guilford (1981,1983) argued that pro- Cross-Modal Representations
cesses, such as visual code manipulation and cross-
modal transformations (which we will discuss pres- Previously, we have posited that these two very
ently), are all part of a general "transformational different processes might produce similar creative
ability" that is a key component of creative products. In this section, we will look at two related
thought. Indeed, one can make a rather strong psy- processes, one perceptual (automatic) and one cog-
chometric argument for that position, as Guilford nitively mediated (executively controlled), that do
has done. On the other hand, perceptual psychol- indeed seem to lead to similar and potentially cre-
ogists who are aware that different neural structures ative representations of the environment.
may be involved in intramodal versus cross-modal Theorists and artists long have recognized the
transformations, and also that common neural correspondences, interrelationships, and interde-
structures may be involved in both voluntary and pendencies of the senses as they are used to capture
involuntary processing of visually coded informa- information about the world. Aristotle and other
tion, are more likely to have interest in specific early thinkers posited various relationships of
comparisons among tasks. One research question cross-modal process or product, and Bishop Berke-
that bears upon the relationship between specific ley added his notion of their ontogeny. More re-
visual information-processing abilities and our con- cently, Stevens (e.g., 1959), Gibson (1966), and
figural model of creativity is whether fluency in Marks (e.g., 1978, 1982) have provided more re-
controlled visual transformations might be related fined theories and hypotheses about the nature,
to the frequency of spontaneous insight in creative workings, and meanings of these correspondences.
thought. We have thus far implied, as has Shepard Particularly for Marks, the evidence of artistic
(1978, 1981), that spontaneous novel constructions awareness and of creative use of these correspon-
are basically the result of involuntary representa- dences adds weight to various laboratory studies
tional processes, decoupled from their normal sen- that are in support of the theoretical ideas of cross-
sory source of data. However, it must be noted that modal equivalences and sensory unity.
Most people are familiar with the use of cepts, numerous products have arisen. According
onomatopoetic words, such as "hiss," "crack," to Marks (1978), synesthesia has enjoyed two peri-
and "woff," to convey the auditory characteristics ods of extensive study, the first 40 years of the
of a sound as well as its semantic meaning. But the twentieth century, and an earlier period during the
sounds of words also seem to convey other informa- nineteenth century. It was during this earlier period,
tion. Kohler (1947) showed that most people when synesthesia was of interest not only to scien-
matched the pseudowords "maluma" and "take- tists and physicians but also to musicians, that cre-
te" with the rounder and the more angular of two ative products were introduced that clearly sprang
line figures, respectively. As a demonstration, from the perceptual phenomena. Perhaps the best
Marks (1978) replicated the finding using a produc- known of these were the multimodal concerts that
tion task in which two children were asked to draw mixed music with colored lights and occasionally
the visual representations of these words. The with odor. Louis-Bertrand Castel built the first light
drawing of takete was clearly the more angular for organ in 1735. This organ and others like it pro-
each person. As another example of how word duced a particular colored light along with each
sounds can supply non auditory information, Sapir note as the keys were depressed. Numerous pieces
(1929) and later Newman (1933) asked subjects were written for and performed with these instru-
about the size of objects referred to by nonsense ments. Laser light shows are a more recent version
syllables and found the words containing the letter of this same artistic use of the close relationship
lal were judged to refer to larger objects than did between color and tonality to produce desired affec-
those containing Iii. Finally, most people in our tive responses.
society are familiar with the relationships between Also, during the nineteenth century and later,
colors and temperatures-red is "hot" whereas there were numerous linguistic expressions of syn-
blue is "cool." esthesia-like experiences, or synesthetic meta-
These types of correspondences are quite differ- phors. The following examples of such poetry are
ent from the wholly involuntary phenomena of true taken from Marks (1978): Charles Baudelaire's
synesthesia, in which "a small minority of people "Correspondences," "perfumes fresh as chil-
experience a curious sensory blending, where stim- dren's flesh, sweet as oboes, green as prairies,"
ulation of a single sense arouses a melange of senso- and Arthur Rimbaud's "Sonnet of the Vowels"
ry images" (Marks, 1978, p. 83). In addition to "A black, E white, I red, U green, 0 blue. " Exam-
truly synesthetic individuals, there are widespread ples of synesthetic metaphor in literature are (from
reports of involuntary synesthesia produced by Marks, 1982): "the sound of coming darkness"
various consciousness-altering drugs, such as LSD (Poe); "a soft yet glowing light, liked lulled mu-
and hashish. sic" (Shelley); and "music suddenly opened like a
As interesting and well-documented as true in- luminous book" (Conrad Aiken).
voluntary synesthesia is, it is not, in itself, a cre- On interesting point concerning these uses of
ative product, just as the novel representations of color-sound correspondence and poetic metaphor
schizophrenic thought are not. Creative products (which is an important difference between these
require the additional processes of, appropriate productions and those of schizophrenics) is that
selection and presentation of those novel represen- none of these authors appear to have been truly
tations. Unlike the schizophrenics' art, which is synesthetic themselves (although one may conjec-
often identified as creative after production, the ture about those authors who had a history of nar-
synesthetics must choose whether or not to capture cotic use, e.g., Poe). Yet they were able to produce
their mental representations and present their novel creative results using these correspondences in
version of the world. ways so compelling that we not only understand and
Furthermore, although loose perceptual organi- agree with their meanings, but often do not even
zation is a potential source of novel representations immediately notice that there are "crossed" or
for synesthetics (as for schizophrenics), documen- "mixed" modality-specific adjectives and nu-
tation of creative products from true synesthetics is ances. Marks provided experimental evidence that
notably lacking. In part, the lack of documented most persons have (or can develop) a strongly inter-
examples simply may reflect the relative infrequen- nalized correspondence of cross-modal relationship
cy of true synesthetic individuals in the population, between certain visual and auditory characteristics
relative to schizotypic or schizophrenic individuals. (primarily brightness and loudness). Thus, we see
But from each of these sources of anomalous per- that an automatic perceptual process-the ca-
pability and tendency toward cross-modal associa- problematic in that individual differences in cog-
tions-is necessary for the appreciation of creative nitive processing that affect these three "routes" to
products generated by an executively controlled creative thought probably are not independent. For
process. example, the looseness in perceptual organization
In this discussion, we have tried to give an exam- that may characterize creative individuals with
ple of how knowledge of an infrequent, perceptual, schizotypal thought patterns may well be negatively
and automatic phenomenon (synesthesia) can pro- correlated with ability to make effortful mental
vide the impetus for the use of executively con- transformations as well as the ability to recognize
trolled processes to create artistic products. Finally, that a novel mental representation is worthwhile.
the appreciation of these creative products greatly We do believe, however, that research efforts
depends upon the ability to "perceptually reso- into individual differences in specific perceptual
nate" with those products, an ability that seems to characteristics potentially can be useful in both in-
be nearly universal and automatic. creasing our understanding of their relationships to
creative behavior, as well as for determining cir-
cumstances that might foster creative thought. As
we have noted, research literature describing indi-
Summary and Conclusions vidual differences in perceptual organization ten-
dencies is notably lacking. The study of individual
We have argued that the generation of creative differences in executively controlled transforma-
behavior can result from a combination of involun- tions of mental representations has received consid-
tary and executively controlled processes. These erably more research effort, due to a long existing
processes rely heavily on neural mechanisms and presumption that such operations are closely related
systems that have evolved primarily as perceptual to measures of basic mental abilities. However, the
systems. However, these systems, which embody "mental measurement" motivation for much of
highly sophisticated computational and inferential this research, has, in our view, directed researchers
mechanisms, also serve to operate on mental codes more toward the study of "common transforma-
actively retrieved from memory (e.g., conscious tional ability factors" (e.g., Guilford, 1981, 1983)
generation and modification of imagery). In addi- rather than toward the properties of specific types of
tion, there is anecdotal evidence that these same mental code manipulations, such as synesthetic
mechanisms can be the source of spontaneously metaphor. Lastly, the literature on spontaneous
generated images or representations, when their generation of novel representations by perceptual
normal driving source of sensory stimulation is de- mechanisms is, at present, highly speculative, and
coupled, as in transitions from dreams or altered consists primarily of self-report anecdotes.
states. Such spontaneous generation, if appropri-
ately selected and recognized as useful, perhaps can
produce the sudden creative insights characterized
by many self-reports from creative individuals. A Flow Diagram of the Creative Process
According to our analysis, creativity thus can and Its Implications
result from some combination of (1) novel percepts
attributable to departures from the normal deter- To summarize our view of how processes related
ministic processes of perceptual organization, (2) to perception influence the creative process, we of-
effortful conscious mental activity involving ma- fer the flow diagram shown in Figure 6. In addition,
nipulation and transformation of codes that gener- Table 1 summarizes some of our conjectures about
ate novel representations, and (3) spontaneous gen- the relationships between various components of
eration of novel representations. Because the the flow diagram and some individual difference
relative contributions of each of these to a specific variables potentially measurable by existing or des-
creative achievement presumably varies markedly ignable assessment instruments and surveys.
across both situations and individuals, the rela- Essentially, Figure 6 is a visual summary of our
tionship between specific cognitive abilities or previous discussions of a configural view of the
characteristics of processing and the likelihood of creative process. It lists as sources of novel repre-
an individual's producing products judged to be sentations (1) atypical involuntary processes of per-
creative is highly configural and thus difficult to ception, including both loose organizational pro-
measure. The configurality offactors is particularly cesses and true synesthesia, (2) spontaneous (and
Involuntary Processes Executivel~ Controlled Processes
Consc~ous Intramodal Cross-MOdal

Anomalous Percepts Perceptual Transccrmation TI"'ansf"ormati
Organ~zation
(Metaphor)
•
"Looss" True
Perceptual S~nesthe5ia
Organizations I
I
I
!
Revise
Pool of Novel
Repr-Bsentations
Figure 6. A schematic representation of our view of how involuntary and executively controlled perceptual processes contribute to
creative thought (see text for details).
involuntary) generation of insight and imagery, (3) ery ability, 1M; (3) general measures of transform a-
conscious perceptual organization, (4) conscious tion ability, T; (4) schizotypal thought, ST; (5) drug
intermodal transformations of representations, and usage, DU; (6) ability to report dreams, DR; and (6)
(5) conscious cross-modal transformations of repre- amount of reported conscious effort expended on
sentations. In addition, it includes the process of the problem for which a creative solution is found,
selection of novel representation and feedback to EF. Note that these items constitute a nonex-
the executively controlled processes for purposes of haustive sample of potentially measurable indi-
refining the potential creative product. vidual differences variables. As we have mentioned
In Table 1, we have chosen to relate the uses of previously, respected psychometric instruments ex-
each of the potential sources of novel representa- ist that presumably tap some of these, whereas for
tions discussed in this chapter together with an "ef- others, instruments would need to be developed.
ficiency" selection mechanism drawn from the fol- Furthermore, measurement of some of the self-re-
lowing set of individual differences variables: (1) port variables (such as DR and EF) would most
strength of perceptual organization, SPO; (2) imag- likely involve both validity and reliability problems
Table 1. Hypothetical Associations between Modes of Generating and Selecting Creative

Representations and Various Measures of Individual Differences a
Source SPO 1M T ST DU DR EF
Loose perceptual organization + +'1 ? -?
Synesthesia ? +? ? -? + ? ?
Spontaneous imagery +? + + -? ? + +
Conscious perceptual organization + + + -? ? +
Intramodal transformation + + + ? +
Cross-modal transformation ? +'1 ? ? ? +?
aSee text for a detailed explanation.
that we will not address here. Given those dis- References

claimers, and with the additional disclaimer that
nearly all the cell entries are based upon our own Brooks, L. R. (1968). Spatial and verbal components of the act
current personal speculations as opposed to existing of recall. Canadian Journal of Psychology, 22, 349-368.
Gibson, J. (1966). The senses considered as perceptual systems.
data, we offer this table of hypotheses as a guide for Boston: Houghton Mifflin.
potential future research. The cell entries them- Goodman, K. J., & Marquart, D. I. (1978). Creativity and per-
selves consist of + and - signs to indicate pre- ception: The neglected theory of Lowenfeld. Journal ofCre-
sumed positive and negative correlations, together ative Behavior, 12, 279.
Guilford, J. P. (1967). The nature of human intelligence. New
with ? signs that specify either that no particular York: McGraw-Hill.
relationship is likely to exist, or, if there is one, we Guilford, J. P. (1981). Higher order structure of intellect abili-
have no basis to speculate its directionality. Addi- ties. Multivariate Behavioral Research, 16,411-435.
tionally, we have used the question mark as a modi- Guilford, J. P. (1983). Transformation abilities or functions.
fier to indicate above average uncertainty about our Journal of Creative Behavior, 17, 75-83.
Hochberg, J. (1981). Levels of perceptual organization. In M.
speculations. Kubovy & J. Pomerantz (Eds.), Perceptual organization (pp.
We should call attention to one pattern among 255-278). Hillsdale, NJ: Earlbaum.
these hypothetical intercorrelations that emphasizes Kahneman, D., & Tversky, A. (1973). On the psychology of
our previous statements about the configural nature prediction. Psychological Review, 80, 237-251.
Keefe, J. A., & Magaro, P. A. (1980). Creativity and schizo-
of creativity. That pattern is that individuals whose phrenia: An equivalence of cognitive processing. Journal of
creativity stems largely from anomalous perceptual Abnormal Psychology, 89, 390-398.
inputs (i.e., individuals with loose perceptual orga- Kellogg, R. (1986). Designing ideal processors for document
nization and perhaps a rare, creative, true syn- composition. Behavior Research Methods, Instruments, and
esthetic) are likely to be very different from creative Computers, 18, 118-128.
Kohler, W. (1947). Gestalt psychology. New York: Liveright.
individuals who rely more heavily upon executively Marks, L. (1978). The unity of the senses. New York: Academic
controlled processes. Conceivably, this might bear Press.
upon differences in the types of creative products Marks, L. (1982). Synesthetic perception and poetic metaphor.
that are developed by individuals who rely pre- Journal of Experimental Psychology: Human Perception and
Performance, 8, 15-23.
dominantly upon anomalous percepts for novel rep-
Newman, S. S. (1933). Further experiments in phonetic sym-
resentations, as opposed to those arising from some bolism. American Journal of Psychology, 45, 53-75.
combination of executively controlled manipUla- Paivio, A. (1971). Imagery and verbal processes. New York:
tions of representations and spontaneous imagery Holt, Rinehart & Winston.
and/or insight. As a possible case in point, one Paivio, A. (1978). Comparison of mental clocks. Journal of
Experimental Psychology: Human Perception and Perfor-
often encounters examples of schizophrenic paint- mance, 4,61-71.
ings and drawings, yet rarely does one encounter Place, E., & Gilmore, G. (1980). Perceptual organization in
mention of schizophrenic poetry or sculpture. schizophrenia. Journal of Abnormal Psychology, 89, 409-
As a final point, we do not wish to imply, by our 418.
apparent emphasis upon individual difference vari- Pomerantz, J., & Schwaitzberg, S. (1975). Grouping by prox-
imity: Selective attention measures. Perception and Psycho-
ables in this final section of this chapter, that re- physics, 18, 355-361.
search on conditions that affect the likelihood of Pomerantz, 1., Sager, L., & Stover, R. (1977). Perception of
creative behavior within a given individual is less wholes and their component parts: Some configural superi-
important than research aimed at predicting patterns ority effects. Journal of Experimental Psychology: Human
Perception and Performance, 3, 422-435.
of creativity among individuals. Clearly, the two Prentky, R. A. (1979). Creativity and psychopathology: A neu-
classes of research complement each other; by rocognitive perspective. In B. Maher (Ed.), Progress in ex-
learning more about relationships between percep- perimental personality research (Vol. 9, pp. 1-39). Academ-
tion-related processes as varieties of creative be- ic Press.
havior, we may discover strategies for optimizing Sapir, E. (1929). A study of phonetic symbolism. Journal of
Experimental Psychology, 12, 225-239.
creative production within individuals. Similarly, Segall, s. J., & Fusella, V. (1970). Influence of imaged pictures
through applied research aimed at development of and sounds in detection of visual and auditory signals. Journal
"thinking aids," "designing aids," and "writing of Experimental Psychology, 83, 458-474.
aids," or through the evaluation of such products, Shepard, R. N. (1978). Extemalization of mental images and the
act of creating. In B. Randhawa & W. Coffman (Eds.), Visual
we may discover additional relationships among learning, thinking, and communication (pp. 133-189). New
perceptual and cognitive processes and the creative York: Academic Press.
process. Shepard, R. N. (1981). Psychophysical complementarity. InM.
Kubovy & J. Pomerantz (Eds.), Perceptual organization (pp. Taylor, D. W. (1960). Thinking and creativity. Annals of the
279-341). Hillsdale, NJ: Earlbaum. New York Academy of Sciences, 91, 108-127.
Stein, M. I. (1956). A transactional approach to creativity. In D. Wells, D., & Leventhal, D. (1984). Perceptual grouping in
W. Taylor (Ed.), Research conference on the identification of schizophrenia: Replication of Place & Gilmore. Journal of
creative talent. Salt Lake City: University of Utah Press. Abnormal Psychology, 93, 231-234.
Stevens, S. (1959). Cross-modality validation of subjective
scales for loudness, vibration and electric shock. Journal of
Experimental Psychology, 57, 201-209.
CHAPTER 9
Memory and Creativity

Barry S. Stein
Although there are many ways to define creativity, inferences people make during learning and by their
creative behavior usually involves a product or re- expectations at the time of test (e.g., Johnson,
sponse both novel and appropriate to the task at Bransford, & Solomon, 1973; Loftus & Palmer,
hand. Creative ideas and discoveries often provide 1974; Pichert & Anderson, 1977). These findings
new information and perspectives that were not ap- suggest that some elements of creativity are in-
parent in the past. In contrast, the concept of memo- volved in remembering.
ry is typically associated with ideas that are not Creative behavior can also involve elements of
novel or original. Indeed, the act of remembering is memory. For example, many creative works are
an attempt to recreate events and experiences that based on the personal experiences of the artists and
have occurred in the past. From this perspective, writers who produce them. The recollection of per-
memory and creativity appear to involve very dif- sonal experiences necessarily involves aspects of
ferent kinds of activities. memory. Creative behavior can also involve more
From other points of view, the differences be- abstract types of knowledge and skills. For exam-
tween memory and creativity may not be as distinct. ple, a writer's linguistic skills or an artist's drawing
For example, Bartlett (1932) noted that the wayan skills reflect knowledge that is acquired through
event is remembered can change over successive experience and represented in memory.
recall tests. He found that information is sometimes The preceding discussion suggests that although
deleted from the recollection of an event, and new memory and creativity often involve different
information is sometimes added to the recollection goals, they may frequently involve similar pro-
of an event that was not present during the original cesses. In order to better understand the factors that
experience. On the basis of these observations, influence creativity, it seems worthwhile to consid-
Bartlett argued that remembering is not simply the er how previous experience affects creativity. This
reinstatement of previously experienced events but chapter explores both th~ positive and negative ef-
rather involves an imaginative reconstruction of the fects of memory on creativity and provides a con-
past. ceptual framework for understanding the con-
The tendency for remembering to involve imagi- straints that influence the creative transfer of
native and interpretive processes is also consistent knowledge. Consideration is also given to the types
with other research. For example, what is remem- of experimental methods that can provide useful
bered about an event can be affected by the kinds of information about the creative transfer of knowl-
edge. In addition, a heuristic framework is dis-
Barry S. Stein • Department of Educational Psychology, cussed that may facilitate the creative transfer of
Tennessee Technological University, Cookeville, TN 38505. knowledge and provide ideas for further research.
163
The Effects of Memory on Creativity use the information presented earlier if they had
attempted to solve the problem first without the
Negative Effects hint. The negative effects of trying to solve the
The idea that previous experience can interfere problem without the hint were found to be problem
with the development of creative solutions is sup- specific; that is, previous experience with a prob-
ported by a variety of empirical studies. For exam- lem only affected subsequent performance on the
ple, Maier (1931) observed how remembering com- same problem and not on other problems presented
mon . properties of objects can adversely affect during the experiment.
creatIve problem solving in the classic "two-cord" The findings of Perfetto et ai. (1983) suggest that
or "two-string" problem. In that study, people memory for previous attempts to solve a problem
can inhibit creative problem solving. Previous ex-
~ere asked to !i~ together two cords that were hang-
perience with a particular problem-solving strategy
mg from a ceIlIng. The cords were positioned far
also may reduce the likelihood of using alternative
enough apart so that a person could not hold one
and more effective problem-solving strategies
cord ",:hile reaching the other. Although the partici-
(e.g., Luchins & Luchins, 1950). These findings
pants m that study were given numerous hints to
demonstrate how memory for recent events or fa-
help them generate a particular solution, Maier
miliar concepts and strategies can interfere with
found that many people failed to attach an available
creativity in problem solving.
set of pliers to one of the cords to enable it to be
swung like a pendulum (making it possible to reach
one cord while holding onto the other cord). Positive Effects
Maier's observations suggest that people fail to pro-
The preceding discussion illustrated some of the
duce the creative pendulum solution because their
ways in which memory can inhibit creative problem
previous experience with pliers as a gripping device
solving; however, there are clearly many situations
pre.vents them from thinking of the pliers as a
in which previous experience facilitates creativity.
weIght (see also Birch & Rabinowitz, 1951).
For instance, it is difficult to imagine any creative
The term functional fixedness is often used to
work in art or science that is not related in some way
describe situations (like those that Maier observed)
to an individual's previous experience. As Weis-
in. which people tend to think about objects only
berg (1986) pointed out, many of the events and
WIth respect to their most characteristic function
characters portrayed in creative works (e.g., Crime
(e.g., Duncker, 1945). This tendency to think about
and Punishment, and Tender Is the Night) appear to
objects in the way that they are most often used
be based on specific personal experiences of the
illustrates one way that our memory for past experi-
artists. These observations suggest that remember-
ences can limit flexibility in thinking and can inhibit
ing can be an important part of the creative process
the production of creative solutions.
(cf. Campbell, 1960).
There are also other ways that memory can ad-
versely affect creativity. For example, Perfetto, In other situations in which a particular experi-
ence is not the focus of creative behavior, the
Bransford, and Franks (1983) found that previous
knowledge that is acquired through past experience
experience with a problem can adversely affect the
is often useful in generating creative solutions
production of creative solutions. They presented
(e.g., Amabile, 1983; Newell & Simon, 1972). For
people with problems such as this:
example, discovery of the double-helix structure
for DNA by Watson and Crick required an under-
A man who lived in a small town in the U.S. married 20 different standing of biological concepts and X-ray diffrac-
women of the same town. All are still living and he has never tion techniques. Similarly, the skills that enable the
divorced one of them. Yet, he has broken no law. Can you
explain?" (p. 25)
artist to paint and the writer to write are also devel-
oped through previous experience. These observa-
tions suggest that the knowledge and skills that are
The participants in the experiment were given clues acquired in a specific discipline provide the concep-
that could help them solve the problems prior to the tual tools that are needed to develop creative solu-
presentation of the problems (e.g., "A minister tions. Without that knowledge, our potential for
marries several people each week"). Perfetto and creative works would be greatly diminished. As
his colleagues found that people were less likely to Amabile (1983) notes, "clearly, it is only possible
solve this problem correctly when given a hint to to be creative in nuclear physics if one knows some-
CHAPTER 9 • MEMORY AND CREATIVITY 165
thing (and probably a great deal) about nuclear rather the appropriateness of that information for
physics" (p. 70). solving a particular problem.
An emphasis on the contextual determinants of
creativity is consistent with other research that dem-
onstrates the important role of contextual factors in
Contextual Constraints on Creative Transfer
learning phenomena (e.g., Bransford, 1979;
One approach for explaining the differential ef- Jenkins, 1974, 1979; Morris, Bransford, & Franks,
fects of memory on creativity is to argue that certain 1977; Rosnow & Georgoudi, 1986; Stein, 1978).
kinds of knowledge are inherently more beneficial For instance, Jenkins (1979) noted that what is
for creative problem solving than are other kinds of learned in any particular situation depends on the
knowledge. For example, one might argue that re- nature of the information presented, the activities
membering unusual facts or experiences (e.g., performed by the learner, and the skills of the learn-
pliers can be used as a weight) will facilitate cre- er (see also Bransford, 1979). Jenkins also sug-
ative problem solving more than remembering fa- gested that what is learned in any particular situa-
miliar properties of objects (e.g., pliers can be used tion will affect the kinds of tasks that can be
to grip things). It could also be argued that abstract performed. For example, Stein (1978) and Morris
knowledge and skills are more beneficial for et at. (1977) found that the effects of any particular
creativity than is knowledge of concrete particulars. learning experience depend on the nature of the test
Indeed, there are a variety of dimensions that could that is given to assess learning. Learning experi-
be considered to explain the varying effects of ences that enhance performance on certain types of
memory on creativity (e.g., episodic vs. semantic criterial tests do not facilitate performance on other
knowledge; procedural knowledge vs. 'declarative types of criterial test (see also Roediger & Weldon,
knowledge). 1987).
Alternatively, one could argue that there are not From a contextualist's perspective, the creative
inherent differences in the effects that different transfer of knowledge is constrained by both the
types of information have on creativity. Instead, the context in which information is to be used and the
effects of any particular memory depend on the context in which information is learned. To under-
context in which it is remembered and how it is stand the factors that affect creative problem solv-
applied to the problem. For example, in Maier's ing, it is necessary to examine the relationship be-
two-cord problem, remembering that a set of pliers tween different learning experiences and perfor-
can be used as a gripping device would not help mance on creative problem-solving tasks.
people generate the creative pendulum solution.
However, in a different problem-solving context,
remembering that same information could prove
beneficial. For instance, in searching for some way Methods for Investigating Creative
to remove a hot pan from the oven, remembering Transfer
that pliers can be used as a gripping device can lead
to a creative solution (e.g., using a pliers as a pot One implication of the contextually senSItive
holder). Within this alternative approach, the cre- nature of learning and transfer is that the contexts in
ative value of any information is only a function of which people are tested need to be carefully consid-
the context in which it is remembered. ered. As Jenkins (1979) notes, "there is no as-
Consider another illustration of how the creative surance that our contrived and artificial situations
value of information can depend on the context in will ever come to relate to anything that we care
which information is remembered. For example, about in the real world" (p. 443). From the stand-
remembering that an igloo is a type of home that can point of investigating creativity, there is no as-
be damaged by heat is not in itself a novel or cre- surance that factors that affect performance on other
ative thought. However, in the context of explain- types of tasks will also affect creative problem solv-
ing the statement, "The home was small because ing. Therefore, it is important to design experimen-
the sun came out," remembering that particular in- tal tasks that tap into the same processes that people
formation about igloos can lead to a creative solu- use when solving problems creatively in natural
tion. What appears to influence the effect of any settings.
particular knowledge on creativity is not the un- There are several important features of creative
usual nature of the information remembered but problem-solving tasks. For instance, in these tasks,
the individual must not only remember information cient force to overcome the fortress but could not
but must also appropriately apply that information bring his entire force as a unit to the fortress because
to a novel problem. Furthermore, creative problem- the enemy had mined the roads, allowing only small
solving tasks usually do not include hints to use groups to travel on the road without setting off the
specific information to solve problems. In fact, de- mines. The general's solution was to divide his
termining what information is relevant for a prob- troops into small units traveling on different roads
lem and how to apply that knowledge is an integral that would eventually converge at the fortress.
part of creative problem solving. In the second phase of the experiment, Gick and
In much of the research that investigates learning Holyoak (1980) presented people with an analo-
and memory, people are given implicit cues or in- gous problem to solve (Duncker's, 1945, "radia-
structions to use information presented earlier in the tion problem"). In the radiation problem, people
experiment at the time of test. For example, sub- are asked to propose a procedure that could be used
jects are generally told at some point in the experi- to destroy a malignant tumor using a special type of
ment that they are to recall the information present- radiation. At a sufficient intensity level, the radia-
ed to them earlier. In contrast, people seldom have tion will destroy the tumor. However, at an inten-
such cues to select relevant experiences in creative sity level sufficient to destroy the tumor, the radia-
problem-solving situations. Indeed, if a person has tion will also destroy healthy tissue. At lower levels
to be explicitly prompted to access and to apply of intensity, the radiation will not destroy the tu-
information to a problem, then the resulting behav- mor. Gick and Holyoak (1980) were particularly
ior is generally not considered creative. For exam- interested in whether people would transfer the so-
ple, if a person has to be explicitly prompted to use lution proposed in the military problem to the radia-
pliers to create a pendulum in Maier's two-cord tion problem. For example, a solution suggested by
problem, his or her solution would not be consid- the military situation is to have different beams of
ered as creative as a solution that is produced with- low-intensity radiation converge at the tumor and
out such prompting. produce sufficient energy to destroy the tumor.
The preceding discussion suggests that ecologi- Gick and Holyoak (1980) found that even when
cally relevant investigations of creative problem people memorized the military story beforehand,
solving may involve different experimental pro- they did not frequently transfer the solution to the
cedures than are typically used in other studies of radiation problem unless they were given a hint to
learning and memory. Specifically, it seems impor- use the prior story to help them solve the problem.
tant to investigate factors that affect the spon- The failure to find evidence of spontaneous trans-
taneous transfer of knowledge to novel problem- fer has been noted in other studies as well. For
solving tasks. In spontaneous transfer experiments, example, Maier (1931) found that people often
people are not given hints to use specific informa- failed to spontaneously transfer construction princi-
tion or experiences. ples presented at the beginning of an experiment to
a novel construction problem unless given specific
directions to do so. Weisberg et at. (1978) also
Investigations of Spontaneous Transfer
reported little spontaneous transfer of clues to a
One method that has been used to investigate the novel problem-solving task.
spontaneous transfer of knowledge to problem- Otherresearch by Perfetto et at. (1983) examined
solving tasks is first to present information relevant the frequency of spontaneous transfer in situations
to the solution of a problem and then, at some later that involved a clear connection between the clues
point, to present the problem (e.g., Gick & presented beforehand and a subsequent problem-
Holyoak, 1980; Maier, 1931; Perfetto et at., 1983; solving task. For example, they presented people
Weisberg, DiCamillo, & Phillips, 1978). For ex- with problems, such as "Uriah Fuller, the famous
ample, Gick and Holyoak (1980) presented people Israeli superpsychic, can tell you the score of any
with a descriptive passage during the initial phase of baseball game before the game starts. What is his
their experiment. This passage described a situation secret?" Prior to solving the problems, people were
(a problem together with a solution) that was analo- given a list of clues that included such statements as
gous to a problem presented later in the experiment. "Before it starts, the score of any game is 0 to 0."
One of the passages presented early in the experi- Perfetto et at. (1983) found that people did not
ment involved a story about a general who was transfer the clues to the problem-solving task unless
trying to capture a fortress. The general had a suffi- they were specifically informed about the relevance
of the information presented earlier in the not perform appropriate elaboration activities that
experiment. are needed for transfer.
The findings discussed above suggest that the In learning environments in which people have
spontaneous transfer of information in memory to a clear goals and expectations for using information,
novel problem-solving situation can be a difficult it is more likely that they will perform appropriate
process. In fact, if these results are indicative of the elaboration activities. For instance, creative prob-
frequency of spontaneous transfer in situations out- lem solvers might think of potential applications for
side of the laboratory, one might expect very little the information they are studying. These elabora-
creative transfer of knowledge. tions could increase the similarity between the in-
formation represented in memory and the informa-
tion available in subsequent problem-solving tasks.
The similarity in appearance between a new situa-
Learning Activities That Facilitate Transfer
tion and a previously experienced event has been
One reason people may have failed to spon- found to facilitate spontaneous remembering (e.g.,
taneously transfer information in the studies de- Gentner & Landers, 1985; Ross, 1984). Indeed,
scribed above is that there was little perceived many mnemonic strategies are based on the princi-
similarity between the original learning task and the ple of elaborating new information with cues that
subsequent problem-solving task. Indeed, other in- are available in the retrieval environment (e.g.,
vestigations of retention demonstrate that the ac- Bransford & Stein, 1984; Hayes, 1981). However,
cessibility of information in memory is influenced the creative transfer of memory to novel problems
by the similarity between information encoded dur- often requires more than just accessing a relevant
ing acquisition and the information available at the concept in an appropriate context.
time of test (e.g., Barclay, Bransford, Franks, Mc- Creative transfer also depends on the appropriate
Carrell, & Nitsch, 1974; Tulving & Thomson, application of concepts (Simon, 1980). For exam-
1973). The latter investigations suggest that the ple, a student may remember an appropriate physics
probability of transfer in problem-solving tasks formula during problem solving but be unable to
should increase as the similarity between the learn- apply it to the problem situation. The successful
ing and transfer tasks increases. transfer of information to novel situations depends
Although creative thinking implies an ability to on understanding the particular relevance of a con-
bridge the gap between previous experience and cept or strategy (e.g., Brown, Bransford, Ferrara,
new situations, it is not clear when and how creative & Campione, 1983). Elaborations that help people
people prepare for bridging that gap. There is abun- understand the relevance of concepts for solving
dant evidence to indicate that people spontaneously specific problems may facilitate the spontaneous
elaborate information that they are trying to under- transfer of knowledge to similar problem situations.
stand and learn (e.g., Franks, Vye, Auble,
Mezynski, Perfetto, Bransford, Stein, & Lit-
An Investigation of Elaboration and
tlefield, 1982; Pichert & Anderson, 1977; Rohwer,
Spontaneous Transfer
1980; Rumelhart & Ortony, 1977; Stein, Brans-
ford, Franks, Owings, Vye, & McGraw, 1982). Stein, Way, Benningfield, and Hedgecough
For example, Stein et al. (1982) found that suc- (1986) examined the effects that different types of
cessful learners tend to elaborate confusing rela- elaborations have on the spontaneous transfer of
tionships in ways that help them better to under- information to novel problem-solving tasks. In this
stand and remember those relationships. There is study, people were asked to provide plausible ex-
also evidence to indicate that the elaborations that planations for seemingly implausible events (e.g.,
people spontaneously generate during learning are "The home was small because the sun came out").
influenced by their interests, goals, and expecta- These problems were adapted from a set of diffi-
tions (e.g., Anderson, Reynolds, Schallert, & cult-to-comprehend sentences used by Auble,
Goetz, 1977; Pichert & Anderson, 1977). The Franks, and Soraci (1979). Prior to the presentation
failure to find spontaneous transfer in certain labo- of these riddles, the participants were presented
ratory experiments may result from people not hav- with clues that could help them solve the problems
ing clear expectations about how information (e.g., "igloo"). These clues were embedded in one
(problem clues) might be used. Without appropriate of four different types of elaborative contexts.
goals for comprehension and transfer, people may The elaborative contexts used by Stein et al.
(1986) differed in terms of whether they used con- relevant properties of the key concepts) primarily
tent words that were similar or dissimilar to those influenced the appropriate application of the key
found in the problem statements (surface structure concepts to the problems. The fact that little or no
similarity). The elaborative contexts also differed spontaneous transfer was found when the elabora-
in terms of whether they emphasized relevant or tions did not emphasize relevant properties of the
irrelevant properties of the key concepts that were key concepts (similar or dissimilar surface struc-
needed to solve the problems (contextual rele- tures) suggests the relative importance of learning
vance). For example, given a problem, such as experiences that help people understand how to ap-
"The home was small because the sun came out, " a ply concepts to novel problems.
statement with a relevant context and dissimilar sur- More recent investigations provide additional
face structure is "An igloo can be damaged by support for the idea that spontaneous transfer in
heat. " This elaborative statement uses the key word problem-solving tasks is affected by the similarities
igloo in a relationship that emphasizes the de- between how information is elaborated during prior
leterious effects of heat on that type of structure. An learning experiences and how the problem is en-
example of a statement with a relevant context and a coded on the subsequent transfer task. For example,
similar surface structure is "An igloo is a home that Holyoak and Koh (1987) found that similarities in
can be damaged by the sun. " The latter statement surface structure and in the deeper relational prop-
emphasizes relevant properties of the key concept erties of clues and problems affect spontaneous
(igloo), and it includes two content words that are transfer. Other research by Adams, Kasserman,
also found in the problem statement (e.g., home, Yearwood, Perfetto, Bransford, and Franks (1988)
sun). In contrast, consider a statement with an irrel- and Lockhart, Lamon, and Gick (1988) indicates
evant context and dissimilar surface structure, such that spontaneous transfer can be enhanced when the
as "Some eskimos live in an igloo." The latter clue information is presented in the same type of
statement does not emphasize properties of an igloo question-answer format that is found in subsequent
relevant to the problem statement, nor does it in- problem-solving tasks.
clude similar surface structure features.
Stein et al. (1986) found that elaborative state-
Differences between Informed and
ments that did not prompt people to consider rele-
Spontaneous Transfer
vant properties of the key concepts resulted in little
or no spontaneous transfer. These findings are con- The findings discussed above seem particularly
sistent with other research that has failed to demon- significant because many studies have been unable
strate spontaneous transfer effects (e. g., Gick & to find evidence of spontaneous transfer in novel
Holyoak, 1980; Perfetto et al., 1983; Weisberg et problem-solving tasks. Investigating the factors
al., 1978). Stein et al. (1986), however, also found that affect spontaneous transfer can also be difficult
that the elaborations that prompted people to con- because laboratory testing environments often pro-
sider relevant properties of the key concepts re- vide implicit cues about the relatedness of initial
sulted in significant levels of spontaneous transfer learning experiences and subsequent transfer tasks.
particularly when the elaborative contexts and the For example, consider a study conducted by Gick
problem statements had similar surface structures and Holyoak (1983). They asked people to solve
(Experiments 1 and 2). These findings demonstrate problems such as Duncker's (1945) "radiation
that the spontaneous transfer of information to problem" after receiving various types of learning
novel problems can be facilitated by elaborative experiences with analogous problems (that in-
activities that are performed during learning. cluded solutions). An important feature of Gick and
Stein et al. (1986, Experiment 3) also examined Holyoak's (1983) study is that it examined transfer
the different effects that elaborations can have on in situations in which people were not informed
accessing information and appropriately applying about the relevance of the analogous stories for
information to problem-solving tasks. For exam- solving the radiation problem.
ple, the similarity in surface structure between elab- Gick and Holyoak (1983) found that transfer in
orations and problem statements primarily influ- the uninformed conditions was facilitated by certain
enced the accessibility of clue information during types of learning experiences. For example, little
problem solving (see also Gentner & Landers, spontaneous transfer was found when the partici-
1985). In contrast, the contextual relevance of elab- pants in the study simply read the analogous stories
orations (i.e., the extent to which they emphasized prior to solving the radiation problem; however,
considerably more spontaneous transfer was ob- trast, problem-solving performance in the spon-
served when people received two examples of anal- taneous transfer conditions was unrelated to
ogous stories and accurately summarized the ab- whether the relevant clues could be recalled prior to
stract relationship that was common to both stories the problem-solving task.
prior to solving the radiation problem. The relationship Stein et al. (1986) observed be-
The findings reported by Gick and Holyoak tween problem-solving performance and free recall
(1983) appear to be relevant for understanding the in the informed transfer conditions may result from
constraints that affect the creative transfer of a particular strategy that people use in informed
knowledge because people transferred the solution transfer tasks. For instance, consider that in most
strategies that were described in the analogies with- laboratory studies of transfer the relevant clues for
out any hints. More recent research by Spencer and solving problems are presented within the context
Weisberg {1986) suggests, however, that people of a relatively small set of irrelevant information. If
may often implicitly assume that information pre- people know that some of the information presented
sented earlier in an experiment is relevant to subse- earlier may facilitate problem solving, they could
quent tasks because of the continuity of the experi- serially recall each clue presented and evaluate its
mental setting. Spencer and Weisberg (1986) used a applicability to the problem. The effectiveness of
procedure similar to that employed by Gick and this strategy would depend on knowing that the
Holyoak (1983) but that also included an experi- information presented earlier in the experiment was
mental condition in which the acquisition and test- relevant to the problem-solving task and on being
ing contexts were made to seem clearly unrelated able to consciously recall that information during
(i.e., they used different experimenters and testing problem solving. In situations in which people use
rooms). Although Spencer and Weisberg (1986) such a strategy, one would expect to find a strong
were able to replicate Gick and Holyoak's (1983) relationship between free recall performance and
findings in situations that involved highly similar problem-solving performance.
acquisition and testing contexts, they did not repli- Although a problem-solving strategy based on
cate Gick and Holyoak's results when the acquisi- free recall may be effective in certain laboratory
tion and testing contexts were made to seem situations, it is doubtful whether such a strategy
unrelated. would be useful in other natural problem-solving
The results of Spencer and Weisberg's (1986) situations that require creativity. For instance, it
study raise the question whether the learning expe- would be difficult to apply a strategy based on serial
riences that Gick and Holyoak (1983) found to af- recall to problem-solving tasks that involve a large
fect uninformed transfer in the laboratory will also set of potentially relevant experiences (i.e., the
affect spontaneous transfer in creative problem- content of a semester course or a college educa-
solving situations outside of the laboratory. An tion). In problem-solving situations in which there
important issue to consider from the standpoint of is a relatively large set of potentially relevant expe-
designing research to investigate the creative trans- riences to consider, it would be difficult, if not
fer of knowledge is whether the processes that un- impossible, to recall each experience and then ana-
derlie transfer in informed conditions are funda- lyze how it might be appropriately applied to the
mentally different from the processes that underlie problem. In fact, the benefits of being prompted to
spontaneous transfer. consider a set of previous experiences in any prob-
From one perspective, it could be argued that lem-solving situation may diminish as the number
both informed and spontaneous transfer tasks proof potentially relevant experiences in that set
vide information about constraints that affect the increases.
creative use of knowledge, as both tasks assess
transfer to novel problems. Alternatively, one
Effects of Set Size and Similarity on Informed
could argue that there are some important dif-
ferences in how information may be accessed in and Spontaneous Transfer
informed and spontaneous transfer situations. For Stein and Hedgecough (submitted) conducted
example, Stein et ai. (1986, Experiment 3) found two experiments to examine differences in how
that problem-solving performance in informed knowledge is accessed in informed and in spon-
transfer conditions was highly related to whether taneous transfer tasks. In the first experiment, the
the problem clues were remembered on a free recall effect of acquisition set size was examined. It was
test given prior to the problem-solving task. In con- hypothesized that increasing the amount of irrele-
vant information presented together with problem- informed conditions with larger acquisition set
solving clues would negatively affect informed sizes (31 %) than in the informed conditions with
transfer because it would make it more difficult to smaller acquisition set sizes (52%). In contrast, ac-
recall each relevant clue. In contrast, acquisition set quisition set size did not significantly affect transfer
size was not expected to affect spontaneous trans- in the uninformed conditions (20% vs. 26%).
fer, as the latter task does not involve explicit The fact that informed transfer was negatively
memory. affected by larger acquisition sets is consistent with
Experiment 1 employed a transfer task similar to the notion that people may use a consciously di-
that used by Stein et ai. (1986, Experiment 1). Each rected recall strategy to access information in in-
of the 80 undergraduates who participated in the formed tasks. As acquisition set size increases, this
study read a list of statements that contained clues recall-based strategy becomes less effective be-
(e.g., "igloo") to help them solve problems pre- cause it is more difficult to remember each episode.
sented later in the experiment (e.g., "The home Spontaneous transfer, on the other hand, is not as
was small because the sun came out. Can you ex- strongly affected by acquisition set size. One expla-
plain?"). The clues were embedded in two types of nation for the latter finding is that spontaneous
elaborative statements that Stein et ai. (1986) found transfer involves an implicit knowledge activation
to influence transfer. The statements either prompt- process that is not dependent on consciously di-
ed people to consider properties of the key concept rected recall. This implicit activation process is pre-
relevant to the problem (e.g., "An igloo can be sumably dependent on the similarity between infor-
damaged by heat") or prompted people to consider mation encoded during problem solving and during
irrelevant properties of the key concept (e.g., previous experience (see also Holyoak & Koh,
"Some Eskimos live in an igloo"). The clue state- 1987).
ments were presented in a list with either 10 or 40 To investigate potential differences in how sim-
unrelated filler statements during the first phase of ilarity might affect informed and spontaneous trans-
the experiment. Before the transfer problems were fer, Stein and Hedgecough (Experiment 2, submit-
presented, half of the participants in the study were ted) varied the surface structure similarity of clues
given hints to consider the information presented and problems. Note that in earlier research it was
earlier. In the spontaneous transfer conditions, the found that manipulating deeper relational sim-
participants were led to believe that the two tasks ilarities between clues and problems significantly
were unrelated. All treatments were manipulated influenced the ability to apply accessible clue infor-
between subjects. mation to the problems. By varying the surface
The mean percentage of problems solved cor- structure similarity of clues and problems (e.g., the
rectly in each condition is presented in Table 1. use of two or three identical content words in the
There was significantly (p < .05) less transfer in the clue and problem statements), Stein and Hedge-
cough hoped to clarify potential differences in the
way similarity affects access to relevant knowledge
Table 1. Mean Percentage of Problems in informed and spontaneous transfer tasks.
Solved a The procedure was similar to that of Experiment
1 except that only relevant clue contexts were used.
Clue statement As mentioned earlier, Experiment 2 varied the sur-
Relevant Irrelevant
face structure similarity of clue and problem state-
ments by using two or three identical content words
Group contexts contexts
in the clue and problem statements. In addition,
Informed Experiment 2 assessed each participant's ability to
15 Items 68 36 recall the clues before the problem-solving task in
45 Items 50 12 order to assess the relationship between the ability
Uninformed to recall clue information and subsequent transfer.
15 Items 40 12 The mean percentage of problems solved cor-
45 Items 32 8 rectly in each condition is presented in Table 2. As
in Experiment 1, larger acquisition sets yielded sig-
aFrom Differences between Spontaneous and Informed Trans-
fer in Problem-Solving Tasks: The Effects of Set Size and Clue-
nificantly less transfer in the informed conditions
Problem Similarity (Experiment 1) by B.S. Stein and C.A. (51.4%) than smaller acquisition sets (65.3%). In
Hedgecough, submitted. contrast, there was little difference in the amount of
Table 2. Mean Percentage of Problems The findings discussed above indicate that there
Solved a are important differences between how knowledge
is accessed in informed laboratory tasks and spon-
Clue statement taneous transfer tasks. The differences between in-
surface structure formed and spontaneous transfer appear to be most
Group Similar Dissimilar dramatic when small acquisition sets are used in an
experiment. These differences may become less
Informed significant, however, as increasing amounts of in-
15 Items 70.8 59.7 formation are learned. For example, in the larger
45 Items 62.5 40.3 acquisition set conditions, Stein and Hedgecough
Uninformed found that surface structure similarity significantly
15 Items 44.5 19.5 affected performance in both the informed and un-
45 Items 40.3 19.4 informed transfer conditions. Furthermore, the dif-
aFrom Differences between Spontaneous and Informed Trans- ference in prolem-solving performance between
fer in Problem-Solving Tasks: The Effects of Set Size and Clue- the informed and uninformed conditions declined
Problem Similarity (Experiment 1) by B.S. Stein and C.A. as acquisition set size increased. The latter results
Hedgecough, submitted. may reflect a change in the type of strategy used by
people in the informed transfer conditions as ac-
quisition set size increases. For instance, it may
spontaneous transfer found in conditions with large become increasingly difficult to use a strategy that
acquisition sets (29.9%) and small acquisition sets is based on consciously directed remembering
(32.0%). The results of Experiment 2 also reveal when there is a large number of potentially relevant
differences in how surface structure similarity af- experiences that must be considered during prob-
fects problem solving in informed and in spon- lem solving. In such cases, people may rely on
taneous transfer tasks. In the small acquisition set more implicit knowledge activation processes that
conditions, surface structure similarity did not sig- depend on the similarity between information en-
nificantly affect informed transfer, a finding that is coded during problem solving and during previous
consistent with the idea that access to relevant infor- experience.
mation in informed transfer tasks is accomplished
through consciously directed recall. In contrast,
surface structure similarity significantly affected Creating Opportunities for Transfer
spontaneous transfer in the small and the large ac-
quisition set conditions. The latter findings suggest The preceding discussion considered the types of
that the similarity between information encoding experimental tasks that can be used to investigate
during problem solving and during previous experi- the effects of different learning experiences on the
ence is an important determinant of access in spon- creative transfer of knowledge. It was noted that
taneous transfer tasks. learning experiences that help people understand
Differences between informed and spontaneous how to apply concepts to novel problem-solving
transfer tasks were also found when the relationship tasks can facilitate spontaneous transfer. In the re-
between recall and subsequent problem solving was search by Stein et al. (1986), the elaborations that
analyzed. For example, the probability of solving facilitated spontaneous transfer to novel problem-
problems in the informed conditions was signifi- solving tasks (e.g., "An igloo can be damaged by
cantly greater if the relevant clue could be recalled heat") were supplied by the experimenter. Stein
ahead of time (84.8 %) than if the clue could not be and his colleagues argued that similar elaborations
recalled (53.1 %). In contrast, there was not a sig- could also be generated by people when learning
nificant difference in the probability of solving new information. For example, it is not difficult to
problems in the spontaneous transfer conditions if imagine people who are learning about igloos ask-
the clue could be recalled ahead of time (37.7%), ing themselves such questions as What would hap-
relative to when the clue could not be recalled ahead pen to igloos if the climate suddenly changed and it
of time (34.9%). The latter findings are consistent became warmer? The latter type of question might
with the idea that consciously directed remember- prompt one to consider the deleterious effects of
ing plays a more important role in informed transfer heat on igloos. Experiences that lead people to ex-
tasks than in spontaneous transfer tasks. plore the applications and limitations of concepts
have often been considered valuable for promoting that problem solving plays in learning is discussed
understanding (e.g., Bransford, Sherwood, & Stur- by Bransford and Stein (1984). They describe five
devant, 1986). The current discussion suggests that problem-solving activities that underlie effective
such experiences may also help facilitate the cre- learning. The acronym, IDEAL, is used to repre-
ative transfer of knowledge. sent these activities. Within this approach, effective
It is important to note that there are many differ- learning depends on identifying the existence of a
ent ways of elaborating new concepts. Further- learning problem and the contextually sensitive
more, different problem-solving contexts may re- nature of transfer; defining the type of criterial task
quire very different types of elaborations to to be performed with the information; exploring
facilitate transfer. For instance, elaborations that appropriate learning strategies to accomplish these
focus learners on the sensitivity of igloos to heat goals; acting on the selected strategies; and looking
may provide little transfer to problems that require back at the effects of using these strategies to deter-
knowledge about the shape of igloos or how to build mine the effectiveness of the problem-solving
them. In other words, the properties of a concept approach.
that are relevant in one context may be itrelevant in Bransford and Stein's heuristic framework em-
another context. In view of this, the benefits of any phasizes the importance of defining one's learning
particular learning experience for creativity are goals before selecting an appropriate learning strat-
contextually limited. egy. Indeed, different learning experiences are
One obvious and yet important implication of often needed to accomplish different learning
arguments about the contextually sensitive nature goals. Note, for instance, that the learning experi-
of transfer is that learning strategies that improve ences that would enable a person to build an igloo
performance on tests of retention may not facilitate are certainly different from the learning experiences
the creative transfer of knowledge. For example, that would enable a person to solve a problem, such
Stein et al. (1986) and Weisberg et al. (1978) found as the home was small because the sun came out.
that memorizing the clues presented in a spon- From the perspective of the current chapter, the
taneous transfer study had little effect on whether actual process of defining one's learning objectives
people spontaneously transferred the memorized may be an important component of creative trans-
information to a novel problem-solving task. Other fer. For example, defining the context for transfer
research also demonstrates the limitations of learn- may help people identify potential applications of
ing strategies that are based on mnemonic tech- the concepts being studied.
niques. For instance, Stein, Brock, Ballard, and It is important to note that, in many laboratory
Vye (1987) found that learning experiences involv- studies of learning, people are not encouraged to
ing visual images can sometimes lower perfor- define their own learning goals and to identify po-
mance on tasks that assess people's understanding tential applications of the information presented.
of verbal relationships. These findings illustrate For instance, in many experiments, people are in-
the potential shortcomings of learning strategies structed to perform tasks (e.g., rating comprehen-
or learning environments that emphasize rote re- sibility) that may prevent or discourage them from
tention. considering realistic learning goals. These labora-
tory investigations may inadvertently prevent peo-
ple from engaging in the types of learning and prob-
Learning Is Problem Solving
lem-solving activities that foster creative transfer.
The contextually sensitive nature of learning and In order to better understand the constraints that
transfer suggests that it is important to select an affect the transfer or knowledge to novel problems,
appropriate learning strategy. Selecting an appro- more attention needs to be given to the decisions
priate learning strategy, however, necessarily in- that people make about learning strategies. It seems
volves some understanding of how information will particularly important to examine differences in the
be used and the effects that different learning strat- way people approach learning problems and how
egies can have on transfer. The importance of estab- these approaches to learning affect spontaneous
lishing goals and planning appropriate strategies transfer to novel problems. Investigations that ad-
during learning indicates that learning is, itself, a dress these issues may reveal important differences
problem-solving process. in the way creative and less creative individuals
One approach that emphasizes the important role define their learning goals.
The Effects of Problem Definitions on cilitate the creative use of knowledge. Given the
Transfer effects that problem definitions can have on
creativity, it seems important to examine factors
The preceding discussion considered how the that can influence the quality and diversity of prob-
creative transfer of knowledge may be influenced lem definitions. For example, consider the research
by the kinds of activities people perform during by Chi, Glaser, and Rees (1982) that examined dif-
learning. The creative transfer of knowledge may ferences between novice and expert problem
also be affected by the activities people perform in solvers within the domain of physics. Chi et al.
the transfer task. For example, how people define (1982) found that novices often have a limited abil-
and interpret a problem situation can affect what ity to generate inferences and relations not ex-
information is accessed during problem solving plicitly stated in the problem because of their less
(e.g., Adams, 1979; Bransford & Stein, 1984). developed knowledge base. In contrast, the more
To illustrate the effects problem definitions can developed knowledge base of experts permits great-
have on the type of knowledge accessed in problem- er elaboration of the problem situation (see also de
solving situations, consider the experiences of a Groot, 1965; Newell & Simon, 1972). An expert's
group working at Tennessee Technological Univer- ability to elaborate new problems could allow them
sity to develop new ways of protecting people from to bridge the gap between novel problems and pre-
injury in automobiles. In the initial stages of prob- vious experience by framing the problem in terms
lem solving, the group found that many of the con- of concepts and relationships that correspond more
cepts people were considering (e.g., safety nets, closely to previous experience. Indeed, one might
automatically activated harnesses, force fields) predict that an expert's ability to transfer knowl-
were derived from a particular way of defining the edge to problems in that domain should be less
problem (e.g., how to restrain people during a affected by surface structure similarities than some-
crash). It was not until people considered alter- one with less expertise in that domain. In fact, re-
native ways of defining the problem (e.g., how to cent research on spontaneous transfer by Novick
make the interior of a car less likely to cause injury (1988) provides some support for the latter idea.
to an unrestrained person or how to help people Novick (1988) found that people with higher levels
avoid crashes) that other types of creative solutions of expertise in a domain were less likely to be af-
were considered. fected by surface structure similarities between new
Adams (1979) also provided examples of how problems and previous experience than people with
problem definitions can affect the type of solutions lower levels of expertise in the problem domain. In
that are considered. In one example, he described contrast, people with higher levels of expertise in
the experiences of a group of engineers who were the problem domain were influenced more by deep-
trying to design an improved method for mechan- er relational similarities between the problems and
ically picking tomatoes that would be less likely to previous experience than were people with lower
damage the fruit. The engineers considered a vari- levels of expertise in the problem domain. These
ety of strategies for improving mechanical pickers, findings suggest that the knowledge and skills ac-
such as increasing the padding on the picking arms quired from past experience can provide the con-
and reducing the speed of the picking arms. An ceptual tools needed to define problems in ways that
alternative way of defining the problem facing the can improve transfer to novel situations.
tomato pickers is How can they make the tomatoes Although the knowledge and skills an individual
less likely to be bruised while being picked mechan- acquires in a content domain can affect how prob-
ically? This alternative definition of the problem lems are defined and elaborated, there are also other
suggested other creative ideas for solving the prob- variables that affect the quality and diversity of
lem, such as developing new strains of tomatoes problem definitions. For instance, it was noted ear-
that are less likely to be bruised. lier that people may fail to consider alternative ap-
proaches to problem solving because of their famil-
iarity with a particular approach. People may also
fail to consider alternate approaches to problems
Generating Alternative Definitions of Problems
because of implicit assumptions about how a prob-
The examples discussed above demonstrate how lem should be defined (e.g., Adams, 1979;
different definitions of a problem situation can fa- Bransford & Stein, 1984).
One way to achieve greater creativity in problem on creative transfer. For instance, can training with
solving is to state explicitly the assumptions and heuristic frameworks like that proposed by
definitions that are guiding one's thinking (Brans- Bransford and Stein (1984) facilitate creative
ford & Stein, 1984; Polya, 1957). For instance, transfer?
Bransford and Stein (1984) suggested that develop- It was noted earlier that the types of cognitive
ing creative solutions to problems involves such activities that are performed during the acquisition
activities as defining the problem in different ways of new information can also affect creative transfer.
and then evaluating the adequacy of those defini- Indeed, learning experiences that optimize the
tions by examining the appropriateness of the as- transfer of information to certain tasks may not pro-
sumptions in each approach. It was noted earlier mote transfer to other tasks. The idea that there does
that creative problem solving often involves look- not appear to be any single best learning strategy
ing back at the assumptions that limit the types of that can be used to optimize transfer to all problem
solutions being considered (e.g., the only effective situations is consistent with theoretical frameworks
way to prevent automobile injuries is by restraining like that which was proposed by Jenkins (1979).
the occupants). By making problem definitions ex- One implication of contextualism is that we should
plicit and by evaluating the assumptions underlying stop searching for an optimal learning activity or
those definitions, people can recognize the limita- strategy that will enhance transfer to all tasks. In-
tions of particular approaches and begin to explore stead, future research should attempt to identify the
other definitions of the problem that may lead to potential benefits of different learning activities and
creative solutions (e.g., How to prevent accidents? delineate the conditions under which such activities
How to reduce the velocity of automobiles immedi- promote transfer. The latter approach promotes re-
ately before a crash? How to position people so that search exploring "spheres of influence" for partic-
they are less likely to suffer severe injuries in a ular learning activities/strategies rather than a
crash?). search for the single "best" learning strategy.
An emphasis on the contextual nature of learning
and transfer also illustrates the value of treating
learning itself as a problem-solving experience. For
Implications for Future Research instance, successful learners must define their
learning goals and select an appropriate learning
The preceding discussion considered how vari- strategy that will optimize transfer to future prob-
ous cognitive activities that are performed during lem-solving tasks. Further research is needed to ex-
the acquisition of new information and during at- plore differences in the way people define learning
tempts to solve problems can affect the creative problems and modify their methods of study to ac-
transfer of knowledge. For example, it was noted complish different goals. For example, my col-
that different definitions of a problem situation can leagues and I are currently exploring differences in
influence the type of knowledge that is accessed the way that successful and less successful students
during problem solving. However, further research approach various learning problems. Our prelimi-
is clearly needed to explore the relationship be- nary findings indicate that successful students tend
tween problem definition activities and the creative to be better able to define learning problems and are
transfer of knowledge. For instance, are creative more flexible in selecting strategies to accomplish
abilities related to the way individuals approach and different learning goals than less successful stu-
define problem situations (e.g., Getzels & Csiks- dents. It is also important to study the types of
zentmihalyi, 1975)? learning activities that creative individuals use in
Further research is also needed to explore how natural learning contexts to optimize transfer. The
people can be prompted to consider alternative latter type of research may help clarify whether
ways of defining problems that might enhance the creative individuals are more likely to identify or
creative transfer of knowledge. For example, pre- anticipate potential problem situations that new in-
liminary findings indicate creative transfer can be formation may help them solve (e.g., Stein et al.,
enhanced by prompting people to consider alter- 1986).
native definitions of problems during problem solv- Finally, an emphasis on the contextual nature of
ing (Stein & Moore, unpublished manuscript). Ad- learning and transfer suggests that to understand
ditional research is needed to evaluate the effects of creative transfer in natural environments, our inves-
more general training in problem definition skills tigations must incorporate testing contexts that re-
flect real-world problem-solving constraints. In A., Bransford, J. D., & Franks, J. J. (1988). Memory access:
most natural problem-solving tasks, the problem The effects of fact-oriented versus problem-oriented acquisi-
solver must spontaneously access and apply appro- tion. Memory and Cognition, 16, 167-175.
Amabile, T. M. (1983). The social psychology of creativity.
priate information to the problem; however, in New York: Springer-Verlag.
many laboratory investigations, the participants are Anderson, R. c., Reynolds, R. E., Schallert, D. L., & Goetz, E.
given direct or indirect hints to use specific clues to T. (1977). Frameworks for comprehending discourse.
solve the problem. There appear to be some impor- American Educational Research Journal, 14, 367-382.
Auble, P. M., Franks, J. J., & Soraci, S. A. (1979). Effort
tant differences between the constraints that govern
toward comprehension: Elaboration of "aha"? Memory and
the spontaneous transfer of knowledge to problem- Cognition, 7(6), 426-434.
solving tasks and the constraints that govern trans- Barclay, J. R., Bransford, J. D., Franks, J. J., McCarrell, N. S.,
fer in situations in which people are informed about & Nitsch, K. E. (1974). Comprehension and semantic flexi-
the relevance of previously studied information. bility. Journal of Verbal Learning and Verbal Behavior, 13,
471-481.
Specifically, transfer in informed problem-solving Bartlett, F. C. (1932). Remembering: A study in experimental
tasks may be directly affected by the ability to inten- and social psychology. New York: Macmillan.
tionally recall previously presented clue informa- Birch, H. G., & Rabinowitz, H. S. (1951). The negative effect
tion. In contrast, the spontaneous transfer of pre- of previous experience on productive thinking. Journal of
viously learned information to problem-solving Experimental Psychology, 41, 121-125.
Bransford, J. D. (1979). Human cognition: Learning, under-
tasks does not appear to be related to factors that standing, and remembering. Belmont, CA: Wadsworth.
influence free recall performance (e.g., set size). Bransford, J. D., & Stein, B. S. (1984). The ideal problem
solver: A guide for improving thinking, learning, and
creativity. New York: W. H. Freeman.
Bransford, J. D., Sherwood, R. D., & Sturdevant, T. (1986).
Conclusion Teaching thinking and problem solving. InJ. B. Baron &R. J.
Sternberg (Eds.), Teaching thinking skills (pp. 162-181).
That previous experience can have both positive New York: W. H. Freeman.
and negative effects on creativity is a consequence Brown, A. L., Bransford, J. D., Ferrara, R. A., & Campione, J.
C. (1983). Learning, remembering and understanding. In J.
of the contextually sensitive nature of learning and
H. Flavell & E. M. Markman (Eds.), Carmichael's manual of
transfer. Although creativity implies an ability to child psychology (Vol. I). New York: Wiley.
bridge the gap between previous experience and Campbell, D. (1960). Blind variation and selective retention in
novel problems, it is not clear when or how creative creative thought as in other knowledge processes.
people prepare for bridging that gap. This chapter Psychological Review, 67, 380-400.
Chi, M. T. H., Glaser, R., & Rees, E. (1982). Expertise in
considered how the activities people perform dur- problem solving. In R. J. Sternberg (Ed.), Advances in the
ing learning and problem solving can affect the psychology of human intelligence (pp. 7-76). Hillsdale, NJ:
transfer of knowledge to novel problems. For ex- Erlbaum.
ample, it was noted that people may enhance the de Groot, A. D. (1965). Thought and choice in chess. The
creative transfer of knowledge during learning by Hague: Mouton.
Duncker, K. (1945). On problem solving. Psychological MQno-
imagining appropriate applications of the informa- graphs, 58, 270.
tion that is being studied. It was also noted that Franks, J. J., Vye, N. J., Auble, P. M., Mezynski, K. J., Perfet-
people's definitions of problem situations can re- to, G. A., Bransford, J. D., Stein, B. S., & Littlefield, J.
strict or enhance the creative application of knowl- (1982). Learning from explicit versus implicit texts, Journal
edge. Clearly, additional research is needed to in- of Experimental Psychology: General, 111, 414-422.
Gentner, D., & Landers, R. (1985, November). Analogical ac-
vestigate these issues and to examine factors that cess: A good match is hard to find. Paper presented at the
affect transfer in natural problem-solving contexts. annual meeting of the Psychonomic Society, Boston, MA.
Getzels, J. W. & Csikszentmihalyi, M. (1975). From problem
ACKNOWLEDGMENTS. I thank Karla Brock and solving to problem finding. In I. A. Taylor & J. W. Getzels
(Eds.), Perspectives in creativity (pp. 90-116). Chicago:
Ada Haynes for their helpful comments on an ear- Aldine.
lier version of this chapter. Gick, M. L., & Holyoak, K. J. (1980). Analogical problem
solving. Cognitive Psychology, 12, 306-355.
Gick, M. L., & Holyoak, K. J. (1983). Schema induction and
analogical transfer. Cognitive Psychology, 15, 1-38.
References Hayes, J. R. (1981). The complete problem solver. Philadelphia:
Franklin Institute Press.
Adams, 1. L. (1979). Conceptual blockbusting (2nd ed.). New Holyoak, K. J., & Koh, K. (1987). Surface and structural sim-
York: Norton. ilarity in analogical transfer. Memory and Cognition, 15,
Adams, L. T., Kasserman, J. E., Yearwood, A. A., Perfetto, G. 332-340.
Jenkins, J. J. (1974). Remember that old theory of memory? Rosnow, R. L., & Georgoudi, M. (1986). The spirit of con-
Well, forget it! American Psychologist, 29, 785-795. textualism. In R. L. Rosnow & M. Georgoudi (Eds.),
Jenkins, J. J. (1979). Four points to remember: A tetrahedral Contextualism and understanding in behavioral science (pp.
model of memory experiments. In L. S. Cermak & F. I. M. 3-24). New York: Praeger.
Craik (Eds.), Levels of processing and human memory (pp. Ross, B. H. (1984). Remindings and their effects in learning a
429-446). Hillsdale, NJ: Erlbaum. cognitive skill. Cognitive Psychology, 16, 371-416.
Johnson, M. K., Bransford, J. D., & Solomon, S. (1973). Mem- Rumelhart, D. E., & Ortony, A. (1977). The representation of
ory use for tacit implications of sentences. Journal of Experi- knowledge in memory. In R. C. Anderson, R. J. Spiro, & W.
mental Psychology, 98, 203-205. E. Montague (Eds.), Schooling and the acquisition of know 1-
Lockhart, R. S., Lamon, M., & Gick, M. L. (1988). Conceptual edge (pp. 99-135). Hillsdale, NJ: Erlbaum.
transfer in simple insight problems. Memory and Cognition, Simon, H. A. (1980). Problem solving and education. In D. T.
16,36-44. Tuma & F. Reif (Eds.), Problem solving and education: Is-
Loftus, G. R., & Palmer, J. C. (1974). Reconstruction of auto- sues in teaching and research (pp. 81-96). Hillsdale, New
mobile destruction: An example of the interaction between Jersey: Erlbaum.
language and memory. Journal of Verbal Learning and Ver- Spencer, R. M., & Weisberg, R. W. (1986). Context-dependent
bal Behavior, 13, 585-589. effects on analogical transfer. Memory and Cognition, 14,
Luchins, A. S. & Luchins, E. H. (1950). New experimental 442-449.
attempts at preventing mechanization in problem solving. Stein, B. S. (1978). Depth of processing reexamined: The effects
Journal of General Psychology, 42, 279-297. of precision of encoding and test appropriateness . Journal of
Maier, N. R. F. (1931). Reasoning in humans. II. The solution Verbal Learning and Verbal Behavior, 17, 165-174.
of a problem and its appearance in consciousness. Journal of Stein, B. S., & Hedgecough, C. A. (submitted) Differences
Comparative Psychology, 12, 181-194. between spontaneous and informed transfer in problem-solv-
Morris, C. D., Bransford, J. D., & Franks, J. J. (1977). Levels ing tasks: The effects of set size and clue-problem similarity.
of processing versus transfer appropriate processing. Journal Stein, B. S., & Moore, S. (1988). The effects ofproblem defini-
of Verbal Learning and Verbal Behavior, 16, 519-533. tions on transfer in problem-solving tasks. Unpublished
Newell, A., & Simon, H. (1972). Human problem solving. manuscript.
Englewood Cliffs, NJ: Prentice-Hall. Stein, B. S., Bransford, J. D., Franks, J. J., Owings, R. A.,
Novick, L. R. (1988). Analogical transfer, problem similarity, Vye, N. J., & McGraw, W. (1982). Differences in the preci-
and expertise. Journal of Experimental Psychology: Learn- sion of self-generated elaborations. Journal of Experimental
ing, Memory and Cognition, 14, 510-5'20. Psychology: General, Ill, 399-405.
Perfetto, G. A., Bransford, J. D., & Franks, J. J. (1983). Con- Stein, B. S., Way, K. R., Benningfield, S. E., & Hedgecough,
straints on access in a problem solving context. Memory and C. A. (1986). Constraints on spontaneous transfer in problem
Cognition, 11, 24-31. solving tasks. Memory and Cognition, 14, 432-441.
Pichert, J., & Anderson, R. C. (1977). Taking different perspec- Stein, B. S., Brock, K. F., Ballard, D. R., & Vye, N. J. (1987).
tives on a story . Journal ofEducational Psychology, 69, 309- Constraints on effective pictorial and verbal elaboration.
315. Memory and Cognition, 15(4),281-290.
Polya, G. (1957). How to solve it. Garden City, NY: Doubleday Tulving, E., & Thomson, D. M. (1973). Encoding specificity
Anchor. and retrieval processes in episodic memory. Psychological
Roediger, H. L., & Weldon, M. S. (1987). Reversing the picture Review, 80, 352-373.
superiority effect. In M. A. McDaniel & M. Pressley (Eds.), Weisberg, R. W. (1986). Creativity: Genius and other myths.
Imagery and related mnemonic processes; theories, indi- New York: W. H. Freeman.
vidual differences, and application (pp. 151-176). New Weisberg, R. W., DiCamillo, M., & Phillips, D. (1978). Trans-
York: Springer-Verlag. ferring old associations to new situations: A nonautomatic
Rohwer, W. D., Jr. (1980). An elaborative conception of learner process. Journal of Verbal Learning and Verbal Behavior, 17,
differences. In R. E. Snow, P. A. Frederico, & W. E. Mon- 219-228.
tague (Eds.), Aptitude, learning, and instruction (pp. 23-46).
Hillsdale, NJ: Erlbaum.
CHAPTER 10
Metacognition in Creativity
Bonnie B. Armbruster
The process of creation is a cognitive process. Per- cognitive move, monitoring and evaluating the ef-
ceiving, learning, thinking, and remembering- fectiveness of cognitive strategies, and revising
this is the stuff of creativity. The creative process cognitive strategies (Baker & Brown, 1984).
involves the acquisition of knowledge and skills, Metacognition is currently a hot topic in cog-
the transformation of knowledge into new forms, nitive psychology. In the past 15 years, literally
and the rendering of these forms into a shareable hundreds of studies have been conducted in the
product. Each stage in the process entails cognition. name of metacognition. This research has had a
It seems appropriate, therefore, to inquire about a significant influence on the area of learning, partic-
cognitive model of creativity. In one way or an- ularly learning from reading. Early metacognitive
other, the chapters in this part of the book all ad- research focused on how, when, and why students
dress this issue. The purpose of this chapter is to have difficulty studying and learning. The research
examine the creative process from the perspective generally indicated that metacognition is a late-de-
of one particular aspect of cognition-metacog- veloping skill, with younger and poorer readers dis-
nition. playing less effective metacognitive skills than
older and better readers. Later research in metacog-
nition has concentrated on intervention research
What Is Metacognition? that is designed to overcome metacomprehension
problems. This research has demonstrated the value
Modem-day cognitive psychology recognizes of including metacognitive training in three areas:
the hierarchical nature of psychological processes (1) skills-training and practice in task-specific
that are involved in cognition. At the top of the strategies; (2) self-reguLation-instruction in the
hierarchy are the executive processes that oversee, orchestration, overseeing, and monitoring of skills;
regulate, and orchestrate the activities of cognition. and (3) awareness-information concerning a
These executive processes are known as metacogni- skill's evaluation, rationale, and usefulness (Baker
tion. Metacognition includes both the knowledge & Brown, 1984). Adding these metacognitive com-
and the controL that individuals have over their own ponents has contributed substantially to the success
cognitive processes. Control, or self-regulation, in- of cognitive skills training programs.
cludes setting goals and subgoals, planning the next Given the impact of metacognition on theory and
practice in the area oflearning, it seems appropriate
Bonnie B. Armbruster • Center for the Study of Reading, to inquire about the relevance of metacognition to
University of Illinois at Urbana-Champaign, Champaign, IL the study of one of the most intriguing of cognitive
61820. processes-creative thinking. Specifically, what is
177
creativity from the perspective of metacognition? pening during this learning phase? One popular
What metacognitive processes might distinguish cognitive theory is schema theory. In schema theo-
creative from less creative individuals? In attempt- ry, information is thought to be stored in the mind in
ing to answer these questions, I will use the frame- the form of abstract knowledge structures called
work of a four-part stage model of the creative pro- schemata. Rumelhart and Norman (1978) sug-
cess as proposed by Wallas (1970). gested three different kinds of learning within sche-
Wall as first proposed the model in his book, The ma theory: (1) accretion, or the encoding of new
Art of Thought, in 1926. Why use such an old information in terms of existing schemata; (2)
model? First, the model seems at least implicit in tuning, or the modification and refinement of a
what many other writers have written about the cre- schema due to its use in different situations; and (3)
ative process. Second, the self-reports of creative restructuring, or the process of creating new sche-
individuals tend to corroborate the model. Third, I mata through patterned generation (patterning by
was unable to find a better model. Finally, the stage analogy on existing schemata) or schema induction
model provides a useful way to organize a discus- (inducing from experience).
sion of the creative process. However, there is a All three types of learning are probably happen-
danger in using W alIas's model, because it implies ing during the preparation stage of creativity. Yet
that the process of creativity is linear. Creativity is the most important type of learning is undoubtedly
much more likely to be interactive and iterative, the learning that takes place prior to or during re-
with much communication among stages. With that structuring, since restructuring seems to be what
caveat in mind, I will now move to a discussion of creativity is all about.
the four stages of creativity as proposed by Wallas. Spiro and his colleagues (Spiro, Vispoel,
At least implicit in Wall as ,s writing is the idea Schmitz, Samarapungavan, & Boerger, 1987) were
that the creative process actually begins with a apparently thinking of restructuring when they re-
problem or a question. The creative act starts with ferred to the importance of a flexible knowledge
"an eager search . . . a strong conscious desire for representation. They speculated that one key factor
something-the solution of a problem or the con- affecting the ability to think independently, produc-
struction of a work of art" (Sinnott, 1970, p. 112). tively, and flexibly (i.e., to think creatively) is the
The recognition of a "strong conscious desire for flexibility with which relevant prior knowledge is
something" is the first instance of metacognition in initially represented in memory. A flexible knowl-
the creative process. The individual is aware of a edge representation is one in which fragments of
goal or purpose; this goal becomes the guiding force knowledge are not represented in a rigid, pre-
behind the rest of the creative endeavor. packaged, compartmentalized schema but can be
Wallas, however, does not include the establish- moved about and reassembled into new knowledge
ment and recognition of a goal or purpose as a spe- structures. A flexible knowledge representation in-
cific stage in his model. creases the potential for the unique recombinations
and reorderings of knowledge fragments that are the
essence of creativity.
Stage 1: Preparation
Spiro and Myers (1984) further suggested how
The first of Wallas's stages of the creative pro- individuals might attain a flexible representation of
cess is "preparation," which includes "the whole knowledge during the preparation stage: by encod-
process of intellectual education" (Wallas, 1970, ing (representing in memory) a lot of information;
p. 92). During this stage, individuals are laying the by encoding the same information in many different
foundation for their later creative acts by acquiring ways; and by using different modes or styles of
the requisite knowledge and skills of their field. thought (e.g., verbal and visual-perceptual) in en-
Creative individuals must work hard to acquire a coding information. In sum, the analysis of Spiro
particularly rich background of knowledge and ex- and his colleagues suggests that, in the preparation
perience (Ghiselin, 1952; Sinnott, 1970). As indus- stage, creative individuals construct very rich, in-
trial designer George Nelson puts it, "people iden- terconnected, flexible cognitive structures that are
tified as creative people don't just walk around and the raw materials for restructuring in the act of cre-
have ideas. The ideas are a result of a considerable ation to follow.
amount of delving into the problem" (Rosner & Metacognition may function in two ways at the
Abt, 1970, p. 256). preparation stage. First, creative individuals may
According to cognitive psychology, what is hap- be especially attuned to the state of their cognitive
CHAPTER 10 • METACOGNITION IN CREATIVITY 179
representations of knowledge. In other words, they which will fulfill the purpose or goal of the creative
may know whether their knowledge is full, rich, enterprise. Creative individuals may have a superi-
and flexible enough to provide the potential for cre- or metacognitive ability to separate the wheat from
ative restructurings. Second, creative individuals the chaff; that is, to control efficiently and effec-
may be particularly adept at controlling or regulat- tively the restructuring of their schemata.
ing the encoding of a flexible representation of Metacognition at the incubation stage of
knowledge. For example, they may know and use creativity, however, is quite unconscious. Indeed,
some of the strategies Spiro and his colleagues men- conscious attempts to guide and control creativity
tion as ways to ensure flexible knowledge. too early in the process seem doomed to failure.
Amy Lowell, for example, writing on the process of
making poetry, stated that "no power will induce it
Stage 2: Incubation
if the subconscious is not ready" (Lowell, 1952, p.
The second of Wallas's stages of the creative 112). And George Nelson claimed, "It is my expe-
process is "incubation." During the incubation rience anyway that unless you tum off the active
stage, or "gestatory period" (Cowley, 1963, p. conscious mind, [the] idea quite possibly won't ar-
256), the problem is not consciously pursued; rive" (Rosner & Abt, 1970, p. 257).
rather, incubation is characterized by the "free
working of the unconscious or partially conscious
Stage 3: Illumination
processes of the mind" (Wallas, 1970, p. 95). In-
trospecting on his own creative process, George Wallas's third stage of creativity is called "il-
Nelson states that "you tum off your conscious lumination." Illumination is "the final 'flash' or
mind with all its buzzing and scurrying, and just 'click' " (Wallas, 1970, p. 96) that is the culmina-
coast. Then the subconscious does the work" tion of the incubation stage. Illumination is inspira-
(Rosner & Abt, 1970, p. 257). tion, revelation, insight; it is the "Eureka!" or
What is the "work" the mind is doing during the "Aha!" experience. At illumination, what has pre-
incubation stage? Some scholars of creativity rec- viously been unconscious suddenly becomes fully
ognize that mental work is somehow controlled by conscious.
inherent organizational processes of the mind. For Many creative individuals have described il-
example, Sinnott (1970) wrote: lumination, yet they are clearly at a loss to explain
how it happened. The new ideas "arise spon-
Here mental work is being done. Here, quite without con- taneously in the mind, often seemingly out of noth-
scious participation, choices are being made and ideas fitted ing and at a time when a person may be thinking of
together into patterns .... One must recognize the operation in something quite different" (Sinnott, 1970, p. 109).
the unconscious of such an organizing factor, for chance alone is
not creative. Just as the organism pulls together random, form-
In the words of Tennessee Williams,
less stuff into the patterned system of structure and function in
the body, so the unconscious mind seems to select and arrange The process by which the idea for a play comes to me has always
and correlate these ideas and images into a pattern. (pp. 111- been something I really couldn't pinpoint. A play just seems to
112) materialize; like an apparition, it gets clearer and clearer and
clearer. (Plimpton, 1984, p. 84)
Apparently, during the incubation stage the inter-

connected network of flexibly organized knowl- Isaac Bashevis Singer described illumination in this
edge acquired during the preparation stage is being way:
restructured into new schemata; that is, elements of
the original representation are being recombined Somehow it [a story] comes to you. It comes "out of the
blue," as you say .... In some cases you have a feeling as if
and reordered to create new mental structures. Of some little imp or devil is standing behind you and dictating to
course, exactly how this process occurs is still you .... How our subconscious works we never really know.
unknown. No man can ever describe how an idea comes to him. (Rosner &
One noteworthy characteristic of creative indi- Abt, 1970, p. 230)
viduals during the incubation stage appears to be a
metacognitive skill-their "mastery or control" Shrouded in mystery, illumination seems to be a
over the reworking of their flexible cognitive repre- metacognitive event. Interestingly, Wallas refers to
sentations (Spiro et at., 1987). Presumably, the in- illumination as a "click," which is the same term
dividual shapes the restructurings ever closer to that Anderson (1979) chose for the awareness of suc-
cessful comprehension. Indeed, I think the clicks of between artist (or scientist) and product that John-
comprehension and of creativity are similar phe- Steiner refers to probably entails two types of meta-
nomena. In both cases, the clicks involve the recog- cognition. The first type involves verifying or mea-
nition of a coherent cognitive representation. In the suring the product against an internal standard-
case of comprehension, it is recognition of the suc- the original purpose of the creative enterprise and
cessful interpretation of information; in the case of the mental image formed during illumination. The
creativity, it is recognition of a mental representa- second type of metacognition involves verifying the
tion that fulfills, or has the potential of fulfilling, product against an anticipated external standard-a
the goal of the creative enterprise. Creative indi- would-be audience.
viduals may be particularly attuned to the click of Rosenblatt (1986) referred to these two types of
illumination. They may have a superior metacog- metacognition in the writing process when she men-
nitive awareness of when an insight is a good one, tioned two kinds of "authorial reading" (the read-
worthy of pursuit. ing by authors of their own writing during the ver-
The creative process may end at the illumination ification stage). In the first "purpose-oriented"
stage; some creations are fully formed at illumina- authorial reading, "the writer tries to satisfy, while
tion. For example, Robert Frost reported that, one refining, a personal conception" (p. 19). In the
winter evening, as he stepped outside into the snow second kind of authorial reading, "the writer dis-
for a breath of air, the entire poem, "Stopping by sociates from the text and reads it with the eyes of
Woods on a Snowy Evening," came into his head potential readers" (p. 19).
(Sinnott, 1970, p. 109). Similarly, Conrad Aiken Creative individuals mention both kinds of meta-
reflected, "It seized me at lunch, the first section, cognition in the verification process. For example,
and I had to leave the table to put it down. Then it author John Cheever seemed to be referring to met-
finished itself. In a way I had little to do with it" acognition regarding an internal standard when he
(Plimpton, 1974, p. 36). The illumination of a fully commented, "It's a question, I guess, of trying to
formed work is not the usual case, however. get it to correspond to a vision. There i~ a shape, a
Ghiselin (1952) stated that "spontaneous ap- proportion, and one knows when something that
pearance of inventions very fully formed is not ex- happens is wrong" (Plimpton, 1981, p. 132). Au-
tremely rare, but it is by no means ordinary. Spon- thor Gabriel Garcia Marquez was clearly referring
taneity is common, but what is given is usually far to metacognition regarding an external standard
from complete" (p. 15). when he said,
In general, I think you usually do write for someone. When I'm

Stage 4: Verification writing I'm always aware that this friend is going to like this or
that another friend is going to like that paragraph or chapter,
In the typical case, then, the incomplete product always thinking of specific people. In the end all books are
of illumination is subjected to a fourth and final written for your friends. (Plimpton, 1984, p. 322)
stage of the creative process, "verification." Ver-
ification involves' 'the work of verification, correc- The conscious metacognition of the verification
tion, or revision that ordinarily follows the more stage is apparently a skill that can be refined and
radical inventive activity and completes or refines perfected over time. Creative individuals may be
its product" (Ghiselin, 1952, p. 28). Although il- able to hone both their sensitivity to the internal and
lumination is a sudden, often joyful experience, external standards and their ability to respond to
verification is frequently sustained and painful. these standards. As author Arthur Koestler put it,
there is "a very curious built-in control which de-
One of the most demanding aspects of creative discipline is the velops as the years go by, with your experience"
revision process: artists and scientists clarify their condensed (Rosner & Abt, 1970, p. 152).
thoughts through the successive drafts (or revisions) of their In sum, creative individuals seem to be es-
work .... The process is like a dialogue between the artist and
his or her product. (John-Steiner, 1985, p. 75)
pecially adept at the conscious metacognitive skills
that are required during the verification stage. They
may be unusually sensitive to both internal and ex-
Verification is a stage in which metacognition ternal standards and particularly able to revise the
clearly plays a significant role. The "dialogue" creative product accordingly. Creative individuals
CHAPTER 10 • META COGNITION IN CREATIVITY 181
may also be especially good at improving these example, planning a compOSItIon, making eval-
abilities with experience and practice. uative judgments about their writing, and diagnos-
ing text problems (Scardamalia & Bereiter,
1983a,b, 1985). Similar research should be carried
out in other areas of creative endeavor.
Concluding Remarks Doubtless there are many other useful lines of
research in the area of metacognition and creativity.
The foregoing analysis suggests that metacogni- The possibilities are limited only by our creativity
tion plays a very important role in creativity. In- as researchers!
deed, metacognition is probably a significant
component of each stage in the creative process.
Therefore, it seems that future studies of creativity
might benefit from a closer look at the role of meta-
References
cognition in the creative process. Likewise, cog- Anderson, T. H. (1979). Study skills and learning strategies. In
nitive psychology might benefit from a closer look H. F. O'Neil, Jr., & C. D. Spielberger (Eds.), Cognitive and
at the creative process as a fascinating and infor- affective learning strategies (pp. 77-78). New York: Aca-
mative example of cognition and metacognition in demic Press.
Baker, L., & Brown, A. L. (1984). Metacognitive skills and
action. reading. In P. D. Pearson (Ed.), Handbook of reading re-
Personally, I would like to see research efforts in search (pp. 353-394). New York: Longman.
three areas. First, I would .like to see studies on Cowley, M. (Ed.). (1963). Writers at work: The Paris Review
expert-novice differences in metacognition and interviews (Second Series). New York: Viking Press.
creativity. What are the differences between an ex- Ghiselin, B. (1952). Introduction. In B. Ghiselin (Ed.), The
creative process: A symposium (pp. 11-31). New York: New
pert creator and a novice creator in their metacog- American Library.
nitive abilities at each stage of the creative process? John-Steiner, V. (1985) . Notebooks of the mind: Explorations of
A related question is the development of metacog- thinking. Albuquerque, NM: University of New Mexico
nitive abilities with practice. For instance, it was Press.
Lowell, A. (1952). The process of making poetry. In B. Ghiselin
suggested that the conscious metacognition of the (Ed.), The creative process: A symposium (pp. 109-112).
verification stage develops over time and with ex- New York: New American Library.
perience. Is this true of metacognition at other Plimpton, G. (Ed.). (1974). Writers at work: The Paris Review
stages? What is it that people are learning over interviews (Fourth Series). New York: Viking Press.
time? Plimpton, G. (Ed.). (1981). Writers at work: The Paris Review
interviews (Fifth Series). New York: Viking Press.
Second, I would like to see a line of research Plimpton, G. (Ed.). (1984). Writers at work: The Paris Review
directed at the purported relationship between interviews (Sixth Series). New York: Viking Press.
"flexible" knowledge and creativity. We need to Rosenblatt, L. M. (1986). Writing and reading: The transac-
research the speCUlations of Spiro and his col- tional theory. Paper presented at the Reading/Writing Con-
ference, University of Illinois at Champaign-Urbana, Illinois,
leagues that flexible knowledge is a prerequisite for
October 19-21.
schema restructuring and, hence, creativity. If this Rosner, S., & Abt, L. E. (Eds.). (1970). The creative experi-
relationship is verified, we then need to know more ence. New York: Grossman Publishers.
about how flexible knowledge is and could be Rumelhart, D. E., & Norman, D. A. (1978). Accretion, tuning,
acquired. and restructuring: Three modes of learning. In J. W. Cotton &
R. L. Klatzky (Eds.), Semantic factors in cognition.
A third potentially profitable research area is in- Hillsdale, NJ: Erlbaum.
tervention studies that are designed to teach meta- Scardamalia, M., & Bereiter, C. (l983a). Child as co-investiga-
cognitive skills in creativity. It seems likely that tor: Helping children gain insight into their own mental pro-
some of the principles of metacognitive skills in- cesses. In S. Paris, G. Olson, & H. Stevenson (Eds.),
struction that are derived from intervention studies Learning and motivation in the classroom (pp. 61-82). Hills-
dale, NJ: Erlbaum.
in learning could be fruitfully applied to creating. Scardamalia, M., & Bereiter, C. (l983b). The development of
Indeed, this potential has already been demon- evaluative, diagnostic, and remedial capabilities in children's
strated by Scardamalia and Bereiter in the area of composing. In M. Martlew (Ed.), The psychology of written
writing. These researchers have completed a language: A developmental approach (pp. 67-95). London,
England: Wiley.
number of studies in which students receive instruc- Scardamalia, M., & Bereiter, C. (1985). Development of dialec-
tion and help in various cognitive and metacog- tical processes in composition. In D. R. Olson, N. Torrance,
nitive skills that are associated with writing; for & A. Hildyard (Eds.), Literacy, language, and learning: The
nature and consequences of reading and writing (pp. 307- Spiro, R. J., Vispoel, W. L., Schmitz, J., Samarapungavan, A.,
329). Cambridge, England: Cambridge University Press. & Boerger, A. (1987). Knowledge acquisition for application:
Sinnott, E. W. (1970). The creativeness of life. In P. E. Vernon Cognitive flexibility and transfer in complex cognitive do-
(Ed.), Creativity (pp. 107-llS). Middlesex, England: Pen- mains. In B. C. Britton (Ed.), Executive control processes.
guin Books. Hillsdale, NJ: Erlbaum.
Spiro, R. J., & Myers, A. (1984). Individual differences and Wallas, G. (1970). The art of thought. In P. E. Vernon (Ed.),
underlying cognitive processes. In P. D. Pearson (Ed.), Creativity (pp. 91-97). Middlesex, England: Penguin Books.
Handbook of reading research (pp. 471-501). New York:
Longman.
CHAPTER 11
The Creative Construction

of Rationality
A PARADOX?
David Moshman and Leslie E. Lukin
Is the development of rationality a creative process? The third section considers empiricist (learning)
Is there something paradoxical about the idea that and nativist (innatist) accounts of the development
creativity can result in increasing rationality? We of reasoning. We argue that neither account is ade-
will argue in this chapter that rationality is indeed quate and that the problems with each are best ad-
creatively constructed by the developing indi- dressed by a constructivist view in which creativity
vidual. If one thinks of creativity as consisting of is not merely an incidental feature but a fundamen-
unconstrained flights of the imagination and of ra- tal characteristic of development. The fourth sec-
tionality as strict adherence to mechanical rules of tion then presents a specific constructivist account
reasoning, there is indeed something paradoxical of the development of reasoning involving the cre-
about this conclusion. The paradox dissolves, how- ative process of reflective abstraction.
ever, upon reconsideration of the definitions of both Finally, we return to the theme of whether
rationality and creativity. creativity can really lead to increasing rationality.
This chapter is organized into five major sec- We conclude that the developmental picture pre-
tions. In the first, we present a protocol from a 13- sented in this chapter, in which development is in-
year-old who is ranking seven arguments. We will herently creative and leads to systematically in-
see (a) that her analysis of the logic becomes increasing rationality, has important implications for
creasingly sophisticated over the course of working the nature of both creativity and rationality.
on the task, and (b) that this improvement is a cre-
ative, constructive process. The second section pro-
vides a more systematic review of the literature on
the development of reasoning, supporting the con- Lori: The Creative Construction of
clusion that there is systematic progress toward in- Logic
creasing rationality.
Lori was a 13-year-old seventh grader who vol-
David Moshman and Leslie E. Lukin • Department of Ed-
unteered to participate in a study of conceptions
ucational Psychology, University of Nebraska, Lincoln, NE about the nature of logic (Moshman & Franks,
68588. 1986, Exp. 1). She was asked to sort and rank argu-
183
ments and to explain the bases for her sortings and typically sorted arguments by content, distinguish-
rankings. The seven arguments, each appearing on ing, for example, arguments about people from ar-
a separate index card, were as follows: guments about animals. She was then asked to rank
the seven arguments from most logical to least log-
1. If elephants are bigger than dogs ical and to explain her ranking. In the course of
And dogs are bigger than mice attempting to generate a ranking that satisfied her,
Then elephants are bigger than mice she produced a series of six separate rankings.
2. If adults are older than babies Lori's initial ranking, from most to least logical,
And children are older than babies was 1 2 3 4 5 (6 7) (parentheses indicate tie). Her
Then adults are older than children explanation for this ranking was primarily in terms
3. If dogs are bigger than mice of the number of true statements in each argument.
And elephants are bigger than mice She considered an argument to be most logical if all
Then dogs are bigger than elephants three statements within it were empirically true,
4. If dogs are bigger than elephants next most logical if it included two true statements,
And elephants are bigger than mice less logical if it included only one true statement,
Then dogs are bigger than mice and least logical if it included no true statements.
5. If babies are older than adults Although she did hint at the concept of validity in
And babies are older than children explaining why she placed Argument 1 above Ar-
Then adults are older than children gument 2, her articulation of validity was vague and
6. If mice are bigger than dogs she saw this as a secondary consideration, useful
And mice are bigger than elephants only as a tiebreaker where two arguments had an
Then dogs are bigger than elephants equal number of true statements within them.
7. If elephants are either animals or plants Before the interviewer could move on, however,
And elephants are not animals Lori's focus shifted to the conclusions, and she de-
Then elephants are plants
cided to change her ranking to 1 2 4 5 (3 6 7),
explaining that the first four all have true conclu-
The key concept in the study was validity of in- sions whereas the latter three are illogical, and
ference. An argument is valid if its conclusion fol- equally so, in that their conclusions are empirically
lows logically from the given premises, regardless false. She still placed no. 1 above no. 2, however,
of the empirical truth or falsity of the premises and explaining more clearly now that no. 1 "says it in
conclusions. Thus, Arguments 1, 4, and 7 above the sentence" whereas in no. 2 you "just know. "
are valid even though no. 4 has a false premise Thinking this over led her to reconsider her ranking
("Dogs are bigger than elephants") and no. 7 has of her first four arguments (those with true conclu-
not only a false premise ("Elephants are not ani- sions) and resulted in a new ranking: (14) (2 5) (3 6
mals") but a false conclusion as well ("Elephants 7). Her explanation referred to the fact that in nos. 1
are plants"). Arguments 2,3,5, and 6 are invalid in and 4 the conclusion' 'has to" be, whereas in nos. 2
that their conclusions do not follow necessarily and 5 it just happens to be "right."
from their premises, even though the conclusions in At this point, she began thinking more seriously
two cases (nos. 2 and 5) happen to be empirically about no. 7 and decided to move it up to a position
true. just below nos. 1 and 4. Shortly, she put it along
Most college students showed a clear under- with them, arguing that its conclusion "has to be
standing of validity in sorting and ranking the above that way" due to the structure of the argument.
arguments, whereas none of the 20 fourth graders in Finally, she collapsed the distinction between (2 5)
the study did so. Seventh graders spanned the entire and (3 6), noting that in none of these cases is the
range, with some performing like fourth graders information sufficient for the conclusion. Thus, she
and some like college students. Of greatest interest made the transition from an initial ranking based
for present purposes are those who initially per- primarily on empirical truth to a final ranking based
formed like fourth graders but, by the end of the entirely on validity. This involved a gradual,
session, were reasoning like the college students. painstaking differentiation of validity from truth
Lori was in this latter group. Given multiple op- and an increasing sense of the former rather than the
portunities to sort sets of arguments drawn from the latter as the essence of logicality.
above seven, she had consistently failed to dis- It is important to note that at no point in this
tinguish valid from invalid arguments. Instead, she episode did the interviewer provide any feedback.
CHAPTER 11 • THE CREATIVE CONSTRUCTION OF RATIONALITY 185
His role was exclusively to write down Lori's rank- conclusions from premises without considering the
ing and explanation. Had Lori been satisfied, this empirical truth or falsity of those premises; it is
portion of the research would have ended with her concerned exclusively with the internal logic link-
explanation of her initial response. Thus, the ining sets of propositions. Induction, on the other
creasingly explicit recognition and articulation of hand, tests hypotheses against reality; it is con-
the concept of validity was an internally generated cerned not with formal relations but rather with
construction rather than a learned response. Al- empirical truth. (Note that here we are using the
though Lori's previous educational experiences term induction in a very broad sense.) We will con-
presumably had an impact on her thinking, it ap- sider development in each of these two areas.
pears that her construction of the concept of inferen-
tial validity was a genuine act of creativity.
It is also worth noting that the concept of validity Development of Deductive Reasoning
did not emerge suddenly from nowhere. Even in her Although the development of deductive reason-
initial ranking, in which Arguments 1 and 2 were ing is a complex, multifaceted process, it can be
seen as the two best because they contained three usefully construed as consisting of a sequence of
true statements, Lori placed no. 1 (the valid argu- four qualitatively distinct stages (Moshman, 1989).
ment) above no. 2, though she could only articulate
vaguely the basis for this decision. As she spon- Stage 1: Explicit Content-Implicit Inference.
taneously reconstructed her rankings, she became Although Stage 1 children (typically preschoolers)
increasingly clear in her account of validity and in can make correct inferences from a wide variety of
differentiating it from empirical truth, leading ulti- premises (Braine & Rumain, 1983; Hawkins, Pea,
mately to her classifying Argument 7 (with a pa- Glick, & Scribner, 1984), they think about content
tently false conclusion) as being just as valid as nos. rather than about the process of making inferences.
1 and 4. Her developmental transition appeared to They do not explicitly grasp the fundamental dis-
involve a continuing self-reflection in which an im- tinction between premises (the given information)
plicit recognition of logical form gradually became and conclusion (the deduced proposition). That is,
the object of explicit awareness and reconstruction, they use inference but do not think about inference.
leading to the sophisticated concept of inferential Consider, for example, the following argument:
validity. Her creativity appears to reside precisely
in this self-reflective process of making the implicit
8. Sprognoids are either animals or plants
explicit. This conception of development as a cre- Sprognoids are not animals
ative reconstruction at a higher level of abstraction Therefore, sprognoids are plants
is a major theme of the chapter to which we will
return later.
Lori's construction of validity is thus consistent Stage 1 children, given the two premises, will cor-
with our general thesis that rationality is neither rectly reach the conclusion that sprognoids are
learned nor innate but, rather, creatively con- plants. They see the conclusion, however, not as a
structed. In the next section, we will support more conclusion but as a new fact. They are not thinking
systematically the claim that reasoning typically de- about deducing a conclusion from premises; they
velops in the direction of greater rationality. We are thinking about sprognoids, animals, and plants.
will then tum to a more systematic examination of We as psychologists, of course, can determine that
the related claim that creativity, rooted in active they have in fact engaged in deductive inference.
self-reflection, is fundamental to this process. The process of inference is merely implicit in their
thinking, however. What they are explicitly aware
of is not inference but content (see Table 1).
The Development of Reasoning Stage 2: Explicit Inference-Implicit Logic.

Children at this stage, typically in the elementary
Reasoning can be divided into at least two major school years, can think about the process of in-
components: deduction and induction. Deductive ference and explicitly grasp the distinction between
reasoning deduces conclusions from given prem- premises and conclusions. In other words, they rec-
ises; inductive reasoning induces conclusions from ognize their conclusions as conclusions, as result-
empirical observations. Thus, deduction draws ing from inference. This can be seen in the fact that
Table 1. Development of Deductive Reasoning

Explicit object of understanding Knowledge implicit in reasoning (subject)
Stage 1 Content Inference: Conclusion deduced and thus dis-
Explicit content tinct from premises
Implicit inference I
t
Stage 2 Inference: Conclusion deduced from and Logic: Form of argument distinct from em-
Explicit inference thus related to premises pirical truth of premises and conclusions
Implicit logic (necessity)
~~----------------------------------~I
Stage 3 Logic: Relation of argument form and em- Metalogic: Formal logical system distinct
Explicit logic pirical truth of premises and conclusion from natural language
Implicit metaIogic (validity)
'"
I
Stage 4 Metalogic: Interrelations of logical systems
Explicit metalogic and natural languages
they, unlike Stage 1 children, distinguish conclu- that some forms of argument are valid and others
sions that are logically necessary from those that are invalid and they sharply distinguish validity from
merely reasonable, plausible, probable, or conven- empirical truth (Moshman & Franks, 1986). They
tional (Fabricius, Sophian, & Wellman, 1987; understand that to claim an argument is valid means
Moshman, 1989; Moshman & Timmons, 1982; that, due to its form, its conclusion would have to be
Pieraut-Le Bonniec, 1980; Somerville, Hadkinson, true if its premises were true, but that in fact the
& Greenberg, 1979) and recognize that inference premises may be false and, if so, the conclusion
can be a source of knowledge (Sodian & Wimmer, may be false as well. Consider, for example, the
1987). following argument:
Consider, for example, the following argument:
10. Elephants are either animals or plants
9. Sprognoids are animals or plants or machines Elephants are not animals
Sprognoids are not animals Therefore, elephants are plants
Therefore, sprognoids are plants
This argument is identical in form to Argument 8.
Stage 2 children will recognize that this argument Whereas Argument 8 presents little difficulty for
differs from the previous one in that, although the Stage 1 or Stage 2 children, however, Argument 10
conclusion has some basis in the premises, it is not does (Moshman & Franks, 1986). To recognize its
logically required by the premises. As psychol- validity in the face of the patent falsity of its conclu-
ogists, we might note that the conclusion in Argu- sion requires an explicit awareness of logical form,
ment 8 is deemed necessary due to the logical form including a sharp differentiation of logic from em-
of the argument: X is p or q; X is not p; therefore, X pirical truth. Studies of children's natural epis-
is q. It can be argued that in distinguishing neces- temologies support the view that logical and em-
sary inferences from merely plausible ones Stage 2 pirical domains of knowledge are not explicitly
children are making use oflogical form. However, differentiated before adolescence (Cummins, 1978;
although Stage 2 children are explicitly aware of Komatsu & GaIotti, 1986; Osherson & Markman,
making inferences, the form of argument, which is 1975). We have already seen, in the example of
the basis for their judgments of necessity, is merely Lori, a sample of the transition from Stage 2 to
implicit in their thinking, rather than an object of Stage 3.
explicit awareness.
Stage 4: Explicit Metalogic. The explicit logic
Stage 3: Explicit Logic--Implicit Metalogic. of Stage 3 may be postulated to require implicit
Stage 3 individuals, typically adolescents or adults, metalogical knowledge about the nature of logical
are expressly aware of logical form. Because they systems. Explicit formalization of logical systems
think about form explicitly, they fully understand involving reflection on the nature of such systems
and on their relations with each other and with natu- Stage 2: Explicit Subjectivity-Implicit Theory.
rallanguages may be defined as Stage 4 (Moshman, Stage 2 children, typically elementary school age,
1989). Systematic reasoning at this level may be explicitly distinguish knowledge and appearance
limited to individuals who are engaged in a formal from reality. They realize the world may be differ-
study of logic, though there is evidence of some ent from what one sees (Flavell, 1986; Flavell
Stage 4 insights among college undergraduates et at., 1986) or thinks (Clinchy & Mansfield,
(Politzer, 1986). 1986). They understand that there is nevertheless a
relation between ideas and reality and, if chal-
lenged, may defend their ideas by citing evidence
Development of Inductive Reasoning
consistent with them and/or seeking more such evi-
Although inductive reasoning, especially as we dence-a verification strategy. Although we can
have defined it, is a broad domain, we believe the view their subjective ideas as theories, they do not
following account captures some of its fundamental think explicitly about theories as theories and are
aspects. thus unsystematic in testing them.
Stage 1: Explicit Content-Implicit Subjec- Stage 3: Explicit Theory-Implicit Metatheory.

tivity. The Stage 1 child, typically a preschooler, Stage 3 individuals, typically adolescents and
thinks about content. Although we can infer subjec- adults, have a more explicit grasp of theory and its
tivity, this is implicit in the child's knowing rather relation to data. They recognize that theories are not
than an object of explicit awareness. Children as old merely more general than data but exist on a differ-
as 3 or 4 seem very limited in their understanding of ent plane, a plane of possibilities as opposed to the
their own subjectivity. For example, they have realm of realities. The need to test theories is ex-
great difficulty distinguishing what something pressly understood. It is recognized that, in an in-
"looks like" from what it "really and truly" is, finite world, theories can never be proven true but
even when researchers have made systematic can be disproved by a single disconfirming datum.
efforts to use familiar items and circumstances, to The Stage 3 thinker understands that the key to
ask simple questions, and to teach the relevant con- genuinely testing a theory or hypothesis is not to
cepts (Flavell, 1986; Flavell, Green, & Flavell, accumulate supportive data but rather to seek poten-
1986; Flavell, Green, Wahl, & Flavell, 1987). tially disconfirming data (falsification strategy)
Young children view knowledge as something di- (Klayman & Ha, 1987; Moshman, 1979a; O'Brien,
rectly absorbed from the world and fail to grasp the 1987; O'Brien, Costa, & Overton, 1986; Overton,
role of subjective construction in creating ideas that Ward, Noveck, Black, & O'Brien, 1987; Popper,
may differ from person to person and may not di- 1959; Skov & Sherman, 1986). The Stage 3 indi-
rectly reflect reality (Broughton, 1978; Clinchy & vidual's sophistication in thinking about theories,
Mansfield, 1986) (see Table 2). data, and the relations between them reveals an
Table 2. Development of Inductive Reasoning

Explicit object of understanding Knowledge implicit in reasoning (subject)
Stage 1 Content Subjectivity: Distinction between appear-

Explicit content ance and reality
Implicit subjectivity I
t
Stage 2 Subjectivity: Relation of appearance and re- Theory: Distinction between theory and
Explicit subjectivity ality (verification strategy) data
Implicit theory I
t
Stage 3 Theory: Relation of theory and data (falsi- Metatheory: Distinction between meta-
Explicit theory fication strategy) theory and theory
Implicit metatheory I
t
Stage 4 Metatheory: Coordination of theories and
Explicit metatheory metatheories
implicit metatheoretical competence (Moshman, phisticated reasoning is simply learned from one's
1979b), though that competence is implicit in the environment or, alternatively, is innate in one's
individual's reasoning about theories rather than it- genes. We will argue against these views and thus
self an object of reflection. set the stage for a more detailed consideration of the
alternative view that the development of reasoning
Stage 4: Explicit Metatheory. Some indi- is indeed a creative construction.
viduals, perhaps in connection with higher educa-
tion, begin to think explicitly about the meta-
theoretical assumptions underlying their theories
Empiricist and Nativist Accounts of the
(paradigms) and about their metatheoretical views
with respect to the testing of theories (philosophy of Development of Reasoning
science). Recognition that all data are theory-laden
and that apparent falsifications of theory can always Even if one accepts the empirical evidence that
be explained away through additional assumptions reasoning improves with age, one might argue that
may lead to epistemological relativism, a transi- this is not the result of a creative process. One alter-
tional phase between Stages 3 and 4 in which the native is that rationality is learned from one's en-
possibility of rationally comparing and testing theo- vironment via relatively direct (noncreative) mech-
ries is rejected. Further reflection, however, leads anisms of transmission (empiricism). Another
to recognition that even if the Stage 3 falsification possibility is that the observed developmental
strategy is inadequate, rational coordination of the- changes are due to the maturation of abilities pro-
ory and data leading to genuine progress in under- grammed in the genes (nativism). We will consider
standing reality is nevertheless possible (Brough- each of these views in tum.
ton, 1978; King, Kitchener, Davison, Parker, &
Wood, 1983; Kitchener & King, 1981; Kitchener& Empiricism
Wood, 1987; Welfel & Davison, 1986).
Empiricist accounts of the development of rea-
soning can be divided into three general categories:
Conclusion: Reasoning and Rationality physical transmission, social transmission, and lin-
The above accounts of stages in the development guistic transmission (Moshman, 1979b). First, we
of deductive and inductive reasoning are both con- will present a rather eclectic empiricist account of
sistent with extensive empirical evidence. They the development of reasoning, including mecha-
suggest that in each area reasoning genuinely devel- nisms associated with all three categories. Then we
ops in the sense that each new stage is not merely will propose that, although learning undoubtedly
different from the previous one but incorporates and plays an important role in the development of rea-
transcends it. With respect to deduction, the in- soning, it is not sufficient for a complete account. In
creasing differentiation and coordination of prem- particular, we will argue that it fails to account for
ises and conclusions lead to a differentiation of ab- (a) conceptions oflogical necessity, (b) progression
stract form of argument from empirical truth. With through distinct levels of knowing, (c) increasing
respect to induction, a similar differentiation and rationality, and (d) the origin of genuine novelties.
coordination of theory and data lead to sophisti- Empiricism thus skirts the central issues of logic,
cated conceptions of truth, falsity, proof, and di~ development, rationality, and creativity.
proof at increasingly abstract levels of meta-
theoretical awareness. In both domains, the stages An Empiricist View of the Development of Rea-
embody a self-reflective trend in which the indi- soning. One aspect of an empiricist view is that
vidual at each level reasons about (reflects on, reasoning develops due to feedback from one's
"knows") the reasoning of the previous level. If physical environment, that is, via physical trans-
we think of rationality as a self-reflective appeal to mission. For example, when new modes of reason-
reasoning (Moshman & Hoover, 1989), the devel- ing are applied to one's problems, they may in some
opment of reasoning appears to move in the direc- cases be maladaptive. This would lead to nonrein-
tion of greater rationality. forcing results and would thus tend to extinguish
We can now return to the question of whether this those modes of reasoning. Some new processes,
is a creative process. In the next section, we will however, would generate adequate solutions lead-
consider the possibility that it is not. Perhaps so- ing to results reinforcing to the individual. Such
reasoning would be used with increasing frequen- cific mechanisms, it seems clear that learning plays
cy. The environment would thereby shape reason- an important role in the development of reasoning
ing in the direction of increasing rationality. (Moshman, 1989). The question is whether learn-
Although environmental selection of this sort ing is sufficient to account for the major phenomena
probably does play a role in the development of that need to be explained. In the next four subsec-
reasoning, sophisticated reasoning processes do not tions, we argue that empiricism fails to account for
lead to consistently reinforcing outcomes. There is four central (and closely interrelated) phenomena:
little evidence that the sorts of contingencies as- (1) the logical necessity inherent in advanced rea-
sumed in this account are sufficiently consistent and soning, (2) the progression of reasoning through
have effects of sufficient magnitude to result in sub- distinct levels of knowing, (3) the systematic in-
stantial increases in rationality in a reasonable peri- creases in rationality, and (4) the creative produc-
od of time. Moreover, the source of new reasoning tion of genuine novelties.
remains obscure.
The empiricist account of the development of Critique of Empiricism: The Problem of Logic.
reasoning is greatly enriched when one adds social Progress through the stages of deductive reasoning
transmission-learning from other people (Band- described above involves increasingly explicit un-
ura, 1977). First, unlike objects in the physical en- derstanding of logical necessity (Moshman, 1989;
vironment, other people actually reason. Although Moshman & Timmons, 1982). In Stage 2, the child
their thinking itself cannot be directly observed, makes accurate judgments regarding whether given
reasoning processes can be and often are exter- premises are sufficient for a conclusion or further
nalized and imitated. Such imitation need not be a information is necessary, thus showing an implicit
rote copying of behavior but can involve abstraction grasp of necessity. In Stage 3, valid arguments are
of generalizable strategies. distinguished from invalid arguments independent
Second, people can provide far more precise and of the truth or falsity of premises and conclusions,
effective feedback on reasoning than can the phys- thus showing a more explicit grasp of necessity as
ical environment. A teacher may consistently and an aspect oflogical form. As Lori eventually recog-
effectively reinforce correct reasoning about math nized, the conclusion of a valid argument, even if it
problems by noting the correctness of each right is empirically false, "has to be that way" (our
answer and the incorrectness of each wrong answer. emphasis). Stage 4 involves further reflection on
Positive or negative evaluation of improper math by the role of necessity in diverse logical systems.
the physical environment, by contrast, would be far Available evidence shows Stage 2 understanding as
less direct and consistent. early as age 6 (Somerville et al., 1979) and indi-
Finally, people can provide direct instruction. cates that Stage 3 understanding is common by ear-
They can explain logical concepts and discuss the ly adolescence and spontaneously applied by most
rationale for various sorts of reasoning. There is college students (Moshman & Franks, 1986).
simply no parallel to this in one's dealings with the Metalogical comprehension at a Stage 4 level is
physical world. probably much less common. Even if it were lim-
Much social transmission occurs via the vehicle ited to logicians, however, we would have to keep
of language: language tremendously facilitates in mind that the logician who devises an explicit
modeling, feedback, and direct instruction. In addi- theory of necessity is a person whose behavior
tion, however, language may itself constitute a ma- needs to be explained.
jor source of reasoning. In fact, language is proba- Although the role of logical necessity in the de-
bly so central to reasoning that it is worth velopment of inductive reasoning is less obvious, it
distinguishing a particular aspect of social transmis- is equally important (Moshman, 1979b). Theories
sion with a label of its own-linguistic transmis- themselves, of course, are distinguished from log-
sion. A strong case can be made that, in learning a ical systems precisely by the fact that they are not
language, a child learns a set of implicit inference logically necessary-they can be disconfirmed by
schemata that underlie reasoning throughout the empirical evidence. Metatheoretical knowledge
lifespan (Braine, 1978; Braine & Rumain, 1983). A concerning the testing of theories, however, does
great deal of our reasoning, including most of our have logical elements. Consider, for example, the
automatic deductive inferences, may be in part a Stage 3 insight that, in an infinite world, a hypoth-
result of learning language. esis cannot be conclusively confirmed by any finite
Although there is room for argument about spe- number of observations but can be conclusively dis-
confirmed by a single falsifying datum. This is not tion, it appears that the changes in reasoning are
simply a theory that happens to work in our uni- developmental, in the strict sense of moving in a
verse. On the contrary, it is a logical insight that systematically progressive direction.
rests on an understanding of hypotheses as univer- Genuine development is, to say the least, diffi-
sal statements and of the logical implications of cult to explain from an empiricist perspective. In-
such statements (Popper, 1959). Of course, pro- stead its existence is typically denied. The em-
gress toward Stage 4 complicates matters by putting piricist account suggests that children's reasoning
the testing of simple hypotheses into a broader, changes in whatever direction it is shaped by the
explicitly metatheoretical context. Nevertheless, it environment. The fact that there are certain con-
remains clear that there is a definite logic underly- sistencies in developmental trends is explained as
ing higher stages of inductive reasoning that pro- due to similarities in children's environments with-
vides a basis for using empirical evidence but is in any culture or even across cultures.
itself based not on empirical evidence but on the Thus, for example, an individual who tests hy-
necessary relations of propositions. potheses by seeking verifying information would be
Empiricist efforts to account for logical necessity expected to use increasingly a falsification strategy
(of which John Stuart Mill's work is probably the to the extent that it is modeled and/ or reinforced in
most famous) generally suggest that it is an illusion. the environment. An individual who uses a falsi-
Some beliefs, it is argued, are based on more evi- fication strategy would be equally likely to reject it
dence than others. The more evidence we have for a in favor of a verification strategy to the extent that
belief, the more strongly we believe it. If the evi- the latter is modeled and/or reinforced by the en-
dence for a belief is sufficiently extensive and con- vironment. In this account, there is no logic to the
sistent, we see it not merely as true but as logically falsification strategy that, once understood, makes
necessary. movement from verification toward falsification
This account does not fare well, however, in ac- more likely than the reverse. Change moves in
counting for the actual phenomena of human belief. whatever direction the environment pulls or pushes;
The daily rising of the sun, for example, is highly there is no genuine development.
consistent, but there is no temptation to see this as The empiricist account is not intuitively very
logically necessary. The truths of arithmetic, on the plausible, nor is it consistent with available evi-
other hand, appear to be disconfirmed every time dence. Class exercises by the first author over many
our checkbooks fail to balance. Nevertheless, we years show that undergraduates who discuss a diffi-
see 2 + 2 = 4 as logically necessary; instead of cult hypothesis-testing task (Wason's, 1983, selec-
rejecting our "theories" of arithmetic, we reject tion task) with no input from the instructor show a
the evidence by searching for calculation errors. strong tendency to move toward falsification, even
Similar empiricist efforts that attempt to account though the verification approach is initially far
for logical necessity as a social or linguistic conven- more popular. Arguments for falsification present-
tion fare no better (Moshman, 1979b; Moshman & ed by fellow students are not initially more frequent
Timmons, 1982). It appears that people do come to or better expressed than arguments for verification,
distinguish logical from empirical matters and that but their inner logic does seem, over a sufficient
learning from the physical, social, or linguistic en- period of time, to have a greater impact.
vironment cannot explain how this distinction is Consistent with this informal observation, re-
made. search in a variety of areas has shown that children
are far more likely to be influenced by a model
Critique ofEmpiricism: The Problem ofDevelop- when what is modeled involves a level of reasoning
ment. Available evidence suggests a clear direc- slightly beyond their own than when it involves a
tionality to age-related changes in reasoning. The level of reasoning below their own (e.g., Walker,
stages earlier described summarize thisdirec- 1982). Such evidence is difficult to reconcile with
tionality. As we have seen, each stage includes an the empiricist view that children are equipotential to
explicit differentiation and coordination of what move in any direction the environment pulls or
was only implicit in the previous stage. In some- pushes. It is more consistent with the view that there
what broader terms, each stage is a metacognitive is a natural directionality in the development of
reflection on the previous stage. Given this progress reasoning.
toward increasing differentiation and hierarchic in- This is not to say, of course, that people move
tegration at higher and higher levels of self-reflec- inexorably through developmental stages re-
gardless of their environments. Learning undoubt- been created by human minds, either individually
edly plays an important role in development. Our or in the context of social interaction. But this puts
argument is that (a) there is a natural, progressive us right back where we started: How does the mind
direction of change, (b) reasoning tends to move in create new and better forms of reasoning? The em-
this direction in any normal environment, and (c) piricist account avoids a psychological answer to
empiricism cannot account for this. this question by throwing the problem to lin-
guistics, sociology, or anthropology. But this is (to
Critique of Empiricism: The Problem of Ra- use the vernacular) a cop-out: There is no reason to
tionality. There is no doubt that people can and think that these disciplines are any better equipped
do learn a variety of modes of reasoning. An em- to explain creativity than is psychology, or that
piricist account, however, fails to explicate pro- creativity can ever be explained without reference
gress toward rationality. At best, it might suggest to psychological considerations.
that most societies teach good strategies of reason- Again, we are not denying that society and lan-
ing, that good reasoning somehow inheres in guage embody considerable rationality and play an
human languages, and/or that physical environ- important role in the learning of reasoning by indi-
ments are sufficiently consistent in reinforcing viduals. We do suggest, however, that the human
good reasoning. There are at least three major mind is capable of genuine creativity in the con-
problems with this account. First, it suggests that a struction of rationality. Further, we suggest that this
society might just as easily teach irrationality as creativity is not limited to brilliant minds that have
rationality. Although there is no doubt that environ- influenced society or to ancient social interactions
mental factors can hinder the development of rea- that produced language. On the contrary, we sug-
soning and sharply limit rationality, the empiricist gest that creativity is central to the cognitive devel-
account, as noted above, fails to explain the natural opment of every individual and is indispensable in
tendency for development to move in the direction explaining the development of reasoning.
of rationality. Second, empiricism has little to say The failure of empiricism to account for impor-
about the self-reflective nature of advanced reason- tant aspects of the development of reasoning does
ing. Individuals at or beyond Stage 2 not only reach not in itself show that a creative process must be
conclusions but reflect on the fact that they are involved. In fact, the problems of empiricism have
doing so. Any explanation of rationality must ad- traditionally been taken as arguments for nativism
dress its self-reflective nature. Third, the empiricist (see Bereiter, 1985; Campbell & Bickhard, 1987).
account merely shifts the need for explanation from With respect to the specific issues of reasoning,
the individual to language and society. If children then, we must now tum to the possibility that ra-
learn to be rational via language and social influ- tionality is innate.
ences, then rationality must be somehow inherent in
language and/ or in society. It remains unclear how
Nativism
it got there, especially if rationality is not spon-
taneously produced by the human mind. This leads A Nativist View of the Development of Reason-
us directly into the final part of our critique. ing. A nativist account of the development of rea-
soning stresses its hereditary basis. The simplest
Critique of Empiricism: The Problem of version of this, prejormationism, would propose
Creativity. A major basis for empiricist views is that mature rationality is directly encoded in the
the difficulty of explaining genuine creation of nov- genes. Modem nativists generally reject preforma-
elty. It is not at all clear how new modes of reason- tionism in favor of epigenesis. The latter is a more
ing or higher levels of rational understanding can be developmental conception in that it involves a pro-
created by the human mind. The empiricist solution cess of change over some portion of the lifespan. It
is to argue that they are not created; rather, they are remains nativist, however, in viewing change as
internalized from the social environment. strongly directed by heredity (Kitchener, 1978,
This assumes, however, that rationality already 1980).
exists in language and/or society. But if natural For example, no modem biologist would suggest
languages include an inherent rationality, how did that a tiny version of the human heart exists in the
that come to be there? If society already possesses DNA of a fertilized human ovum. Rather, the DNA
sophisticated modes of reasoning, where did they contains genetic instructions that will lead to the
come from? It appears the rationality must have development of a heart in any normal (species-ap-
propriate) environment. Although the environment We suggested earlier that, in explaining cog-
plays a critical role in providing necessary nutri- nitive novelties as due to learning from society, we
ments, it has no effect on the final structure of the are merely foisting on sociology the task of ac-
heart (except in pathological cases). The structure counting for novelty. Nativism is often an equally
of the human heart is essentially the same for all unhelpful appeal to biology. It assumes, usually
humans; its development follows a predetermined without basis, that evolutionary mechanisms can
path guided by the genes. better explain the creation of novelties than can
A nativist account of the development of reason- developmental mechanisms. There is no evidence
ing would follow this general framework. Ra- that sophisticated reasoning of the sort observed in
tionality is not directly encoded in the genes; how- older children and adults has come about through
ever, human genes do contain precise instructions natural selection or any other phylogenetic process
that, in any normal environment, will result in a or that its development is in any direct sense pro-
sequence of stages of reasoning leading ultimately grammed in human genes. We cannot count on bi-
to mature rationality. Only in pathological cases ology, any more than on sociology, to explain the
will there be deviations from this course. Although emergence of rationality. We would do better to
epigenesis does result in something that was not presume that rationality is creatively constructed by
literally in the genes, the course of development is the developing individual and to try to explicate this
too directly programmed to be considered creative. process.
Predetermined epigenesis probably plays an
important role in anatomical and physiological de-
Interactionism
velopment and may have important implications for
psychological development. More broadly, we Before we tum to accounting for the develop-
have no quarrel with the view that heredity, like ment of reasoning as a creative construction, it is
learning, plays an important role in development. worth saying a word about interactionism. In psy-
The question before us is whether heredity is chological writing, it is common to assert that nei-
sufficient to account for available evidence con- . ther heredity alone nor environment alone can ac-
cerning the development of reasoning. Again, we count for development; rather, development
think not. involves an interaction of both. In a general sense,
this is surely correct. With respect to our current
Critique of Nativism. Heredity is often in- concern, however, the concept of interactionism is
voked when we cannot account for something in not particularly helpful.
terms of learning. If the observed developmental One form of interactionism is statistical interac-
course of reasoning cannot be explained in terms of tionism. This involves the assertion that heredity
the environment, it is tempting to conclude that it and environment interact in the statistical sense:
was programmed in the genes. The effects of the environment depend on one's
To say something is innate, however, does not genes and the effects of one's genes depend on
explain it. It simply redirects our attention from the one's environment. This is no doubt true but oflittle
ontogenetic (developmental) history of the indi- use for our present purpose in that it says nothing
vidual to the phylogenetic (evolutionary) history of about the process of development.
the species. It is not obvious that evolutionary From a developmental point of view, a more
mechanisms are any more powerful in accounting useful approach to interactionism is dynamic in-
for logic, development, or rationality than are de- teractionism (Lerner, 1986). This approach in-
velopmental mechanisms. Natural selection, for cludes the concept of statistical interaction ism but
example, is simply an environmental feedback stresses the actual processes by which the interact-
mechanism analogous to operant conditioning. In ing forces of heredity and environment determine
fact, at the level of phylogenesis, we run into pre- the course of development. Such an approach is
cisely the same difficulty we encountered at the particularly helpful in explaining the partially
level of ontogenesis: How can empirical experience unique course of each individual's development in
result in logical necessity? William James noted the terms of unique and continuing interactions of
problem a century ago: "The experience of the race genes and experience.
can no more account for our necessary or a priori With respect to our current concerns, however,
judgments than the experience of the individual even dynamic interactionism is of limited use. Ini-
can" (1890!l950, pp. 617-618). tially, we are concerned primarily with the general
course of development in the area of reasoning, not qualitatively distinct stages. The constructivist as-
with individual differences. Although the existence pect of his theory revolves around the concept of
of individual differences is beyond dispute, evi- "equilibration. " Most of Piaget' s research and the-
dence suggests there are also important com- orizing prior to the mid-1960s, however, dealt not
monalities in the course of development, summa- with equilibration but with the structure of the vari-
rized in the stages presented earlier, that are worthy ous stages of development (e. g., the groupements
of explanation. Dynamic interactionism has little to of concrete operations and the complete com-
say about this, focusing instead on individual binatorial system and INRC group of formal opera-
differences. tions). Although Piaget did discuss the process of
Perhaps more important, dynamic interactionism equilibration and made it clear that the structural
is not itself constructivist, though it is not inconsis- stages were the outcome of this process, he was far
tent with constructivism. In at least some versions, more clear and detailed in his stage descriptions
dynamic interactionism provides an image of the than in his accounts of the developmental process.
person as a passive space within which heredity and Beginning around the mid-1960s, Piaget's em-
environment interact. The interaction may produce phasis shifted (see Campbell & Bickhard, 1986).
novelties, but there is no self actively and purposely The process of development became a more central
producing those novelties. Creation may occur, but focus and its mechanisms (especially equilibration
no one is responsible for that creativity. It is far and reflective abstraction) received detailed em-
from clear that any such process, however interac- pirical and theoretical attention. Among the most
tive it may be, can result in rationality. important of his late works was a 1975 reworking of
Finally, it is important to note that our critique of his theory of equilibration (Piaget, 1985). Piaget
empiricism was not at all based on its neglect of proposed that knowing, at any level of develop-
hereditary factors, nor was our critique of nativism ment, involves a complex coordination of subject
based on its neglect of the individual's environ- and object. All knowing is subjective in the sense
ment. Thus, in terms of the question of how reason- that the reality cannot be directly known; it can only
ing develops, interaction between heredity and en- be known by assimilation to the subject's (mental or
vironment does not resolve the fundamental physical) actions. Nevertheless, increasingly ob-
inadequacies of each. Once again, we are led to the jective knowledge is possible to the extent that one
view that only a conception of reasoning as a cre- becomes aware of and reconstructs one's subjec-
ative construction by an active self can explain the tivity. Thus, self-understanding and objective
progress of rationality. knowledge are inextricably linked.
It is not sufficient, of course, for a constructivist Awareness of one's subjectivity is not an all-or-
account of reasoning to show the inadequacies of none matter, however. Campbell and Bickhard
empiricism, nativism, and interactionism. What is (1986) have proposed a general theory of cognitive
needed is a positive explication of how rationality is development that postulates a series of knowing
creatively constructed. It is to this task that we now levels, each related to the previous one by a process
tum. of reflective abstraction. Knowing Levell interacts
with the environment. Knowing Level 2 reflects on
knowing Levell. It thus makes explicit what was
previously only implicit in the process of knowing
A Constructivist Account of Rationality the environment. To do this, however, requires
construction and application of new know ledge that
In general terms, constructivism maintains that is only implicit at Level 2. Explicit reflection on
development is driven by neither heredity nor en- knowing Level 2 can only be done from knowing
vironment, nor even by an interaction of the two. In Level 3, which, in tum, involves a newly con-
fact, it is not driven at all but, rather, constructed by structed implicit knowledge knowable only from
the active mind. This view is not a specific theory the perspective of Level 4, and so forth. Ascent
but rather a general metatheoretical approach, or through these levels is a very gradual constructive
paradigm, within which specific theories can be process that takes place in each of various domains
formulated (Moshman, 1982). of knowing. Rate of progress in the various do-
The best-known constructivist theory is that of mains may differ; the sequence of stages within any
Jean Piaget. Piaget proposed that development con- domain must be consistent, but there need not be
sists of the active construction of a succession of consistency across domains.
The knowing levels approach suggests the the next higher level. As the child refines his or her
emergence at each stage of cognitions that are genu- use of argument form to make judgments about con-
inely novel (for the individual) but do not simply clusions, he or she increasingly grasps the distinction
appear mysteriously out of nowhere. This approach between form and content, until form of argument is
suggests that the development of reasoning may be sharply distinguished from empirical truth of con-
a genuinely creative process. Cognitive novelties tent. Reflection on the validity of argument forms
are actively constructed by the mind but may never- involves a grasp of formal logic and thus an ability to
theless lead to systematically increasing rationality use a formal logical system (as distinct from a natural
in that each new stage is a systematic reflection on language). This requires some degree of (implicit)
and reconstruction of the stage before. In the next metalogical knowledge. The construction of such
two subsections, we will reconsider the stages of knowledge is central to the construction of Stage 3,
deductive and inductive reasoning proposed earlier and, in tum, may become an object of further reflec-
from the perspective of a knowing levels approach. tion, leading to the construction of an explicit grasp
of metalogic at Stage 4.
There is no reason in principle for any upper
Construction of Deductive Reasoning
bound on this process, though there may be limits
Young children think about their environments, due to inherent information-processing constraints
including, from a very early age, the content of of the human mind. The explicit Stage 4 reflection
linguistic propositions. They routinely make a vari- on metalogic presumably takes place from the per-
ety of inferences and thus reach conclusions that go spective of an implicit subjectivity (call it "meta-
beyond the given information. As psychologists, metalogic" if you like), which, in tum, can become
we can reflect on their reasoning and determine that an object of reflection at Stage 5. Not being at Stage
they are using a variety of inference schemata to 5 ourselves, we have nothing explicit to say about
reach their conclusions. They themselves, how- this.
ever, are not initially aware of their reasoning. They
think about content, not about the process of
Construction of Inductive Reasoning
inference.
As their inference schemata become sufficiently A similar account can be provided for the con-
efficient and consolidated, however, children begin struction of inductive reasoning. Stage 1 children
to reflect on their conclusions, on the distinction think about content. As psychologists we can see
between conclusions and premises, on the relation the SUbjectivity in their thinking and can note the
of conclusions and premises, and on the process of distinction between appearance and reality but chil-
inference by which they arrive at conclusions from dren themselves simply know what they know.
premises. Their coordination and reconstruction of Knowledge is seen as direct; questions about how it
premises and conclusions, including an explicit relates to reality, even if put in simple language and
awareness of the inference process linking them, contexts, are likely to be viewed with wonderment
does not come from nowhere but rather from reflec- or simply misconstrued (Broughton, 1978; Flavell,
tion on and coordination of what was already im- 1986; Flavell et ai., 1986).
plicit in their reasoning. This reflection includes the With rich enough experience, children begin to
construction of a new level of understanding in note that people do not always respond to things the
which form of argument is distinguished from con- same way as each of them do and that things may
tent. Only such awareness of form will allow a even seem different to them at different times. Re-
genuine grasp of the nature of inference, including a flecting on their own subjectivity, children consider
sense of the necessity of certain conclusions. the distinction between appearance and reality and
Awareness of form, however, is only implicit in the relations between them. They realize that some-
Stage 2 thinking. It is part of the framework that times they need more evidence before they can
allows an explicit grasp of inference but is not itself know that they are correct. Implicit in this Stage 2
the object of awareness. It is a newly constructed orientation toward verification is the idea of testing
subject that is used to understand the previous sub- one's theories.
ject, which, accordingly, now becomes the object Further reflection on the subjectivity of knowl-
of understanding. edge yields an increasingly explicit distinction be-
Although the Stage 2 subject is not an object of tween theory and data, including the Stage 3 con-
reflection, it can become an object of reflection from cept that theories are purely hypothetical and must
be tested by gathering data that could disconfirm invariant sequence of stages. Because each level of
them. This is a metatheoretical view and can itself knowing is a reflection on the previous level, the
become the object of further reflection leading (at stages must emerge in the same order for all indi-
Stage 4) to explicit conceptions of metatheories and viduals. Environmental or hereditary factors may
their relation to theories and data. Again, there is no determine how rapidly one moves through the
reason in principle why Stage 4 must be the highest stages or how far one gets, but, because each stage
stage, but we know of no evidence that anyone is a reflection on the previous one, it would be
proceeds beyond this and, in any event, have noth- impossible for the stages to occur in a different
ing substantive to say about "metametatheory." order or for a stage to be skipped.
With respect to rationality, the knowing levels
approach accounts for the systematic increase sug-
Conclusions
gested by the empirical literature. Each stage con-
The specific stages of inductive and deductive stitutes a higher level of rationality than the pre-
reasoning differ. Although a case can be made for a vious one because it is a reflection that' systemati-
conceptual parallel (e.g., in each sequence, Stage 3 cally reconstructs the previous stage. Each stage is a
is precisely two reflective abstractions away from reflective awareness of one's reasoning at the pre-
the level that interacts directly with reality), typical vious stage and thus, in an important sense, pro-
ages of emergence may not be comparable. What is vides reasons for one's previous reasoning. In high-
central for our purposes is that the process is identi- lighting the increasingly self-reflective nature of
cal in each case. The individual begins with direct advanced reasoning, the knowing levels approach
(and epistemically naive) interactions with reality, justifies the identification of such reasoning as con-
including no awareness of his or her implicit subjec- stituting rationality.
tivity in observing, interpreting, or drawing in- Finally, with respect to creativity, the knowing
ferences. Through reflection, not on the environ- levels approach shows how new ideas can be cre-
ment but on his or her own knowing, the individual ated that are based on, but nevertheless transcend,
constructs a new level of understanding that ex- earlier knowledge, and that then serve as the basis
plicitly knows the previous level and thus creates for further creativity. This conception of creativity
new knowledge via differentiation and coordination allows for genuine novelty but can nevertheless ac-
of what was only implicit in his or her interactions count for systematic progress in rationality.
with the environment. This new level of under-
standing, however, itself includes knowledge that
can become the object of explicit awareness, thus
setting the stage (so to speak) for the construction of The Subjective Creation of Objectivity
a still higher level.
Creativity and Rationality
This analysis, we believe, accounts for the four
phenomena that were so difficult for empiricism Can creativity lead to increasing rationality? The
and nativism-namely, logic, development, ra- traditional paradigm of rationality is the scientist,
tionality, and creativity. With respect to logic, the who uses the scientific method to gain objective
problem was to account for how the individual de- knowledge of the real world. Thus, rationality may
velops a sense that certain things are logically nec- be intuitively seen as mechanical conformity to
essary if everything is simply learned from the en- fixed rules of reasoning that guarantee objectivity.
vironment (of the individual or of the species). The A traditional model of creativity, in contrast, is the
constructivist account proposes that not everything eccentric and misunderstood artist, whose creative
is abstracted from the environment. Some knowl- insights spring from the uncharted depths of his or
edge is constructed via reflection on and coordina- her unconscious mind. Creativity, then, may be
tion of the workings of the mind. This allows not construed as consisting of unconstrained flights of
merely observation of and generalization from real- fancy deep in the mists of our subjectivity. Given
ities but systematic anticipation of all possibilities, these conceptions, it is indeed difficult to see how
thus permitting the construction of a sense of neces- rationality could develop via a creative process.
sity with regard to one's mental operations (Piaget, From a constructivist perspective, pure objec-
1986). tivity is impossible not only in fact but in principle.
With respect to development, the knowing levels There may indeed be a real world with actual char-
account does explain convincingly the basis for the acteristics, but it can be known only through assim-
ilation to one's subjective knowledge and processes individuals, nor is it a set of skills or abilities that
of knowing (schemes, schemata, frames, scripts, emerge at some point in development. On the con-
ideas, theories, strategies, structures, and so forth). trary, it is a fundamental aspect of development in
Piaget (1985) viewed equilibration, the central pro- all normal individuals. The process of develop-
cess of development, as a continuing renegotiation ment, from this perspective, is inherently creative.
of the relation between subject (knower) and object This is true even in areas of development, such as
(reality as construed by the subject). The object of reasoning, where progress toward rationality and
knowing always remains a function, however, of objectivity is of central concern.
the knower (subjectivity) as well as a function of
reality (objectivity). At no point does it become
Directions for Research
possible to know reality from a perspective outside
one's own knowledge and thus achieve final objec- The suggestion that development is a genuinely
tivity (von Glasersfeld, 1979). creative process suggests avenues of research that
It is, however, possible to construct a new frame- might not be considered if one simply assumed that
work from which to know one's previous processes any outcome of development can be entirely ex-
or structures of knowing. This process, reflective plained as a mature version of what was already in
abstraction, makes it possible to evaluate the contri- the genes (maturation), an internalization of what
bution of one's earlier subjectivity to one's experi- was already in society (learning), or an interaction
ence and thus to attain a more objective understand- of maturation and learning. As a metatheoretical
ing of the contribution of reality to that experience. approach, constructivism is less fully reductionist
The reflection on one's own subjectivity does not, and less committed to an antecedent-consequent
of course, take place from a completely objective sort of mechanical causality than are nativism and
perspective. On the contrary, it takes place from a empiricism. It even goes beyond most versions of
higher level of sUbjectivity. Nevertheless, the re- interactionism in these respects. Nevertheless, it
sulting metasubjectivity, in rendering one's pre- can generate specific, testable accounts of develop-
vious subjectivity an object of knowing and thus ment and may even go deeper than alternative ap-
reconstructing and understanding it, constitutes a proaches in genuinely explaining the underlying
genuine increase in objectivity. To the extent that process.
the previous knowing involved reasoning, the new In the present case, we have described specific
stage involves reasoning about that reasoning and stages of development in two areas of reasoning and
thereby constitutes a higher level of rationality. Ra- explained their emergence via reflective abstrac-
tionality is thus not a mechanical following of rules tion. One direction for research would be to provide
but a self-reflective appeal to and coordination of increasingly fine-grained accounts of the stages
reasoning at levels of understanding that are one or themselves. This should include microanalysis of
more steps beyond the level that knows the substages and of the sequence of transitional states
environment. leading from each major stage to the next. The more
The construction of each higher level is a genu- detailed our understanding of the route(s) develop-
inely creative act in that the framework for reflect- ment takes and the steps along the way, the better
ing on one's previous subjectivity is given neither we can guess at plausible mechanisms of transition
by heredity nor environment, and the resulting ex- (Keil, 1984).
plicit reconstruction of what was implicit in that There can be no substitute, however, for careful
sUbjectivity involves the creation of genuine novel- analysis of reasoning during the developmental pro-
ty. Such creation does take place within the subjec- cess itself. Through systematic presentation of
tive domain in that it involves reflection on one's challenging problems, it is possible to generate pro-
own subjectivity. It is not a mere flight of fancy, tocols that show how individuals construct and ap-
however, but, on the contrary, a systematic abstrac- ply new modes of reasoning (Kuhn & Phelps,
tion and reconstruction that, by increasing the self- 1979). Analysis of such protocols (cf. the earlier
reflective use of reasoning, leads to greater ra- example of Lori) may not be as straightforward and
tionality in one's knowing of reality and thus to objective as statistical manipulation of quantitative
increasing objectivity. data in a design with clearly distinguishable inde-
In this view, creativity is not a trait associated pendent and dependent variables. Such analysis
primarily with particular (e.g., gifted or talented) can, however, be reasonably systematic and can
surely yield general and testable conclusions about empirical science. The fundamental issues of
what goes on in the creative mind. creativity are, ultimately, metaphysical questions
about the nature of mind.
Creativity and the Limits of Science
Although constructivism thus constitutes a viable ACKNOWLEDGMENTS. We are grateful to Mark
research program, it also raises difficult issues. The Bickhard and Robert Campbell for detailed and
specific question we have addressed in greatest de- thoughtful comments on an earlier draft of this
tail is the question of how rationality can be cre- chapter.
atively constructed. A full answer must explain (a)
how novel reasoning can be produced by the devel-
oping individual, and (b) how this new reasoning
can be not merely different from earlier reasoning References
but in fact more rational, better able to direct the Bandura, A. (1977). Social learning theory. Englewood Cliffs,
knower toward objective knowledge. NJ: Prentice-Hall.
On a psychological level, we have suggested that Bereiter, C. (1985). Toward a solution of the learning paradox.
there is no inherent paradox that renders scientific Review of Educational Research, 55, 201-226.
Braine, M. D. S. (1978). On the relation between the natural
research along these lines pointless. We have ar-
logic of reasoning and standard logic. Psychological Review,
gued that, through repeated reflective abstractions, 85, 1-21.
knowledge implicit in each level of subjectivity can Braine, M. D. S., & Rumain, B. (1983). Logical reasoning. InJ.
be explicitly differentiated and coordinated at a H. Flavell & E. M. Markman (Eds.), Handbook of child psy-
higher level. The resulting metasubjectivity is more chology: Vol. 3. Cognitive development (pp. 263-340). New
York: Wiley.
rational due to the self-reflection inherent in its con- Broughton, J. (1978). Development of concepts of self, mind,
struction and can thus serve as a better basis for reality, and knowledge. In W. Damon (Ed.), Social cognition
objectivity. There is no reason to assume that scien- (pp. 75-100). San Francisco: Jossey-Bass.
tific research cannot further clarify the specifics of Campbell, R. L., & Bickhard, M. H. (1986). Knowing levels
and developmental stages. Basel: Karger.
these processes.
Campbell, R. L., & Bickhard, M. H. (1987). A deconstruction
Nevertheless, constructivism raises challenging of Fodor's anticonstructivism. Human Development, 3D, 48-
questions. Why does reflective abstraction increase 59.
rationality? Why is it that metasubjective knowl- Clinchy, B. M., & Mansfield, A. F. (1986, April). The child's
edge is increasingly adapted to objective reality? discovery of the role of the knower in the known. Paper pre-
sented at the meeting of the Jean Piaget Society, Philadelphia.
These are questions that, in some version, have Cummins, J. (1978). Language and children's ability to evaluate
bedeviled every attempt to account for the relation contradictions and tautologies: A critique of Osherson and
of logical knowledge and empirical reality. As Markman's findings. Child Development, 49,895-897.
William James (1890/1950, p. 652) put it, one can Fabricius, W. V., Sophian, c., & Wellman, H. M. (1987).
imagine worlds in which "our logic ... would Young children's sensitivity to logical necessity in their in-
ferential search behavior. Child Development, 58, 409-423.
form a merely theoretic scheme and be of no use for Flavell, J. H. (1986). The development of children's knowledge
the conduct of life. But our world is no such world. about the appearance-reality distinction. American Psychol-
It is a very peculiar world, and plays right into ogist, 41, 418-425.
logic's hands." Constructivism may provide a bet- Flavell, J. H., Green, F. L., & Flavell, E. R. (1986). Develop-
ment of knowledge about the appearance-reality distinction.
ter approach to these issues than nativism and em- Monographs of the Society for Research in Child Develop-
piricism in that it does not evade the key questions ment, 51(1, Serial No. 212).
right from the start, but, at present, it is at best a Flavell, J. H., Green, F. L., Wahl, K. E., & Flavell, E. R.
general approach to looking for the answers. (1987). The effects of question clarification and memory aids
Although constructivism provides a meta- on young children's performance on appearance-reality tasks.
Cognitive Development, 2, 127-144.
theoretical framework for scientific research, the Hawkins, 1., Pea, R. D., Glick, J., & Scribner, S. (1984).
questions raised here are not merely psychological "Merds that laugh don't like mushrooms": Evidence for de-
questions about certain mental processes. They go ductive reasoning by preschoolers. Developmental Psychol-
far deeper in seeking explanations for the funda- ogy, 20, 584-594.
James, W. (189011950). The principles of psychology (Vol. 2).
mentally creative nature of thought and the creative New York: Dover.
adaptation of thought to reality. Full inquiry into Keil, F. C. (1984). Mechanisms of cognitive development and
creativity must, at some point, test the limits of the structure of knowledge. In R. J. Sternberg (Ed.),
Mechanisms of cognitive development (pp. 81-99). New ing: An iffy proposition. In H. W. Reese (Ed.), Advances in
York: Freeman. Child Development and Behavior (Vol. 20, pp. 61-90). Or-
King, P. M., Kitchener, K. S., Davison, M. L., Parker, C. A., lando, FL: Academic Press.
& Wood, P. K. (1983). The justification of beliefs in young O'Brien, D. P., Costa, G., & Overton, W. F. (1986). Evalua-
adults: A longitudinal study. Human Development, 26, 106- tions of causal and conditional hypotheses. Quarterly Journal
116. of Experimental Psychology, 38A, 493-512.
Kitchener, K. S., & King, P. M. (1981). Reflective judgment: Osherson, D. N., & Markman, E. (1975). Language and the
Concepts of justification and their relationship to age and ability to evaluate contradictions and tautologies. Cognition,
education. Journal of Applied Developmental Psychology, 2, 3,213-226.
89-116. Overton, W. F., Ward, S. L., Noveck, I. A., Black, J., &
Kitchener, K. S., & Wood, P. K. (1987). Development of con- O'Brien, D. P. (1987). Form and content in the development
cepts of justification in German university students. of deductive reasoning. Developmental Psychology, 23, 22-
International Journal of Behavioral Development, 10, 171- 30.
185. Piaget, J. (1985). The equilibration of cognitive structures: The
Kitchener, R. F. (1978). Epigenesis: The role of biological mod- central problem of intellectual development. Chicago: Uni-
els in developmental psychology. Human Development, 21, versity of Chicago Press.
141-160. Piaget, J. (1986). Essay on necessity. Human Development, 29,
Kitchener, R. F. (1980). Predetermined versus probabilistic epi- 301-314.
genesis: A reply to Lerner. Human Development, 23, 73-76. Pieraut-Le Bonniec, G. (1980). The development of modal rea-
Klayman, J., & Ha, Y. (1987). Confirmation, disconfirmation, soning: Genesis of necessity and possibility notions. New
and information in hypothesis testing. Psychological Review, York: Academic Press.
94, 211-228. Politzer, G. (1986). Laws of language use and formal logic.
Komatsu, L. K., & GaIotti, K. M. (1986). Children's reasoning Journal of Psycholinguistic Research, 15, 47-92.
about social, physical, and logical regularities: A look at two Popper, K. (1959). The logic of scientific discovery. New York:
worlds. Child Development, 57, 413-420. Basic Books.
Kuhn, D., & Phelps, E. (1979). A methodology for observing Skov, R. B., & Sherman, S. J. (1986). Information-gathering
development of a formal reasoning strategy. In D. Kuhn processes: Diagnosticity, hypothesis-confirmatory strategies,
(Ed.), Intellectual development beyond childhood (pp. 45- and perceived hypothesis confirmation. Journal of Experi-
57). San Francisco: Jossey-Bass. mental Social Psychology, 22, 93-121.
Lerner, R. M. (1986). Concepts and theories of human develop- Sodian, B., & Wimmer, H. (1987). Children's understanding of
ment (2nd ed.). New York: Random House. inference as a source of knowledge. Child Development, 58,
Moshman, D. (1979a). Development of formal hypothesis-test- 424-433.
ing ability. Developmental Psychology, 15, 104-112. Somerville, S. c., Hadkinson, B. A., & Greenberg, C. (1979).
Moshman, D. (1979b). To really get ahead, get ametatheory. In Two levels of inferential behavior in young children. Child
D. Kuhn (Ed.), Intellectual development beyond childhood Development, 50, 119-131.
(pp. 59-68). San Francisco: Jossey-Bass. von Glasersfeld, E. (1979). Radical constructivism and Piaget's
Moshman, D. (1982). Exogenous, endogenous, and dialectical concept of knowledge. In F. B. Murray (Ed.), The impact of
constructivism. Developmental Review, 2, 371-384. Piagetian theory: On education, philosophy, psychiatry, and
Moshman, D. (1989). The development of metalogical under- psychology (pp. 109-122). Baltimore: University Park Press.
standing. In W. F. Overton (Ed.), Reasoning, necessity, and Walker, L. J. (1982). The sequentiality of Kohlberg' s stages of
logic: Developmental perspectives. Hillsdale, NJ: Erlbaum. moral development. Child Development, 53, 1330-1336.
Moshman, D., & Franks, B. A. (1986). Development of the Wason, P. C. (1983). Realism and rationality in the selection
concept of inferential validity. Child Development, 57, 153- task. In J. St. B. T. Evans (Ed.), Thinking and reasoning:
165. Psychological approaches. London: Routledge & Kegan
Moshman, D., & Hoover, L. M. (1989). Rationality as a goal of Paul.
psychotherapy. Journal of Cognitive Psychotherapy, 3,31- Welfel, E. R., & Davison, M. L. (1986). The development of
51. reflective judgment during the college years: A 4-year longitu-
Moshman, D., & Timmons, M. (1982). The construction of dinal study. Journal of College Student Personnel, 27, 209-
logical necessity. Human Development, 25, 309-323. 216.
O'Brien, D. P. (1987). The development of conditional reason-
CHAPTER 12
Dialectical Thinking and

Adult Creativity
Suzanne Benack, Michael Basseches, and Thomas Swan
Perhaps because most work on creativity has origi- the acquisition of expertise in specific domains) af-
nated in personality and social psychology, there fect creative functioning? Other chapters in this vol-
has been little attention given to transformations of ume (see Moshman and Lukin, Chapter 11) have
creativity across the lifespan. Researchers have addressed this last question by examining the ways
generally taken one of two approaches. Those in- in which cognitive development from childhood to
terested in deriving nomothetic tests of creative adolescence might be related to changes in
ability or in studying social factors affecting cre- creativity. In this chapter, we will consider one
ative performance have focussed on very general particular point in the lifespan at which cognitive
features of the creative process equally applicable development and the growth of creativity might be
to people of a wide range of ages and levels of related: namely, the ways in which the development
expertise in a domain. Other researchers have been \ of dialectical thinking in adulthood might facilitate
interested in describing the creative process in the development of mature forms of creativity.
adults who have made notable creative achieve-
ments in public fields, for example, creative artists
and scientists. Although both of these approaches Adult Cognitive Development: The
have been fruitful to the study of creativity, neither Move from Formal to Postformal
lends itself to the investigation of developmental
changes in creative functioning. Operations
Several questions are central to the developmen-
tal study of creativity: How does adult creativity Piaget's series of stages of cognitive develop-
differ from creativity in childhood and adoles- ment ends in adolescence with the consolidation of
cence? What experiences contribute to the maturing formal operations. Formal operational thinking al-
of creativity? How do developments in other realms lows the late adolescent to manipulate the logical
(cognitive development, affective development, relations among abstract propositions, to think
about all logically possible states of affairs, and to
use the experimental method to test hypotheses.
Suzanne Benack and Thomas Swan • Department of Psy- The formal operational thinker solves problems by
chology, Union College, Schenectady, NY 12308.
Michael Basseches • Massachusetts School of Professional
modeling them as "closed systems" that are made
Psychology, and Clinical Developmental Institute, Belmont, up of a given number of variables that can be related
MA 02178. in a finite number of ways. By systematically ma-
199
nipulating one variable while controlling the others, tions; these three aspects are coordinated in the idea
all permutations and combinations of the variables' of the dialectical process, that is, the transformation
states can be observed and the relations among the of wholes through interactive and constitutive
variables deduced. relations.
Although this kind of formal analysis is a very Basseches's Dialectical Thinking Framework
powerful cognitive tool, many theorists have ar- consists of 24 schemata, or "moves in thought,"
gued that it is not an adequate description of mature reflecting different aspects of the dialectical world-
adult thinking. The important problems of adult view (as summarized in Table 1). There are four
thought, it is claimed, are typically ill-defined and major groups of schemata: those reflecting each
open-ended; much of the solution consists of decid- major aspect of the idea of dialectic-change,
ing upon a way to model the problem, rather than on wholeness, and relation-and those schemata or-
applying logical algorithms to the parameters of a ganizing these three aspects in the notion of the
given model. In the last decade, a number of au- dialectical process. We will briefly describe each
thors have described forms of reasoning that devel- aspect of the dialectical world view and the schema
op "beyond formal operations," in late adoles- corresponding to them.
cence and adulthood. These "postformal" cogni-
tive structures are hypothesized to retain the
Emphasis on Change
analytic power of formal analysis while going be-
yond its limitations. Although formal operations A model of reality begins with the question,
structure the relations of propositions within a "What exists?" Many worldviews hypothesize
logical system, postformal operations operate on that there are fundamental, unchanging elements
logical systems, relating multiple systems to making up reality; these elements can enter into
each other, transforming systems, and evaluating various relations with each other, but their funda-
systems. mental nature remains unchanged. In contrast, di-
The claim of this chapter is that by freeing the alectical world views assume that reality is made up
person from the constraints of formal reasoning, of fundamental processes of change or becoming,
while retaining the power of formal analysis, in which old forms give way to new emergent
postformal thinking may facilitate the development forms. Thus, from a dialectical perspective, what
of mature forms of creativity. Specifically, we will currently exists is seen not as fundamental, un-
examine the one particular model of postformal changing reality but as the current moment of an
thinking, Basseches's (1984a) depiction of dia- ongoing process of change.
lectical thinking and its potential relations to "Motion-Oriented Schemata" reflect the em-
creativity. phasis of dialectical thought on change. They de-
scribe moves in thought that function to preserve
fluidity in thought, to draw the attention of the
thinker to processes of change, or to describe such
Dialectical Thinking processes. Schemata 1 and 4 describe a process in
thought of moving from reflecting on one idea to
Dialectical thinking is a mode of cognition that reflecting upon its negation-something apart
attempts to comprehend phenomena in the world by from, contrary to, or excluded from the first idea-
seeing them as instances of dialectic. In this sense, and then moving again toward a synthesis, a more
to think dialectically is to use a particular model of inclusive idea that relates the original idea and its
reality, one that is exemplified in the dialectical negation. Schemata 2,3,5,6, 7, and 8 all function
philosophical tradition, as manifested in the works to direct the thinker's attention to processes of
of a range of great thinkers, including Hegel change and to counteract the tendency to treat pro-
(1967), Marx (1967), Darwin (1962), Von Ber- cesses as though they were static, fixed "things."
talanffy (1968), and Piaget (1952). The claimofthe For example, Schema 2 affirms that change is fun-
dialectical thinking model is that some adults exhib- damental in the nature of things; Schema 6 applies
it patterns of thought that share the underlying as- this same notion to thought and asserts that knowl-
sumptions and descriptive categories of the dialec- edge is fundamentally active and changing. Simi-
tical philosophical outlook. The dialectical larly, Schema 8 understands a present event or sit-
world view is characterized by an emphasis on uation as part of a larger historical process of
change, on wholeness, and on constitutive rela- change, and Schema 7 involves the rejection of
CHAPTER 12 • DIALECTICAL THINKING AND ADULT CREATIVITY 201
Table 1. The Dialectical Schemata Table 1. (Continued)

Framework
A. Motion-oriented schemata 23. Criticism of formalism based on the
I. Thesis-antithesis-synthesis movement in thought interdependence of form and content
2. Affirmation of the primacy of motion 24. Multiplication of perspectives as a concreteness-
3. Recognition and description of thesis-anti thesis- preserving approach to inclusiveness
synthesis movement
4. Recognition of correlativity of a thing and its
other seeing things as fixed and static when they are actu-
5. Recognition of ongoing interaction as a source of ally processes involving change.
movement
6. Affirmation of the practical or active character of
knowledge Emphasis on Wholeness
7. Avoidance or exposure of objectification, Dialectical world views reject the separating of
hypostatization, and reification reality into fundamentally independent individual
8. Understanding events or situations as moments (of elements. Instead, reality is organized into forms,
development) of a process which are coherent wholes. From a dialectical per-
B. Form-oriented schemata spective, one cannot think of the parts as separate
9. Location of an element or phenomenon within the elements that come together to make up wholes,
whole(s) of which it is a part for the very nature of the parts is determined by the
10. Description of a whole (system, form) in wholes in which they participate. Similarly, the
structural, functional, or equilibrational terms nature of a whole is more than simply the sum of
II. Assumption of contextual relativism the properties of its parts.
The "Form-Oriented Schemata" reflect the di-
C. Relationship-oriented schemata alectical worldview's emphasis on wholeness.
12. Assertion of the existence of relations, the limits These schemata describe moves in thought that (a)
of separation, or the value of relatedness direct the thinker's attention to organized or pat-
13. Criticism of multiplicity, subjectivism, and terned wholes (forms), and that (b) enable the
pluralism thinker to recognize and describe such forms. Sche-
14. Description of a two-way reciprocal relationship ma 9 directs a thinker's attention from an element or
15. Assertion of internal relations phenomenon being considered to the larger wholes
of which it is a part; Schema 10 describes the struc-
D. Metaformal schemata
tural or functional organization of wholes. Schema
16. Location (or description of the process of
11 applies the notion of wholeness to thought and
emergence) of contradictions or sources of
knowledge, by directing the thinker to see particu-
disequilibrium within a system (form) or between
lar ideas and values within the context of the larger
a system (form) and external forces or elements
conceptual frameworks of which those ideas are
which are antithetical to the system's (form's)
parts.
structure
17. Understanding the resolution of disequilibrium or
contradiction in terms of a notion of Emphasis on Constitutive Relations
transformation in developmental direction
If one views reality as made up of independent
18. Relating value to (a) movement in developmental
elements, then one will view the relations those
direction and/or (b) stability through
elements enter into as "external" to the elements
developmental movement
themselves, that is, as not determining what the
19. Evaluative comparison of forms (systems)
elements are. In contrast, dialectical outlooks em-
20. Attention to problems of coordinating systems
phasize constitutive relations. The relations among
(forms) in relation
parts within a whole help make the parts what they
21. Description of open self-transforming systems
are, and they also make up the internal structure of
22. Description of qualitative change as a result of
the whole.
quantitative change within a form
The "Relationship-Oriented Schemata" serve to
direct the thinker's attention to relationships and to
enable the thinker to conceptualize relationships in forms to each other, either by coordinating them
ways that emphasize their constitutive nature. (Schema 20) or evaluatively comparing them
Schemata 12 and 13 describe moves in thought that (Schema 19); Schema 24 specifically describes the
turn the thinker's attention to relationships through coordination of systems of thought.
the assertion of the existence of relationships, of the The term dialectical thinking is used to refer to a
limits associated with seeing things as separate, and holistic, organized worldview that irtterprets phe-
of the value of bringing things into relation. Schema nomena in terms of dialectic; this kind of thinking
14 brings out the interactive aspect of relationships, seems not to develop until at least young adulthood,
and Schema 15 describes moves in thought recog- following the consolidation of formal operations
nizing that relationships help determine the nature (Basseches, 1984a). However, a person can use
of the things being related. many ofthe individual dialectical schemata prior to
developing full dialectical thinking. Basseches
(1984b) has described four possible phases in the
The Notion of Dialectic development of dialectical thinking. In the first
The notion of a dialectical process ties together phase, schemata appear that describe initial under-
the notions of change, wholeness, and constitutive standings of the ideas of form, movement, and rela-
relations. In a dialectical process, forms or struction (Schemata 1, 2, 6, 9, 10, 12, and 16). These
tures exist consis.ting of elements related in certain schemata often appear in conjunction with or even
ways. Interactive relationships exist both within the prior to the development of formal operational
form (between the parts of the whole) and between thinking in adolescence. Next, "intermediate"
the form and that outside of it (i.e., with the en- level schemata appear in the second phase (Sche-
vironment). The interactions among the parts and mata 3, 11, and some forms of 16). In the third
between the whole and its environment eventually phase, "clusters" of more advanced schemata ap-
lead to change in the whole itself; the whole is pear. These are oriented around the ideas of form,
transformed, the relations within it are transformed, change, and value. Often people develop two of
and thus the nature of the elements is also trans- these clusters while failing to use the schemata that
formed. Moreover, this transformation is not ran- are related to the third. Finally, true dialectical
dom change; it goes in a developmental direction, thinking appears; the notions of movement, form,
that is, the whole develops toward greater in- and constitutive relation become organized into a
clusiveness, differentiation, and integration. It be- coherent dialectical worldview. At this phase, all
comes, then, a bigger, more complex, more richly three of the clusters of schemata are well repre-
organized whole. We can summarize this model of sented and Schemata 21 and 22 appear.
a dialectical process by saying that dialectic is de- Most studies of dialectical thinking have used
velopmental transformation (i.e., developmental Basseches's (1984a) original assessment pro-
movement through forms) occurring via con- cedure. People are interviewed about complex,
stitutive and interactive relationships. open-ended topics, such as the nature of education.
The "Metaformal Schemata" describe aspects The interviews are transcribed, and then use of the
of this notion of dialectical process. Thus, they inte- 24 schemata is coded. A score of 0-3 is assigned to
grate the categories of relationship, motion, and each schema, indicating whether it is clearly pre-
form. The use of metaformal schemata, then, present, probably present, probably absent, or clearly
supposes an understanding of most of the motion- absent. The scores on all 24 schemata are summed
oriented, form-oriented, and relationship-oriented to create a Dialectical Index, which can range from
schemata. Schema 16 describes moves in thought 0-72. Some researchers also have used measures of
that point out the limits of a system's stability by comprehension of and preference for arguments
pointing to contradictions either within the system using the dialectical schemata (e.g., Basseches,
or between the system and its environment. Sche- 1984b; Irwin, in press).
mata 17 18, 21, and 22 are most clearly related to We have described dialectical thinking as an
the process of transformation of forms. They direct ontology, a way of modeling phenomena in the
thought toward instances of systems that are in open world. Much of what the dialectical thinker does is
interaction with their environments, toward change to see various objects and events in terms of dialec-
in developmental directions, and toward the value tic. For example, one might take a dialectical view
of developmental chance. Schemata 19, 20, and 24 of a family, seeing it as a system in which members'
describe moves in thought that relate systems or identities are determined partly by their relations
within the family system, and seeing the family as creativity combine these two elements of novelty
evolving through the interactions of its members and value. Rothenberg and Hausman (1976), for
and its interaction with the outside world. example, defined a creative response as one having
If one applies a dialectical model to the under- both "newness" and "value" (p. 6); Amabile
standing of the nature of thought and knowledge, (1983) argued that a creative response "is both a
one derives a dialectical epistemology. Dialectical novel and appropriate, useful, correct, or valuable
thinkers generally see knowledge as an active pro- response to the task at hand" (p. 33). Bruner (1962)
cess of conceptually organizing and reorganizing saw creative products as those producing "effec-
phenomena rather than as the accumulation of fixed tive surprise" in the beholder. Barron (1955) pro-
truths. Knowledge, then, is seen as evolving, as posed that creative responses must have "a certain
changing in its fundamental structure rather than as uncommonness" and be "adaptive to reality" (pp.
simply increasing. Dialectical epistemologies also 478-479).
emphasize the structure and functioning of concep- The criteria of novelty and value focus on the
tual systems and collective wholes rather than indi- nature of the creative product; other commonly
vidual, separate facts and ideas. Several of the di- mentioned features of creativity derive from a focus
alectical schemata are directly concerned with this on the creative process. First, creativity is often
application of the idea of dialectic to epistemology described as a response to an ill-defined problem
(i.e., Schemata 6, 11, 13, and 24). rather than a well-defined problem-one in which
In the following sections, we consider ways in the nature of a solution and the path to a solution are
which dialectical thinking might promote cre- unclear. Amabile (1983) termed these two types of
ativity. Specifically, we argue (1) that a dialectical problems "heuristic" and "algorithmic" and
ontology facilitates the creative generation of nov- claimed that, in a creative response, the problem
elty and the creation of new relations, (2) that a must be defined by the person as heuristic. Similar-
dialectical epistemology supports the kinds of ly, one of the four characteristics distinguishing
moves in thought necessary for the unconventional, creative problem solving for Newell, Shaw, and
set-breaking, and the synthesizing aspects of Simon (1962) was "difficulty in problem formula-
creativity, and (3) that the evolution of a meta- tion. " Second, many authors see creative thought
systematic dialectical perspective facilitates specif- as involving the ability to move away from past
ic forms of mature creativity that are seen in highly ways of thinking, to "break mental sets." Newell
functioning creative adults. Before turning to this et al. (1962), for example, saw creative problem
task, however, it is necessary to have before us a solving as "unconventional"; Henle (1962) re-
characterization of creativity. ferred to this aspect of creativity as "freedom";
Stein (1974) described creativity as a "leap" away
from what has previously existed. Third, creativity
Characteristics of the Creative Process is often seen as the forming of relations among
things formerly disconnected. Koestler (1964) saw
There seems to be no one commonly accepted the essence of creativity as "bisociation," the asso-
definition of creativity. There are, however, core ciation of two self-consistent but normally incom-
characteristics that recur regularly in descriptions of patible frames of reference; Henle (1962) called this
the creative process or product. The first of these is aspect of creativity "harmony." Finally., some
novelty; nearly all definitions of creativity start theorists gave particular importance to the role of
from the assumption that the creative product is contradictions in the creative process. Creativity is
novel, atypical, unusual. This is the dimension seen as the ability to hold together or bring into
most often assessed in tests of creativity, which relation elements that were previously seen as con-
typically ask the person to generate a large number tradictory. Kuhn (1963), for example, saw diver-
of unusual responses to standard stimuli (e.g., the gent thinking in science as a response to anomalies,
Torrance Test of Creative Thinking, 1962, and the data contradicting the existing paradigm. Rothen-
Unusual Uses Test of Guilford, 1967). Most defini- berg (1976) described a process of "lanusian think-
tions of creativity also stipulate that the creative ing," the holding together of apparently contradic-
response should meet some criterion of value. A tory views as essential to creativity.
creative response must not merely be unusual; it The current literature, then, describes creativity
must be effective or useful or an appropriate solu- as a response to ill-defined problems, involving the
tion to a problem. The most common definitions of breaking away from existing ways of thinking,
creating relations among dissociated or even con- the cake in all these ways and in all these relations
tradictory elements, and resulting in novel and val- helps the dialectical thinker generate a much richer
uable products. Given this characterization of the and more novel set of factors to consider in potential
creative process, how might dialectical thinking cake designs.
contribute to creativity? In the following discus- To take an example in the realm of adult scien-
sion, we consider the relations among creativity and tific thought, consider how two psychologists, one
three aspects of dialectical thinking: (a) the dialec- a nondialectical and the other a dialectical thinker,
tical understanding of reality (dialectical ontology), might try to explain why adolescents exhibit re-
(b) the dialectical understanding of thought (dialec- bellious behavior. The nondialectical psychologist
tical epistemology), and (c) the dialectical under- begins by elaborating a set of truths about the entity
standing of the interrelation and evolution of sys- "adolescents": they have erratic hormone levels
tems (metasystematic perspective). associated with pubertal changes, they show high
conformity to peer group norms, and they are capa-
ble of some forms of abstract reasoning unavailable
Dialectical Ontology and Creativity to children. This psychologist might then consider
which set of these factors could cause rebellion. In
Many authors posit that, in a creative response, contrast, the dialectical psychologist will consider
one perceives a given object or event in multiple the act of rebellion in several contexts and relations:
and novel ways, or brings it into new relations to as a part of the overthrowing of childhood ways of
other objects or events. Using a dialectical perspec- functioning, as a strategy in the process of identity
tive to model or understand a phenomenon should formation, as a chapter in the changing parent-
support this aspect of creative functioning. In a non- child relationship, as a way of negotiating one's
dialectical perspective, an object or event is seen as relation with the wider society, and as an act within
essentially static and separate from other phe- the adolescent peer culture in relation to changes in
nomena; a thing unto itself. The same object, moral thinking in this period. In all these contexts,
viewed from a dialectical perspective, is likely to be rebellion is likely to be seen not as a static phe-
seen as a process rather than a thing, changing nomenon in itself but as part of the ongoing pro-
rather than static, as part of a larger whole, and as cesses of change.
fundamentally in relation to other phenomena. Di- The use of dialectical schemata to interpret real-
alectical schemata facilitate creative responses, ity, then, provides paths in thought leading from the
then, by directing attention from "the thing itself" "present object" to a wider set of realities: its past
to its history, its future, the systems in which it is and future history, its contexts and relations. Thus,
embedded, and its relations. dialectical ontology "opens up" one's perception
Consider, for example, how a nondialectical of reality, lends to it more variety and more connec-
thinker and a dialectical thinker might approach de- tedness-both central features of the creative
signing a birthday cake. The nondialectical thinker process.
starts with an image, perhaps, of a completed birth- Another aspect of dialectical thinking that
day cake and considers a number of variables that should support the generation of novel responses is
might be manipulated: the shape, the color, the that dialectical thinkers tend to move from a given
size, and the type of candles. The dialectical perception or thought to its antithesis: that which
thinker, on the other hand, might think about the stands outside or in contradiction to it (Schema 1).
birthday cake as a process: its creation, its presenta- For example, if it occurs to the dialectical thinker
tion at the party, its consumption, the memories that birthday cakes generally are done in bright,
people will have of it after the party. Moreover, the contrasting colors, she or he is likely to consider
dialectical thinker might think about the role the making a cake in dark, muted colors. If she or he
cake plays in various contexts and relations: as an recognizes that, as part of a larger cultural tradi-
act in the life of the person who creates it, as a tion, the candles on birthday cakes should reflect
moment in the process of the party, as an aesthetic the person's age, she or he might generate an an-
object in the scheme of party decorations, as an tithesis by saying, "Perhaps this tradition isn't
important object to the birthday child, as food for necessary . . . maybe the candles could symbolize
the other guests at the party, as part of a cultural something else." Moreover, the dialectical thinker
tradition of birthday cakes, as part of a sequence of responds to these contradictions by trying to form a
birthday cakes across the child's life. Considering synthesis, a new higher order relationship reconcil-
ing or relating the opposing ideas or perceptions. liefs and observations is highly valued. The formal
The tendency of dialectical thinkers to generate operational thinker interprets problems, then, as
contradictions and then find syntheses reconciling "algorithmic" in Amabile's terms, as having de-
them should support both the novelty-generating fined paths leading to clearly and uniquely correct
and the relationship-creating functions of creati- solutions.
vity. This kind of formal analysis is an extremely
powerful tool; it is seen by many philosophers of
science as the basis of theory testing in science. On
Dialectical Epistemology and Creativity the other hand, formal operational analysis does not
seem like an adequate description of the creative
In adolescence and adulthood, our epis- aspects of scientific thought: that is, of theory cre-
temologies-the models we create of the nature of ation rather than theory testing. Creativity in fields
our own thought-come to play an important role like science, which are based upon formal analysis,
in regulating thought. Our epistemological beliefs seems to require cognitive operations that, on the
tell us how to proceed in solving problems, when to one hand, retain the power of systematic thinking
accept solutions as true, and how to feel about while, on the other hand, transcend its limitations.
events in our mental life, such as discovering con- Dialectical thinking may be particularly useful in
tradictions to our beliefs, being uncertain, or find- fostering creativity in these kinds of fields, because
ing our beliefs changing. In one sense, then, hold- the dialectical perspective subsumes formal,
ing a dialectical view of knowledge is simply one closed-system analysis within a broader model de-
more application of the dialectical world view to scribing transformations and interrelations of for-
reality; in another sense, however, it has particular mal systems. How, then, might holding a dialec-
importance for directing one's thinking. Basseches tical epistemology, a model of the evolution of
(1984a) reflects the special role of a dialectical epis- thought systems, support creativity in these kinds of
temology in directing thought by including several fields?
schemata that specifically reflect the application of First of all, the dialectical thinker sees his or her
dialectical perspectives to the topic of knowledge own thought as in a process of evolution. Where the
(Schemata 6, 11,23, 24). A dialectical model of formal thinker sees the necessity to change his or
knowledge is likely to foster creativity, we argue, her thought only if the old view is "in error," the
because it represents the process of thinking as cre- dialectical thinker sees change in his or her thinking
ative. In other words, the features of thought a di- as natural, expectable, and valuable. Thus, a di-
alectical model emphasizes and encourages are pre- alectical view of knowledge encourages persons to
cisely those described as important to the creative be ready and willing to move away from their past
process. points of view, to do the kind of "set breaking" or
Consider how a nondialectical thinker, using the "leaping" away from an old tradition that has been
tools of formal operational analysis, approaches a seen as characteristic of creative thinkers.
problem. First, the person already holds some sys- Second, in dialectical epistemology, particular
tematic framework from which to approach the beliefs and values are seen as part of larger thought
problem. This framework will specify a finite systems. Differences of opinion exist, then, not be-
number of variables to be considered and will de- cause one answer is "right" and the others are
fine other aspects of the problem as irrelevant to the "wrong," but because there are many frameworks
solution. The relevant variables are then related in or perspectives from which to view any problem.
all combinations and permutations of their various This awareness of multiple systematic ways of
values, finding their interrelations through logical viewing reality makes one's own view more per-
deduction and/ or experimental manipulation of one meable, more influenceable by other perspectives,
variable while holding the others constant. The for- perspectives in which the problem is defined in fun-
mal thinker expects this method to produce a single damentally different ways. This tendency of dialec-
right answer that will hold in all similar circum- tical thinkers to be aware of and look to other per-
stances and across time. Contradictions-observa- spectives than their own should be a source of
tions inconsistent with this answer or disagreement greater diversity and novelty.
from other people-are regarded as a sign that The dialectical thinker sees the evolution of
something is wrong with one's solution; the goal is knowledge as resulting from contradictions within a
to eliminate them. Internal consistency of one's be- thought system or between a thought system and
that which is outside it. For the dialectical thinker, cess, and (b) that a dialectical view of knowledge
then, contradictions play a key role in intellectual will provide ways of retaining the power of formal
growth. One might expect, therefore, that holding a systematic analysis, while fostering the creative
dialectical epistemology would alter one's response evolution of thought systems. There is one other
to contradictions that are encountered in problem aspect of dialectical thinking that may be of particu-
solving: rather than being signs of trouble, irritants lar importance to what one might think of as the
to be ignored when possible and eliminated when optimal functioning of creativity, that is, the kind of
necessary, contradictions are opportunities to be sustained creative work carried on by adults who
sought out and developed. are eminent in their fields. Gruber (1981) con-
Finally, a dialectical epistemology directs the ducted an intensive study of this kind of sustained
person, when encountering contradictions, to look creative achievement in the life of Charles Darwin
for ways of resolving them in higher order syn- and presented an "evolving systems approach'; to
theses, by creating new, more complex systems en- creativity (Gruber, 1984). He claimed that signifi-
compassing the old contradictory elements. For cant creative achievement is not so much a matter of
those authors who see creativity as involving hold- a single brilliant insight, but represents the slow
ing together or relating contradictory ideas or development of a "novel point of view. " He saw
frameworks (e.g., Rothenberg's, 1976, "lanusian creative production as resulting not from a few dra-
thinking" or Koestler's, 1964, "bisociation"), di- matic moments of intuition or genius, but from a
alectical epistemology serves as a "roadmap" for lifelong organization of work and life leading to-
the creative process. ward and supporting the evolution of this point of
In sum, then, a dialectical understanding of view. He described the creative person as being
knowledge gives both cognitive and affective sup- comprised of three organizational subsystems: the
port to the processes many authors see as central to organization of knowledge, the organization of af-
creativity. On the cognitive side, dialectical epis- fect, and the organization of purpose. These sub-
temology might be seen as providing a set of direc- systems overlap and interact, but are to an extent
tives to thought: Expect your way of thinking to independent. The organization of knowledge con-
change. Consider the problem from multiple persists of a complex hierarchy of beliefs that are con-
spectives: how would people with different frame- stantly being altered in response to new data and
works see it? Look out for contradictions and, when new points of view. The organization of affect con-
you find them, pay close attention to them. Create sists of a "symphony" of affective themes and
ways of relating and synthesizing things that seem tones, relating to people's work and to their person-
to be in opposition or inconsistent. All these "hab- allife and needs. The organization of purpose is a
its of thought" foster creativity. On the affective complex network of tasks, goals, projects, and en-
side, a dialectical understanding of the evolution of terprises that the person intends to carry out. Each
knowledge helps a person to support the emotional of these subsystems includes several smaller sys-
tensions of the creative process: the tension of hold- tems that are themselves complexly interrelated,
ing opposing views simult'meously, of sustaining with each subsystem evolving over time. More-
uncertainty, of breaking away from an established over, Gruber sees the creative person as con-
way of seeing things, and of tolerating ambiguity. sciously monitoring and directing the evolution of
A dialectical view of knowledge helps to support the three systems and their interrelation.
these affective tensions in creative thought, by as- Far from being an irrational or intutive "leap,"
serting that these are natural and valuable parts of then, significant creative achievement, for Gruber,
the evolution of thought, and that they do arises from an individual's ability to conceptualize
eventually lead to the creation of new and more and direct the evolution of a complex organization
adequate knowledge. ofthoughts, feelings, and purposes in life, centered
around the telos of the creative achievement.
It would seem that formal operations, which act
Metasystematic Thinking and Creative on the relations of propositions and variables within
Work: The Evolving Systems Approach a single system, but cannot manipulate systems as
wholes, would be an inadequate cognitive structure
So far, we have suggested (a) that a dialectical for this kind of task. The implication of Gruber's
view of objects and events will foster the attention model is that mature, sustained creative effort re-
to novelty and relation central to the creative pro- quires exactly the kinds of cognitive abilities de-
scribed by the metaformal dialectical schemata: the One question future research in this area might
ability to understand, anticipate, and direct the evo- address is whether the kinds of processes that
lution of a system (Schemata 16, 17,21,22) and the Gruber observed in the development of Darwin's
ability to relate multiple systems to each other thinking are seen in other creative scientists and in
(Schemata 19, 20, 24). The advanced stages in the creative persons in other fields. Does artistic
development of dialectical thinking, then, make it creativity, for example, also rely on conscious or-
possible for a person to consciously direct his or her ganization of systems of ideas and beliefs and pur-
own creative process in projects that are extended in poseful construction of networks of plans and pro-
both time and scope, such as are characteristic of jects? Case studies of creative achievers in different
mature creative achievement in public, organized disciplines could explore whether they use dialec-
fields of endeavor. tical models to explain and direct the evolution of
their thought and work.
A second line of inquiry would be to look across
the lifespan at the development of ways in which
Summary and Implications for Future people become aware of and transcend the limits of
Research various systematizations of experience. Fully de-
veloped dialectical thinking, we have seen, enables
We have suggested that the development of di- one to become aware of and transcend the limita-
alectical thinking might facilitate the growth of tions of formal thinking in ways that might foster
creativity in several ways. First, a dialectical view adult creativity. It may be, however, that children
of objects and events (dialectical ontology) should and young adolescents also have ways of coming to
foster awareness of novelty and of relations among understand and go beyond the systems that they use
things. Many of the schemata that direct attention to to organize experience and solve problems, and that
change and relation appear concurrent with or even fostering this kind of critical awareness of their cog-
prior to the acquisition of formal operations. These nitive systems may be important in supporting
early dialectical understandings of change and con- creativity. For example, one might look at chil-
stitutive relation may be related to creativity, then, drens' attention to and understanding of contradic-
even in childhood and adolescence. Second, a ditions, their sense of their own thought as static or
alectical view of knowledge seems likely to foster changing, their tendency to construe problems as
habits of thought that will promote set-breaking, closed- or open-ended.
attention to contradiction, and attempts at syn- Finally, creativity has generally been studied in
thesis, all of which are important features of the arenas where the product is public and discrete-a
creative process. Dialectical epistemologies may be work of art, a scientific theory, a short story. Often
particularly important in promoting creative the assessment of creativity rests upon an evalua-
thought in disciplines that rely heavily on formal tion of these kinds of products, as though the
analysis. In these contexts, a dialectical epis- creativity lay in the thing rather than in the process
temology provides the person with a model of cre- that produced it. If we take seriously Gruber's per-
ative thought that does not involve relinquishing the spective, that the creation of a novel point of view is
power of the formal system, but rather involves fundamental to the creative process, then it might
integrating the present formal system in a larger be relevant to explore people's formation of novel
view of multiple systems evolving through interre- points of view in their conduct of their everyday
lation with each other and with data. Finally, the life, even when this process does not result in pub-
development of a coherent, metasystematic dialec- lic, discrete products. For example, people are
tical perspective may provide the cognitive opera- "authors" of their moral systems, their sense of
tions that are necessary to carry out the kind of personal identity, their views of politics, and their
conscious management of interrelated evolving ways of understanding and relating to their chil-
systems which Gruber (1984) suggests is charac- dren. Surely these are realms in which people are
teristic of mature, sustained creative efforts. In more or less creative, in which some people do and
sum, then, early dialectical schemata may foster the others do not create novel, valuable, and unconven-
most basic and general functions of creativity, such tional approaches. Structural-developmental psy-
as the generation of novelty, whereas the full devel- chologists have developed the techniques of semi-
opment of dialectical thinking may make possible clinical interviewing and structural analysis to
mature, specifically adult forms of creativity. study people's thinking about these kinds of topics;
we suggest this methodology might prove useful in Bruner, 1. S. (1962). The conditions of creativity. In H. E.
beginning to look at creativity in people's thought Gruber, G. Terrell, & M. Wertheimer (Eds.), Contemporary
about and conduct of their lives. approaches to creative thinking (pp. 1-30). New York: Athe-
rton Press.
In general, the approach to creativity taken in this Darwin, C. (1962). Origin of species. New York: Macmillan.
chapter is to move from a view of creative activity Gruber, H. (1981). Darwin on man: A psychological study of
as rooted in a trait of creativity, where some people scientific creativity. Chicago: University of Chicago Press.
have more of it than others, to seeing creative ac- Gruber, H. E. (1984). The emergence of a sense of purpose: A
cognitive case study of young Darwin. In M. L. Commons, F.
tivity as reflecting the development and application A. Richards, & c. Armon (Eds.), Beyondformal operations:
of specific cognitive and affective capacities. These Late adolescent and adult cognitive development (pp. 3-27).
capacities may be seen as growing out of the in- New York: Praeger.
teraction of prior cognitive structures (e.g., con- Guilford, 1. P. (1967). The nature of human intelligence. New
York: McGraw-Hill.
crete operations, formal operations) with experi-
Hegel, G. W. F. (1967). The phenomenology of mind. New
ences that confront concrete and formal operational York: Harper & Row.
models of phenomena with their limits. Unfortu- Henle, M. (1962). The birth and death of ideas. In H. E. Gruber,
nately, people often tend to avoid and defend G. Terrell, & M. Wertheimer (Eds.), Contemporary ap-
against experiences that would challenge the mod- proaches to creative thinking (pp. 31-62). New York: Ather-
ton Press.
els on which they rely to make sense of themselves Irwin, R. (in press). A critique of the proposed stage of dialec-
and the world, and therefore the development of tical thinking. In M. L. Commons, 1. D. Sinnott, F. A. Rich-
these dialectical capacities may be rare. However, ards, & C. Armon (Eds.), Beyondformal operations: Vol. 2.
these capacities can be fostered through adequate Comparisons and applications of adolescent and adult devel-
opmental models. New York: Praeger.
educational experience and personal support for the Koestler, A. (1964). The act of creation. New York: Macmillan.
painful disequilibrium and uncertainty of having Kuhn, T. S. (1963). The essential tension: Tradition and innova-
one's trusted models challenged. Although this tion in scientific research. In F. Barron & C. W. Taylor
poses a major challenge to educationists and other (Eds.), Scientific creativity: Its recognition and development
supporters of human development who are con- (pp. 341-354). New York: Wiley.
Marx, K. (1967). Writings of the young Marx on philosophy and
cerned with creativity, it opens the door to doing society. L. D. Easton & K. H. Guddat (Eds.). Garden City,
more for creativity than merely marveling at it or NY: Anchor.
measuring it. Newell, A., Shaw, I., & Simon, H. (1962). The processes of
creative thinking. In H. Gruber, G. Terrell, & M. Wertheimer
(Eds.), Contemporary approaches to creative thinking (pp.
63-119). New York: Atherton Press.
References Piaget, 1. (1952). The origins of intelligence in children. New
York: Norton.
Amabile, T. M. (1983). The social psychology of creativity. Stein, M. I. (1974). Stimulating creativity (Vol. 1). New York:
New York: Springer-Verlag. Academic Press.
Barron, F. (1955). The disposition toward originality. Journal of Rothenberg, A. (1976). The process of lanusian thinking in
Abnormal and Social Psychology, 51, 478-485. creativity. In A. Rothenberg & C. R. Hausman (Eds.), The
Basseches, M. (l984a). Dialectical thinking and adult develop- creativity question. Durham, NC: Duke University Press.
ment. Norwood, Nl: Ablex. Rothenberg, A., & Hausman, C. R. (Eds.) (1976). The
Basseches, M. (1984b). Dialectical thinking as ametasystematic creativity question. Durham, NC: Duke University Press.
form of cognitive organization. In M. L. Commons, F. A. Torrance, E. P. (1962). Guiding creative talent. Englewood
Richards, & c. Armon (Eds.), Beyond formal operations: Cliffs, Nl: Prentice-Hall.
Late adolescent and adult cognitive development (pp. 216- Von Bertalanffy, L. (1968). General system theory. New York:
238). New York: Praeger. Braziller.
PART III
Personalogical Variables
and Creativity
Part III is comprised of four chapters, each emphasizing a different element of how personalogical
variables influence creativity. Martindale provides a general perspective for thinking about personalogical
variables and creativity. Dowd explores a difficult and often controversial topic in his chapter on creativity
and the self. Prentky examines the evidence surrounding the notion that creativity may be related to
psychopathology. Finally, Heppner, Fitzgerald, and Ball discuss the role of creativity in the therapeutic
process.
CHAPTER 13
Personality, Situation, and Creativity

Colin Martindale
Introduction be taken into account: men and women create in

different ways. Again, Helson (e.g., 1973b, 1978)
Because creativity has to do with the production of has taken the lead in exploring these differences.
new ideas, one might think that its study rightly Across the last several decades, it has become clear
falls within the domain of cognitive psychology. Of that personality cannot be considered in isolation,
course, creativity involves cognition, but it in- but social context must also be taken into account.
volves a type of cognition that seems only to occur Although I touch on this issue as it relates to
within a matrix of associated motivational, at- creativity, much more information may be found in
titudinal, and personalogical traits. Thus, to under- Amabile (1983a), Martindale (1975), and Simon-
stand creativity, the person as a whole must be con- ton (1984).
sidered. Because of this, theories about the creative
process have traditionally been personality theories
Definitions of Creativity
rather than purely cognitive theories. In 1949,
Guilford (1950) pointed out that we did not know A creative idea is marked by three attributes: It
enough about creativity. We can never know too must be original, it must be useful or appropriate for
much about the creative personality, but we cer- the situation in which it occurs, and it must actually
tainly know more than I could hope to cover in this be put to some use. To be counted as creative, it is
chapter. For more information, the reader may con- not enough that an idea be original. Original ideas
sult the reviews of the literature by Dellas and Gaier are common enough, but few of them qualify as
(1970), Wallach (1970), Stein (1974), Taylor and creative. The word salad of a hebephrenic is cer-
Getzels (1975), and Barron and Harrington (1981). tainly original, but for a variety of reasons it cannot
As a point of departure, it is useful to consider- serve as the basis for a poem. By the same token,
as I have in this chapter-a "composite pho- the delusions of a paranoid are original but hardly
tograph" or prototype of the creative personality. serve as a fruitful source of scientific hypotheses.
This gives a general outline but with a good deal of Perfectly normal people also produce original ideas
blurring. Thanks to the pioneering work of Helson that are useless for a variety of reasons-they are
(e.g., 1973a, 1977), it is known that there are really ridiculous, incorrect, and impractical. Even if an
several rather distinct types of creative person- original idea is useful, it does not really count as
alities. There are also important sex differences to creative unless it is actualized or at least communi-
cated. If I had an original and useful idea about the
Colin Martindale • Department of Psychology, University causes of creativity, we would not count the idea as
of Maine, Orono, ME 04469-0140. creative unless I communicated it to others. Or, to
211
212 PART III • PERSONALOGICAL VARIABLES AND CREATIVITY
take another example, were I to think of the perfect then we do not need different explanations of
plot for the Great American Novel, this would be of creativity for different areas of endeavor but only
little import unless I actually wrote the novel. one general explanation. The idea that creativity is a
Creative ideas are always new combinations of general rather than a domain-specific trait is dis-
old ideas. As Poincare (1913, p. 386) put it, "to cussed in detail by Root-Bernstein (1984), Wad-
create consists of making new combinations of as- dington (1969), and Tang (1984).
sociative elements which are useful." He went on Combining elements in a new way generally in-
to remark that creative ideas "reveal to us un- volves perception of an analogy. Consider how
suspected kinships between other facts well known Cyrus McCormick invented the grain reaper
but wrongly believed to be strangers to one another. (Weber, 1969). He noted that grain is like the hair
Among chosen combinations the most fertile will on a person's head. Because mechanical clippers
often be those formed of elements drawn from do- can be used to cut hair, they could also be used to
mains which are far apart" (p. 386). Consider Ein- cut shafts of grain. If McCormick had been a poet
stein's equation, E = me 2 . Einstein did not invent rather than an inventor, he might have composed a
the components of the equation de novo nor did he simile rather than creating a machine. I have sug-
invent the concept of energy or the operation of gested the following possibility (Martindale, 1981):
raising a quantity to a power. By the same token, a
poet does not generally invent new words but puts Your hair is like the golden grain,
old words together in new ways. And the dandruff thereupon is but the chaff.
Most investigators have defined creativity in the
way stated above: There must be a product of some I am quite sure that a better poem could be wrung
sort and it must be both novel and appropriate or out of the analogy. The point, however, is that the
useful in some sense (Barron, 1955; Mednick, creative process in poetry and science is really the
1962; Newell, Shaw, & Simon, 1962; Stein, 1974). same thing. In this example, the only difference is
As Bruner (1962) put it, a creative product produces what is derived from the newly discovered analogy.
"effective surprise" and a "shock of recognition" Generally, poets and scientists combine different
that the idea is correct. Koestler (1964) made it sorts of elements. Energy and mass do not make for
explicit that creativity often involves not a com- good poetry, after all. However, the thought under-
bination of isolated elements but a connection lying scientific and poetic creativity seems to be
of two entire "matrices of thought." Amabile quite similar if not identical. A creative physicist,
(1983b) added the requirement that the idea be pro- for example, seems to think, talk, and behave more
duced in a heuristic rather than an algorithmic fash- like a creative poet than like an uncreative
ion. An algorithmic task is one in which there is a physicist.
clear and specifiable method of solution; for exam-
ple, how differential equations are solved. A
heuristic task is one in which there is no known Prerequisites for Creativity
"recipe" for a solution. This is an important point.
Capacity to Produce Creative Combinations
Many ideas that appear creative to laymen may not
be so at all. Experts would recognize that the idea Before consulting any of the research literature,
rose from already known algorithms. For example, we can infer some of the necessary conditions for
if one knows nothing about mathematics, it will be creativity. If creativity involves new combinations
impossible to differentiate between creative and un- of mental elements, then it would certainly seem to
creative mathematical ideas. be the case that the more mental elements a person
had, the more creative he or she should be. Besides
having a lot of mental elements, they should also be
Creativity as a General Trait
distributed across a wide spectrum of domains, if
Ultimately, all creative products have this quali- Poincare is correct that remote associations are
ty: old ideas or elements are combined in new ways. most likely to give rise to creative ideas. This does
This is the case for all domains of creativity. A seem to be the case. Being an expert in a given
composer puts together musical tones in new ways, specialty does not guarantee that one will be cre-
whereas a poet puts together words in new ways. If ative in that area (Simon, 1983).
the ability to recombine elements is a general one, We might guess that intelligence should be a
CHAPTER 13 • PERSONALITY, SITUATION, AND CREATIVITY 213
good predictor of creativity, because the more intel- difference between more and less creative people is
ligent one is, the more mental elements one should in the way that they think? We know that this is not
be able to acquire. However, this may not be the the case. As shall be seen, creative cognition tends
case. Beyond an IQ level of 120 or so, intelligence to occur only within a certain configuration of per-
and creativity are often held not to be closely related sonality traits. Even if we did not know this, we
(e.g., Barron, 1955). It would seem to be the case could be moderately sure that creativity would be
that a minimal IQ of 120 or so is necessary for one to connected with certain personality traits. Thomas
be creative in any meaningful sense of the term. Alva Edison remarked that genius is 1% inspiration
However, beyond that level, intelligence does not and 99% perspiration. Because inspiration is more
predict creativity as well as might be expected. One interesting than perspiration, it has been studied
doubts, however, that an IQ of 120 would be suffi- much more intensively. However, the 1% versus
cient for creative work in all disciplines. Different 99% partitioning of the "variance" in creativity is
areas of endeavor most certainly require different probably close to the mark.
minimal levels of intelligence. Once this minimum Consider, for example, the production of a scien-
level is reached, however, creativity and intel- tific article. The idea behind the article-whether
ligence are not highly correlated (Wallach & Ka- creative or not-as likely as not came with little
gan, 1965). By elimination, then, it must be the effort. In any event, the effort was slight compared
case that creativity is dependent upon a particular with the time and effort involved in operationaliz-
type of cognition, a particular method of combining ing the idea, performing the experiment, analyzing
mental elements. In order to understand creativity, the data, and writing the article. A good bit of in-
we need to discover what this type of thought is. terest, perseverence, and/or ambition must be pre-
sent or the article would never be completed. Given
Domain-Relevant Skills that most highly creative scientists produce an aver-
age of about 200 articles across the course of their
A necessary but not sufficient condition for
careers (Dennis, 1954), it must be the case that they
creativity is that one have certain skills or knowl-
.are driven by very high levels of motivational fac-
edge relevant to the area in which one is working. It
tors, such as interest, curiosity, or ambition.
is clear enough, for example, that one cannot think
Although I am sure it may happen occasionally,
of a creative idea about physics if one does not
it seems unlikely that one would choose a career
know anything about physics. You cannot very well
goal of being a mediocre poet. Rather, if one
combine mental elements in a new way if the ele-
decides to be a poet, this is almost always a decision
ments are not known to you in the first place. It is
to be a great poet. Let the reader consider the com-
also fairly clear that certain aptitudes or special abil-
petition: one is going up against Shakespeare,
ities not directly connected with creativity are nec-
Goethe, Milton, and others like them. Once this
essary (Amabile, 1983b; Feldman, 1980). For ex-
fact occurs to most people, they are likely to decide
ample, to be a creative composer, one needs not
upon a rather less challenging line of work. Cre-
only ability for creative thinking but also musical
ative people must have extremely high levels of
talent. Although there are certainly many notable
self-confidence. This does seem to be the case.
exceptions, creativity is generally confined to a sin-
Creative people might also be expected to have
gle domain. Michelangelo and Dante Gabriel
high levels of self-confidence if we consider the
Rossetti were both poets and visual artists, but they
are exceptional. In general, poets are not good most likely reaction to creative ideas. This reaction
is, of course, often extremely negative. Without a
painters and painters are not good poets. In fact,
many visual artists are rather deficient in their ver- good deal of self-confidence, one would hardly be
expected to venture toward a goal (production of a
bal skills. Even though I shall not focus on the issue
of domain-specific skills in this chapter, it should creative idea) that if reached would quite likely re-
be kept in mind that such skills are probably crucial sult in derision, hostility, and so on. Most people
in creative achievement. simply do not like novelty. It must be the case that
creative people do like it, otherwise they would take
no pleasure in producing creative ideas and, indeed,
Motivational and Personality Factors
would produce none. Except under rather unusual
Is is possible that creativity could be a purely circumstances, people do not, of course, do things
cognitive matter? That is, is it possible that the only that bring them displeasure.
Situational Variables The Creative Process

Studies of creativity have tended to focus upon
Stages of the Creative Process
cognitive and personalogical variables. In light of
Mischel's (1979) demonstrations that, at the very Based upon the self-reports of eminently creative
least, situation is often as good a predictorofbehav- people, there is general agreement that the creative
ior as are personality traits, it would seem that more process consists of four successive stages. Wallas
attention should be paid to situational factors affect- (1926), drawing heavily upon the observations of
ing creativity. Common sense tells us that certain Helmholtz (1896), labeled them preparation, in-
situations must be more conducive to creativity than cubation, illumination, and verification. Helmholtz
others. It is difficult to imagine creative thought in a (1896) noted that, when confronted with a problem,
situation in which such thought is severely punished he as often as not worked intensively on it but came
or in which, by dint of circumstances, the mental up with no solution. This is the preparation stage:
elements one thinks about are not susceptible to elements presumed to be relevant to the problem are
creative combinations. If the task at hand were bur- learned and/ or manipulated in an intellectual man-
ying one's grandmother, one could think of novel ner. When progress was not made, Helmholtz set
methods, but they would probably not be creative, the problem aside. This is the incubation stage.
because they would likely be inappropriate, given After some period of time, often with no clear
that funeral ceremonies are tightly constrained by cause, the solution simply came to his mind. This is
social and legal restrictions. the stage of illumination or inspiration. As often as
Everyone knows that the chance visit of "a per- not, it was not the elements that he worked with
son from Porlock" considerably shortened one of during the preparation stage that were combined in
the greatest poems ("Kubla Khan") in the English the flash of illumination, but, rather, some element
language. Interruption is one of the lethal enemies not before considered relevant that provided the
of creativity (compare Tchaikovsky, 187811906). key. After inspiration, the verification stage in-
At least some minimal amount of time for solitary volved his subjecting the idea to scrutiny and put-
contemplation must be a necessary factor in ting it into its final form. For a scientist, this might
creativity. As we shall see below, other less ob- involve devising and conducting an experiment.
vious situational factors also have effects on For a poet, it might involve putting an image into a
creativity (Amabile, 1983a; Simonton, 1984). form consistent with stylistic rules of the genre
within which he works.
Conclusions
Creative Inspiration
By necessity, creative people must possess a cer-
tain set of traits. On the cognitive level, they must The sequence of stages outlined above seems to
be able to think in some unusual way unavailable to be very general. Neither scientists nor artists get
most people, otherwise they would be unable to their creative ideas from purely logical, intellectual
conceive of creative ideas. In most areas of endeav- work. After reviewing a large number of self-re-
or, specialized talents, whether learned or innate, ports of creative people, Ghiselin (1952, p. 5) con-
must also playa role. On the motivational level, cluded that' 'production by a process of purely con-
creative people must have high levels of self-confi- scious calculation seems never to occur. " This is a
dence and ambition, perseverance, or interest, oth- very strong statement, but almost everyone who has
erwise, they would be unable to bring their ideas to written about the creative process has drawn a sim-
fruition. Finally, situational factors must foster or ilar conclusion (compare Harding, 1965; Weber,
hinder creative accomplishment. Note that there is 1969). About the only serious dissenter is Edgar
no obvious a priori reason to expect the cognitive, Allan Poe, who held that writing is a purely rational
special talent, motivational, and situational factors enterprise. Given that much of Poe's output is-to
to covary. Yet all these factors must be present if a borrow a phrase from Jung- "slime from the un-
person is to be creative. Perhaps this is why conscious" and that Poe was probably more often
creativity is so rare. It depends upon the simul- inebriated than rational, it is difficult to take his
taneous presence of a set of traits and factors-none theory very seriously.
of which is especially rare-that may not, in gener- Poets and other writers are quite explicit about
al, be highly correlated with each another. the effortless and nonintellectual nature of in spira-
tion. Creative inspiration seems to occur in an al- ative idea. One may presume that something caused
tered state of consciousness (Martindale, 1981). the novel combination of elements, but the creators
Nietzsche (1908/1927) described the composition do not make clear what it was. Let us consider a
of Thus Spoke Zarathustra as follows: final example, which concerns how I thought of my
theory of aesthetic evolution (Martindale, 1984a).
Everything occurs quite without volition, as if in an eruption of Let us leave aside the question of how creative the
freedom, independence. power and divinity. The spontaneity of theory is; the issue is how it was conceived. The
the images and similies is most remarkable. (p. 897) initial problem concerned the history of French po-
etry-specifically why its content had become
William Blake's (1803/1906, p. 115) comment more primary process in nature and its metaphors
about the composition of his poem on Milton is and similes more remote across the course of the
more extreme: "I have written this poem from im- last several centuries. The trend in metaphors could
mediate dictation, twelve or sometimes twenty or be ascribed to a need for novelty, which was ram-
thirty lines at a time without premeditation, and pant among the poets in question. But need for
even against my will." Nietzsche and Blake were novelty did not explain the regression toward pri-
rather eccentric to put it mildly. Thus, it is reassur- mary process cognition, and need for novelty itself
ing that even the perfectly normal English novelist, needed to be explained. Although the problem had
William Makepeace Thackeray (1899, p. 97), de- been laid aside without a satisfactory answer, the
scribed a similar possession: "I have been surprised solution occurred immediately while I was reading
at the observations made by some of my characters. an article by Mednick (1958) on schizophrenia.
It seems as if an occult power was moving the pen. Mednick argued that schizophrenics have avoid-
The personage does or says something, and I ask, ance gradients around anxiety-arousing words and
how the dickens did he come to think of that?" that-for reasons I do not know, since I never
It is surprising to many that scientists and mathe- finished reading the article-these gradients ex-
maticians give very similar descriptions of their expand over time. The complete solution to the prob-
periences, as this example from the French mathe- lem concerning the history of poetry leaped imme-
matician, Henri Poincare (19l3): diately to my mind: of course, poets are faced with
continually expanding avoidance gradients around
One evening contrary to my custom. I drank black coffee and words. Once a simile about a word has been used, it
could not sleep. Ideas rose in crowds; I felt them collide until cannot be reused; a more remote associate must be
pairs interlocked, so to speak, making a stable combination. By
the next morning I had established the existence of a class of found. How this related to primary process content
Fuchsian functions, those which come from the hypergeometric was apparent as well but need not concern us here.
series; I had only to write out the results, which took but a few The inspiration was quite effortless and automatic
hours. (p. 387) and, in this case, its cause is quite clear. The ele-
ments in the Mednick article-avoidance of words,
Scientists who engage in what Kuhn (1962) called regression (of schizophrenics)-were similar
revolutionary science uniformly seem to arrive at enough to the elements of the poetry problem to
their ideas in this way. By revolutionary science is draw the latter into consciousness and, at the same
meant the sort of theorizing that establishes a new time, to solve it. Note a rather embarrassing point:
paradigm, for example, Einstein's replacement of all the elements necessary for the solution were
the older Newtonian paradigm. Kuhn contrasts this already in my long-term memory. I had simply not
with normal science. Normal scientists engage in simultaneously thought about them before. That is,
"puzzle solving" and do not think in a creative I knew that poets cannot keep on reusing the same
manner. Suppose that an animal runs a maze faster simile, that creativity and regression are related
if it has been deprived of food for 24 hours than if it (Kris, 1952), and that creativity and associative
has been deprived for 12 hours. The normal scien- gradients are related (Mednick 1962). I expect that
tist may hypothesize that running speed will be in- all cases of inspiration are similar to this. The dif-
termediate after 18 hours of food deprivation. ference may be that some third-and forgotten or
Rather clearly, such a hypothesis most likely arose unnoticed-stimulus served to get the elements
from purely intellectual thought rather than from combined into a creative idea into consciousness at
any blinding flash of inspiration. the same time. In Poincare's case, for example, let
The examples of inspiration quoted above do not us suppose that combining Elements 1 and 3 led to
make explicit exactly what stimulus elicited the cre- the insight. Assume that Element 1 was in con-
sciousness. For whatever reason, Element 2 be- nition is important for creativity. 1 Lombroso (1901)
came conscious. It may have been completely unre- commented on the "unconsciousness and in-
lated to Element 1 but related enough to Element 3 stinctiveness" of the creative genius: production of
to bring the latter at least to the fringe of awareness. creative works seems not to be under conscious
At that point, the similarity of Elements 1 and 3 control. He also commented on their "somnam-
immediately became apparent. bulism": composition often occurs in an almost
dreamlike state. Nordau (1895) noted a tendency to
fall into what he called "inane reverie" and as-
Theories of the Creative Process cribed it to an inability to order, correct, or control
the flow of associations or to suppress "irrelevant"
associations. He held that the creative genius is
Primary Process Cognition
subject to the' 'tyranny of the association of ideas. "
Kris (1952) hypothesized that creative indi- It should be noted that a number of cognitive
viduals are more able to alternate between primary psychologists (e.g., Newell et al. 1962; Perkins,
process and secondary process modes of thought 1981; Weisberg, 1986) argue that creativity in-
than are uncreative people. The primary process- volves only secondary process cognition. Their ar-
secondary process continuum is the main dimen- gument is that creative ideas are conceived of in a
sion along which consciousness varies (Fromm, manner that does not differ from everyday problem
1978). Primary process cognition occurs in normal solving. This disagreement with Kris seems to be
states, such as dreaming and reverie, and in abnor- due in part to a misunderstanding of what is meant
mal states, such as psychosis and hypnosis. It is by primary process cognition. Weisberg (1986), for
autistic, free-associative, and analogical and tends example, argues that creative ideas almost never
to operate on concrete images rather than abstract occur in dreams, and that creative ideas do not
concepts. Secondary process thought is the ab- emerge fully formed from "the unconscious." But
stract, logical, reality-oriented thought of everyday theorists such as Kris never said that creative prod-
waking consciousness. According to Kris, creative ucts come from dreams or the unconscious. Dream-
inspiration involves a "regression" to a primary ing is an extreme primary process state quite un-
process state of consciousness. Because primary likely to produce creative ideas. Secondary process
process cognition is associative, it makes the dis- and primary process cognition form the poles of a
covery of new combinations of mental elements continuum rather than two discrete types of
more likely. Creative elaboration or verification in- thought. It is not necessary to dream in order to
volves a return to a secondary process state. Be- think in a primary process manner. Furthermore, in
cause uncreative people are "stuck" at one point on its modem usage, primary process thought is not
the primary process-secondary process continuum, confined to the unconscious; rather, it is a term used
they are unable to come up with creative ideas. to describe the quite conscious thought found, for
Creation of a very simple creative product may in- example, in fantasy and reverie. Thus, it is proba-
volve on one inspiration-elaboration cycle. More bly true enough that creative thought does not differ
complex productions are likely to require a number qualitatively from the thought of normal problem
of such cycles. solving. However, the latter function probably in-
A good deal of evidence is supportive of Kris' s cludes a good bit of what we have called primary
(1952) theory that creative people have easier ac- process cognition.
cess to primary process modes of thought (see Sul-
er, 1980, for a review). They report more fantasy
Defocused Attention
activity (Lynn & Rhue, 1986; Singer & McCraven,
1961), remember their nighttime dreams better Mendelsohn (1976) hypothesized that individual
(Hudson, 1975), and are more easily hypnotized differences in how attention is focused are the cause
than uncreative people (Aston & McDonald, 1985;
Bowers & van der Meulen, 1970; Lynn & Rhue, I Throughout this chapter, I have occasion to refer to the writings
1986). Wild (1965) showed directly that they are of Lombroso (1901) and Nordau (1895). Although their expla-
better able to shift between use of primary process nations of creativity were incorrect, and their phraseology was
and secondary process cognition. Although they often intemperate, their observations were particularly acute.
They were perfectly well aware of the majority of creative
used other terms, case studies of eminently creative personality traits that we have "discovered" across the course
individuals led nineteenth-century theorists to con- of the last several decades (see Martindale, 1971) as well as of a
clude that what we would call primary process cog- number of traits that we have not yet rediscovered.
of differences in creativity: "The greater the atten- whereas uncreative people have relatively steep as-
tional capacity, the more likely the combinatorial sociative hierarchies. This hypothetically accounts
leap which is generally described as the hallmark of for the ability of the creative person to make the
creativity" (p. 363). To become aware of a creative remote associations that are the basis of creative
idea, one must obviously have the elements to be ideas. According to Mednick's theory, the ordering
combined in the focus of attention simultaneously. of elements on associative hierarchies is propor-
If one could attend to only two things at the same tional to the relative strength of the responses. Re-
time, only one possible analogy could be dis- search with continuous word association supports
covered at a time. If one could attend to four things this contention (Mednick, 1962). At first, creative
at once, six possible analogies could be discovered. and uncreative people tend to give similar responses
Mendelsohn's theory is most relevant to creative in a similar order. However, creative people con-
inspiration. Creative people may prefer or be prone tinue to respond at a steady rate, whereas uncreative
to states of defocused attention, but they must be people run out of responses.
capable of focused attention, because the elabora-
tion or verification stage of the creative process A Cognitive Integration
calls for this type of attention. In fact, there is a
good deal of evidence that uncreative individuals The theories of Kris (1952), Mednick (1962),
have more narrowly focused attention than do cre- and Mendelsohn (1976) are really identical theories
ative ones (Dewing & Battye, 1971; Dykes & expressed in very different vocabularies. De-
McGhie, 1976; Mendelsohn & Griswold, 1966). focused attention is one of the properties of primary
Based upon case studies of highly creative writers process cognition (see Martindale, 1981). De-
and artists, Nordau (1895) concluded that one of focused attention and flat associative hierarchies
their most common traits was a lack of ability to are cognitive and behavioristic ways of describing
focus attention. In his typical unsympathetic tone, exactly the same phenomenon (Mendelsohn &
Nordau remarked about the creative artist that Griswold, 1966). The three theories can be restated
in cognitive terms (Martindale, 1981).
his want of attention makes him incapable of apprehending the The consensus is that mind may be represented as
real connecting links between the simplest and most obviously a vast set of nodes and relationships among these
related phenomena, and leads him to deduce them from one or
another of the hazy, intangible presentations wavering and wan- nodes. The nodes may be activated to varying de-
dering in his consciousness. (p. 69) grees. The connections between nodes are either
excitatory or inhibitory. An activated node influ-
ences the activation of nodes with which it is con-
Flat Associative Hierarchies
nected via spreading activation or inhibition. Nodes
Ideas are associated with one another to varying may be identified with neurons or groups of neurons
degrees. For example, on a word association task, if in the neocortex and relationships with the axonal
the stimulus word is "table," the most likely re- and dendritic connections among these neurons.
sponse is "chair." "Food" is a somewhat less The nodes may be seen as being partitioned into
probable response, but "victory" is far less proba- various "analyzers"; for example, there seem to be
ble. Because people are fairly consistent in the separate analyzers for perception of printed words,
probabilities of the responses they will make to a faces, spoken words, and so on. The nodes in these
stimulus, an associative hierarchy may be plotted analyzers are activated by the presence of the rele-
for that stimulus. People differ in the steepness of vant stimulus. Semantic memory hypothetically
their associative hierarchies. Those with steep hier- contains nodes that code concepts and relationships
archies have only a few responses to make to a among these concepts. The relationships code at-
stimulus: The mental representation of the stimulus tributes of concepts and also connect related con-
is strongly bonded to only a few other mental repre- cepts to nodes coding superordinate categories.
sentations. On the other hand, people with flat asso- Nodes in all of the analyzers also hypothetically
ciative hierarchies have more associations to a stim- receive nonspecific input from the reticular activat-
ulus. In this case, closer associates are less strongly ing system (Martindale, 1981).
connected to the stimulus, and more remote associ- Consciousness corresponds to the set of currently
ates are more strongly connected to the stimulus activated nodes. It may be divided into attention
than is the case for people with steep associative (the several most activated nodes) and short-term
hierarchies. Mednick (1962) theorized that creative memory or the fringe of awareness (the remaining,
people have relatively flat associative hierarchies, less-activated nodes). Having a flat associative hi-
erarchy or defocused attention corresponds to hav- that inspiration will occur, because more paths
ing a relatively large number of nodes activated- among nodes can be simultaneously activated.
to about the same degree-at the same time. Hav- Campbell (1960) ascribed creativity to a quasi-
ing a steep associative hierarchy or focused atten- random combination of mental elements. This
tion corresponds to having fewer nodes activated to makes a good deal of sense if we think of inspiration
differing degrees. In a secondary process state, as an idea in the focus of awareness "pulling" a
which corresponds with focused attention and steep primed set of ideas into consciousness. Our experi-
associative gradients, a few nodes are highly acti- ences are not random, but they may be more or less
vated and these nodes inhibit other nodes. In a pri- random with respect to whatever problem we are
mary process state, entailing defocused attention trying to solve. One might think that creative people
and flat associative gradients, more nodes are acti- simply have a lot of original ideas and discard the
vated but not activated enough to exert much inhibi- ones that are not useful. This, however, is definitely
tion. As a rough generality, we may say that inhibi- not the case. Poincare (1913) pointed out that "ster-
tion is predominate in secondary process states and ile combinations" very seldom occur to the creative
disinhibition is predominate in primary process genius. This is easy to explain. If the primed ele-
states. The total amount of activation is lower in ments are not very similar to the elements in the
primary process states of consciousness, but it is focus of attention, they will not become sufficiently
spread over a larger number of cognitive nodes. activated to enter consciousness. Because of this,
Further, these nodes are not capable of inhibiting sterile combinations are automatically filtered out.
each other very much.
Creative inspiration consists of noticing an anal-
Creativity and Cortical Arousal
ogy between at least two things. Another way of
putting this is that creative inspiration consists of In order for a person to be creative, as many
discovering that these two things share a common nodes as possible must be simultaneously activated.
superordinate category. The more nodes and rela- How can this state be reached? The best way it can
tionships that are activated at a given time, the be reached is by being in a state of low overall
greater the likelihood of this is (Findlay & cortical arousal. In such a state, more nodes will be
Lumsden, 1988; Martindale, in press). Of course, activated and to a more equal degree than in a state
this is essentially just a cognitive restatement of the of higher arousal (Martindale, 1981). The reason
theories described above. Note, however, that the for this is that input from the reticular activating
implication is that the capacity of short-term memo- system affects activation of nodes in a multi-
ry and attention-rather than just of attention-is plicative rather than an additive fashion: All nodes
held to be greater in creative people than in un- receive the same amount of nonspecific input, but
creative people. The cognitive restatement sheds this input increases the activation of the more acti-
light on the sequence of preparation, incubation, vated nodes much more than the activation of the
and inspiration. During preparation, attention is less activated nodes. The more activated nodes then
probably too focused. One attends to ideas pre- inhibit the less activated ones. There is also good
sumed to be relevant to the problem at hand. The evidence that primary process thought, defocused
difficulty is that the solution lies in ideas thought to attention, and flat associative hierarchies are asso-
be irrelevant. During incubation, the nodes coding ciated with states of low cortical activation (see
the problem remain primed or partially activated on Martindale, 1981, for reviews). It must be the case,
the fringe of awareness. As the creator goes about then, that creativity is related to the level of cortical
his business, many other nodes are activated. If one activation.
of these nodes happens to be related to the nodes Virtually anything that increases arousal causes
coding the problem, the latter will become activated decrements on tests of creativity. This has been
and leap into attention. This corresponds to inspira- shown to be the case for stress (Dentler & Mackler,
tion, the discovery of the creative analogy. This 1964), mere presence of other people (Lindgren &
would not have occurred if the nodes coding the Lindgren, 1965), noise (Martindale & Greenough,
problem were not already partially activated, be- 1973), extremes of temperature (Lombroso, 1901),
cause the path connecting the two sets of nodes is and even reward (Amabile, 1983a). It is not the
presumably long and circuitous. Furthermore, the case, however, that creative people are generally in
more nodes that can be simultaneously activated a state of low physiological arousal. They may be
(the more creative the person), the more likely it is more anxious than uncreative people (Maddi & An-
drews, 1966) and tend to show slightly higher levels Cognitive Traits
of basal arousal on physiological measures (Martin-
dale, in press). However, as compared with less It is not enough to be able to combine mental
creative people, those who are more creative do elements in a primary process fashion; one must
show low levels of cortical arousal while per- also have the right elements to combine. It takes no
forming creative tasks (Martindale & Hines, 1975). research to tell us that most people could not possi-
Martindale and Hasenfus (1978) found that the low bly be creative. No possible combination of the
levels of arousal occur exactly where they should: ideas with which the ordinary person concerns him-
during creative inspiration rather than during the self (e.g., what little Johnnie did in school today;
elaboration stage. It is unlikely that these dif- where to plant the tulips this fall; the price trajecto-
ferences are due to self-control, because creative ries of gasoline and lettuce; which team will win the
subjects perform worse than uncreative subjects on Superbowl) could yield a creative insight. Creative
biofeedback tasks (Martindale & Armstrong, 1974; geniuses often exhibit an amazing lack of interest in
Martindale & Hines, 1975). It may be the case that or concern with the mundane details that constitute
creative people are more variable in their level of virtually the entire fabric of the ordinary person's
arousal than are uncreative people; that is, they may mental life. Indeed, creative people, in general,
show more extreme fluctuations. This is merely a tend to be uninterested in details or facts for their
physiological restatement of Kris' s (1952) conten- own sake (MacKinnon, 1975). To take an example
tion that creative people are more variable on the of another sort, no matter how much a person
primary process-secondary process continuum. knows about linguistics, he or she will not be able to
There are no direct studies of this hypothesis. How- explain the historical sound changes that seem to
ever, there is evidence that at least some highly occur constantly in all languages unless that person
creative people tend to fluctuate between states of has mental elements drawn from other domains.
excessive energy and excessive apathy, abulia, or The reason is that modem linguistics consists al-
depression (Lombroso, 1901). most wholly of statements of rules, and none of
these rules pertain to or could cause changes in the
phonetic realization of the elements of speech. To
solve the problem, our hypothetical linguist would
The Creative Personality have to know something about psychology, so-
ciology, or physiology, because these neighboring
Necessary Traits versus Linked Traits disciplines do have constructs that could explain
phonetic change. In short, the more diverse and
It was argued above that certain cognitive and general a person's store of ideas, the greater is the
motivational traits are necessary if one is to be cre- chance that a creative idea will emerge.
ative. For example, creative people must be able to As a rule, creative people do have a very wide
combine mental elements in a different way than range of interests. Among the eminently creative,
uncreative people. We have just seen that they do diversity of interests is rather astounding. Prentky
have this ability. Can the necessary traits exist in an (1980) gives the example of the geneticist The-
otherwise "normal" personality or are they linked odosius Dobzhansky, whose interests included mu-
to other traits or personality configurations that are sic, history, anthropology, philosophy, religion,
not, in a logical sense, necessary for creativity. natural science, and linguistics. Other examples are
Intuition tells us that they must be linked to other not difficult to find. Goethe was probably as in-
traits. After all, the stereotype is that creative ge- terested in science as in literature. Edgar Allan Poe
niuses are eccentric and bizarre creatures. It is im- believed that his lasting fame would rest not upon
possible to picture the mad, sad, and thoroughly his literary works but upon his contributions to cos-
bad Lord Byron trying to earn a living as an insur- mological theory. A large number of empirical
ance salesman. But one may immediately recall that studies have shown that creative people con-
Wallace Stevens, whose immense creativity cannot sistently describe themselves as having a wide
be questioned, did in fact earn his keep as an insur- range of interests (see Barron & Harrington, 1981).
ance company executive and, so far as is known, It is very common for creative ideas to arise from
led a thoroughly normal life. Thus, the question combining ideas from different disciplines. Hard-
comes down to one of probabilities. Are the Stevens ing (1965) has compiled an extensive list of exam-
or the Byrons more likely? ples. Based upon her studies, she formulated her
"law" of creative thought: an idea will eventually artists, Getzels and Csikszentmihalyi (1976) found
be applied to everything that it is applicable to. that creativity was not related to speed or efficiency
As well as having diverse interests, creative peo- of problem solving but that it was related to the
ple also seem to categorize ideas in a different way ability to discover problems. If creative people have
than do less creative people. It would appear that very wide categories, they are likely to see excep-
their categorizations are both broader and more id- tions to previously accepted ideas. These excep-
iosyncratic than those of the uncreative. On object- tions provide new problems; for example, a scien-
sorting tasks, the categorizations of creative people tific theory may apply given only certain boundary
are as "bizarre" as those of schizophrenics (Dykes conditions. For a narrow categorizer, these bound-
& McGhie, 1976). Pettigrew (1958) developed a aries may be so constrained that anomalies literally
category-width test. The test consists of items such cannot occur: the theory applies only to the things to
as the following: "The average religious de- which it applies. For a wide categorizer, or for
nomination in the United States has 551 churches." someone who thinks on a more abstract level,
The subject is then asked to guess what the number boundary conditions may be looser. Such a person
of churches is for the largest and smallest de- can discover that a theory does not work in a broad-
nomination. A good number of studies have shown er context (problem finding) or, perhaps, that it
that creative people exhibit broader categories on does work (simultaneous problem finding and prob-
this test (Pettigrew, 1982). It is reasonable to expect lem solving).
that broader categories would allow one to see more More thought and research are needed on the
items as similar. Perception of similarity where related issue of problem selection. Creative people
none had been seen before is, of course, the basis of seem to have an intuitive knack for selecting impor-
creative insight. It should be noted that broad, dif- tant problems. A chemist once described his latest
fuse, or hazy categories are an attribute of primary idea to James Bryant Conant with great enthusiasm.
process states of consciousness (compare Werner, The idea had all the marks of a creative break-
1948). through, save one. Conant's response was that
The concept of category width is somewhat am- whoever solved the problem would certainly merit a
biguous. The concept of "animal" is not only footnote to a footnote in any book on the history of
broader but also more abstract than the concept of chemistry. Creative people not only need to find
"bird. " Research on conceptualization in more and problems, they also need to find important ones.
less creative people has not addressed the issue of Wide categorization, ability for abstraction, and
whether capacity for abstraction or category width wide interests may help in this respect, but they do
per se is the crucial variable. At least in the sci- not provide a complete explanation. It would be
ences, one is led to suspect that ability for and in- interesting to disentangle motivational and cog-
terest in abstractions must be absolutely crucial for nitive determinants of problem selection. Are un-
creativity. The reason for thinking this is that the creative people really unable to discriminate impor-
importance-and, hence, the creativity-of a sci- tant from trivial problems, or do they avoid tackling
entific idea is closely related to how general or ab- important problems because of low self-confidence
stract it is: No matter how ingenious it was, a new or other motivational factors?
theory concerning the causes of creativity in, say,
Cross-Sexual Interests
Guatemalans of British descent would not be as-
sessed as very creative unless it could be readily Studies have uniformly found that creative peo-
generalized to some wider population. It is surpriz- ple have more cross-sexual or androgynous in-
ing that there is no coherent body of research on terests and traits than less creative people; that is,
abstract ability and creativity. Casual observation they have interests generally held by members of
certainly suggests that uncreative people are often the opposite sex. They hold these interests in addi-
uninterested in and/or incapable of abstraction. tion to, rather than in place of, interests ster-
Simply put, they cannot see the forest for the trees. eotypically associated with members of their own
It could well be argued that problem finding is sex. Thus, a creative woman might be interested in
just as creative as problem solving (Mackworth, both carpentry and cooking. This pattern of wide-
1965). After all, there are no explicit problems to be ranging interests has been found for both creative
solved in the arts. In science, as often as not, the men (Barron, 1957; Kanner, 1976; MacKinnon,
creator solves not an already existing problem but 1962) and creative women (Crutchfield, 1962;
one that has been discovered. In a study of student Helson, 1971). Because of this wider range of in-
terests, creative people may respond in an unusual ones were foreign born. Landry (1972) produced
manner to problems as compared with less creative evidence that just learning a foreign language en-
members of their sex. hances creativity. In that study, grammar school
Whether this pattern is just an artifact of the wide children signed up to learn a foreign language. One
range of interests in creative people or reflects group began instruction immediately, whereas an-
something deeper is not clear. There is some evi- other group (comparable in terms of intelligence
dence that creative people tend to have less clear- and other possible confounding variables) had to
cut sex-role identities than uncreative people. Bes- wait. After a period of time, the children taught a
dine (1968) examined the biographies of a number second language scored higher on several paper-
of highly creative men. He found evidence for what and-pencil tests of creative ability. Similar results
he called "Oedipal victory" in a large number of were reported by Lambert, Tucker, and d' Anglejan
these people. By Oedipal victory he means a situa- (1973). Carringer (1974) found that "balanced"
tion where the son has replaced the father in the bilingual children (i.e., those who had learned two
affections of the mother. Freud held that the normal languages about equally well) were more creative
pattern of development involves Oedipal rivalry; than monolingual children. Learning a second lan-
for example, around the age of five, the son wants guage also involves learning associative hierarchies
to replace the father in the eyes of his mother. Most that are "deviant" as compared to those learned for
children lose this struggle and this leads, as Freud the first language; that is, the connotations of words
argued, to the establishment of male sex-role identi- in different languages are not exactly the same. This
ty based upon identification with the father. Identi- factor apparently enhances creativity. But it hardly
fication with the father is also hypothetically the seems likely that learning a second language would
basis for the development of the superego. In the induce very large increases in creativity. However,
case of Oedipal victory, this would not occur, and these results are of interest because they imply that
we should expect the child to maintain his initial even factors that only slightly increase a person's
infantile identification with his mother and not to store of mental elements and associations are
develop a normal superego. The disinhibition or positively related to creativity.
lack of control often found in creative people (see
below) would be consistent with weak superego
Creativity, Age, and Expertise
functioning. Another way that a male child can fail
to establish a firm masculine sex-role identity is if Creativity exhibits a curvilinear relationship to
the father is absent from the home during this cru- age (Lehman, 1953; Simonton, 1984). In general, a
cial period. This would lead to a sort of de Jacto person's most creative work is done at a fairly early
Oedipal victory. In a sample of eminent French and age, and this age of peak productivity varies from
English poets, I found a father-absence rate of field to field. It is fairly early in lyric poetry, mathe-
30%-far higher than would be expected for the matics, physics, and chemistry (ages 25-35) and
general population (Martindale, 1972b). Albert somewhat later in psychology and the social sci-
(1971) and Goertzel, Goertzel, and Goertzel (1978) ences (ages 30-40). Only a few specialties, such as
have found similar results for other domains of architecture and novel writing show peak perfor-
creativity. In spite of popular stereotypes, homo- mance at later ages (40-45). Not only the best work
sexual subjects are not more creative than hetero- but the most work is produced at these ages. Output
sexual subjects. In fact, several studies have found rises rather abruptly to this peak and then declines
trends in the opposite direction (e. g., Domino, throughout the rest of a person's career.
1977). Thus, cross-sexual interests and, perhaps, Although age per se may be a crucial factor in
cross-sexual identification may be related to creativity, "age-within-specialty" might also be
creativity but homosexuality is not. It seems to be important: One's most creative work may be done
cognitive orientation rather than sexual orientation relatively early in one's career within a given spe-
that is crucial. However, more thorough research is cialty, regardless of actual age. A person cannot
needed before any firm conclusions can be drawn. have any creative ideas until at least some of the
elements relevant to a field have been learned. But
it would seem that one's best ideas come early-
Second-Language Learning and Creativity
before one becomes an "expert." Once you have
In a study of creativity in women mathemati- become an expert, you may be less likely to be
cians, Helson (1971) found that more of the creative creative. Why? Because being an expert means
knowing which elements are important and (the po- or extreme form. Creative people are indeed self-
tentially disastrous part) which elements are irrele- confident. A large number of studies have shown
vant and inappropriate. In other words, the expert that, when asked to describe themselves, creative
may be less likely to make "inappropriate" re- people pick adjectives, such as confident, ego-
sponses and is consequently less likely to make tistical, and self-confident (Gough, 1979;.. Har-
creative ones. If this line of reasoning is correct, rington, 1975). Arthur Rimbaud's remark, "I am
then the way to maintain one's creativity with in- he who would be God," suggests that self-confi-
creasing age is to shift fields. In fact, many con- dence is often carried a bit far. Recall also Jean
sistently creative people do just this. The chemist, Cocteau's remark that "Victor Hugo was a mad-
Linus Pauling, provides us with a good example of man who thought he was Victor Hugo," or Paul
this strategy. Another example would be the psy- Valt~ry' s comment that" A man who has never tried
chologist, Leon Festinger. After formulating and to make himself like the gods is less than a man. "
testing several important theories in social psychol- Lombroso (1901) commented on the "morbid van-
ogy, Festinger gave up his work in this area com- ity," excessive preoccupation with self, and
pletely and began research on visual perception. "megalomania" of the creative genius. Nordau
There is no empirical research on the question of (1895) went even further and used the term "ego-
whether age per se or age-within-specialty is the mania. " Perhaps because of their self-confidence,
main determinant of creativity. Now that mid-life creative people are often not overbearing or obnox-
career changes are more common, such a research ious. Lombroso (1901, p. 46) quotes Alexandre
project is possible. If it turned out that age-within- Dumas as observing that Victor Hugo" did not say,
specialty is the crucial determinant, this would, of 'I am Genius,' but he began to believe firmly that
course, have profound implications for both indi- the world would say so."
vidual and social planning. Creative people also describe themselves with
Even the expertise involved in formal education adjectives, such as ambitious, curious, and enthusi-
may, beyond a certain point, be detrimental to astic (Gough, 1979; Harrington, 1975). At high
creativity. In a study of 301 geniuses in various levels of creativity, interest sometimes takes on an
fields, Simonton (1976) found an inverted-U almost obsessive quality. Prentky (1980) provides
relationship between creativity and amount of edu- interesting examples of Sir Isaac Newton's absent-
cation. Those with a moderate amount of education mindedness. For example, Newton occasionally
were most creative. This was the case even in the left his lodgings having forgotten to get dressed.
sciences (Simonton, 1984). Simonton's sample of We may presume that his intense interest in some
scientists was small, however. Now that a Ph.D. is intellectual problem left no time for such trivial
more or less required for work in most areas of details as putting on his clothes.
science, it seems unlikely that an inverted-U Although creative people often describe them-
relationship would be found. Success in school or selves as energetic (Barron & Harrington, 1981),
college does not seem to be related to creative this contrasts with findings that they tend to be
achievement (MacKinnon, 1960). However, more physically inactive (Maddi, 1965) and with a
studies are required before this assertion can be number of well-documented self-reports. Indeed,
accepted with much confidence. Simonton (1984) Lombroso (1901) and Nordau (1895) held, on the
gave a number of examples of eminently creative basis of such reports, that abulia, apathy, and lack
people whose scholastic work was mediocre or of will are among the most common traits shared by
poor. He pointed out that time devoted to achieving creative artists. Lombroso (1901) argued that the
scholastic excellence may be time made unavail- creative genius alternates between erethism and
able for acquiring the "irrelevant" knowledge nec- atony-that is, between excitement and inspiration
essary for creativity. He also cited evidence that on the one hand and exhaustion and apathy on the
extremely creative people often dislike formal edu- other. The journals of Charles Baudelaire and Gus-
cation and prefer self-education. tave Flaubert are filled with reports of extreme lack
of energy to the point that getting out of bed or
beginning to write were seen as almost herculean
Motivational Traits tasks. Flaubert's behavior is especially enlighten-
Creative people appear to have the motivational ing. When he was in bed, he did not have the energy
factors that would seem necessary. It is interesting to get up. Once he began writing, he persisted hour
that these factors are often present in an exacerbated after hour-often on the same sentence. Other ex-
amples could readily be adduced. We might specu- nally seen the very blurred pictures, these subjects'
late that creative people have a tendency to persist guesses were quite inaccurate. It would seem that
in whatever state they are in. If they are doing noth- their early incorrect guesses led them to fit the
ing, they find it difficult to begin a task; but if they "data" to their "hypotheses." Order imposed too
are performing a task, they find it difficult to stop early thus led to deformed perceptions. It would be
doing it; they either persevere or perseverate. One interesting to know if creative people are less prone
wonders whether creative people are capable of to such distortion effects on this task, but I am not
sustaining attention on a topic for a long period of aware of any research on this topic. Greenwald,
time or incapable of shifting it away from the topic. Pratkanis, Leippe, and Baumgardner (1986) dem-
There is no systematic research on these questions, onstrated that a quite analogous process can retard
but it would be well worth undertaking. In- progress in science. Amabile (1985) pointed out
terestingly, creative subjects show very slow rates that keeping options open and suspending judgment
of habituation. Rosen, Moore, and Martindale would seem to be useful if not necessary skills for
(1983) measured habituation of skin potential re- creativity. Thus, though a need for order or sim-
sponses to a 60 dB tone. More creative people took plicity must be present in the creative person, it
twice as long to habituate than did less creative must be counterbalanced by forces that prevent pre-
people. This suggests that less creative people may mature imposition of order.
tire quickly of a problem that is not easily solved, In a series of studies, Amabile (1983a,b, 1985)
whereas the problem may continue to hold the in- showed that intrinsic motivation is important for
terest or attention of more creative people. creativity. Intrinsic rewards are those that arise
There is also some confusion concerning the rela- from performing a task; that is, the task is interest-
tionship between creativity and preference for sim- ing and pleasurable in itself. Extrinsic rewards are
plicity versus complexity. If asked to choose be- those that come from beyond the domain of the task;
tween pairs of visual designs, creative people are for example, being paid or praised for working on
generally found to prefer the more complex, asym- the task. It would seem that creativity is enhanced
metric, or ambiguous design (e.g., Barron, 1953; by intrinsic rewards and decreased by extrinsic re-
MacKinnon, 1962). On the other hand, creative wards. In one of her studies, Amabile (1985) had
people have a high need for order (Barron, 1958) poets compose poems under two conditions. In one
and prefer consonant over dissonant tone pairs condition, poets were first led to think about extrin-
(Nardi & Martindale, 1981). On the Allport-Ver- sic rewards that could be important in writing. In
non-Lindzey Study of Values Scale (1951), cre- the other condition, they were led to ponder the
ative people, whether artists or scientists, score intrinsic rewards of writing. Rated creativity of
high on the aesthetic value scale (MacKinnon, poems written in the extrinsic-reward condition was
1975). These seemingly contradictory findings significantly lower. Amabile (1979) had other sub-
make some sense if we consider what it means to jects construct collages. One group was told that
have a well-developed aesthetic sense. Almost all their products would be evaluated for creativity,
aestheticians agree that beauty arises when both whereas the control group was led to believe that the
unity and variety are maximal (see. Martindale, collage was not even the experimenter's primary
1984b). That is, a beautiful object is one that unifies interest. The control group produced significantly
a set of diverse elements. It would seem that the more creative collages. It would seem that extrinsic
goal of the creative person is to produce such an reward takes the joy (intrinsic reward) out of a task.
object. For example, a creative scientific theory is These findings fit with some of what we know about
beautiful, because it unifies elements that, on the creativity in the real world. Creative scientists tend
surface, are diverse and unrelated. As the poet Sam- to prefer "pure" science and are often reluctant to
uel Taylor Coleridge (181711907) long ago noted, undertake, or are even disdainful of, practical pro-
creativity requires the ability to "exist in ambigu- jects that might yield a good deal of financial re-
ity" or to tolerate disorder. However, the ultimate muneration. However, Amabile's findings should
goal is to arrive at an overall synthesis or order. not be generalized too far. A poet wants to create a
Wyatt and Campbell (1951) and Bruner and Pot- beautiful poem (an intrinsic reward), but this act of
ter (1964) showed subjects blurred photographic creation very often occurs in the context of seeking
images. The images were gradually focused, and fame or immortality (extrinsic rewards). A scientist
subjects were then asked to guess what they de- aims to solve important problems not only for their
picted. As compared with people who had not origi- own sake but also in order to obtain some extrinsic
reward-for example, a Nobel prize. Indeed, if in- temperamental or motivational rather than a cog-
trinsic rewards alone were important to creative nitive cause.
people, it would be difficult to explain why scien- One problem with the construct of openness to
tists would bother to publish their findings and why experience is that the adjectives that are used to
artists would exhibit and sell their paintings. measure it in self- and peer-ratings are not, on the
face of it, directly related to openness per se. For
example, of the adjectives loading on what McCrea
Openness to Experience and Creativity and Costa (1987) labeled the openness to experi-
I mentioned above that creative people must like ence factor, four refer directly to creativity (origi-
novelty. There is some empirical evidence that they nal, imaginative, creative, and artistic), six refer to
do (e.g., Houston & Mednick, 1963). Again, this traits often used by creative people to describe
trait sometimes tends to be carried a bit far in those themselves (complex, independent, daring, analyt-
people of very high creativity. Consider Baude- ical, liberal, and untraditional), and only three are
laire's remark that "the beautiful is always directly or indirectly related to openness (broad in-
bizarre" or George Moore's (1885/1959) com- terests, curious, and prefer variety). It is not clear to
ment, "I am ... morbid, perverse. But above all me why the factor is labeled "openness" rather
perverse, almost everything perverse fascinates than just "creativity." In other words, openness to
me ... the commonplace, the natural is constitu- experience and creativity would seem to be syn-
tionally abhorrent" (pp. 59, 61). Salvador Dali onyms that are used to describe the same set of
complained that the world is not a very interesting traits. If so, openness cannot be said to explain
place because, among other things, he had never anything about creativity. Costa and McCrea
once been served a baked telephone when he or- (1985) have developed a questionnaire measure of
dered lobster in a restaurant. In describing scientific openness that is less subject to this criticism. The
geniuses, Koestler (1959) noted that they have measure does correlate with paper-and-pencil mea-
sures of creativity (McCrea, 1987).
on the one hand skepticism, often carried to the point of icon-
oclasm, in their attitude toward traditional ideas, axioms, and
dogmas but an open-mindedness that verges on naive credulity Disinhibition and Creativity
towards new concepts. (p. 518)
Psychoanalytic theorists (e.g., Kris, 1952;
A number of factor-analytic studies of person- Kubie, 1958) have used the term "regression in the
ality traits have yielded five-factor solutions (e.g., service of the ego" to describe the creative person's
Norman, 1963). The first four factors-neurot- access to primary process modes of thought. Such
icism, extraversion, agreeableness, and conscien- thought is generally pathological, as in schizo-
tiousness-are fairly unambiguous. The fifth fac- phrenia, or driven by primitive drives and emo-
tor, which has been labelled variously as culture tions, as in dreaming. Their idea was that creative
(Norman, 1963), intelligence (Peabody, 1987), and people can use primary process thinking to deal
openness to experience (McCrea & Costa (1985a), with neutral ideas. Is it really the case that creative
seems to be related to creativity. McCrea and Costa people have free access to the beneficial aspects of
(1985b) defined openness to experience as an in- such thought without having to pay the price of
terest in varied experience for its own sake in a suffering from its pathological consequences? It
variety of domains, such as fantasy, aesthetics, would seem not. Creative people consistently ob-
feelings, actions, ideas, and values. It is argued that tain high scores on the Psychotic ism scale of the
such openness is conductive to creativity, and Mc- Eysenck Personality Inventory (Eysenck, 1983;
Crea (1987) has shown that openness to experience, Woody & Claridge, 1977) and on related scales on
measured by both self-ratings and peer-ratings on the Minnesota Multiphasic Personality Inventory
both questionnaires and adjective checklists, is cor- (MMPI) (Barron, 1969). They either do not or can-
related with creativity. Hypothetically, those who not use the defense mechanism of repression (Bar-
are not open to experience prefer familiar stimuli ron, 1955; Fitzgerald, 1966; Myden, 1959). On ob-
and, therefore, are not motivated either to experi- ject-sorting tasks, highly creative people cannot be
ence or to produce novelty. Thus, to the extent that differentiated from schizophrenics, although the
openness is seen as a cause of creativity, it is a behavior of both groups is markedly different from
that of normal uncreative subjects (Dykes & deed, Anderson and Cropley (1966) have argued
McGhie, 1976). that creativity can be explained as due to a general
In the above discussion, I connected primary pro- inability or unwillingness to internalize the "stop
cess cognition with a relative lack of inhibition. It rules" that people generally learn as they are
would seem that the disinhibition of creative people socialized.
is not confined to cognition but is, rather, a general There is an emerging consensus that personal
trait. This is to be expected if creativity is related to identity or the self is not unitary (e.g., Gergen,
low levels of cortical arousal. One of the functions 1971; Kihlstrom & Cantor, 1983; Rosenberg &
of the cortex is to inhibit lower brain centers. If the Gara, 1985; Watkins, 1978). The idea is that every-
cortex is underaroused, it performs this function one is to one extent or another a multiple person-
poorly and lower centers are dis inhibited (Eysenck, ality. Lombroso (1901) went so far as to list "dou-
1967). The result is a general lack of both cognitive ble personality" as one of the traits of creative
and behavioral inhibition. Creative people are less genius. I have speculated that the "subselves" of
compliant in laboratory tasks that measure confor- creative people may be more dissociated than those
mity (Barron, 1961), and they are more willing to of uncreative people (Martindale, 1980), and that
engage in risk-taking behavior (Anderson & Cropley, this dissociation arises from the disinhibition found
1966; Pankove & Kogan, 1968). When they are in primary process states of consciousness (Martin-
asked to describe themselves on adjective check- dale, 1981). The reasons for this contention are
lists, probably the most striking thing is that many several. Creative people are not especially con-
of the traits chosen share the common factor of sistent in their traits. Their autobiographical reports
disinhibition. Examples of adjectives picked by highly often make clear that this inconsistency is not sim-
creative people are gloomy, loud, unstable, bitter, ply disorganization but, rather, a contrast among
dissatisfied, pessimistic, irritable (Barron, 1952); organized but dissimilar subselves. Also, there is
original, emotional, enthusiastic, impUlsive (Van the "dictated" quality of much creative inspira-
Zelst & Kerr, 1954); and inventive and industrious tion. William Blake is not alone in saying that his
(MacKinnon, 1962). In contrast, uncreative people poetry was dictated to him. Robert Louis Stevenson
describe themselves as gentle, conservative, patient, was quite serious when he said that his stories were
peaceable (Barron, 1952); contented, conventional dictated by "brownies" or "little people." Robert
(Van Zelst & Kerr, 1954); and virtuous, good, ra- Graves (1966) was firm in his contention that the
tional, and concerned with others (MacKinnon, muse is not at all a metaphor but a real psychic
1962). In this case, the common trait is inhibition. entity and the source of all true poetry. Such state-
Based upon studies such as the above, Har- ments make sense if we think of the poet as a scribe
rington (1975) developed a Composite Creative who copies down the "utterances" of an alien or
Personality Scale that includes the adjectives that dissociated subself that has access to "inner
creative people most consistently use to describe speech." If something like this occurs, there is no
themselves. Note that no matter how narrowly or reason to expect the scribe to understand what he
broadly one defines disinhibition, over half of the has transcribed and, in fact, he very often does not.
adjectives included in this scale have to do with lack Plato noted that poets are "like soothsayers and
of inhibition in one sense or another: active, alert, prophets, who say many fine things but who under-
ambitious, anxious, argumentative, artistic, assert- stand nothing of what they say." To take but one
ive, capable, clear thinking, clever, complicated, example, T. S. Eliot claimed not to know the mean-
confident, cynical, demanding, egotistical, ener- ing of his more obscure images and depended upon
getic, enthusiastic, hurried, idealistic, imaginative, literary critics for their explication. These, of
impulsive, independent, individualistic, ingenious, course, are extreme cases. It is not clear at present
insightful, intelligent, interests wide, inventive, how far, if at all, we should generalize from them.
original, practical, quick, rebellious, reflective, re- Systematic research on self-identity in more and
sourceful, self-confident, sensitive, sharp-witted, less creative people needs to be undertaken.
spontaneous, unconventional, versatile, not con-
ventional, and not inhibited. Gough (1979) devel-
Creativity and Psychopathology
oped a similar scale based on 12 samples of subjects
including 1,700 people. Again, adjectives with Some theorists have gone so far as to argue that
connotations of disinhibition are prominent. In- schizophrenia and creativity-both of which in-
volve primary process cognition-are based upon needed, it is of interest that alcohol abuse is often
identical cognitive processes that differ only in de- classified as a disinhibition syndrome (e.g.,
gree (Forrest, 1976; Hasenfus & Magaro, 1976; Gorenstein & Newman, 1980).
Keefe & Magaro, 1980; Prentky, 1980). This argu-
ment would seem to be based upon firmer ground
than might at first be thought. Several very well- Conclusions
controlled studies (Heston, 1966; Karlsson, 1968, In order to be creative, one needs access to a
1978; McNeil, 1971) have shown that the biolog- primary process state of consciousness involving
ical relatives of schizophrenics exhibit higher levels associative thought, defocused attention, flat asso-
of creativity than do matched subjects who are not ciative gradients, and cognitive disinhibition. It
relatives of schizophrenics. It would seem that the would seem that one cannot usually gain access to
same gene (or genes) may transmit predisposition such a state in an isolated compartment of mind;
for a common cognitive and personalogical config- rather, it is an all-or-none affair. If disinhibition is
uration (compare Eysenck, 1983; Jarvik & Chad- found in one place, it tends to be found everywhere.
wick, 1973). We might speculate that pre-, peri-, If your thoughts are uninhibited, then it is likely that
and postnatal environments determine whether this your behavior is also uninhibited. If you will not or
configuration is actualized as creativity, schiz- cannot conform to the conventional beliefs of sci-
otypal personality, or schizophrenia. ence, you are not likely to be any more able to
Several studies have found a higher than normal conform to the conventional rules of social behav-
incidence of bipolar or manic-depressive illness in ior. Thus, your disinhibition and unconventionality
creative people (Andreason & Cantor, 1974; Jam- are likely to be general rather than specific traits.
ison, Gerner, & Hammen, 1980). In one study of30 Finally, it is possible that these general traits have
creative writers, 80% had been treated for mood brought not only creativity but also all sorts of less
disorders as compared with 30% of subjects in a desirable traits, ranging from hangovers to bouts of
matched control group (Holden, 1987). Both peo- depression. It should be made explicit that the hy-
ple exhibiting bipolar symptoms and their asymp- pothetical linkage between creativity and general
tomatic relatives have been found to be more cre- disinhibition is not perfect. There are certainly cre-
ative than control subjects (Holden, 1987). In this atiye people who are not at all uninhibited. Further-
study, creativity was highest in subjects with mild more, if creative people tend to be disinhibited, this
as opposed to severe bipolar mood swings. It is in no way implies that people who are dis inhibited
difficult to distinguish periods of intense creative are necessarily creative.
activity from hypomanic episodes (Jamison et at.,
1980). Indeed, Isen, Daubman, and Nowicki
(1987) have found that induction of positive affect Situational Variables
in normal subjects leads to dramatic increases in
creative problem-solving abilities. Heredity, Environment, and Creativity
Predisposition to bipolar mental disorder is ge- Attempts to train people to be more creative have
netic, but schizophrenia and bipolar illness are not not been very successful (see Stein, 1974, 1975 for
determined by the same gene or genes. It is not clear reviews). This is rather surprising, as it is clear
at present how findings concerning the relationship enough what one needs to do in order to be creative.
between creativity and these two disorders can be The problem is not that creativity is difficult to
tied together aside from the fact that both disorders teach but that creativity is impossible to learn. This
may be described as involving primary process cog- difficulty in training people to be more creative, as
nition and disinhibition. Lombroso (1901) pro- well as the physiological differences that have been
duced a large amount of anecdotal evidence that found between creative and uncreative people,
creative people overuse alcohol and other drugs. leads one to doubt that creativity could be deter-
Recent studies have found high rates of alcoholism mined wholly by environmental factors. The rela-
in creative writers (Holden, 1987) and positive cor- tionship between creativity and genetically trans-
relations between creativity and marijuana use mitted disorders, such as schizophrenia, also leads
(Victor, Grossman, & Eisenman, 1973) and drink- us to suspect that creativity might be genetically
ing to the point of drunkenness (Martindale, determined. I am aware of four studies that have
1972a). Although more extensive studies are investigated the heritability of creativity (Barron,
1969; Olive, 1972; Reznikoff, Domino, Bridges, & sion of role models for already potentially creative
Honeyman, 1973; and Vandenberg, 1968). The re- people. As mentioned above, attempts to teach peo-
sults of these studies are weak and inconsistent. ple to be creative have not been notably successful.
However, all these studies examined paper-and-
pencil tests rather than actual creative accomplish-
Oversensitivity and Withdrawal
ments. At present, the extent to which creativity is
environmentally determined versus genetically de- The stereotype that creative people are oversen-
termined is not clear. sitive at both the sensory and the emotional level
would seem to be no stereotype at all. The Belgian
poet Emile Verhaeren disconnected his doorbell be-
Creative Environments cause its ringing caused him physical pain. Marcel
Several theorists (e.g., Crutchfield, 1962; Proust withdrew into his cork-lined room because
Rogers, 1954) have argued that freedom from ex- normal levels of light and sound were, he said,
ternal pressure or control and a warm, supportive painful to him. Lombroso (1901) collected a
environment are necessary or at least quite helpful number of similar self-reports. In fact, creative peo-
for creativity. Harrington, Block, and Block (1987) ple do seem to be physiologically overreactive to
have produced empirical evidence that child-rear- stimuli. For example, Martindale and Armstrong
ing practices producing such an environment are (1974) found more electroencephalogram (EEG)
correlated with creative potential in both preschool blocking to onset of a tone in more creative as com-
children and adolescents. Amabile's (1983b) con- pared with less creative subjects. Rosen et at.
tention that extrinsic reward decreases creativity is (1983) found much larger skin potential responses
related to this notion. She also mentions sur- to noise in more as compared with less creative
veillance and externally imposed deadlines as being people. Martindale (1977) asked subjects to rate the
detrimental to creativity. Naroll, Benjamin, Fohl, intensity of electric shocks: intensity ratings were
Fried, Hildreth, and Schaefer (1971) and Simonton positively correlated with creativity. Nardi and
(1975), in historical studies, have found that epochs Martindale (1981) elicited preference judgments
characterized by political fragmentation produce for tones varying in intensity. The more creative a
more creative individuals. It would seem that any subject was, the lower the intensity of his max-
sort of control emanating from the environment has imally preferred tone. It has also been found that
negative effects. Given that creative people tend to creativity is correlated with augmentation (" ampli-
be disinhibited, one would expect that environmen- fication" of the intensity of stimulation) on the kin-
tal control of any sort should be especially noxious esthetic aftereffect task (Martindale, 1977).
to them. If a creative person is stuck in such an The most common response to oversensitivity
environment, one would expect lessened creativity. would seem to be withdrawal in order to escape
One would also suppose that, if possible, creative overstimulation. The image of the creator in some
people would avoid such environments. As shall be analogue of Alfred de Vigny's ivory tower is ubiq-
seen below, there are other reasons to expect that uitous; for example, Flaubert immured at Croisset,
creative people may tend to minimize environmen- Algernon Swinburne at Putney, and Friedrich
tal inputs of all sorts-not merely those of a con- Holderlin in the castle tower at Tiibingen. Similar
straining or controlling sort. withdrawal is found among scientists; the clearest
Kroeber (1944) argued that the reason that inno- example certainly being Henry Cavendish (see Pre-
vations tend to come in waves has to do with emula- ntky, 1980). The extreme withdrawal practiced by
tion: One is more likely to be creative if there are such creators is a sort of self-imposed sensory dep-
creative role models to emulate or compete with. rivation. We know that ~ensory deprivation causes
Simonton (1984) provided strong quantitative evi- decreases in cortical arousal and increases in pri-
dence for this point of view. There is a very high mary process thinking (Schultz, 1965). Thus, it
correlation between the number of eminently cre- should facilitate creative inspiration. Although cre-
ative people in a given generation and the number in ative people do not seem generally to have low
the prior generation. Zuckerman (1977) found that levels of arousal, their oversensitivity may drive
over half of the people who win Nobel prizes had them to withdraw or to restrict sensory input. This,
previously studied under another Nobel laureate. It in tum, would put them in the low-arousal state
seems likely that these findings are due to the provi- necessary for creative inspiration.
Research Questions I have argued that inhibition on a number of lev-

els is weaker in creative than in uncreative people.
Throughout this chapter I have tried to point out More direct evidence for this now somewhat spec-
areas in which research should be undertaken and ulative contention is needed. It is agreed there are a
areas in which further research is needed in order to large number of phenomena that are based upon
clarify ambiguities or contradictions in the liter- inhibition. Forgetting, for example, is almost uni-
ature. In this section, three research topics that versally explained as being due to a process of inhi-
seem especially promising are outlined. We have bition. Do creative people forget less rapidly? If so,
seen that the major theories concerning the creative this would go a long way toward explaining how
process all involve some variant of the idea that they come to possess a large number of mental ele-
production of creative ideas involves an oscillation ments and associations. Pavlovian extinction and
between primary process and secondary process passive avoidance (suppression of a formerly re-
cognition. Although there are a number of studies warded response that has come to elicit punish-
suggesting that creative people do have more access ment) are supposedly due to active inhibition. Cre-
to primary process modes of thought, it is surprising ative people should extinguish slowly and exhibit
that there are no quantitative studies of verbal re- deficits on passive avoidance tasks. In short, dif-
ports during problem solving and creative activity. ferences between more and less creative subjects
Is there an oscillation between secondary process should be found on virtually any task involving
and primary process content in such reports? Is the inhibition if the generalized disinhibition model of
oscillation more extreme for more creative sub- creativity presented above is correct.
jects? Is it more extreme for creative tasks (e.g.,
composing a poem) than for standard laboratory
puzzles (e.g., the Tower of Hanoi problem)? The Conclusions
existence of validated content analysis measures of
primary process versus secondary process content Creativity consists of combining previously un-
(e.g., Martindale, 1975) would make such research related mental elements in a new and useful fash-
relatively straightforward. As well as being of in- ion. In order to do this, it is necessary that a person
terest for its own sake, such research might correct have a large stock of such elements drawn from
the (I believe fallacious) notion of many cognitive diverse domains. Intelligence does not interfere
psychologists that creative thought does not differ with the acquisition of these elements, but, without
from routine problem solving and that both ac- the presence of certain motivational factors, they
tivities employ purely secondary process cognition. will not be learned or remembered in the first place.
There are several studies showing attentional dif- Purely intellectual or secondary process thought
ferences between more and less creative people. does not seem to produce creative ideas. Rather, a
The robustness of these findings needs to be tested person must have access to an opposite primary
using other techniques for measuring attentional ca- process type of cognition that is marked by de-
pacity. Because differences in capacity of short- focused attention and associative, undirected think-
term memory are probably even more important ing. At cognitive and neural levels, primary process
than differences in attentional capacity, it is surpriz- thought may be explained as involving disinhibi-
ing that there are-to my knowledge-no studies tion. It would appear that access to primary process
of differences in short-term memory processes be- thought can only occur within a matrix of person-
tween more and less creative people. Invention and ality traits many of which also involve disinhibi-
creative insight were ascribed above to the ability of tion. Creativity, then, seems to be a general person-
creative people to maintain the cognitive units that alogical trait rather than an isolated cognitive skill.
code a problem in a primed or partially activated
state. If activation of cognitive units generally de-
cays less quickly in creative people, they should
show exacerbated priming effects. There are a vari- References
ety of priming techniques that would be of interest
to explore. For example, a tachistoscopically pre- Albert, R. S. (1971). Cognitive development and parental loss
among the gifted, the exceptionally gifted and the creative.
sented word is recognized more quickly and more Psychological Reports, 29. 19-26.
accurately if it is preceded by a related word. Is this Allport, G. W., Vernon, P. E., &Lindzey,G. (1951). Study of
effect larger for creative people? values: Manual of directions. Boston: Houghton-Mifflin.
Amabile, T. M. (1979). Effects of external evaluation on artistic Crutchfield, R. (1962). Conformity and creative thinking. In H.
creativity. Journal of Personality and Social Psychology, 37, Gruber, G. Terrell, & M. Wertheimer (Eds.), Contemporary
221-233. approaches to creative thinking (pp. 120-140). New York:
Amabile, T. M. (1983a). The social psychology of creativity. Atherton Press.
New York: Springer-Verlag. Dellas, M., & Gaier, E. L. (1970). Identification of creativity:
Amabile, T. M. (1983b). The social psychology of creativity: A The individual, Psychological Bulletin, 73,55-73.
componential conceptualization. Journal of Personality and Dennis, W. (1954). Bibliographies of eminent scientists.
Social Psychology, 45, 357-376. Scientific Monthly, 79, 180-183.
Amabile, T. M. (1985). Motivation and creativity: Effects of Dentler, R. A., & Mackler, B. (1964). Originality: Some social
motivational orientation on creative writers. Journal of Per- and personal determinants. Behavioral Science, 9, 1-7.
sonality and Social Psychology, 48, 393-399. Dewing, K., & Battye, G. (1971). Attention deployment and
Anderson, C. c., & Cropley, A. J. (1966). Some correlates of nonverbal fluency. Journal ofPersonality and Social Psychol-
originality. Australian Journal of Psychology, 18, 218-227. ogy, 17, 214-218.
Andreason, N. C., & Cantor, A. (1974). The creative writer: Domino, G. (1977). Homosexuality and creativity. Journal of
Psychiatric symptoms and family history. Comprehensive Homosexuality, 2, 261-267.
Psychiatry, 15, 123-131. Dykes, M., & McGhie, A. (1976). A comparative study of
Aston, M. A., & McDonald, R. D. (1985). Effects of hypnosis attentional strategies of schizophrenic and highly creative nor-
on verbal and non-verbal creativity. International Journal of mal subjects. British Journal of Psychiatry, 128, 50-56.
Clinical and Experimental Hypnosis, 33, 15-26. Eysenck, H. J. (1967). The biological basis of personality.
Barron, F. (1952). Personality style and perceptual choice. Springfield, IL: Charles C Thomas.
Journal of Personality, 20, 385-401. Eysenck, H. J. (1983). The roots of creativity: Cognitive ability
Barron, F. (1953). Complexity-simplicity as a personality di- or personality trait? Roeper Review, 5(4), 10-12.
mension. Journal of Abnormal and Social Psychology, 68, Feldman, D. (1980). Beyond universals in cognitive develop-
163-172. ment. Norwood, NJ: Ablex.
Barron, F. (1955). The disposition toward originality. Journal of Findlay, C. S., & Lumsden, C. J. (1988). The creative mind:
Abnormal and Social Psychology, 51, 478-485. Toward an evolutionary theory of discovery and innovation.
Barron, F. (1957). Originality in relation to personality and af- Journal of Social and Biological Structures, 11, 3-55.
fect. Journal of Personality, 25, 730-742. Fitzgerald, E. T. (1966). Measurement of openness to experi-
Barron, F. (1958). The needs for order and disorder as motives in ence. Journal of Personality and Social Psychology, 4, 655-
creative activity. In C. W. Taylor (Ed.), The 1957 research 663.
conference on the identification of creative scientific talent Forrest, D. V. (1976). Nonsense and sense in schizophrenic
(pp. 119-128). Salt Lake City: University of Utah Press. language. Schizophrenia Bulletin, 2, 286-301.
Barron, F. (1961). Creative vision and expression in writing and Fromm, E. (1978). Primary and secondary process in waking
painting. In Conference on creative person (pp. 1-19). Berke- and in altered states of consciousness. Journal of Altered
ley: Institute of Personality Assessment and Research, Uni- States of Consciousness, 4, 115-128.
versity of California. Gergen, K. J. (1971). The concept of self. New York: Holt,
Barron, F. (1969). Creative person and creative process. New Rinehart & Winston.
York: Holt, Rinehart & Winston. Getzels, J. W., & Csikszentmihalyi, M. (1976). The creative
Barron, F., & Harrington, D. M. (1981). Creativity, intel- vision: A longitudinal study of problem finding in art. New
ligence, and personality. Annual Review of Psychology, 32, York: Wiley.
439-476. Ghiselin, B. (Ed.), (1952). The creative process. Berkeley: Uni-
Besdine, M. (1968). The Jocasta complex, mothering, and ge- versity of California Press.
nius. Psychoanalytic Review, 55, 259-277. Goertzel, M. G., Goertzel, V., & Goertzel, T. G. (1978). Three
Blake, W. (180311906). Letter to Thomas Butts. In A. G. B. hundred eminent personalities. San Francisco: Jossey-Bass.
Russell (Ed.), The letters of William Blake (p. 115). London: Gorenstein, E. E., & Newman, J. P. (1980). Disinhibitory psy-
Methuen. chopathology: A new perspective and a model for research.
Bowers, R. S., & van der Meulen, S. (1970). The effect of Psychological Review, 87, 301-315.
hypnotic susceptibility on creativity test performance. Journal Gough, H. G. (1979). A creative personality scale for the Adjec-
of Personality and Social Psychology, 14, 247-256. tive Check List. Journal of Personality and Social Psychol-
Bruner, J. (1962). The conditions of creativity. InH. Gruber, G. ogy, 37, 1398-1405.
Terrell, and M. Wertheimer (Eds.), Contemporary ap- Graves, R. (1966). The white goddess: A historical grammar of
proaches to creative thinking (pp. 1-30). New York: Atherton poetic myth. New York: Noonday Press.
Press. Greenwald, A. G., Pratkanis, A. R., Leippe, M. R., &
Bruner, J., & Potter, M. C. (1964). Interference in visual recog- Baumgardner, M. H. (1986). Under what conditions does
nition. Science, 144, 424-425. theory obstruct research progress? Psychological Review, 93,
Campbell, D. T. (1960). Blind variation and selective retention 216-229.
in creative thought as in other knowledge processes. Guilford, J. P. (1950). Creativity. American Psychologist, 14,
Psychological Review, 67, 380-400. 469-479.
Carringer, D. C. (1974). Creative thinking abilities in Mexican Harding, R. (1965). An anatomy of inspiration. New York:
youth. Journal of Cross-Cultural Psychology, 5, 492-504. Barnes & Noble.
Coleridge, S. T. (181711907). Biographia literaria. Oxford: Harrington, D. M. (1975). Effects of explicit instructions to "be
Clarendon Press. creative" on the psychological meaning of divergent thinking
Costa, P. T., Jr., & McCrea, R. R. (1985). TheNEO Personality test scores. Journal of Personality, 43, 434-454.
Inventory manual. Odessa, FL: Psychological Assessment Harrington, D. M., Block, J. H., & Block, 1. (1987). Testing
Resources. aspects of Carl Rogers's theory of creative environments:
Child-rearing antecedents of creative potential in young ado- Kroeber, A. L. (1944). Configurations of culture growth.
lescents. Journal of Personality and Social Psychology, 52, Berkeley: University of California Press.
851-856. Kubie, L. S. (1958). Neurotic distortion of the creative process.
Hasenfus, N., & Magaro, P. (1976). Creativity and schizo- Lawrence, KS: University of Kansas Press.
phrenia: An equality of empirical constructs. British Journal Kuhn, T. S. (1962). The structure of scientific revolutions.
of Psychiatry, 129, 346-349. Chicago: University of Chicago Press.
Helmholtz, H. von (1896). Vortriige und Reden. Brunswick: Lambert, W. E., Tucker, G. R., & d'Anglejan, A. (1973).
Friedrich Viewig und Sohn. Cognitive and attitudinal consequences of bilingual schooling:
Helson, R. (1971). Women mathematicians and the creative The St. Lambert project through grade five. JournalofEduca-
personality. Journal of Consulting and Clinical Psychology, tional Psychology, 65, 141-159.
36, 210-220. Landry, R. G. (1972). The enhancement of figural creativity
Helson, R. (1973a). The heroic, the comic, and the tender: Pat- through second language learning at the elementary school
terns of literary fantasy and their authors. Journal of Person- level. Foreign Language Annals, 4,111-115.
ality, 41,163-184. Lehman, H. C. (1953). Age and achievement. Princeton: Prince-
Helson, R. (1973b). Heroic and tender modes in women authors ton University Press.
of fantasy. Journal of Personality, 41, 493-512. Lindgren, H. C., & Lindgren, F. (1965). Brainstorming and
Helson, R. (1977). The creative spectrum of authors of fantasy. orneriness as facilitators of creativity. Psychological Reports,
Journal of Personality, 45, 310-326. 16, 577-583.
Helson, R. (1978). Creativity in women. In J. Sherman & F. Lombroso, C. (1901). The man of genius (6th ed.). New York:
Denmark (Eds.), The psychology of women: Future directions Charles Scribner's Sons.
in research (pp. 553-604). New York: Psychological Lynn, S. J., & Rhue, J. W. (1986). The fantasy-prone person:
Dimensions. Hypnosis, imagination, and creativity. Journal of Personality
Heston, L. L. (1966). Psychiatric disorders in foster home reared and Social Psychology, 51, 404-408.
children of schizophrenic mothers. British Journal ofPsychia- MacKinnon, D. W. (1960). The highly effective individual.
try, 112, 819-825. Teachers College Record, 61, 367-378.
Holden, C. (1987). Creativity and the troubled mind. MacKinnon, D. W. (1962). The personality correlates of
Psychology Today, 21(4), 9-10. creativity. A study of American architects. In G. S. Nielson
Houston, J. P., & Mednick, S. A. (1963). Creativity and the (Ed.). Proceedings of the 14th International Congress of Psy-
need for novelty. Journal of Abnormal and Social Psychol- chology (pp. 11-39). Copenhagen: Munksgaard.
ogy, 66, 137-141. MacKinnon, D. W. (1975). IPAR's contribution to the concep-
Hudson, L. (1975). Human beings: The psychology of human tualization and study of creativity. In I. A. Taylor & J. W.
experience. New York: Anchor. Getzels (Eds.), Perspectives in creativity (pp. 60-89). New
Isen, A. M., Daubman, K. A., & Nowicki, G. P. (1987). York: Aldine.
Positive affect facilitates creative problem solving. Mackworth, N. H. (1965). Originality. American Psychologist,
Journal of Personality and Social Psychology, 52, 1122- 20, 51-66.
1131. Maddi, S. R. (1965). Motivational aspects of creativity. Journal
Jamison, K. R., Gerner, R. H., & Hammen, C. (1980). Clouds of Personality, 33, 330-347.
and silver linings: Positive experiences associated with pri- Maddi, S. R., & Andrews, S. (1966). The need for variety in
mary affective disorders. American Journal of Psychiatry, fantasy and self-description. Journal of Personality, 34, 610-
137, 198-202. 625.
Jarvik, L. F., & Chadwick, S. B. (1973). Schizophrenia and Martindale, C. (1971). Degeneration, disinhibition, and genius.
survival. In M. Hammer, K. Salzinger, & S. Sutton (Eds.), Journal of the History of the Behavioral Sciences, 7, 177-
Psychopathology: Contributions from the social and behav- 182.
ioral sciences (pp. 57-73). New York: Wiley. Martindale, C. (1972a). Anxiety, intelligence, and access to
Kanner, A. D. (1976). Femininity and masculinity: Their rela- primitive modes of thought in high and low Remote Associ-
tionships to creativity in male architects and their indepen- ates Test scorers. Perceptual and Motor Skills, 35, 375-381.
dence from each other. Journal of Consulting and Clinical Martindale, C. (1972b). Father absence, psychopathology, and
Psychology, 44, 802-805. poetic eminence. Psychological Reports, 31, 843-847.
Karlsson, J. L. (1968). Genealogical studies of schizophrenia. In Martindale, C. (1975). Romantic progression: The psychology
D. Rosenthal & S. S. Kety (Eds.), The transmission of schizo- of literary history. Washington, DC: Hemisphere.
phrenia (pp. 85-94). Oxford: Pergamon Press. Martindale, C. (1977). Creativity, consciousness, and cortical
Karlsson, J. L. (1978). Inheritance of creative intelligence: A arousal. Journal of Altered States of Consciousness, 3, 69-
study of genetics in relation to giftedness and its implication 87.
for future generations. Chicago: Nelson-Hall. Martindale, C. (1980). Subselves: The internal representation of
Keefe, J. A., & Magaro, P. (1980). Creativity and schizo- situational and personal dispositions. In L. Wheeler (Ed.),
phrenia: An equivalence of cognitive processing. Journal of Review ofpersonality and social psychology (Vol. I, pp. 193-
Abnormal Psychology, 89, 390-398. 218). Beverly Hills: Sage.
Kihlstrom, J. F., & Cantor, N. (1983). Mental representations of Martindale, C. (1981). Cognition and consciousness. Home-
the self. In L. Berkowitz (Ed.), Advances in experimental wood, IL: Dorsey.
social psychology (Vol. 17, pp. 1-47). New York: Academic Martindale, C. (1984a). The evolution of aesthetic taste. In K.
Press. Gergen & M. Gergen (Eds.), H istorial social psychology (pp.
Koestler, A. (1959). The Sleepwalkers. New York: Macmillan. 347-370). Hillsdale, NJ: Erlbaum.
Kris, E. (1952). Psychoanalytic explorations in art. New York: Martindale, C. (1984b). The pleasures of thought: A theory of
International Universities Press. cognitive hedonics. Journal ofMind and Behavior, 5, 49-80.
Martindale, C. (in press). Creative imagination and neural ac- Newell, A., Shaw, J., & Simon, H. (1962). The process of
tivity. In R. Kunzendorf & A. Sheikh (Eds.), creative thinking. In H. Gruber, G. Terrell, &M. Wertheimer
Psychophysiology of mental imagery: Theory, research, and (Eds.). Contemporary approaches to creative thinking (pp.
application. Amityville, NY: Baywood. 63-119). New York: Atherton Press.
Martindale, C. (1988). Innovation, discovery, and evolution. In Nietzsche, F. (190811927). Ecce Homo. In The philosophy of
E. O. Wilson (Ed.), Special issue on the creative mind: To- Friedrich Nietzsche. New York: The Modern Library.
ward an evolutionary theory of discovery and innovation. Nordau, M. (1895). Degeneration (5th ed.). London: William
Journal of Social and Biological Structures, 11, 120-122. Heinemann.
Martindale, c., & Armstrong, J. (1974). The relationship of Norman, W. T. (1963). Toward an adequate taxonomy of per-
creativity to cortical activation and its operant control. Journal sonality attributes: Replicated factor structure in peer nomina-
of Genetic Psychology, 124, 311-320. tion personality ratings. Journal of Abnormal and Social Psy-
Martindale, C., & Greenough, J. (1973). The differential effect chology, 66, 574-583.
of increased arousal on creative and intellectual performance. Olive, H. (1972). Sibling resemblances on divergent thinking.
Journal of Genetic Psychology, 123, 329-335. Journal of Genetic Psychology, 120, 155-162.
Martindale, C., & Hasenfus, N. (1978). EEG differences as a Pankove, E., & Kogan, N. (1968). Creative ability and risk-
function of creativity, stage of the creative process, and effort taking in elementary school children. Journal of Personality,
to be original. Biological Psychology, 6, 157-167. 36, 420-439.
Martindale, C., & Hines, D. (1975). Creativity and cortical Peabody, D. (1987). Selecting representative trait adjectives.
activation during creative, intellectual, and EEG feedback Journal of Personality and Social Psychology, 52, 59-71.
tasks. Biological Psychology, 3, 71-80. Perkins, D. (1981). The mind's best work. Cambridge: Harvard
McCrea, R. R. (1987). Creativity, divergent thinking, and open- University Press.
ness to experience. Journal ofPersonality and Social Psychol- Pettigrew, T. (1958). The measurement and correlates of catego-
ogy, 52, 1258-1265. ry width as a cognitive variable. Journal of Personality, 26,
McCrea, R. R., & Costa, P. T., Jr. (l985a). Updating Norman's 532-544.
"adequate taxonomy": Intelligence and personality dimen- Pettigrew, T. (1982). Cognitive style and social behavior: A
sions in natural language and in questionnaires. Journal of review of category width. In L. Wheeler (Ed.), Review of
Personality and Social Psychology, 49, 710-721. Personality and Social Psychology (Vol. 3, pp. 199-224).
McCrea, R. R., & Costa, P. T., Jr. (1985b). Openness toexperi- Beverly Hills: Sage.
ence. In R. Hogan & W. H. Jones (Ed.), Perspectives in Poincare, H. (1913). Thefoundationsofscience. Lancaster, PA:
personality (Vol. I, pp. 145-172). Greenwich, CT: JAI Science Press.
Press. Prentky, R. A. (1980). Creativity and psychopathology: A neu-
McCrea, R. R., & Costa, P. T., Jr. (1987). Validation of the rocognitive perspective. New York: Praeger.
five-factor model of personality across instruments and ob- Reznikoff, M., Domino, G., Bridges, C., & Honeyman, M.
servers. Journal of Personality and Social Psychology, 52, (1973). Creative abilities in identical and fraternal twins.
81-90. Behavior Genetics, 3,365-377.
McNeil, T. F. (1971). Prebirth and postbirth influences on the Rogers, C. (1954). Towards a theory of creativity. ETC: A Re-
relationship between creative ability and recorded mental ill- view of General Semantics, 11, 249-263.
ness. Journal of Personality, 39, 391-406. Root-Bernstein, R. S. (1984). On paradigms and revolutions in
Mednick, S. A. (1958). A learning theory approach to schizo- science and art: The challenge of interpretation. Art Journal,
phrenia. Psychological Bulletin, 55, 316-327. 44, 109-118.
Mednick, S. A. (1962). The associative basis of the creative Rosen, K., Moore, K., & Martindale, C. (1983, September).
process. Psychological Review, 69, 220-232. Creativity and rate of habituation. Paper presented at the
Mendelsohn, G. A. (1976). Associative and attentional pro- Eighth International Colloquium on Empirical Aesthetics,
cesses in creative performance. Journal of Personality, 44, Cardiff, Wales, United Kingdom.
341-369. Rosenberg, S., & Gara, M. A. (1985). The multiplicity of per-
Mendelsohn, G. A., & Griswold, B. B. (1966). Assessed cre- sonal identity. In P. Shaver (Ed.), Review of personality and
ative potential, vocabulary level, and sex as predictors of the social psychology (Vol. 6, pp. 87-114). Beverly Hills: Sage.
use of incidental cues in verbal problem solving. Journal of Schultz, D. P. (1965). Sensory restriction: Effects on behavior.
Personality and Social Psychology, 4,423-431. New York: Academic Press.
Mischel, W. (1979). On the interface of cognition and person- Simon, H. A. (1983). Discovery, innovation, and development:
ality: Beyond the person-situation debate. American Psychol- Human creative thinking. Proceedings of the NationalAcade-
ogist, 34, 740-754. my of Sciences, 80, 4569-4571.
Moore, G. (188511959). Confessions of a young man. New Simonton, D. K. (1975). Age and literary creativity: A cross-
York: Capricorn Books. cultural and transhistorical survey. Journal of Cross-Cultural
Myden, W. (1959). Interpretation and evaluation of certain per- Psychology, 6, 259-277.
sonality characteristics involved in creative production. Simonton, D. K. (1976). Biographical determinants of achieved
Perceptual and Motor Skills, 9, 139-158. eminence: A multivariate approach to the Cox data. Journal of
Nardi, K., & Martindale, C. (1981, April). Creativity and pref- Personality and Social Psychology, 33, 218-226.
erence for tones varying in dissonance and intensity. Paper Simonton, D. K. (1984). Genius, creativity, and leadership:
presented at the Eastern Psychological Association conven- Historiometric inquiries. Cambridge: Harvard University
tion, New York. Press.
Naroll, R., Benjamin, E. Fohl, F., Fried, M., Hildreth, R., & Singer, J. L., & McCraven, V. G. (1961). Some character-
Schaefer, J. (1971). Creativity: A cross-historical pilot sur- istics of adult daydreaming. Journal of Psychology, 51,
vey. Journal of Cross-Cultural Psychology, 2, 181-188. 151-164.
Stein, M, (1974). Stimulating creativity (Vol. 1). New York: Wallach, M. A. (1970). Creativity. In P. H. Mussen (Ed.),
Academic Press. Carmichael's manual of child psychology (Vol. I), pp. 1211-
Stein, M. (1975). Stimulating creativity (Vol. 2). New York: 1272). New York: Wiley.
Academic Press. Wallach, M. A. & Kagan, N. (1965). Modes of thinking in young
Suler, J. R. (1980). Primary process thinking and creativity. children. New York: Holt, Rinehart & Winston.
Psychological Bulletin, 88, 144-165. Wallas, G. (1926). The art of thought. New York: Harcourt,
Tang, P. C. C. (1984). On the similarities between scientific Brace & World.
discovery and musical creativity: A philosophical analysis. Watkins, J. G. (1978). The therapeutic self. New York: Human
Leonardo, 17, 261-268. Sciences Press.
Taylor, I. A., & Getzels, J. W. (Eds.). (1975). Perspectives on Weber, J. P. (1969). The psychology of art. New York:
creativity. Chicago: Aldine. Delacorte.
Tchaikovsky, P. (1878/1906). A letter of 1878. InR. Newmarch Weisberg, R. W. (1986). Creativity: Genius and other myths.
(Ed.), Life and letters of Peter Tchaikovsky (pp. 274-275). New York: W. H. Freeman.
London: John Lane. Werner, H. (1948). Comparative psychology of mental develop-
Thackeray, W. M. (1899). The works ofW. M. Thackeray (Vol. ment. New Y::>rk: International Universities Press.
12). London: John Murray. Wild, C. (1965). Creativity and adaptive regression. Journal of
Vandenberg, S. G. (Ed.). (1968). Progress in human behavior Personality and Social Psychology, 2, 161-169.
genetics. Baltimore: Johns Hopkins University Press. Woody, E., & Claridge, G. (1977). Psychotic ism and thinking.
VanZelst, R. H., & Kerr, W. A. (1954). Personality self-assess- British Journal of Social and Clinical Psychology, 16, 241-
ment of scientific and technical personnel. Journal of Applied 248.
Psychology, 38, 145-147. Wyatt, D. F., & Campbell, D. T. (1951). On the liability of
Victor, H. R., Grossman, J. c., & Eisenman, R. (1973). Open- stereotype or hypothesis . Journal of Abnormal and Social
ness to experience and marijuana use in high school students. Psychology, 46, 496-500.
Journal of Consulting and Clinical Psychology, 41, 38-45. Zuckerman, H. (1977). The scientific elite. New York: Free
Waddington, C. H. (1969). Beyond appearance: A study of the Press.
relationship between painting and the natural sciences in this
century. Edinburgh: Edinburgh University Press.
CHAPTER 14
The Self and Creativity

SEVERAL CONSTRUCTS IN SEARCH OF A THEORY
E. Thomas Dowd
I was invited by the editors of this volume to write a bly line worker could be said to be creative. True
chapter on "creativity and the self." Given the creativity is invention, or the process of making
vague and slippery nature of these concepts, this is something new. This is not as obvious as it seems.
not an easy task-attempting to be heuristic in my What passes for new is often nothing more than a
creative thinking about creativity. I shall first define variation on an already existing theme; for exam-
creativity and selfhood, then relate creativity to sev- pIe, our assembly line worker may "create" a new
eral person variables and psychological processes. I method of fastening a fender to an automobile
will then critique the literature and offer some sug- body. Although management might value this
gestions for a new program of research. Last, I will change (and even give the worker a bonus), this
discuss various ways of fostering creativity in type of activity is not normally what we mean by
people. creativity. Likewise, what is called new is often
nothing more than the same technique with a differ-
ent label or rationale. It is more accurate to refer to
Definition of Creativity these variations as innovative rather than creative.
The term creativity should be reserved for activities
Two constructs seem to emerge from the diction- or products that are truly original and break new
ary definitions of creativity: the act of making or ground, even though they build to some extent on
producing and, more specifically, the act of making previous activity. Creative endeavors generally re-
something new. Unless one produces something, sult in strikingly new formulations that are not de-
one cannot be creative. Thus, pure mental activity rived from what preceded them and in that sense
without a resulting product is not creativity. This represent a discontinuity.
distinction is important, because people often as-
sume that thought is in itself creative and are willing
to pay large sums of money for think tanks from Definition of the Self
which the product is often minimal. Simply produc-
ing is not enough, however; if it were, every assem- The dictionary definition of self consists of a
number of meanings and is then followed by almost
E. Thomas Dowd • Department of Educational Psychology, three pages (in small print) of various self-attributes
Kent State University, Kent OH 44242. (e.g., self-conscious, self-directed, self-praise,
233
self-satisfied). Two general constructs seem to The concept of self has had a tangled history that
emerge out of this chaos, however: unity and con- in part may account for its fuzziness. It belongs to
sistency. Because these two constructs seem es- the domain of philosophy as well as to psychology.
pecially important, they have been selected for fur- Adam Smith originally spoke of society as a mirror
ther discussion. The self possesses unity in that the in which individuals could see themselves as spec-
experiencing individual is reflexively aware of only tators of their own behavior and who therefore
one identity. Although we speak colloquially of a could acquire knowledge about themselves.
"better self" or a "weaker self," it is tacitly under- William J ames and Charles Horton Cooley, in elab-
stood that these are only metaphors expressing vari- orating on this concept, stressed that individuals
ous aspects of a unified self. Only in the case of true develop a self-feeling based on how they think they
multiple personalities does the unity of selfhood as appear to other people. This has been termed the
experienced by the individual break down. There social self or the reflected self and implicit in this
does seem to be a deep human fear, however, that construct is the assumption that people develop
the unit of the self is fragile and can be easily de- their sense of self out of their ongoing interactions
stroyed. The tremendous fascination with Robert with other people.
Louis Stevenson's Dr. Jekyll and Mr. Hyde, which Carl Rogers is perhaps the psychologist most
magnificently expresses this dilemma, may be due identified with the concept of self. Rogers (1959)
in part to this fundamental fear. referred to the self as an organized and consistent
The self also possesses consistency in that people set of perceptions of the characteristics of "I" or
seem to themselves and others to be the same person "me" as well as the perceptions of the relationships
across situations and over time. Although we often of this "I" to other people. Thus, although Rogers
speak of' 'not being ourself today," it is understood stresses the internal knowledge and consistency of
that this is only a metaphorical figure of speech and the sense of self, he also acknowledges that it is at
does not reflect reality. Likewise, we assume that least partially formed through our interactions with
people will behave reasonably consistently in a va- others.
riety of situations and we may regard it as a sign of
emotional distress if they do not. Again, as a spe-
cies, we seem to have a deep fear that this consisten-
cy of the self may be fragile, because behavioral Creativity and the Self-Concept
and emotional volatility in ourselves or others
frightens us. We also become concerned if we On the surface, it would seem logical that highly
change our personality patterns too readily, as we creative individuals would also have a positive
feel that we do not know "who we are" any longer. self-concept and high self-esteem. However, some
This attitude is very likely a major source of re- people also have argued, usually on the basis of
sistance to psychotherapeutic change. People have selected cases, that there is a strong correlation be-
a strong need to maintain the consistency of experi- tween creativity and neuroticism. Therefore, this
enced selfhood, even those aspects of the self that question warrants empirical investigation. Felker
are unhelpful or even destructive. Guidano and and Treffinger (1971) found that high self-concept
Liotti (1982), in particular, have referred to the fourth-grade students scored significantly higher
sense of confusion and disorientation of personal than low self-concept students on self-evaluation
identity and resulting fright that often accompanies of creative abilities and verbal fluency, flexibility,
profound therapeutic change. People strive to main- and originality. Jaquish and Ripple (1981) found a
tain consistency of self-experience, even when that relationship between measures of divergent think-
consistency may be damaging to them. ing (to be discussed later) and self-esteem across
Central also to an understanding of the concept of different age groups. Nabi (1979) argued that the
self is its self-reflexive nature. In other words, only creative person possesses self-acceptance to a
by self-reflexively examining our own thinking greater extent than less creative people. On the
process, that is, thinking about our own thinking, other hand, Whiteside (1977) found that highly
can we arrive at a sense of self at all. Thus, the self creative females had lower self-esteem than either
is inextricably intertwined with self-awareness and men or less creative women. Furthermore, she
self-identity. In this capacity for examining the self, found that highly creative men had self-esteem
humans seem to be unique. scores similar to women of minimal creative abil-
CHAPTER 14 • THE SELF AND CREATIVITY 235
ity. Williams, Poole, and Lett (1977) found that Creativity and Self-Actualization
there were no significant differences between the
self-esteem scores of high and low creative indi- Abraham Maslow described self-actualized per-
viduals. Wright, Fox, and Noppe (1975) found no sons as unusually flexible and tolerant individuals
significant relationship between creativity and self- who are able to hold contradictory concepts in their
esteem. They did, however, find a significant rela- minds simultaneously. They have a rich sense of
tionship between creativity and creative self- humor and are able to look at the complexities of
concept. life. Although they enjoy the company of others,
Gifted students may have an especially difficult they tend to be inner-directed. Following Maslow's
time forming a positive self-concept. Torrance model, one might expect that creative people would
(1971) argued that gifted children often have diffi- be high on measures of self-actualization; after all,
culty in evolving a self-concept because parents and the same attributes could be present in both. Again,
teachers may suppress their new and often unusual the evidence is equivocal. Schubert and Biondi
and deviant ideas. By their negative attitudes, these (1977), concluded that there is support for a
adults may communicate to the gifted child that positive relationship between creativity and self-
commonly accepted ideas, values, and concepts are actualization. Similarly, Damm (1970) found that
not to be questioned. This problem may affect the subjects high in both creativity and intelligence
vocational development of creative children as scored significantly higher on self-actualization
well. Holland and Nichols (1964) found significant than subjects who scored high on either creativity or
differences between vocationally decided and un- intelligence. There was no significant relationship,
decided boys on overall creativity. The undecided however, between creativity and self-actualization
boys not only showed greater creativity but also only. On the other hand, Murphy, Dauw, Horton,
showed "unusual potential for artistic and per- and Fredian (1976) found no relationship between
suasive achievement." They concluded that cre- creativity and self-actualization, a finding also re-
ative boys may need to defer vocational choice be- ported by Mathes (1978).
cause of a slow and complex rate of personal Thus, the literature suggests that there is no rela-
development. This is an important point and one to tion between creativity and self-actualization, at
which I will return later. least as measured by such paper-and-pencil tests as
What can we conclude from the research on the Personal Orientation Inventory. The concept of
creativity and self-concept? Certainly the evidence self-actualization is an especially murky one, how-
is far from clear, in that some studies found a ever, and this may contribute to the nonsignificant
positive relationship whereas others found a nega- findings. At any rate, the qualities that contribute to
tive one. Many of the discrepancies may be due to self-actualization do not appear to contribute to
the use of different measures of creativity and of creativity.
self-esteem or self-concept. The latter two concepts
are not precisely the same, and some of the dif-
ferences in the results may be due to using the terms
interchangeably. In addition, people of different Creativity and Locus of Control
ages were used as subjects in different studies. It
may not be appropriate to compare studies using Locus of control, which is a concept developed
people of different ages because the relationship by Julian Rotter, refers to the degree to which peo-
between creativity and self-esteem/self-concept ple perceive that they are controlled internally (i.e.,
may vary over the life span. In any event, it is not at by their own efforts or abilities) or externally (i.e.,
all clear that creativity is necessarily related to a by environmental forces). Individuals possessing
good self-concept. But, in a review of the literature, an internal locus of control believe that what hap-
Schubert and Biondi (1977) found no relationship pens to them is largely governed by such internal
between creativity and maladjustment either. attributes as their intelligence, tenacity, or attrac-
Therefore, it is probably best to say that creativity tiveness. Those possessing an external locus of con-
and the self-concept are moderately positively relat- trol believe that what happens to them is largely
ed at best. Even this conclusion is suspect, how- governed by such external forces as luck, fate, or
ever, because of the variety of measures used to other people's actions. Although locus of control is,
assess self-concept. to some degree, situation-specific, there are never-
theless individual differences in the degree to which Creativity and Divergent Thinking
people attribute outcomes to internal or external
forces. Divergent thinking is characterized by thought
It may seem intuitively obvious that people who processes that radiate outward and explore new
possess an internal locus of control would be more ideas that are generated from the original notion. By
creative than those having an external locus of con- its very nature, divergent thinking is tentative, ex-
trol. After all, creative individuals are commonly ploratory, and creative and is oriented to the devel-
assumed to spend hours, perhaps days, in cognitive opment of possibilities rather than data, to specula-
ruminations and creative activity and appear to be tion rather than conclusions. Convergent thinking,
self-contained and self-oriented. One can scarcely on the other hand, is characterized by reasoning that
imagine an externally oriented creative person. brings together the relevant data and arrives at a
What, after all, would be the point of intense cre- firm conclusion based on these data. It tends to be
ative activity if outcomes were indeed determined deductive rather than inductive. Thus, divergent
by fate, chance, or luck? Bamber, Jose, and Boice thinking can be thought of as more intuitive and less
(1975) indeed found that internals had significantly data-based (or data-bound) than convergent think-
higher scores on the flexibility and fluency mea- ing. Although both types of thinking are necessary
sures of the Unusual Uses subtests of the Torrance under different conditions, it is reasonable to sup-
Tests of Creative Thinking, whereas externals had pose that the ability to engage in divergent thinking
significantly higher elaboration scores. Glover and should be related to creativity. Therefore, it is frus-
Sautter (1976) found very similar results in that trating to find that almost no studies have been con-
internals had higher flexibility and originality ducted to examine this relationship. Jaquish and
scores, whereas externals had higher elaboration Ripple (1981) found that self-esteem predicted di-
scores. They found no differences between inter- vergent thinking across groups. However, in view
nals and externals on fluency. Cohen and Oden of the apparent lack of relationship between self-
(1974) found more mixed results, because esteem and creativity (discussed above), it would
creativity was related to internal locus of control be premature to make much of these results.
only for female kindergarteners. For male students, Meichenbaum (1975) developed and modeled sets
the reverse was true. DuCette, Wolk, and Friedman of self-statements that were operationalized from
(1972) found that internal subjects were more cre- three conceptualizations of creativity and then
ative than externals regardless of race. On the other asked college students to rehearse these statements.
hand, Bolen and Torrance (1978) found no dif- The students showed a significant increase in origi-
ferences in creativity as measured by the Unusual nality and flexibility on tests of divergent thinking
Uses subtests of the Torrance Tests of Creative as well as an increase in preference for complexity
Thinking between internals and externals. Rich- and higher self-concept. Meichenbaum concluded
mond and de la Serna (1980) found that Mexican that creativity may be enhanced by getting people to
externals were more creative than internals and sug- talk to themselves differently.
gested that creative college students of the 1970s Despite the lack of research evidence, I would
have a more external orientation than those of a suggest that to engage in divergent thinking com-
decade ago. fortably, people need to be able to do several things.
Thus, even though the majority of the studies First, they must be able to tolerate ambiguity. In
support the proposition that internals are more cre- general, people feel most comfortable when they
ative than externals, there is some doubt as to just can reach cognitive closure on a subject as soon as
how firm these findings are. In particular, it is in- possible. Their world is therefore more predictable
triguing to speculate that perhaps there are shifts and, hence, safer. Second, they must be able to hold
over time in the relative numbers of internals and contradictory ideas simultaneously in their minds.
externals in the population and, therefore, in the Most people search for answers that are "either-
relative numbers of creative people who are inter- or"; creative individuals should be able to think in
nals or externals. The results of Richmond and de la terms of "both-and. " Again, this removes some of
Serna would point to such a conclusion. In any the predictability and safety from life and may re-
event, it appears that, in general, creativity may be sult in psychological confusion. Third, they must
associated with an internal locus of control, al- be able to readily incorporate new ideas into their
though the evidence is somewhat shaky. cognitive systems and to modify the constructs by
which they organize and make sense of the informa- of the outer world. If this is indeed true, then hyp-
tion which they receive from their environment. notic ability might be one indicator of creative
People must categorize sensory information into potential.
regular and recursive patterns or else they would not Hypnotic ability, or the capacity to enter and
be able to generalize and predict behavior across experience a hypnotic trance readily, is a relatively
situations and time. Creative people, however, are stable individual-difference variable that is nor-
able to maintain flexible constructs and avoid a pre- mally distributed in the population (Bowers, 1976).
mature "hardening of the categories" and "im- It is difficult to determine the correlates of hypnotic
maculate perception. " In this regard, it is notewor- ability, because most investigations to date have
thy that Meichenbaum (1975) was able to train found little or no relationship of hypnotic ability to
individuals to be more original and flexible in their anything else. It does appear, however, that hyp-
thinking by having them make self-statements that notic capacity is correlated with the capacity to be-
were based on three conceptualizations of cre- come imaginatively involved in mental activity and
ativity. Obviously, this is an area that could benefit to concentrate on one idea or stimulus while block-
from future research. ing out competing stimuli (Kroger, 1977). People
who possess good hypnotic ability report losing all
track of time when absorbed in an idea, a capacity
Creativity and Intrinsic/Extrinsic attributed to creative people as well. There are also
Motivation some indications that hypnotic ability may be, in
part, genetically based, so that the same might be
Some evidence indicates that creativity is related said of highly creative individuals. In any event,
to intrinsic motivation. Koestner, Ryan, Bernieri, this is another area that is ripe for further
and Holt (1984) evaluated the impact on intrinsic exploration.
motivation of informational and controlling limit-
setting. They found that controlling limits reduced
the quality and creativity of artistic production, but Creativity and Complex Thinking
that limits did not undermine intrinsic motivation if
these limits were informational in nature. Amabile There are some hints that creative thought may be
(1982, 1985) found that crea'tive writings done un- more complex than less creative thought. Recall
der an extrinsic motivation orientation were signifi- that Holland and Nichols (1964) concluded that cre-
cantly less creative than those written under an in- ative boys may be characterized by a slow and com-
trinsic motivation condition or even in a control plex rate of personal development. In a similar
condition. These results tend to support the com- vein, Dellas (1978) suggested that the sense of iden-
mon idea that creative activity is generated and sus- tity of highly creative individuals is tentative, ten-
tained "from within." uous, and complex. Nabi (1979) argued that the
creative person is more complex than the less cre-
ative. Meichenbaum (1975) found that subjects
Creativity and Openness to Inner who were trained to make creative self-statements
Experiences showed an increase in preference for complexity. In
Piagetian terms, the structural differentiation (dif-
It is widely supposed that creative people are ferentiation-integration-subsequent integration
unusually open to inner experiences in a sort of at a higher level) that characterizes the development
mystical fashion, and there is research support for of every person may, in the case of creative people,
such a notion. Schaeffer, Diggins, and Millman proceed more slowly in a more complex fashion and
(1976) found that male creativity correlated most may involve more layers. This is again a topic that
strongly with openness to theoretical and aesthetic could benefit greatly from further research.
experiences. Female creativity was correlated with
openness to inner experiences and sensation seek-
ing. MacKinnon (1964) examined a variety of cre- Creativity and Autonomy/Independence
ative individuals in different fields and concluded,
among other things, that creative individuals are There is a popular stereotype of the creative per-
especially open to experience of the inner self and son as an isolated, socially maladjusted, noncon-
forming loner. Although this view may provide mation is matched and filtered. Thus, selfhood is
some comfort for the noncreative individual, it, like built out of the recursive and dynamic interaction
all stereotypes, may contain a grain of truth. Nabi of two opponent processes; self-perception and
(1979) concluded that the creative person demon- perception of the world. It would seem that any
strates autonomy and independence. MacKinnon definition of the self should take these cognitive
(1964) argued that creative people are independent processes into account.
in thought and action. On the other hand, Williams Second, there is a problem with the measurement
et al. (1977) found that creative individuals did not of creativity. Although standardized tests, such as
appear to value highly impulsivity, desire for the Torrance Tests of Creative Thinking, have often
change, and nonconformity, but valued such been used, other assessment devices have ranged
qualities as obedience, diligence, attentiveness, from teacher descriptions (Sisk, 1972) to judges'
and cooperation. The literature discussed earlier ratings (Amabile, 1982). With such divergent mea-
under locus of control is also relevant here. It is sures, including different types of standardized
likely that the creative person, far from being mal- tests, it is difficult to compare results or to carry out
adjusted, may actually be better adjusted than the a sustained line of research. We may be faced with a
less creative. paradox here, however. If creativity involves the
production of something new and not merely some-
thing different, then it becomes difficult to measure
it in advance of an actual creative product. In other
Critique and Suggestions for Future
words, creativity, by its very nature, may be ac-
Research cessible only after the fact.
Third, the literature on creativity is characterized
What can we conclude from the review of the largely by a collection of one-shot studies that do
research on creativity and the self? The initial an- not seem to be informed by any significant higher-
swer is-not much! The reasons for this sad state order theoretical constructs or programmatic thrust.
of affairs can be classified as deficiencies in three The result is that it is difficult to draw firm conclu-
areas. First, there is a problem with definitions of sions regarding the relationship of some common
key concepts. As noted earlier, creativity itself has definition of creativity to any other psychological
been poorly defined, having been confused with variables. It is absolutely impossible, however, to
innovation or simple extrapolation from existing determine any causal directions between creativity
information or products on one hand and pure and a host of psychological variables or environ-
mental activity without a definite product on the mental manipulations. This is a problem shared by
other. Until we can agree on the definitional pa- other topics of psychological investigation. Thus, it
rameters of creativity, we will be seriously handi- is hoped that it is not too much to expect that jour-
capped in our research efforts on the topic. As a nals specializing in studies of creativity (such as the
starting point, I would suggest that the definition Journal of Creative Behavior) will themselves dis-
of creativity focus on the process of divergent play creativity by fostering the systematic explora-
thinking. Similarly, the concept of self has been tion of programmatic research using common defi-
variously defined and is often confused with the nitions, assessment instruments, and theoretical
concept of identity. Certain of the more humanistic bases.
writers have not helped the situation by discussing However, some tentative conclusions can be
the self in obscure, almost mystical, terminology. drawn from the existing research. First, it is appar-
Perhaps a proper definition of this concept might ent that creativity is only weakly related, at best, to
begin with a consideration of the self-reflexive such person variables as self-concept and self-actu-
nature of human thinking (e.g., thinking about alization. Creativity may be somewhat correlated
one's thinking) and the contribution that that pro- with internal locus of control, suggesting that cre-
cess can make to the experience of selfhood. ative individuals tend to believe that they are in
Guidano (1987) has recently described increasing control of their own destiny and fate. What is not
self-knowledge and, therefore, one's experience of clear, however, is the direction of causality. Are
selfhood as derived from interactions with other people creative because they are internally con-
people and a consequent differentiation from those trolled or are they internally controlled because they
others. This increasing sense of self then becomes are creative? The answer has implications for fos-
a template against which incoming external infor- tering creativity. Second, there appears to be a
moderate relationship between creativity and such There have been a few attempts to foster
psychological processes as intrinsic motivation, creativity. The Meichenbaum (1975) study referred
openness to inner experiences, preference for com- to earlier found that divergent thinking and a prefer-
plex thinking, relative autonomy, and the capacity ence for complexity can be increased by self-state-
for divergent thinking. Unless creativity is an in- ment training. Sisk (1972) identified a group of
born trait, giving rise to these processes, it is likely gifted children who were also considered by their
that creativity can be fostered by training people, teachers to be "low creatives" and put them
especially children, in such activities as divergent through ten weekly three-hour sessions that in-
thinking, complex thinking, openness to inner excluded role-playing, creative writing, open discus-
periences, and the like. sion, and artistic expression. At the end of the ses-
What should an ideal program of research look sions, the teachers reported that the students had
like in this area? First, a common definition of become "more aware of their own strengths."
creativity should be developed so that investigators They were also getting better grades. Teacher rat-
are dealing with common concepts. I would suggest ings are notoriously unreliable, however, and
that the process of divergent thinking should playa awareness of strengths seems only tangentially re-
central role in this definition. It is intriguing as well lated to creativity. Schempp, Cheffers, and
as frustrating (despite Guilford's efforts) that al- Zaichowsky (1983) tested 208 first through fifth
most no research has investigated the relation be- graders who were in physical education classes in
tween creativity and divergent thinking. Likewise, which the teacher dominated all classroom deci-
some attention should be paid to a common defini- sions or in which the students were encouraged to
tion of seltbood. This topic, in particular, has been share in decision making. They found that the
plagued with much woolly and flatulent excres- shared decision-making group scored significantly
cences and is badly in need of clear thinking with a higher than the teacher-dominated group on
solid conceptual base. Newer developments in cog- creativity, motor skills, and self-concept. In addi-
nitive psychotherapy (e.g., Guidano, 1987) may tion, students allowed to assist in making their own
provide the integrated theoretical insights needed. decisions regarding their learning had more positive
Second, common assessment instruments should be attitude scores. Amabile (1985) found that students
developed from these definitions, utilizing multiple who wrote poems for extrinsic reasons wrote signif-
measures. For example, behavioral assessments as icantly less creative poems than those written for
well as paper-and-pencil instruments should be intrinsic reasons. In the same vein, Koestner et al.
used. Third, overarching theoretical viewpoints (1984) found that information limits on first and
should be used to guide research directions, perhaps second graders did not undermine intrinsic moti-
involving the comparisons of studies that are de- vation whereas controlling limits did. These results
rived from differing theoretical orientations. suggest that creativity can be fostered by reducing
Fourth, leadership should be provided by a signifi- authoritarian directions, by involving children and,
cant journal in the field, perhaps by devoting a se- presumably, adults in decision-making processes,
ries of special issues to the presentation of studies and by deliberately training people in those skills
reflecting a common theoretical approach. Fifth, that are thought to be related to creativity.
more attention should be paid to the investigation of There are other ways to foster creativity as well.
methods of fostering and increasing creativity. In- We can encourage people to avoid "premature
deed, it could be argued that this is a crucial national evaluation" of an issue until they have collected
need on which our survival as a nation and a society information and taken time to think through the
may ultimately depend. implications of each side. Organized debates are
useful for this purpose. Brainstorming seeks to
avoid premature evaluation as well. We can encour-
Fostering Creativity age complex thinking by encouraging people to be-
come comfortable with ambiguities, unresolved is-
As part of their initial charge, the editors of this sues, and paradoxes. In fact, explicit training in
volume have given me liberty (or perhaps license) paradoxical thinking (Dowd & Milne, 1986), in
to speculate freely regarding the nature of creativity which concepts are true if they are false and false if
as it relates to the self. I therefore wish to offer some they are true, might be useful in fostering creativity
suggestions for fostering creativity in oneself and in by giving individuals experience in holding contra-
others. dictory information in their minds simultaneously.
This is not as easy as it sounds. The psychological we not only make mistakes but view them simply as
press of most people demands that new information progress-markers toward ultimate success. It is well
be fitted into preexisting categories as rapidly as to remember that the easiest way of avoiding mis-
possible and that a fundamental examination of takes is to do nothing. Those who aspire to height-
their conceptual categories be avoided. It is very ened creativity should be told to increase their
uncomfortable to suspend judgment or to reevaluate failure rate in order to increase their success rate.
one's habitable patterns of thinking. In addition, Humor has been used by many approaches to
most social institutions-family, school, and psychotherapy as a distancing technique-distanc-
church-strongly encourage people to follow the ing from our problems and from our tendency to
rules and avoid questioning of authority and soci- take ourselves too seriously. Humor can also foster
etal norms. The press, both from within and with- creativity in that it can help us distance from an
out, is toward the maintenance of existing cognitive overinvolvement in our rules of life, our ways of
concepts and behavior patterns. However, cre- making sense of the world, and our automatic as-
ativity demands that we constantly reevaluate our sumptions. Humor can help us to see the ambiguity
existing cognitive categories and remain willing to in many situations and avoid a premature cognitive
modify or even suspend them on occasion. Thus, closure.
there is an ongoing tension between maintaining Activities such as meditation may foster
one's concepts, which is necessary for functioning creativity as well. During meditation we tem-
in the world, and revising these categories, which is porarily suspend the" if-then" linear patterned way
necessary for creativity. of thinking, in which every event has a cause and a
Most people engage in convergent thinking, in result. This linear causal way of thinking is so in-
which they search for the right answer to a problem. grained in Western culture that it is difficult to real-
Although this may be useful in many circum- ize that other ways of thinking exist. Creativity may
stances, creative people recognize that there are be enhanced by adopting a more circular way of
often many right answers or no right answers. Ask- thinking in which the focus is on relationships, pos-
ing ourselves or others, "What are some other ways sibilities, and recursive patterns, rather than on lin-
oflooking atthis?" or "What else might I do?" can ear causality and single-outcome events. Medita-
be useful in fostering divergent thinking and, tion can help free the mind of old thinking patterns
hence, creativity. Simply asking, "What if .. ? and allow new patterns to develop. In this context,
can get one in the habit of thinking divergently. I am it is interesting to realize how many truly creative
reminded here of Robert Kennedy's famous state- activities were the result of serendipity when the
ment, "Some men look at what is and ask why? I individual's mind was on something else. Cre<;ltive
look at what is not and ask why not?" Looking for solutions may require that one not think about the
additional possibilities and relationships in every- problem rather than think about it.
thing can foster increased creativity. Engaging in appropriate leisure activities can
From childhood on we are taught to "follow the also enhance creativity. I have noted (Dowd,
rules." There are times, however, when we should 1984) that leisure activities can provide one of the
purposefully discard the rules in order to see few opportunities for people to engage in explora-
through to a new solution. Rules constrain thinking tory, divergent, and creative activities, rather than
and channel thought processes in certain directions. problem-solving, goal-directed, and convergent
Without any rules, we would have no way of using activities. McDowell (1984) argued that leisure ac-
past experience to guide and direct our future activities can provide a heightened sense of self-ex-
tions, thus forcing us to treat every situation as a pression and autonomous control and allow one to
totally new experience. Obviously, this approach explore one's outer ranges of tolerance for novelty,
would be extremely inefficient as well as psycho- complexity, unfamiliarity, and competence. Iso-
logically unsettling. However, without the ability Ahola (1984) stated that intrinsic motivation is at
to break rules occasionally, we increasingly find the heart ofleisure behavior, not unlike creativity,
ourselves bound by the rigidity of our preexisting and advocated the deliberate cultivation of intrinsic
rules. Because many of these rules are not con- motivation through the development of perceived
sciously known, however, the help of another per- competence and self-efficacy by means of leisure
son is often needed. experiences. Thus, engaging in leisure experiences
Although we are taught to avoid making mistakes that are freely chosen may significantly enhance
whenever possible, creative thinking demands that creativity in all spheres of life.
In one sense, however, creativity cannot be fos- Cohen, S., & Oden, S. (1974). An examination of creativity and
tered at all. Because creativity is, by definition, locus of control in children. Journal of Genetic Psychology,
124, 179-185.
unplanned, spontaneous, and divergent, then any Damm, V. J. (1970). Creativity and intelligence: Research im-
planned activity to foster it will render impossible plications for equal emphasis in high school. Exceptional
that which is desired. It is similar to the "Be Spon- Children, 36, 565-569.
taneous" paradox, in which an attempt to comply Dellas, M. (1978). Creative personality and identity. Psycho-
with the order makes spontaneity impossible. For logical Reports, 43, 1103-1110.
Dowd, E. T. (1984). Leisure counseling with adults across the
if the order is obeyed, it cannot be spontaneous, life span. In E. T. Dowd (Ed.), Leisure counseling: Concepts
but if it is spontaneous, it must be disobeyed. Only and applications (pp. 214-233). Springfield, IL: Charles C
by disobeying the order can it be obeyed. Similar- Thomas.
ly, we cannot create creativity; we can only set up Dowd, E. T., & Milne, C. R. (1986). Paradoxical interventions
in counseling psychology. The Counseling Psychologist, 14,
the conditions for it is hoped its spontaneous 237-282.
occurrence. DuCette, J., Wolk, S., & Friedman, S. (1972). Locus of control
I want to offer one caveat, however. Many peo- and creativity in black and white children. Journal of Social
ple do not take kindly to the demonstration of Psychology, 88, 297-298.
creativity in others and may work to discourage its Felker, D. W., & Treffinger, D. 1. (1971, February). Self con-
cept, divergent thinking abilities, and attitudes about
expression. Creative individuals have been accused creativity and problem solving. Paper presented at the Ameri-
of being maladjusted, antisocial, and antireligious. can Educational Research Association convention, New
Creative children have earned the animosity of York.
some teachers. Creative people have been villified Glover, J. A., & Sautter, F. (1976). An investigation of the
relationship of four components of creativity to locus of con-
(Einstein) or even killed (Socrates). Thus, indi- trol. Social Behavior and Personality, 4, 257-260.
viduals should know that creativity is practiced at Guidano, V., & Liotti, G. (1982). Cognitive therapy and emo-
some risk to themselves. tional disorders. New York: Guilford Press.
I want to close with a final suggestion. People are Guidano, V. F. (1987). Complexity of the self: A developmental
often urged to be reasonable and logical in their approach to psychopathology and therapy. New York:
Guilford Press.
thinking processes, an attitude which is certain Holland, J. L., & Nichols, R. C. (1964). The development and
death to creativity. Instead, I would like to suggest validation of an indecision scale: The natural history of a
an attitude that may be more conducive to creativity problem in basic research. Journal ofCounseling Psychology,
but that is likely, at the same time, to earn the 11, 27-34.
Iso-Ahola, S. E. (1984). Social psychological foundations of
enmity of others. leisure and resultant implications for leisure counseling. In E.
T. Dowd (Ed.), Leisure counseling: Concepts and applica-
The reasonable man adapts himself to his environment. tions (pp. 97-125). Springfield, IL: Charles C Thomas.
The unreasonable man persists in adapting his environment to Jaquish, G. A., & Ripple, R. E. (1981). Cognitive creative
himself. abilities and self-esteem across the adult life span. Human
Therefore, all progress depends on the unreasonable man! Development, 24, 110-119.
Koestner, R., Ryan, R. M., Bemieri, F., & Holt, K. (1984).
May each of you spend a certain amount of time Setting limits on children's behavior: The differential effects
each day being unreasonable! of controlling vs. informational styles on intrinsic motivation
and creativity. Journal of Personality, 52, 233-248.
Kroger, W. S. (1977). Clinical and experimental hypnosis (2nd
ed.). Philadelphia: J. B. Lippincott.
MacKinnon, D. W. (1964, March).ldentification and develop-
References ment of creative abilities. Paper presented at the conference on
Creativity and the Mentally Gifted, Fresno, CA.
Amabile, T. M. (1982). Children's artistic creativity: Detrimen- Mathes, E. W. (1978). Self-actualization, metavalues, and
tal effects of competition in a field setting . Personality and creativity. Psychological Reports, 43,215-222.
Social Psychology Bulletin, 8, 573-578. McDowell, C. F. (1984). Leisure: Consciousness, well-being,
Amabile, T. M. (1985). Motivation and creativity: Effects of and counseling. In E. T. Dowd (Ed.), Leisure counseling:
motivational orientation on creative writers. Journal of Per- Concepts and applications (pp. 5-51). Springfield, IL:
sonality and Social Psychology, 48, 393-397. Charles C Thomas.
Bamber, R. T., Jose, P. E., & Boice, R. (1975). Creativity as Meichenbaum, D. (1975). Enhancing creativity by modifying
affected by differential reinforcements and test instructions. what subjects say to themselves. American Educational Re-
Bulletin of the Psychonomic Society, 6, 361-363. search Journal, 12, 129-145.
Bolen, L. M., & Torrance, E. P. (1978). The influence on Murphy, J. P., Dauw, D. c., Horton, R. E., & Fredian, A. J.
creative thinking of locus of control, cooperation, and sex. (1976). Self-actualization and creativity. Journal of Creative
Journal of Clinical Psychology, 34, 903-907. Behavior, 10, 39-44.
Bowers, K. S. (1976). Hypnosisfor the seriously curious. New Nabi, K. S. (1979). Personality and self: The creativity discus-
York: Norton. sion. Indian Psychological Review, 18, 5-8.
Richmond, B. 0., & de la Serna, M. (1980). Creativity and health: III. Creativity and adjustment. Journal of Creative
locus of control among Mexican college students. Behavior, 11,186-197.
Psychological Reports, 46, 979-983. Sisk, D. (1972). Relationship between self-concept and
Rogers, C. R. (1959). A theory of therapy, personality, and creativity: Theory into practice. Gifted Child Quarterly, 16,
interpersonal relations, as developed in the client-centered 229-234.
framework. In S. Koch (Ed.), Psychology: A study of science, Torrance, E. P. (1971). Identity: The gifted child's major prob-
Vol. 3 (pp. 184-256). New York: McGraw-Hili. lem. Gifted Child Quarterly, 15, 147-155.
Schaeffer, C. E., Diggins, D. R., & Millman, H. L. (1976). Whiteside, M. (1977). Self-concept differences among high and
Intercorrelations among measures of creativity, openness to low creative college students. Journal of College Student Per-
experience, and sensation-seeking in a college sample. sonnel, 18, 224-227.
College Student Journal, 10, 332-339. Williams, A. J., Poole, M. E., & Lett, W. R. (1977). The
Schempp, P. G., Cheffers, J. T., &Zaichowsky, L. D. (1983). creativity/self-concept relationship reviewed: An Australian
Influence of decision making on attitudes, creativity, motor longitudinal perspective. Australian Psychologist, 12, 313-
skills, and self-concept in elementary children. Research 317.
Quarterly for Exercise and Sport, 54, 183-189. Wright, R. J., Fox, M., & Noppe, L. (1975). The interre-
Schubert, D. S., & Biondi, A. M. (1977). Creativity and mental lationship of creativity, self-esteem, and creative self-con-
cept. Psychology, 12, 11-15.
CHAPTER 15
Creativity and Psychopathology

GAMBOLING AT THE SEAT OF MADNESS
Robert Prentky
This chapter addresses a topic that has been the rightfully so. It certainly would not be evolu-
subject of intense curiosity throughout civilized his- tionarily adaptive to wed the curse of mental illness
tory. What are the roots of genius? Are geniuses to the gift of creativity. This dilemma was raised by
divinely inspired? Are they mad? What is the nature Barron (1958):
of their gift? Is it something that can be cultivated?
Or perhaps harnessed like a beast of draught? Do But one cannot readily abandon the idea that to create is in some
geniuses possess a genetic loading for their gift? If sense-perhaps in the best sense-to be healthy in mind. Yet
the germ of inspiration has a biological component, how can the maladjustment of many great creative minds be
it would seem that the cornerstone of the creative reconciled with the assertion that they are in some respects un-
usually healthy? (p. 163)
process remains inaccessible to most of us. If this is
the case, it poses an interesting dilemma: Why
should a feature of human behavior so critical to the Barron determined, however, that the seeming il-
advancement of the species be restricted to so small logic of associating maladjustment with creativity
a segment of the species? After more than 2,500 can be understood from the standpoint of the source
years of curiosity, we can conclude very little that is of creative inspiration-the unconscious. The cre-
definitive. We cannot state with certainty that some ative person is said to be at higher risk for disorder
genetic contribution enhances creative potential. by virtue of gamboling at the seat of disorder.
We cannot state with certainty that creativity can be In the past, theoreticians contributed little in the
encouraged, though we undoubtedly can state how way of explanation for this farfetched relationship,
it may be discouraged. Despite reams of stories, little, that is, except for the psychodynamic for-
some farfetched and some true, relating creativity mulation. More recently, investigators havt;: been
to mental illness, the idea that one must be slightly seriously and productively pursuing other angles
mad to be creative is so counterintuitive that it is (e.g, Cropley & Sikand, 1973; Dykes & McGhie,
rejected by most sober-minded scientists. And 1976; Hasenfus & Magaro, 1976; Karlsson, 1978;
Keefe & Magaro, 1980; Martindale, 1977-1978,
and with his colleagues [with Hasenfus, 1978; with
Robert Prentky • Research Department, The Massachusetts
Treatment Center, Bridgewater, MA 02324; and Department of Hines, 1975; and with Hines, Mitchell, & Covello,
Psychiatry, Boston University School of Medicine, Boston, MA 1984]; Rothenberg, 1976, 1979, 1983). The crucial
02118. question, however, is not whether creativity and
243
mental illness are causally linked but whether our aspects of nature into one. In Behrens's terms, Don
knowledge of mental illness offers us any insight Quixote's penchant was to see hidden likeness
into the mechanics of creativity. Does the alle- where likeness clearly did not reside. What we seek
gation, be it true or false, that an unusually high to understand is not Quixote's madness nor Sancho
percentage of extraordinarily creative people were Panza's straight vision of reality ("those which ap-
disturbed help us in any way to comprehend the pear yonder are not giants, but windmills"), but
creative process? I have argued elsewhere that the the genius of those who first thought of harnessing
coincidental association between creativity and real the wind on land using the same sails that harness
emotional turmoil in a number of highly visible, the wind at sea. Such constructive "likenesses" are
highly celebrated cases may provide, in a seren- discussed in detail elsewhere (Bronowski, 1965;
dipitous fashion, an important key for unlocking Prentky, 1980).
one of the secrets of their creativity-specific The aim of this chapter is twofold: (1) to high-
modes for processing information (Prentky, 1979, light some of the most well-known cases of reputed
1980). It is not my thesis that mental illness is a association between creativity and mental illness,
prerequisite for creativity. It is my thesis that a and (2) to review the empirical literature that has
fruitful theoretical model for understanding the cre- examined the personalities and cognitive styles of
ative process may be in terms of cognitive style those defined as creative. Because other chapters in
(i.e., mode of information processing), and that a this volume focus on the latter issues (personality
genetic predisposition to certain forms of mental and cognitive style), treatment in this chapter will
illness may be associated with a cognitive style sim- be abbreviated and address the specific question of
ilar to that which promotes creativity. If there is one whether the scientific literature provides any sup-
thing that may characterize all highly creative peo- port for an incidence of mental illness among cre-
ple, it is the unique manner in which they apprehend ative people that exceeds chance expectation, and,
their world (i.e., they see things, experience things, if so, how that increased incidence may be
or conclude things that the rest of us do not). A understood.
disorder of cognition (or information processing)
must be present for someone to receive a diagnosis
of schizophrenia (Callaway & Naghdi, 1982). An Historical Overview
Thus, a cardinal feature of both creativity and men-
tal illness is information processing. Highly cre- Alliances or misalliances between "genius" and
ative people tum into a distinct and productive asset psychosis have been noted in the literature for at
what mentally disturbed people blindly grope for- least 2,300 years. One of the earliest references
the ability to absorb and efficiently process infor- dates back to Aristotle's Problemata (ca. 360 B.C.):
mation from the environment. The majority of us "Those who have become eminent in philosophy,
seem to find sufficient light in the tunnel to navigate politics, poetry, and the arts have all had tendencies
without excessive fumbling, though we are rarely toward melancholia." Aristotle may have had his
blessed with that sudden illumination which, for a eminent predecessor Socrates in mind. The first
split second, makes everything clear. Severely case study of genius was written about Socrates in
mentally ill people seem to fall at one of two ex- 1836 by the psychiatrist Louis-Fran<;ois Lelut (Du
tremes, a constant blinding illumination or a black- Demon de Socrate). Becker (1978) stated that
out condition. The obvious parsimony of this analo- "Socrates' inclination 'to take the inspirations of
gy is intended only to provide a simple heuristic his conscience for the voice of a supernatural agent
model for looking at two otherwise unrelated phe- (his demon),' confirmed, in Lelut's opinion, that
nomena-creativity and mental illness. Socrates suffered from a 'most undeniable form of
In a delightful article on "lunatics, lovers, and madness' " (p. 28). Gibson's (1889) observation
poets," Behrens (1975) concluded that "the dif- was similar. He stated, "Socrates, heathen though
ference between the schizophrenic act and the act of he was, was a great philosopher; yet it is credited of
creation is the difference between Don Quixote (as him that he acknowledged a tutelary genius or de-
madman) and author Cervantes (as poet). Madmen mon, who informed him of coming events and di-
take metaphors literally, and the difference is one of rected-probably in a measure controlled-his be-
awareness" (p. 232). Bronowski (1965) referred to havior" (p. 4).
the Pythagorean "Geistesblitz" as "explosions of Despite the often-noted association between
a hidden likeness" in which the creator fuses two "enthousiasmos" (enthusiasm or "mania")-de-
CHAPTER 15 • CREATIVITY AND PSYCHOPATHOLOGY 245
monic possession or melancholia-and genius dur- hospital for the insane. Fourth, and least frequently, he startles
ing antiquity, Greek" madness" is not equatable to the world by an invention or discovery in science or by an origi-
nal compOSItion of great merit in art, music or literature. He is
modem notions of insanity. According to Aristotle, then styled a genius. (p. 752)
the "homo melancholic us , " depending upon hu-
moral balance, could be divinely distinguished or
Among many others who supported this degeneracy
mad. When black bile was not properly counter-
theory was the Italian criminologist Cesare Lom-
balanced by blood, phlegm, and yellow bile, the
result was depression and anxiety. Being of melan- broso.
The first systematic study of the relationship be-
cholic temperament, however, did not require that
tween genius and insanity was published by
one be acutely depressed or anxious. Rosen (1969)
cited Phaedrus, a first century A.D. Roman author Lombroso in 1864. Over a quarter of a century
later, Lombroso (1891) published his research in
of fables and reputed translator of some of Aesop's
book form, concluding that genius was often a "de-
fables, as writing:
generative psychosis of the epileptic group." In a
The greatest blessing came by way of madness, indeed of mad- subsequent edition of the same work, Lombroso
ness that is heaven sent. It was when they were made that the (1910) stated that
prophetess at Delphi and the priestess at Dodona achieved so
much for which both states and individuals in Greece are anyone who has had the rare fortune to live with men of genius is
thankful: when sane they did little or nothing. As for the Sibyl soon struck by the facility with which they misinterpret the acts
and others who by the power of inspired prophesy have so often of others, believe themselves persecuted, and find everywhere
foretold the future to so many, and guided them aright, I need not profound and infinite reasons for grief and melancholy.
dwell on what is obvious to everyone. (p. 84)
Nisbet (1912) related genius to no fewer than 40

This same distinction was made fourteen cen-
~epara~e maladies, from apoplexy to vanity, and
turies later, during the Renaissance, where "gen-
mcludmg such varied afflictions as gout, rheu-
io" was described in terms of melancholia and
~atism, i~becility, ne'er-do-wellism, opium eat-
"pazzia" in terms of madness. Pazzia as applied to
mg, prodIg~lity, scrofula (tuberculous disorder),
a sane melancholic genius usually implied such at-
sexual passIOn, and skull shape. Approximately
tributes as eccentricity, sensitivity, moodiness, and
1.0~ to 125 pa¥es are included in the three separate
solitariness (Becker, 1978). None of these qualities IIstmgs of "msanity," "insane temperament,"
was pejorative. The most distinguished artists pos-
and "hallucinations." In a somewhat more recent
sessed pazzia. With the arrival of the Enlighten-
examination, Tsanoff (1949) claimed that "in cre-
~ent, the first real change occurred in the concep-
ative activity, genius is at the ultimate limit of ten-
tIon of genius. Genius was seen as "rational"
s.ion, the utmost of reach, of intensity, penetra-
rooted primarily in the imagination. Rather th~n
tIon .... But no wonder if it overtakes the physical
humors being balanced, imagination, judgment,
and mental powers and unbalances the high-strung
and aesthetic "taste" had to be balanced. This
genius" (p. 29).
change explains why seventeenth-century artists
The literature is replete with anecdotal and
and scholars (e.g., Rubens, Rembrandt, Bernini)
c1inic~1 descriptions of the putatively pathological
generally were not described as melancholic (Witt-
behaVIOr of hundreds of gifted artists, musicians,
kower, 1973).
novelists, poets, and scientists (see Table 1). Grant
Some 2,100 years after Aristotle, the prevalent
(1968), in his treatment of "great abnormals " de-
conception of genius was that it was allied with
sCri?ed Ka~a as suicidally depressed, pr~bably
degeneracy. This notion was first set forth by Bene-
SC.hIz~phren~c, van Gogh as a neurotic-hysteric
dict Augustin Morel (ca. 1850), who argued that
WIth ImpulsIve and unreasoning rages, Strindberg
dege.nerac~ was. a state of biological inferiority and
as a. paranoid .schizophrenic, and Poe as psycho-
that It was mhented. It was conceived that geniuses
pathIc, paranOId, and megalomaniacal. According
evol~ed from the same maladaptive gene pool as the
to Lombroso (1910), Schopenhauer and Tasso
lowlIest elements of society-criminals and luna-
~ight be diag?osed, in contemporary nosology, as
tics. Babcock (1895) enumerated the dire conse-
bIpolar mallIc-depressives. Chopin, Newton
quences of being born with degenerate genes:
Gogol, Linnaeus, Pascal, and Swift were all said t~
First, and most prominent in the order of frequency is an early su~fer from paralytic dementia, whereas Coleridge,
death. Second, he may swell the criminal ranks. Third, he may MIchelangelo, Cromwell, Johnson, Mill, Rossini,
become mentally deranged and ultimately find his way into a Raphael, and Schiller were said to be melancholic
Table 1. Presumed Psychiatric Disturbances of Some Eminent Writers, Artists, Composers, and
Scientists a
Schizophrenia Affective disorders (unipolar and bipolar) Personality disorders b
Baudelaire Barrie Lamb E.B. Browning

H6lderlin Balzac London c R. ,Browning
Johnson Berryman c Lowell Carlyle
Kant Blake Mill Comte
Pound Boswell Maupassant Eddy
Strindberg Byron O'Neill Eliot
Swift Chatterton c Plath C Gogol
Coleridge Poe Heine
Writers, Collins Roethke Huxley
poets and Conrad Rossetti Proust
philosophers Cowper Rousseau Rimbaud
Crane c Saroyan Spencer
Fergusson Schiller Tennyson
Fitzgerald Schopenhauer Zola
Frost Sexton c
Goethe Shelley
Hemingwayc Smart
Hopkins Tasso
Kafka Woolfc
Beil d Crevel c
Brendel d Michelangelo
Cellini Modigliani c
Da Vinci Pollock c
El Greco Raphael
Goya Rothko c
Klotz d Van Goghc
Knupfer d
Artists
Moogd
Neter d
Orth d
Pohl d
Rembrandt
Sell d
Wain
Welz d
Donizetti Berlioz Beethoven

Mendelssohn Chopin Schubert
Rimsky-Korsakov Elgar Wagner
Handel
Mahler
Rachmaninoff
Composers
Rossini
Schumann
Scriabin
Tchaikovsky
Wolf
Wood
Table 1. (Continued)
Schizophrenia Affective disorders (unipolar and bipolar) Personality disorders b
Cantor Darwin Ampere

Copernicus De Forest Einstein
Descartes Mayer Freud
Faraday Kammerer c Heaviside
Hamilton Mendel
Scientists Legrange Nightingale
Linnaeus
Newton
Pascal
Swedenborg
Weierstrass
aThese "diagnostic classifications" are based almost entirely on presumptive judgments made from archival and anecdotal evidence
and should be construed as inferential "best guesses."
bThis category undoubtedly includes individuals who were highly eccentric but otherwise not psychiatrically disturbed. It also
includes others (e.g., Comte, who was described by Nisbet, 1912, as schizophrenic) who rightly belong in a different category.
"Suicides.
dThese ten artists were designated as schizophrenic by Prinzhom (1972); however, case descriptions strongly suggest that several
were affective disorders.
(Tsanoff, 1949). Lombroso (1910) further claimed given to violent tantrums, occasionally very de-
that many great musicians, including Mozart, pressed, and often "muddled" or forgetful; how-
Schumann, Beethoven, Handel, Pergolesi, and ever, there is little evidence that he was delusional:
Donizetti, suffered from" attacks of insanity," in-
cluding delusions, hallucinations, depression, and The picture then is of persistent ill-health, of a prevailing mood
mania. Lombroso also included Rousseau, New- of depression, a highly strung, suspicious, "persecuted" man,
unstable under stress, hypomanic at times, impUlsive to the point
ton, Comte, and Ampere in this latter category. of violence, perfectionist, deaf, irritable. Nevertheless, he had
Lange-Eichbaum (1932), who looked at the tem- enormous charm, and although he might one day drive a visitor
poral relationship of mental illness to creativity, away with brutal discourtesy, he would put himself out to be
found that many artists become psychotic only after attentive on another occasion. Here is a man who tried to make
up for his failings in temperament, breeding, and education,
their major contributions were made. He included whose tremendous achievements were in spite of cruel handicaps
in this group Baudelaire, Donizetti, Kant, Faraday, and who did not flounder in self-pity, though not above occasion-
Newton, Copernicus, and Linnaeus. Lange- ally being pathetic in letters to women. (p. 460)
Eichbaum placed the majority of highly creative
artists in the "psychopath" category, a term not Schumann believed that Beethoven and Men-
referring to psychopathy in the Cleckley (1964) delssohn dictated musical compositions to him
sense but to a poorly socialized, schizoid, or schiz- from their tombs (Lombroso, 1910). In a recent
othymic personality. In this group he included Bee- article about Schumann, Murphy (1979) described
thoven, Michelangelo, Schopenhauer, Byron, and the composer's battle with mental illness:
Heine.
Beethoven was said to have experienced a variety Robert had been mentally unstable all his life, haunted by fears of
insanity since the age of 18.
of subsidiary delusions, ranging from persecution,
In 1854, he complained of a "very strong and painful attack"
hero-rescue, fraticide, and sacrifice to birth fan- of ear trouble that had bothered him before. This was followed by
tasies (Solomon, 1975). Grillparzer, an Austrian illusions, such as dictation by angels of a theme on which he
playwright and frequent companion of Beethoven, wrote some piano variations.
described Beethoven as "half-crazy" and "a wild About 16 days later, he asked to be taken to a lunatic asylum,
and the next day attempted to drown himself.
beast" when angry (Reichsman, 1981). Larkin's In March 1854, he was put in a private asylum in Endenich,
(1970) thorough treatment of Beethoven's medical where he lived for about 2Y2 years. Clara was kept from him for
history suggests that the composer was eccentric, fear that her visit would be too disturbing.
She did visit him in July 1856, when it was realized that death creator of the immortal Peter Pan, was beset by
was near. He seemed to recognize her, but was unable to speak obsessions, compulsions, childhood amnesia, and
intelligibly.
He died on July 29, 1856. He was only 46. (p. 5C)
deep depressions (Alston, 1972). William Cowper
was said to be manic-depressive, though he com-
Chopin was described by George Sand as "shut- posed his verse during periods of quiescence, and
ting himself up in his room for whole days, weep- Swift "was haunted by the spectre of insanity"
ing, walking, breaking his pens, repeating and al- throughout his life (Pickering, 1974).
tering a bar a hundred times, spending six weeks on Gibson (1889) commented that:
a single page" (Storr, 1976, pp. 65-66). Storr Swift was a man whose soul could be attuned to the softest,
(l971) also cited the case of Wagner: gentlest music, but which could be aroused to the most tem-
pestuous energy. This unhappy condition of Swift's mind was
Wagner had one of the most remarkable creative imaginations of palpably connected in the latter part of his time with diseased
any artist who has ever lived. His originality, the assurance with brain: gid<!iness and deafness assaulted him; his memory failed;
which he handled his material, and the enonnous scale on which ungovernable rage followed; and this was succeeded by furious
the world he created was constructed, are all unparalleled .... madness. One phase more, and we see him quiet, passionless,
And yet Wagner, as an individual, was deplorable: Unreliable, idiotic, in which condition death gently overtook him. (p. 45)
selfish, arrogant; a sponger who was constantly in debt, he was
quite unscrupulous both sexually and financially. It seems un-
likely that this poet of passionate love ever achieved a really
Pickering (1974) quoted Nicholson as stating that
satisfactory relationship with a woman; more especially as we
know that he retained a strong fetishistic interest in silk and Shelley's hallucinations were vivid and detailed. On one occa-
perfume. The personality of Wagner remained infantile; but his sion, while walking near Pisa, he saw a man coming slowly
creations contained all that he himself lacked. (p. 123) towards him dressed in a long cloak and with a hood or cowl
hiding his face. When the man was only a few yards distant he
raised his hood, and Shelley was much startled to see that the
British playwright Peter Schaffer (1981) drama- man was in fact himself. (p. 292)
tized the relationship between Wolfgang Amadeus
Mozart and Antonio Salieri in Amadeus. Schaffer Gibson (1889) stated that
portrayed Mozart as a childish, conceited, uncouth
rival to the more successful Salieri. Salieri, court with all his faults, Shelley was, perhaps, the grandest meta-
composer to the Austrian Emperor Joseph II, alone physical poet, as he is said to have been one of the best classical
recognized Mozart's genius. Driven by jealousy scholars, in Europe. Nevertheless, his life was a failure in almost
every particular. His nearest relatives rejected him; he suffered
and aided by Mozart's rude court behavior, Salieri tonnents of mental and bodily pain, and was the subject of ex-
managed to effectively squeeze Mozart out of royal traordinary illusions. No wonder, then, that he had difficulty in
commissions. Mozart died believing that Salieri bearing up against the mental dejection which fitfully attacked
poisoned him. Mozart's fears were dismissed. Thir- him. (p. 48)
ty-two years later, Salieri attempted suicide, at
which time· he confessed to having poisoned It was suggested by Pickering (1974) that Eliz-
Mozart. Salieri's confession was given no more abeth Barrett Browning was "psychoneurotic,"
credence than Mozart's accusation. In fact, though her chronic lung problems (possibly pulmo-
Mozart's death has been attributed to kidney dis- nary tuberculosis), her life-long invalidated condi-
ease and not to poisoning (Reichsman, 1981). tion, and her addiction to morphine certainly com-
Reichsman quoted from two letters that Mozart plicate any attempt to extract a clear psychiatric
wrote to his wife shortly before his death: diagnosis.
A number of French writers are reputed to have
If people could see into my heart, I should almost feel ashamed. been afflicted by varying degrees of instability. The
To me, everything is cold-cold as ice [and] I can't describe brilliant Marcel Proust certainly was eccentric and,
what I have been feeling-a kind of emptiness which hurts me
dreadfully-a kind of longing which is never satisfied, which in later life, a victim of bronchial asthma, a slave to
never ceases, and which persists, nay, rather increases daily. (p. drugs, and a recluse. However, he also has been
293) charged with possessing an implaqble superego,
evidencing sadistic cruelty and narcissistic self-ab-
Yet another composer, Donizetti, "was during sorption (Bychowski, 1973). Another apparently
his latter years confined in a lunatic asylum. He disturbed French writer was Guy de Maupassant.
died at fifty, leaving a son, like himself, insane" Within two years after completion of "Le Horla"
(Nisbet, 1912, p. 170). in 1887, Maupassant (l900) was "obviously in-
Sir James Barrie, the English writer and the sane," though he did not end up in an asylum until a
suicide attempt in 1892 (Dale, 1952). Dale stated psychiatrically disturbed, the disturbance is likely
that Maupassant's "mental breakdown was at- to be affective (i.e., unipolar or bipolar depression)
tended by migraine, melancholia, obsessions, per- (e.g., Andreasen, 1980; Andreasen & Powers,
versions, hallucinations, and outbreaks of violent, 1975; Andreasen & Canter, 1974; Jamison, 1988;
ill-tempered behavior" (p. 237). Louis-Fran<;ois Jamison, Gerner, Hammen & Padesky, 1980). An-
Lelut wrote his second case study on Pascal in 1846 dreasen and Canter (1974) studied 15 writers from
(L'Amulette de Pascal). Lelut concluded that Pas- the University of Iowa Writers' Workshop and a
cal was the victim of mental illness, an opinion later matched control group of 15 individuals who were
supported by Tsanoff (1949). Yet another French not involved in the arts. Of the 15 writers, four had
writer, Emile Zola, submitted himself to psychi- been hospitalized for a psychiatric problem, com-
atric examination. The majority of the panel of 15 pared with none of the controls. Ten of the 15 writ-
psychiatrists concurred that Zola's "genius had its ers met the diagnostic criteria for affective disorder,
source in the neurotic elements of his tempera- compared with two of the controls. Two of the writ-
ment" (Trilling, 1950). Zola agreed entirely with ers were diagnosed as bipolar, compared with none
their conclusion. of the controls. For eight of the nine writers who
Of the wide variety of psychiatric conditions at- had seen a psychiatrist, the presenting complaint
tributed to the aforementioned individuals, the clus- involved affective symptoms. Andreasen and Can-
ter of conditions subsumed under the category of ter (1974) concluded that
affective disturbance clearly predominates. This is
especially true for writers, poets, philosophers, and creative writers tend to have a high prevalence of what would
composers (see Table 1). Edel (1975) quoted Emer- normally be called psychiatric illness, that this illness is usually
son as saying that "after thirty a man wakes up sad affective disorder, and that their primary relatives also have a
high prevalence of affective disorder. (p. 129).
every morning ... until the day of his death." A
nadir of depression has resulted in the suicide of
many well-known writers, including the Americans In a subsequent study Andreasen and Powers
Sylvia Plath, Hart Crane, and Ernest Hemingway (1975) compared 15 writers from the Writers'
and the Britons Virginia Woolf and Thomas Chat- Workshop with 16 manic and 15 schizophrenic in-
terton. Woolf's reflections suggest an association patients on the Goldstein-Sheerer Object-Sorting
between her depression and her literary inspiration: Test. The authors reported that the writers re-
sembled the manic patients in conceptual style (i.e.,
If I could stay in bed another fortnight (but there is no chance of overinclusive thinking) more than they resembled
that) I believe I should see the whole of "The Waves" ... I the schizophrenic patients. The schizophrenics ex-
believe these illnesses are in my case-how shall I express it?- amined in this study tended to be underinclusive.
partly mystical. Something happens in my mind. It refuses to go
on registering impressions. It shuts itself up. It becomes a Andreasen has expanded her original sample of
chrysalis. I lie quite torpid. often with acute physical pain-as writers to 30, reporting that the proportion of writ-
last year; only discomfort this. Then suddenly something ers treated for mood disorders in this larger sample
springs. (Bell, 1972, p. 142) is 80%, compared with 30% for the comparison
group (Holden, 1987). Forty-three percent of this
Although Rousseau did not take his own life, larger sample evidenced symptoms of manic-de-
Ellrich (1974) recounted Rousseau's "psychologi- pressive illness, compared with 10% of the
cal crisis" (suicidal melancholia). Franz Kafka, controls.
Schopenhauer, and Tasso (the greatest Italian poet Jamison (1988) provided an excellent overview
of the late Renaissance) all had psychological crisis of the clinical and empirical literature on the asso-
similar to that of Rousseau. Edel' s (1975) discus- ciation of major affective disorder with creativity.
sion of depression in the writings of highly creative She reported on a study conducted in 1983 on 47 of
poets and novelists brings to mind Dr. Benjamin Britain's foremost artists and writers. Jamison
Rush, who viewed the tragedy of artistic madness found that 38% of her sample had sought treatment
as "tristimania," a mania of sadness or, tech- for mood disorders, and that approximately three-
nically, agitated depression. Edel found that' 'noth- fourths of those treated had received antidepres-
ing is more chronic among writers than their sants, lithium, or hospitalization. The most severe
sadness. " cases were found among the poets, one-half of
As an aside, there is some empirical evidence to whom required medical intervention in the form of
support the observation that if creative writers are hospitalization and/ or medication. Jamison pointed
out that these rates are even more striking when Scientists do not appear to be exempt from re-
compared with the base rates in the general popula- ports of mental illness. At a Donner Laboratory
tion of 1% for bipolar illness and 5% for unipolar Seminar in Berkeley, California, Karlsson identi-
illness. According to those base rates, one would fied a number of highly creative scientists alleged to
have expected, by chance alone, no cases of bipolar be mentally ill (Robinson, 1977). Karlsson said of
illness and 2 cases of unipolar illness in the full Gregor Mendel, "He wasn't a man of religion. He
sample. joined a monastery because he had a nervous break-
Goertzel, Goertzel, and Goertzel (1978) exam- down when he was seventeen years old. He con-
ined a sample of noted twentieth-century leaders tinued to have breakdowns and had to get away
and creators, finding that 5% suicided or attempted from the rigors of having to make a living, as he put
suicide. Most of the subsample of suicidal cases it" (p. 10). Of Charles Darwin, Karlsson com-
were manic-depressive. Goertzel et al. noted that mented that "psychiatrists who have analyzed his
psychiatric disturbance was more common among life history have concluded that he was manic de-
the creators than among the leaders and, within the pressive and not the victim of some exotic disease
group of creators, more common among the artists he may have picked up in South America" (p. 10).
than among the scientists. In a related study, Karlsson also related that Sir Isaac Newton' 'wrote
Tomlinson-Keasey, Warren, and Elliott (1986) delusional letters which have been preserved, and
sought to predict suicide in a sample of 40 women psychiatrists who read them today say these are the
from the Terman Genetic Studies of Genius data productions of a paranoid schizophrenic" (p. 10).
base. A total of eight women committed suicide as Included in Karlsson's roster of eminent individuals
of 1964. These eight cases were compared with two who exhibited signs of mental illness are Michael
control groups, 15 women who died of natural Faraday, Jack London, and Robert Frost.
causes and were matched on age at death and 17 Karlsson (1978) elaborated upon this theme in
women who were still living in 1964. More than his book on creativity and psychosis:
three-quarters of the entire sample of 40 female
"geniuses" experienced transient depression. As part of the present research, a survey was made of famous
Among the eight who suicided, the depression was individuals who were born in the United States west of the Mis-
"severe and apparently unremitting." sissippi River before 1890. Since the population of that area was
then still relatively small, the total number of men of world fame
Phillips (1982) reported on eight cases of manic- was found to be rather limited. The ten individuals so identified
depression that he treated with lithium. Contrary to were Lee DeForest (inventor), T. S. Eliot (poet), Robert Frost
the findings of Jamison, Gerner, and Goodwin (poet), Edwin Hubble (astronomer), Sinclair Lewis (novelist),
(1979), Phillips observed that the lithium seemed to Jack London (novelist), Edgar Lee Masters (poet), Ezra Pound
have a beneficial effect on the creativity of his pa- (poet), Josiah Royce (philosopher), and Mark Twain (humorist).
In this group it is reported that T. S. Eliot suffered a nervous
tients. He properly noted, however, that no conclu- breakdown at age thirty-three, Jack London exhibited manic and
sions could be drawn about creative output and depressive episodes before committing suicide at age forty, and
lithium prophylaxis. It is doubtful that a profound Ezra Pound was diagnosed as schizophrenic in his mid-fifties. In
disturbance of affect, such as manic-depressive ill- addition, it is recorded that Lee DeForest and Robert Frost,
although apparently never considered psychotic, had episodes of
ness, could facilitate the expression of creativity. moderately severe depression. Frost's sister and his children
Indeed, at least one study has provided evidence for suffered from mental illness. Edgar Lee Masters had some kind
a decrement in the functioning of the right cerebral of nervous exhaustion after writing his famous poems. Mark
hemisphere that is associated with depressive affect Twain also became rather melancholic in his later years, and his
(Tucker, Stenslie, Roth, & Shearer, 1981). How- children exhibited mental disorders. (pp. 135-136)
ever, it may be, as Mitchell (1972) proposed, that
creative writers and artists respond to their experi- Not surprisingly, Isaac Newton's mental status
ences with greater sensitivity and are thus at higher has been the subject of much consideration. In dis-
risk for suffering. Answers to these questions await cussing several biographies of Newton, Holton
additional studies with appropriate control groups, (1978) concluded that "these notes and self-im-
a tight operational definition of creativity, and some posed exercises [Newton's diaries] are a record of
estimate of base rates. Otherwise it will be impossi- Newton's anxiety, confessions, loneliness, repres-
ble to conclude that the observed rate of psychiatric sion of instinctual desires, shame, fantasies of
symptoms in experimental groups exceeds the ex- bringing death to his stepfather, mother, and him-
pected rate of those symptoms in the general self-and the dread of punishment for all these" (p.
population. 270). Anthony Storr found that
Newton, particularly, is an apt subject for psychiatric charac- ventive brilliance with the label of madness than
terization. As you know, Newton was born prematurely, some that which comes from a group of men who met
three months after his father's death. His mother remarried just regularly in Birmingham, England, in the late eigh-
after his third birthday, leaving him in the care of his grand-
mother; a betrayal which he never forgave. Newton grew up to be teenth century. Erasmus Darwin (physician),
a recluse and a very hypochondriacal, anxious, insecure recluse Josiah Wedgwood (pottery manufacturer), James
into the bargain. He was intensely suspicious and quarrelsome, Watt (Scottish engineer), and Matthew Boulton
and in 1693 became sufficiently disturbed in mind for rumours of (manufacturer and engineer) were four of the most
his insanity to gain wide acceptance . . . . The great creators
often seem to be more the prey of intrapsychic conflict than most
inventive minds of eighteenth-century England. In
of us; and when we consider those men of genius whose achieve- fact, it has been said that these four men were the
ments depend especially on a power of abstraction and dissocia- fathers of the Industrial Revolution and, indirectly,
tion from the immediate impact of the physical, like Descartes, of all modem technology. The collective met on the
Newton and Schopenhauer, it is not difficult to show that their night of a full moon and, hence, earned the name of
genius and their psychopathology are intimately related. (Krebs
& Shelley, 1975,p.llO)
the Lunar Society. The members of the society, not
surprisingly, called themselves Lunaticks. Even
The incident of 1693 has been described else- without evidence that any of these men were in fact
where as "a complete nervous collapse" in which "lunatics", there was an irresistible inclination to
Newton ended relationships, infrequently slept, re- associate their inventive genius with "lunacy."
ported nonexistent conversations, and accused Nordau (1900) had his own thoughts on the mat-
friends of plotting against him (Broad, 1981). Par- ter. In referring to the Industrial Revolution, he
enthetically, Newton held such undignified titles as concluded that "we stand now in the midst of a
the "autocrat of science," "the dictatorial Presi- severe mental epidemic" characterized by "degen-
dent of the Royal Society, " and "the authoritarian eration and hysteria. " Nordau believed that
Master of the Mint. " Schopenhauer, Nietzsche, Tolstoy, Ibsen, Wagner,
Charles Darwin has been the subject of lively and Zola were products of this degenerate, hys-
debates as to whether he was afflicted by physical or terical period.
mental illness. The fact that Darwin was healthy Another scientist of prolific accomplishment,
and vigorous in his twenties and a recluse in his typically overlooked in discussions of mental disor-
early thirties was to set a pattern for the remainder der, is Sigmund Freud. Despite the libraries written
of his life. His illness emerged while he was await- by and about Freud, it is rarely suggested that the
ing the departure of the Beagle and disappeared man himself was disturbed. One of the best au-
during the six years of the voyage. If nothing else thorities on the subject, Ernest Jones, stated quite
can be concluded about Darwin, he was either a unequivocally that "there is ample evidence that
hypochondriac or very infirm. Between the years for ten years or so-roughly comprising the nine-
1890 and 1895, he was said to suffer from prostra- ties-[Freud] suffered from a very considerable
tion, deafness, neuralgia, bronchitis, influenza, psychoneurosis" (Pickering, 1974, p. 220). Freud
lumbago, and liver problems (Winslow, 1971). was well aware of his neuroses, as is apparent in a
Darwin also was the victim of chronic insomnia, letter written to Fliess on July 7, 1897. In the letter
headaches, "nervousness," and biliousness. All he comments, "What has been going on inside me I
these symptoms, thought to be common to many still do not know. Something from the deepest
prominent English, Victorians such as Thomas depths of my neurosis has been obstructing any
Huxley, George Eliot (a.k.a. Mary Ann Evans), progress in the understanding of neuroses" (cited in
Robert Browning, Herbert Spencer, and Thomas Pickering, 1974, p. 224). The relationship between
Carlyle, defined a loosely conceived malaise that Freud's most original contributions and his psychic
Winslow (1971) called dyspepsia. Strictly speak- distress is an interesting one. Freud's writings
ing' dyspepsia is an old-fashioned word for indiges- were, in effect, an insightful reflection of his own
tion. However, nervous dyspepsia is indigestion turmoil. He stated, "I have long known that I can't
brought on by emotional turmoil and usually occurs be industrious when I am in good health; on the
in high strung, anxious, nervous individuals. Thus, contrary, I need a degree of discomfort which I
it is not surprising that nervous dyspepsia would be want to get rid of" (Pickering, 1974, p. 222).
accompanied by anyone of a host of psycho- One last case is the remarkable nineteenth-cen-
physiological disorders. tury English reformer and nurse, Florence Night-
Although it is strictly a historical aside, there is ingale. The case of Nightingale is a most unusual
no more easily demonstrable association of in- one, rarely cited in psychiatric quarters, typically
relegated to romantic tales of a gallant young wom- after the person was alive, are unreliable, im-
an slipping silently from one trench or foxhole to precise, and certainly not compatible with modem
another, holding a flickering paraffin lamp and nosology. Occasionally, writers even disagree as to
tending to the agony of the wounded. This image is the mere presence of pathology. Such a case was
not entirely incorrect, but it only scratches the sur- Michelangelo, who was given a clean bill of health
face of a truly remarkable person. During her stay in by Lombroso (1910) but was said to be a melan-
the Crimea, she undoubtedly endured no less the cholic by Tsanoff (1949).
privations and pain than any of the young men she A quick glance through the roster of those alleged
treated. She was overcome by dysentery, rheu- to have been mentally disturbed might well leave
matism, and almost died from Crimean (typhoid) one with the distinct impression that "everyone
fever. Despite all these distresses, her energy and who was anyone" was afflicted. As we will see in
devotion were boundless. That much of her life is the second half of this chapter, available research
common knowledge. simply does not support such a conclusion. In
Her foreign service lasted roughly three years, Havelock Ellis's (1904) well-known study of Brit-
from the time she took up residence at 1 Harley ish "geniuses," for example, he found that out of
Street (1853) to her return from the Crimea after the 1,030 names drawn from the Dictionary ofNational
war (1856). During this time, all her afflictions Biography, only 44 individuals (or 4.2%) had a
were organic, and she endured them with enormous demonstrable mental disorder. Obviously, this per-
courage and strength. From 1857 until the end of centage would change somewhat if contemporary
her life, however, she was invalided with what has DSM-III criteria were applied. To play the devil's
been described as neurasthenia (nervous exhaustion advocate, one may argue, of course, that the ac-
or prostration) (Pickering, 1974). Nightingale's at- complishments of the individuals discussed in the
tacks of breathlessness and giddiness confined her first half of this chapter are of an order of magnitude
to a couch or bed but in no way affected the pro- greater (more creative, more exceptional, more en-
digious amount of work she turned out. Indeed, her during, and more revolutionary) than the accom-
"illness" insulated her from all unwanted intru- plishments of most of the 986 sane entries that Ellis
sions from the outside world, allowing her to devote examined. This tends to be a moot point as it is
every moment to her work. difficult enough to define creativity, let alone to
Apparently, there was no organic basis to the rank order degrees of creativity.
debilitating "nervous" symptoms that incapaci- Many of Ellis's 986 "normals" may have been
tated Nightingale throughout her ninety years. She more or less deviant in cognitive style, though un-
was a fiery, relentless warrior who constantly gave deniably sane. Constable held a similar notion in
battle and achieved victory-particularly where her 1905, "I deny the miraculous in genius; it is noth-
male predecessors had failed. In a thoroughly male- ing more than a very exceptional deviation from the
dominated world, she invariably proved to be "a average" (quoted in Becker, 1978, p. 61). Too
better man than their best" (Pickering, 1974). Al- great a deviance in cognitive style will, of course,
though the conflicts that initiated or contributed to cast the ominous shadow of mental illness, and
her neuroses are speculation, her expenditure of there is no reason to suspect that there should not be
energy seemed to relate to her symptoms. When, in at least as many genuine cases of psychosis among
the latter years of her life, the symptoms abated, her the highly creative as is found in the general popula-
extraordinary drive also declined. Her decline was tion. Similarly, if we take the incidences of non-
not due exclusively to advanced age because it was psychotic depression, personality disorder, and
visible in 1880 with her mother's death, thirty years neurosis found in the general popUlation and in-
before Nightingale died. clude them among the highly creative, we would
The cases and examples discussed thus far are expect to find that a respectable fraction of creative
illustrative of the long history of association be- people had emotional problems.
tween exceptional conduct (howsoever it was de-
fined) and abnormal behavior, ranging from eccen-
tricity to psychosis. It is essential to keep in mind Research on the Personality Correlates
that even informal diagnostic classification is inad- of Creativity
visable, and in most instances impossible. Writers
have had to rely on archival documents, third-per- Over the past several decades, a number of inves-
son accounts, and anecdotes. Thus, resulting "di- tigators have attempted to identify the personality
agnoses," sometimes made several thousand years traits most often associated with creativity. Prin-
cipal research efforts in this regard have been con- (cited in Cattell & Drevdahl, 1955). A similar study
ducted by Cattell and his colleagues using the Six- by Wright (1954) found that men of high scholastic
teen Personality Factor Test (16 PF), MacKinnon, achievement could be distinguished from those of
who used the California Psychological Inventory low scholastic achievement by greater skepticism,
(CPI) and the Minnesota Multiphasic Personality aloofness, withdrawal, emotional maturity, self-
Inventory (MMPI), and Roe, who used the confidence, and tendency to be tense and driven
Rorschach and the Thematic Apperception Test (cited in Cattell & Drevdahl, 1955). Drevdahl
(TAT). (1956) found that creative students were more with-
The seminal research of Roe (e.g., 1951, 1952, drawn and quiescent than noncreative students, as
1953, 1956) sought an answer to the question well as superior in verbal facility, fluency, flexibil-
"What makes a scientist?" through in-depth case ity, and originality. Drevdahl and Cattell (1958)
histories of scientists who were nominated as found large differences when comparing creative
eminent by peers. Roe's (1951) first study of 22 artists and writers with the general popUlation. The
physical scientists was later expanded to include "creatives" appeared to be stable but living under
biologists, social scientists-psychologists, and an- considerable strain. They were more skeptical, ex-
thropologists (1952). Among the multitude of find- traverted, radical, self-sufficient, bohemian (un-
ings drawn from her research, the following points conventional, imaginative, absentminded), and
are noteworthy: (1) as children the scientists were emotionally sensitive than the noncreative controls.
highly independent and self-sufficient; they typ- Of particular interest was that the artists were even
ically enjoyed school and were avid readers; they more tense, extraverted, and radical than the writ-
were slow in social development, shy, introverted; ers. Cross, Cattell, and Butcher (1967) also clearly
and they admired and respected their fathers; (2) in differentiated artists from a matched control group.
the child's home there was stability and harmony; The artists were low in emotional and superego
the father was usually a professional with ad- strength and high in bohemian tendency.
vanced education; and there was an intellectual Cattell and Drevdahl (1955) contrasted the
ambience in the home that projected a love of learn- general adult population and a university under-
ing and a respect for knowledge; and (3) as adults graduate population with three groups of scientists:
the scientists were indifferent to religion; they often researchers, teachers, and administrators. Re-
had the fortune of experiencing one or two out- searchers, compared with teachers and admin-
standing mentors who were highly influential and istrators, were more skeptical, aloof, self-
supportive; and, although they typically performed sufficient, emotionally unstable, bohemian, and
well on aptitude and intelligence tests, their success radical. They were also significantly more assert-
was more a function of hard work and dedication ive, competitive, suspicious or jealous, and with-
than sheer brilliance. Roe (1952) noted that a high drawn. The scientists as a whole were significantly
proportion (39 out of 64) of the scientists were first- more skeptical, aloof, self-sufficient, intellectually
born. In her discussion of birth order, Roe stated, adaptable, cautious, withdrawn, and radical than
"It seems probable that all this may point to the the undergraduate student group and significantly
most important single factor in the making of a more competitive, assertive, adventurous, sen-
scientist-the need and ability to develop personal sitive, tolerant, intellectually adaptable, self-confi-
independence to a high degree" (p. 25). Since first- dent, emotionally mature, self-indulgent, frivo-
born children tend to be more dependent than later- lous, and controlled or compulsive than the general
born children, Roe suggested that the scientist's population.
pronounced independence may be in reaction to These results generally agree with the Drevdahl
overdependence. (1954) and Wright (1954) studies. The results of
The influential work of Roe and of Terman these studies are further supported by a subjective
(1926) was followed, and in good measure sup- assessment of the biographies of 100 distinguished
ported, by the extensive research efforts of Ray- researchers (mostly U.S. and British), in which
mond Cattell and his colleagues using the 16 PF. Cattell (1966) concluded that these researchers
Drevdahl (1954) compared "creative" and "non- would be more skeptical, aloof, intellectually
creative" students on the 16 PF (Cattell & Stice, adaptable, self-sufficient, competitive, assertive,
1957), finding that creative students were signifi- suspicious or jealous, and disciplined or conscien-
cantly more skeptical, aloof, critical, withdrawn, tious than less-eminent members of the same disci-
disciplined, insightful, cautious, radical (experi- plines. Cattell and Butcher (1968) stated that emi-
menting, analytic, free-thinking), and resourceful nent scientists are skeptical, aloof, emotionally
unstable, self-sufficient, and bohemian. The emo- more skeptical, detached, and withdrawn than con-
tional instability of the scientist derives from a high trols judged to be less creative or low achievers.
I factor (emotional sensitivity) and not a low C There is a tendency for creative subjects to be more
factor (ego weakness). Cattell and Butcher (1968) radical and unconventional than noncreative sub-
also cited a study by Tollefson (1961) that provided jects. Findings regarding emotional stability, ego
an equation using "beta weights" that represented strength, and anxiety are more equivocal.
correlations between primary factors on the 16 PF A similar approach to the 16 PF research deserv-
and ratings of research contribution. The equation ing mention is the well-known work of the late
was based on a sample of 53 Ph.D. chemists who Donald MacKinnon (1960, 1962a,b, 1965). Mac-
were employed in the research department of a large Kinnon used the CPI and the MMPI to examine
oil company. groups of architects rated by professionals in the
field as being highly creative. MacKinnon (1960)
Research performance = 0.258 + 0.46C + 0.32£ reported that his creative architects were more in-
- 0.461 + 0.33N + 0.45QI + 0.29Q2 - 0.35 terested in symbolic ideas than in concrete ideas and
in implications of detail rather than the detail itself.
where, B = general intelligence, C = ego strength, Even though the architects scored high on the psy-
E = dominance, I = emotional sensitivity, N = chopathic-deviate (Pd) and schizophrenia (Sc)
shrewdness, QI = radicalism, Q2 = self-sufficien- scales of the MMPI, this was interpreted as indicat-
cy, and 0.35 = a constant. ing greater unusualness of thought process, less
Stein (1971) also reported findings on a sample inhibition, and freer expression of impulse and im-
of chemists. He found that the creative chemists agery (MacKinnon, 1962a). Another prominent
were more autonomous, integrative, more "ori- feature of the MMPI profiles was a very high peak
ented to achievement and acceptance of their own on the femininity (Mf) scale.
inner impulses," "placed more emphasis on har- In brief, MacKinnon's research also supports the
mony and form and less on mystical values," and work of Roe and Terman. He found that creative
"gave more evidence of psychological well- scientists are introverted, nonconforming, autono-
being" than a comparison group of less-creative mous, yet adaptive, striving and assertive, open-
colleagues. minded or flexible in working style, and intuitive
Thus far we have spoken primarily of research (MacKinnon, 1967, 1972).
scientists and college students. At least one study The extensive research to come out of Berkeley's
clearly demonstrated that 16 PF profiles of 153 Institute of Personality Assessment and Research
writers of fiction ("imaginative literature") bore a (IPAR) cannot be done justice here, but it would be
striking similarity to profiles of creative scientists inexcusable to omit mention of that massive report
(Drevdahl & Cattell, 1958). A similar profile was prepared by MacKinnon, Crutchfield, Barron,
also found of artists selected from Who's Who in Block, Gough, and Harris in 1958 for the Air Force.
American Art. Drevdahl (1956) reported the same The essence of that report was translated into book
finding when looking at liberal arts graduate stu- form by Barron in 1963. Close to 200 personality
dents who were chosen as highly creative. In a 1966 assessment variables were related to an overall
review chapter, Cattell wrote, "originality composite" score. The score on the
Concept Mastery Test was used to partial out the
For the present I am inclined to see the general evidence as effect on intelligence. The full sample of 343 of-
agreeing with Terman that emotional stability may be low for
literary geniuses, but that the average level of ego strength and ficers (all captains) was examined according to an
emotional stability is distinctly higher for effective scientific abbreviated originality composite (four of the eight
researchers than for the general popUlation. (p. 122) measures) and the Concept Mastery Test score.
Two groups were contrasted-one with scores fall-
This is not surprising if we recall that the origin of ing one standard deviation above the mean on origi-
the Cattell and Butcher (1968) finding of instability nality and one sigma (standard deviation) below the
in scientists was emotional sensitivity (high I). mean on intelligence (N = 15) and a comparison
The consistent findings from the 16 PF research group with scores falling one sigma below the mean
are that scientists are more skeptical, detached, crit- on originality and one sigma above the mean on
ical, intellectually adaptable, and self-sufficient intelligence. The former ("original") group was
than their comparison groups and that students or characterized by the following adjectives: affected,
artists judged to be creative or high achievers are aggressive, demanding, dependent, dominant,
forceful, impatient, initiative, outspoken, sar- It may be of limited value to demonstrate the
castic, strong, and suggestible; the latter ("intel- presence of eccentricity in creative people. First of
ligent") group was characterized as mild, op- all, we have no reasonable way of ascertaining the
timistic, pleasant, quiet, and unselfish. amount of comparable eccentricity found in the
Another psychometric contribution to the assess- general population. Second, eccentricity does not
ment of the creative personality is a 30-item scale necessarily provide even prima facie evidence for
devised by Gough (1979). The Adjective Check the presence of major mental illness. Eccentricity
List (Gough & Heilbrun, 1965) was administered to denotes a peCUliarity of behavior, and any behavior
12 separate samples comprising 1,701 subjects. that deviates sufficiently from the norm or from
Based on an item analysis of the 300-item Adjective convention may be viewed as eccentric. In the
Check List, 18 positive and 12 negative items were above examples, Einstein pointed to introversion,
chosen (positive: capable, clever, confident, ego- Scholander to irresponsibility, and Ogilvy to la-
tistical, humorous, individualistic, informal, in- bility of affect. Perhaps the one least-common de-
sightful, intelligent, wide interests, inventive, orig- nominator, which would be of interest, is the find-
inal, reflective, resourceful, self-confident, sexy, ing that eccentricity could be operationally defined
snobbish, and unconventional; negative: affected, in terms of deviant mood states.
cautious, commonplace, conservative, conven- Although it is most challenging to speculate as to
tional, dissatisfied, honest, narrow interests, man- what personality traits reliably tum up among high-
nerly, sincere, submissive, and suspicious). The ly creative individuals, a note of caution is worth
resulting Creative Personality Scale was then relat- mentioning. As Taylor (1960) pointed out, the rela-
ed to six measures of creativity as well as criterion tionship between a personality trait and creativity
evaluations. This new scale correlated positively may be dependent upon other traits, perhaps the
and significantly with all six creativity measures but gestalt of traits, that make up the individual. Fur-
surpassed them with respect to validity. thermore, the trait-creativity relationship may be
For a comprehensive discussion of research on task- or situation-specific. In other words, we
personality correlates of creativity as of 1980, the should keep in mind that the interaction of creativity
reader is directed to Barron and Harrington's (1981) with critical environmental factors is at least as
excellent review article as well as the relevant important-maybe more important-than the
contributions to this volume for more recent "main effect" of creativity.
work. From my own perspective, the single most
Albert Einstein's remark that scientists are important methodological problem in research on
"rather odd, uncommunicative, solitary fellows" creativity concerns the criterion itself. As Hudson
typifies one common stereotype of the highly cre- (1970) noted, creativity is now so much a part of our
ative individual-eccentricity. In the writings of jargon that it has come to refer to everything from
Cattell and his colleagues, "bohemian tendency" performance on a psychological test to one's rela-
comes to mind; in the writings of MacKinnon, tionship with one's spouse. The criterion problem
"nonconforming" and "autonomous." It is easy was discussed in detail by Shapiro (1970), who
to collect anecdotes and personal statements that spoke of "ultimate" and "concurrent" criteria of
seem to support such an idea. Scholander (1978), a creative activity. An ultimate criterion is based on
noted physiologist, remarked: "Like a vulgar an individual's total output over a lifetime. This
streaker I shall expose myself in the limelight. I obviously can only be assessed retrospectively or
frequently ask myself how has it been possible to longitudinally in the prospective style of Terman's
live in such an exciting and happy life being such an study. In either case, it does not permit a current
irregular and irresponsible person?" (p. 1). In an assessment of ongoing activity. Thus, concurrent
entirely different context, Ogilvy (1971) com- validation became a favored approach. Shapiro dis-
mented: cussed three types of concurrent methodology. The
first type is exemplified by Guilford's research
Few of the great creators have bland personalities. They are (e.g., 1950, 1959, 1966, 1968). Guilford assumed
cantankerous egotists, the kind of men who are unwelcome in the that the factor analysis of his tests provided con-
modem corporation. Consider Winston Churchill. He drank like struct validity. A second type uses tests to divide a
a fish. He was capricious and willful. When opposed, he sulked.
He was rude to fools. He was wildly extravagant. He wept on the sample into "high" and "low" creatives and then
slightest provocation. His conversation was Rabelaisian. He contrasts those subgroups on personality tests. If
bullied his subordinates. (p. 208) the subgroups separated out as predicted, one could
assume that the original tests used for dichotomiza- concerned fluency of thinking, which was broken
tion (usually of an aptitude nature) were valid. The down into "word fluency" (the ability to produce
third type employs a battery of instruments and cor- words comprising a certain letter or set of letters),
relates the resulting scores with some external "associational fluency" (the ability to generate
criterion. All three methods lack one essential in- words of similar meaning), "expressional fluen-
gredient-a truly valid external criterion. Disparate cy" (the ability to generate phrases or sentences),
groups fall under the experimental umbrella of and "ideational fluency" (the ability to generate
creativity, for example, high scorers on a multitude ideas to fulfill specified requirements in a given
of different instruments, peer nominees, Nobellau- time). Another primary trait concerned flexibility
reates, and subjects chosen through "consensual of thinking. This was divided into "spontaneous
validation" (i.e., those "assumed" to be creative) flexibility" (the ability to generate a wide variety of
and subjects chosen as a result of some arbitrarily ideas rapidly and without perseveration) and
established level of accomplishment (i.e., number "adaptive flexibility" (the ability to generate un-
of journal articles or books published and contest usual solutions for problems). Guilford also pro-
awards). Studies have examined preschoolers, col- posed, as primary traits, "originality" (the ability
lege students, subjects at the apex of their careers, to perceive remote associations, generate responses
and subjects at the twilight of their careers. Surely, rated as "clever," or produce responses of low
the creative activity of a preschooler does not fall in frequency in the population), "redefinition" (the
the same category as the creative activity of a fifty- ability to reconceptualize a familiar interpretation
year-old seasoned veteran, just as the creative per- and apply it in a unique situation), and "elabora-
formance of a college sophomore does not fall in the tion" (the ability to extend a simple design to a
same category as a Nobel laureate. more complex or intricate design). The reader may
Creativity is so complex, multifaceted, and het- well be correct in surmising that there is little dif-
erogeneous that it has all but lost any utility as an ference between originality and adaptive flexibil-
empirical construct. It would seem, therefore, that ity. Guilford commented that originality may be a
there is an urgent need for empirically derived tax- case of adaptive flexibility when verbally mean-
onomic differentiation, wherein we can begin to ingful material is involved, whereas nonverbal in-
establish the fundamental attributes of the various put is more in keeping with the intention of adaptive
products or processes we label as creative. All areas flexibility.
of psychological science have, as Melton (1964) Guilford identified three factors that were based
noted, a taxonomic burden; however, the impor- upon factor analyses of inventories assessing
tance of taxonomic differentiation is maximal when "thinking interests": tolerance of ambiguity (will-
the area of inquiry is diverse and amorphous, as is ingness to accept uncertainty and avoidance of
certainly the case with creativity. rigidity), convergent thinking (thinking through to
one correct answer), and divergent thinking (a
search that uncovers several answers). The general
line of research that followed pursued the notion
Research on Cognitive Aspects of
that the original person should be one who is toler-
Creativity ant of ambiguity, flexible (as opposed to rigid), and
divergent in thinking. In his 1956 review article,
Just as the preceding section on personality traits Guilford (1956b) concluded that convergent and di-
inquires as to the temperamental or motivational vergent thinking are complementary aspects of in-
aspects of creativity, this section examines the cog- tellectual ability, that the two can be easily dis-
nitive and intellectual features of creativity. Un- tinguished factorially, and that convergent thinking
doubtedly, the single most important impetus in this comprised the bulk of most general intelligence
area was provided by Guilford (1950, 1956a,b, tests.
1959, 1966, 1968, 1970). Thirty-seven years have The next step in the development of Guilford's
elapsed since Guilford (1950) first presented his theory of intellect was reported in 1955 (see
hypotheses regarding creative thinking. The fecun- Guilford, 1966). He grouped 42 factors of intellect
dity of those hypotheses may be measured in more into six categories, according to the type of thinking
than two decades of fruitful research. involved. By 1957, the number of intellectual fac-
Guilford's mission was to identify primary cog- tors had increased to 46 and the number of major
nitive traits related to creativity. One primary trait categories were reduced to five (cognition
[discovery of information], memory, divergent gence (Pezzullo, Thorsen, & Madaus, 1972; Van-
thinking, convergent thinking, and evaluation denberg, 1968), thus lending support to the influ-
[testing of information]). These five mental ac- ence of social learning.
tivities or "operations" form one dimension of a A study that compared Guilford's theory with
three-dimensional cubical model that Guilford Mednick's (1962) associative theory of creativity
called the "structure of intellect." A second di- found striking support for the Guilford position
mension consists of six "products" into which in- (Korb & Frankiewicz, 1979). In brief, Mednick
formation may be converted after processing, and a argued that one cognitive process was responsible
third dimension consists of four broad types of input for creativity, whereas Korb and Frankiewicz found
that the individual must discriminate between (fig- two significant discriminant functions with differ-
ural, symbolic, semantic, and behavioral). The ent variates on each. The investigators looked at
cube consists of 120 (5 x 6 x 4) unique cells, each 146 college students (juniors or above) who were
representing an independent feature of intellectual majoring in art, creative writing, music, or mathe-
ability. For present purposes, this complex model matics. The judges (there were 12) determined that
will serve only a limited function. A rather small 77 of the students were creative, whereas the re-
slice of the cube represents creative thinking. Ac- mainder of the students constituted the noncreative
cording to the three dimensions, that slice would be comparison group. The creative writers were
the divergent-thinking operation cut across most "original in a semantic context, fluent and flexible
products and including three of the contents (figur- in both the semantic and figural content areas,
al, symbolic, and semantic). This means that, at imaginative, emotionally undisciplined, self-indul-
most, 18 of the 120 cells are involved with creative gent, interested in reflective thinking and willing to
thinking. However, that small slice represents only disregard rules" (p. 20). The musicians turned out
"classic" creativity (artists, composers, and writ- to be very similar to the writers. The creative artists,
ers). Guilford (1956a,b, 1966, 1968) has spelled however, lacked "fluency and flexibility in both
out the range of creative-thinking abilities found in the semantic and figural content areas, while being
the structure-of-intellect model. The aforemen- conscientious, controlled and disinterested in re-
tioned classic type provides fluency of verbal ex- flective thinking" (p. 20). The mathematicians
pression, spontaneous flexibility, adaptive flexibil- were similar to the artists, though the authors indi-
ity, and elaboration. Two other types define cated that the mathematicians did possess an in-
scientific or mathematical creativity: convergent terest in reflective thinking.
thinking, cutting across figural, symbolic, or se- In sum, convergent- and divergent-thinking
mantic transformation and yielding "flexibility of styles blend together at lower levels of ability and
closure" or redefinition, and cognition, cutting separate at higher levels of ability. Creativity may
across figural, symbolic, or semantic implications derive from either thinking style, and it may do so
and yielding "sensitivity to problems." with little or no regard to profession (i.e., the diver-
One interesting note on the relationship between gence of artists and the convergence of scientists).
convergent and divergent thinking was provided by At the outer reaches of sophisticated problem solv-
Haddon and Lytton (1968). These investigators re- ing and artistic creation, the mode of endeavor is
ported findings that "strongly confirm" the com- probably an individual characteristic rather than a
plementary relationship between convergence and professional one. As Nicholls (1972) pointed out,
divergence, as well as the association between inde- the lack of association between tests of divergent
pendence and high ability level; that is, con- thinking and high levels of creativity may well be
vergence and divergence more or less overlap at the failure of the test (i.e., to take convergence into
lower ability levels and become increasingly inde- account).
pendent at higher ability levels. Additionally, the Similarities in cognitive styles between normals
authors found strong evidence in support of their and psychotics have been discussed in detail by
hypothesis that informal schools develop qualities McConaghy (1960, 1961; McConaghy & Clancy,
of personality that facilitate divergent thinking. The 1968). McConaghy (1960) distinguished two
tight rein, firm governance, and rigid schedule of modes of thinking that reflect predispositions to
the formal school were more conducive to develop- psychosis. In one case, filtering mechanisms are
ment of convergent thinking. The sparse literature impaired, thereby permitting the intrusion of irrele-
on the heritability of divergent thinking provides vant associations. Thought processes are vague,
little evidence to support the genetic basis of diver- dominated by intuition. This disorder is said to af-
fect about 10% of the population and carries a warn- nitive styles characterizing creative thought that
ing sign of schizophrenic vulnerability. In the other Cattell and his colleagues sought to define and the
case, the capacity for making logical attributions is psychotic loosening (or tightening) of ideational
enhanced. Thus, once a conclusion is reached, it is boundaries that McConaghy described. In the latter
adhered to with greater-than-normal devotion. In case (McConaghy), the disorder may be called
this instance logic has been said to triumph over over- or underinclusion, and in the former case
common sense, a condition that may predispose to (Cattell), the gift is called creativity. The apparent
paranoia. difference between the divergent thinking, loose
McConaghy (1961) identified the former mode associations, and irrelevant themes of psychotics
of thinking as "allusive" (considered to be less and the amazing conceptual leaps, cognitive flexi-
pejorative than "loose") and proposed that it may bility, and serendipitous discoveries of creative art-
be observed in a wide variety of people, many of ists and scientists is one of control. Psychotic think-
whom never become schizophrenic. Allusive thinking is unbridled and capricious, whereas creative
ing is characterized by imprecise (overabstract) and thinking is rationally directed and purposeful. In
inappropriate speech (similar to overinclusion). sum, it appears that there is reason to suspect a
McConaghy (1960) found that normal college stu- continuity between normal and pathological think-
dents with allusive thinking had less difficulty ing, a point convincingly argued by Claridge
learning the central portion of a word list than stu- (1972). Thought disorder is not "present or ab-
dents not showing allusive thinking. Because cen- sent" but measured as a single dimension with nu-
tral nervous system (CNS) inhibition impedes serial merous shades of gray. Thus, typical modes ofpsy-
learning, it was concluded to be weak in allusive chotic thinking are also found in the general
cases. Normally, the inhibitory process produces (normal) popUlation, only in less severe or debilitat-
attentional constriction. When inhibition is weak ing fashion. The existence of such cognitive styles
(like blinders removed from a horse), focusing or may facilitate creativity as well as reflect a genetic
"tunnel" thinking is impaired. An intrusion of pe- predisposition to psychosis. Claridge (1972) com-
ripheral and irrelevant information produces over- mented that the traits on which creative people dis-
abstraction. McConaghy and Clancy (1968) admin- tinguish themselves are similar to those traits which
istered a test for allusive thinking to university possess an increased loading on personality profiles
students and their parents, finding that significantly associated with schizophrenia.
more students with high scores (presence of allusive There is sparse, though interesting, evidence that
thinking) had a parent with a high score than stu- associates the Claridge research with the Guilford
dents with low scores. It was concluded that high research. Claridge, Canter, and Hume (1973)
test scores among family members, as evidenced in looked at personality variation in monozygotic and
samples of both normals (1968) and parents of dizygotic twins, using an extensive battery of psy-
schizophrenics (1960), supported the hypothesis chophysiological tests. The investigators defined a
that the test is measuring allusive thinking in both personality dimension of "psychoticism," based
groups. The latter finding was later replicated by on a principal-components analysis of the psycho-
Lidz, Wild, Schafer, Rosman, and Fleck (1963) physiology data. They discovered, among other
and Rosman, Wild, Ricci, Fleck, and Lidz (1964). things, that psychoticism was related to scores on
McConaghy's conceptualizations are based, in tests of Guilford's divergent thinking. Farmer
good measure, upon the work of Pavlov (1955), (1974) also reported that the psychoticism score on
who also distinguished two personality types. The Eysenck's Personality Questionnaire (EPQ) was re-
artistic type, prone to hysterical or manic-de- liably associated with the originality score on one of
pressive reactions, is intense, vivid, and highly re- Guilford's tests (cited in Woody & Claridge, 1977).
sponsive to external stimulation. The thinking type, Woody and Claridge (1977) administered the EPQ
prone to obsessive-compulsive and schizophrenic and tests of divergent/convergent thinking to uni-
reactions, is quiet, ruminative, and hypersensitive versity students. The association of divergent think-
to excessive stimulation. It was Pavlov's early writing with psychoticism (on the EPQ) was strongly
ing that led McConaghy (1961), Claridge (1972), confirmed, though the hypothesis of an inverse re-
and others to attribute allusive thinking to weak lationship between psychoticism and convergent
levels of CNS inhibition resulting from low cortical thinking was rejected. As Woody and Claridge
excitability. pointed out, these clear findings may be interpreted
There is marked resemblance between the cog- two ways. If indeed the psychoticism scale on the
EPQ is valid, the results support the association of of the creative students enabled them to opt for a
disposition to psychosis with divergent thinking. more convergent strategy when that was deemed
The authors go one step further and suggest that the more efficient. The authors concluded with an
results relate psychosis to creativity, though it is far important comment. They noted that narrow atten-
from certain that divergent thinking is equatable tional focus on key elements of a problem might
with creativity. Another way of looking at the same facilitate problem solving; however, the strategy of
results would be to interpret the data as providing encompassing and assimilating a wide range of in-
validity for the psychoticism scale. Logically, a formation is more advantageous when it comes to
heritable predisposition to psychosis should have as "original" thinking. Again, it is apparent that the
a primary feature some distinctive cognitive style. difference, at least on a superficial level, between
According to the Woody and Claridge study, high- the creative students and the schizophrenics is one
psychoticism students showed marked divergent of voluntary control over processing of input. An-
thinking, though none of the predicted cognitive other interesting point emerged from the Dykes and
slowness that is characteristic of many schizo- McGhie study. Presumably the schizophrenics, all
phrenics. inpatients, were medicated, though this fact is not
Most of the foregoing research is predicated upon stated. In the event of medication, one would ex-
two fundamental assumptions. To reiterate: (1) pect "truncated" performance. Free of medica-
there is a genetic predisposition to psychosis that tion, schizophrenic performance might be consider-
manifests itself as a nonpsychotic schizoidlike per- ably more extreme, allowing one to envision a more
sonality in the normal popUlation; and (2) this schiz- Gaussian distribution, with schizophrenics not
oid profile is, in part, characterized by a thinking using the "same" strategy as normals but em-
style similar to that found in psychosis. Because ploying an extreme form of that strategy.
highly creative individuals may display thinking Cropley and Sikand (1973) looked at the ways in
styles similar to psychotics, it has been theorized which the information processing of creatives and
that such individuals possess a predisposition to schizophrenics differed from normals. Five aspects
psychosis. This might explain why creative people of information processing (general complexity,
are frequently perceived as "eccentric," like fish flexible complexity, flexibility, differentiation, and
out of place in the normal ocean. This may also general incongruity adaptation level) were exam-
explain why so many highly distinguished indi- ined in 80 adults (40 controls, 20 schizophrenics,
viduals have been the apparent victims of some and 20 creatives). The creative people were in the
form of mental illness. arts (as opposed to the sciences) and were so de-
The second assumption-that similar cognitive, fined according to Eiduson's (1958) criteria-ac-
attentional, or information-processing strategies tive participation in some field acknowledged as
may be found in schizophrenics and creatives-has creative and peer recognition. The investigators
been examined in several investigations. Dykes and found that there were useful and reliable ways to
McGhie (1976) predicted that similarities between differentiate the chronic schizophrenics from the
creativity and psychosis were attributable to com- creatives. The creatives were superior to schizo-
mon attentional (overinclusive) strategies. Three phrenics and normals in tolerance for incongruity in
groups were contrasted: highly creative students, information input. On the other hand, schizo-
uncreative (equally intelligent) students, and acute phrenics differed from creatives and normals in
nonparanoid schizophrenics. The measure of complexity, flexibility, and differentiation. It was
creativity was Cropley's (1968) paper-and-pencil speculated that "creative thinkers can tolerate
test. According to two object-sorting tasks and a strain (drive for incongruity reduction) and con-
dichotic-shadowing task, similar strategies were tinue to employ broad categorizing strategies in
employed by the schizophrenics and the high- concept formation, whereas non-creatives cannot"
creativity students. Both groups took in a wider (p. 467).
range of input than the less-creative students. The Over the past several decades, there has been a
broad attentional focus of the schizophrenic seemed strong attraction to the concept of divergent think-
to be involuntary and unharnessed, resulting in im- ing as an operational construct for defining
paired performance, whereas the creative indi- creativity. It is a clean, simple concept and can be
viduals possessed a cognitive flexibility that per- measured easily with a variety of different tests.
mitted effective processing of stimulus overload Additionally, anecdotal reports seem to suggest that
without, as it were, blowing a fuse. The flexibility something like divergence captures the "essence"
of creativity. This enthusiasm survives despite Greenough, 1973; Martindale, Hines, Mitchell, &
Guilford's own note of caution, as well as the re- Covello, 1984). Martindale accumulated evidence
search, indicating that divergent thinking does not that relates creativity (as measured in college stu-
correlate highly with creativity (e.g., Nicholls, dents using the Remote Associates Test and the
1972). There are at least three possible explana- Alternate Uses Test) to low levels of cortical activa-
tions: (1) divergent thinking is only one of several tion, variability in level of activation, and disinhibi-
independent cognitive styles that facilitate cretion. The conclusion drawn by Martindale and
ativity; (2) divergent thinking interacts with another Hines (1975) associating disinhibition with
style, such as convergent thinking; and (3) the crite- creativity appears to be consistent with expectation.
ria used to define or measure creativity failed to Disinhibition, or "dehabituation," is the use of a
capture the intended behavior. novel stimulus to revive a habituated orientation
Hasenfus and Magaro (1976) found, in their re- reaction. It was a phenomenon ascribed by Pavlov
view of the literature, that "the schizophrenic 'def- to dissipation of inhibition. An adjunctive, albeit
icit' does in some instances equal creativity" (p. not incidental, feature of this phenomenon may be
348). These instances included ideational fluency the recultivating and reconceptualizing of past asso-
and overinclusion, as well as word association. In ciations and experiences, giving rise to fresh in-
the latter case, it was observed that both schizo- sights and novel ways of looking at old ideas.
phrenics and creatives express remote and unusual Martindale hypothesized that primary process
associations. It was also noted that remote associa- cognition is associated with either high or low
tions are more typical of chronic, nonparanoid arousal, whereas the secondary process is associ-
schizophrenics, whereas ideational fluency is more ated with medium arousal. Efficient task perfor-
characteristic of paranoid schizophrenics. Keefe mance is usually optimized at medium levels of
and Magaro (1980) administered two creativity arousal, with performance dropping off when
tests to 10 paranoid and 10 nonparanoid schizo- arousal is too low or too high. It has been assumed
phrenics, 10 nonpsychotic psychiatric controls, and in the past that, as task difficulty increases, the
10 normal subjects. The nonparanoid schizo- optimal level of arousal decreases. This would sug-
phrenics were significantly more creative than the gest that low arousal favors creativity. Such a con-
paranoid schizophrenics and the psychiatric conclusion assumes that creativity has something to do
trols on the Alternate Uses Test (AUT). In addition, with task complexity. It makes more sense to sup-
the nonparanoid schizophrenics produced signifi- pose that, in cases where "need for solitude" is a
cantly more "highly creative" responses than the prerequisite for creative inspiration, low arousal is
normal subjects. No group differences relating to facilitatory. As Martindale (1977-1978) pointed
creativity were found on the Revised Art Scale out, even the often-noted listlessness of creative
(RAS). Thus, ideational fluency (as measured by individuals can be attributed to either high or low
the AUT) and not perceptual complexity (as arousal. These findings were integrated by Martin-
measured by the RAS) provided support for the dale in the following remarks:
Hasenfus and Magaro hypothesis that the empirical
constructs of schizophrenia and creativity are It is hypothesized that highly creative individuals have a high
equivalent. Keefe and Magaro were appropriately resting level of activation and are thus overreactive to weak
stimuli. Highly creative individuals also have a low threshold for
prudent in their discussion of the results, pointing protective inhibition so that slightly stronger stimuli elicit a para-
out that an equivalence of cognitive processes does doxical reaction; that is, further stimulation elicits decreases in
not imply an equivalence of mental status. In other arousal. (1977-1978, p. 80)
words, the Keefe and Magaro findings suggest that
certain types of schizophrenics exhibit cognitive In a more recent report, Martindale et at. (1984)
processes that are similar to the ones employed by again hypothesized that creativity involved the use
creative people. The study does not provide evi- of primary process cognition which would be ac-
dence that creative people are schizophrenic. companied by right-hemisphere activation. Three
A most notable contribution to our understanding studies supported the hypothesis that highly cre-
of the relationship between creativity and informa- ative students would exhibit greater right-hemi-
tion processing comes from the research of Martin- sphere than left-hemisphere EEG activity during
dale (1977-1978) and with his colleagues (Martin- creative activity. It is not possible in this chapter to
dale & Hasenfus, 1978; Martindale & Hines, 1975; do justice to Martindale's research; however, the
Martindale & Armstrong, 1974; Martindale & findings from his studies draw support from many
quarters, including the literatures on evoked re- acter" (p. 68). This conclusion illustrates a "tran-
sponses, lateralization and hemispheric deficit, and scendentalism" that requires clarification. Person-
input regulation. ality traits may be confounded with styles of
A different conception of cognitive process in thinking. For instance, personality traits, such as
creativity was provided by Rothenberg (1976, introspective, ruminative, analytical, and cerebral,
1979, 1983). Two processes operative in creativity also suggest a clear style of thinking. The style may
were defined as 1anusian and homospatial thinking. be deviant enough to be noteworthy, though ob-
The 1anusian process "consists of actively conceiv- viously not deviant enough to be psychotic. Taken
ing two or more opposite, contradictory, or anti- to an extreme, that same style might be labeled
thetical concepts, images, or ideas simul- underinclusive thought disorder. Logically, where
taneously" (1976, p. 18). The term derives from deviance in thinking has not approached the thresh-
the Roman god 1anus, who looked in two directions old of psychopathology, one would not expect to
at the same time. The homospatial ("same space") find a relationship between creativity and mental
process "consists of actively conceiving two or disorder. Many of the personality and cognitive fea-
more discrete entities occupying the same space, a tures that Dellas and Gaier found associated with
conception leading to the articulation of new identi- creativity may be defined alternatively as modes of
ties" (1976, p. 18). Thus, the homospatial process processing information (such as openness to inter-
integrates 1anusian thoughts. In the former case, nal and external stimuli, ideational fluency, dis-
logically unrelated entities are conceived of at the criminative observation, and intuitiveness). In fact,
same point in time, and, in the latter case, those the authors found that the data support the conclu-
discrete entities are "superimposed, fused, or oth- sion that the personality traits of creatives "develop
erwise brought together in the mind and totally fill fairly early. " This would seem to suggest that these
its perceptual space" (p. 18). Thus, the initial and traits, whether they relate to personality or cogni-
highly distinctive characteristic of creative thought tion, have an endogenous component.
process is the simultaneous conception of two or
more antithetical concepts. For this very reason, the
process must transcend logic, because the inte- Concluding Thoughts and Speculations
grated elements are contradictory. The homospatial
process is the creation of harmonious coexistence
out of opposing and disharmonious ideas. A Metatheoretical Bridge
The putative relationship of these thought pro- The intent of this chapter has been to explore the
cesses to psychosis is an interesting one. Homo- hypothetical relationship between extraordinary
spatial thinking is a process of sophisticated spatial creativity and mental illness. At this point, we may
abstraction; it is nonconcrete. At the same time, fairly conclude that creativity and major mental ill-
such thinking does not violate the laws of logic. In ness (i.e., psychosis) are not causally related. The
fact, Rothenberg calls it "translogical." Psychotic crucial question, however, is not whether creativity
thinking may be characterized by the illogical asso- and mental illness are causally linked, but whether
ciation of unlike entities. Two points are worth not- our understanding of mental illness offers any in-
ing. First, creativity often transcends logic, placing sight into the creative process. The question is in-
emphasis on intuition or speculation. Second, the spired by the seemingly large number of highly
illogic of psychotic thinking may represent decom- creative individuals who have been sufficiently dis-
pensation from a formerly high level of abstraction turbed to come to the attention of historians. Out of
to a translogical plane devoid of logic. Support for the total pool of, for instance, highly creative writ-
the association between 1anusian thinking and ers, if the number who were demonstrably dis-
creativity and the dissimilarity between the perfor- turbed exceeds the base rate for that disturbance in
mance of 12 Nobel laureate scientists and 18 men- the general population, then the (possibly) coinci-
tally ill patients was provided in a more recent study dental association between creativity and distur-
by Rothenberg (1983). bance in those cases invites plausible explanation.
A critical review of the creativity literature by I have sought to provide (it is hoped) a plausible
Dellas and Gaier (1970) concluded with the state- explanation in the form of a theoretical model based
ment that' 'the roots of creativity do not seem to lie upon a unidimensional, neurocognitive view of
in convergent or divergent thinking, but rather in thought process (Prentky, 1979; 1980). I proposed
the personality and motivational aspects of char- that there is an input (information processing) con-
A TYPE C TYPE
Ambicognitive thought process
,,
I
Extra- Intro-
I
I
I
A+ : cognitive cognitive : C+
thought : thought thought I thought
disorder : process I process : disorder
i
'PSYCHOTIC: NONPSYCHOTIC:
Labile affect, : Characterological "Normal" range
I
I
'
NONPSYCHOTIC:: 'PSYCHOTIC:
: Characterological I Flat affect,
Depression, I I : Withdrawal,
Agitation : : : Apathy
Figure 1. A neurocognitive, unidimensional model for the distribution of thought process in the general population; *illustrative
clinical descriptors.
tinuum ranging from extreme constriction to ex- I further proposed that the hypothetical optimum
treme expansion, and that the tendency toward one deviation toward constriction or expansion of sen-
mode or the other is normally distributed in the sory input defined two distinct profiles, each of
population (see Figure 1). The overlap between which possessed identifiable cognitive, physiologi-
creativity and mental illness blends somewhere cal, and clinical characteristics. Each profile was
along this continuum. Excess deviation toward ei- intended to capture a unique constellation of char-
ther extreme results in thought disorder, whereas acteristics that, when taken together, promoted
some hypothetical "optimum deviation" promotes creativity. One profile was labeled A-type
creativity without risk of major mental illness. (abstract). A-type is characterized by extensive
When this optimum deviation slides farther out into scanning that often incorporates much peripheral,
one of the tails of the distribution, normality slips extraneous information, erratic mental "thresh-
into abnormality. Because creative people "live" ing" of large amounts of information, and a hyper-
farther from the hypothetical norm and, hence, are alertness that facilitates the whole process. Such
closer to the fringes of deviation, there may be a individuals, possessing "strong nervous systems"
higher incidence of psychopathology among this (e.g., Gray, 1967) that tend to resist (reduce or
group. This increased incidence of psycho- dampen) stimulation, habituate to stimuli so rapidly
pathology would not be manifested as a major men- that they often pay little or no attention to prior
tal illness but rather as a variety of subclinical be- stimuli. In a task situation, one would expect rapid
havioral "eccentricities." Such eccentricities, performance with many errors. It was predicted that
depending on their nature, may reflect a genetic the A-type would evidence a weak attentional
predisposition to major mental illness; however, focus, high distractibility, low input registration,
such a contention remains in the realm of specula- easy attentional shift, and relatively poor (diffuse
tion. The notion that disordered thinking falls along and broad) selection. Clinically, affective symp-
a continuum with normal thinking and that tomatology should predominate, including lability
"thought disorders" are not discrete entities bear- and depression. A-type creativity is characterized
ing no relation to orderly thinking is certainly not by the accumulation of the widest possible array of
new (Harrow & Quinlan, 1977). information. This smorgasbord of data is then har-
vested, with a keen eye to uncovering previously begins to intrude, creativity typically recedes into
unobserved relations. If the search is fruitful, and the background. In fact, the stress occasionally as-
subsequent research (or work) bears out the validity sociated with creativity may help to precipitate
of the observation (or contribution), the discovery breakdown. Kuhn (1970) observed that "some men
(or product) is often relegated to the mystical realm have undoubtedly been driven to desert science be-
of intuition because no systematic program of re- cause of their inability to tolerate crisis" (pp. 78-
search (or work) immediately preceded it. Robin- 79). Finally, it should be emphasized that a
son and Sabat (1977) metaphorically equated the "thought process continuum" remains a hypo-
nervous system to radar. In their terms, the A-type thetical notion, and that there are no agreed upon
functions in a broadband "coasting" mode, re- anchor points for such a continuum. Thus, it is not
sponding with reduced resolution to a wide variety possible to quantify "deviation" from the hypo-
of input. thetical mean of such a continuum.
The second profile was labeled C-type (con-
crete). C-type is characterized by the tendency to
Genetic and Eugenic Considerations
underinclude or constrict the attentional field. Such
individuals, possessing weak nervous systems that The following dilemma was posed at the begin-
tend to augment or magnify stimulation, habituate ning of this chapter: Why should a feature of human
to stimuli so slowly that they may appear, at the behavior that is seemingly so critical to the ad-
behavioral level, obsessive-compulsive. In a task vancement of the species be restricted to so small a
situation, one would expect slow performance with segment of the species? The tendency of the human
few errors. It was predicted that C-type would species is to resist change, both genetically and
evidence a strong attentional focus, low or normal socioculturally. To let up on the biological brakes
distractibility, high input registration, and difficult that regulate extraordinary creativity might threaten
attentional shift. Clinically, schizoidlike sympto- the delicate adaptational balance of the species.
matology should predominate, including blunted or This would tend to explain the otherwise inexplica-
flat affect and withdrawal. C-type creativity is char- ble eruption of creativity at isolated moments in
acterized by a microscopic, dissectional focus on history, each moment being separated by at least as
the separate constituent elements of a problem. The long a period of adjustment. This line of reasoning
approach to problem solving is to zero in on the implies, of course, that there is some biological or
absolutely critical relations or the unexpected but genetic underpinning to extraordinary creativity. It
meaningful anomalies. Again, the "magic" of has been argued, in this regard, that an entire epoch
intuition arises when no "logical" explanation suf- of human inspiration and talent, the golden age of
fices. I have argued that the fortuitous and accurate Athens, was attributable to genetics (Darlington,
association between formerly unrelated elements is 1969). The forces of heredity alone contrived to
no less important to C-type than to A-type. In initiate (and conclude) that remarkable period of
Robinson and Sabat's terms, the C-type functions intellectual development. Darlington may actually
under the most band-limited conditions and is thus be correct, but it is all speculation. There is no way
least variable in response to sequential input. of determining the influence of Athenian genes in-
It was hypothesized that, within the general pop- dependent of the current zeitgeist. Indeed, it has
ulation, there is an optimum regulation of input that been argued that the achievements of Periclean Ath-
provides for maximally efficient information pro- ens were attributed to "the peace of mind allowed
cessing, that insulates against psychiatric disorder, by good government ... [and] intellectual free-
and, tangentially, that reduces the likelihood of ex- dom and the consequent encouragement to entertain
traordinary creativity. As argued, there is also an new ideas, and new art forms" (Pickering, 1974, p.
optimum deviation from this normative regulation 276). In sum, evidence supporting higher rates of
of input that increases the likelihood of A-type or C- illness in the biological parents or offspring of
type cognition while still permitting adequate filter- gifted subjects may imply a diathesis for that illness
ing of stimuli. Too much deviation, in the form of and suggest nothing about the (possibly) coinciden-
extreme abstraction or concretization, may man- tal appearance of creativity.
ifest itself as psychotic thought disorder. It is More radical arguments have been made re-
important to note that I did not propose that cently, however, suggesting "that the schizo-
creativity and mental illness coexist harmoniously. phrenia homozygote may some day have the most
As mental illness (i.e., psychotic thought disorder) effective brain" (Karlsson, 1978, p. 188). Jarvik
and Chadwick (1973) noted that so maladaptive a If the relevance of the foregoing discussion on
condition as schizophrenia, with a high probability genetics has escaped the reader, I will try to recap-
of a genetic component, should have declined ture a logical progression of thought. The focus of
through natural selection unless it was, in some this chapter has been on the presumed similarities
way, adaptive. Jarvik and Deckard (1977) elabo- between creative cognition and deviant cognition
rated upon this theme, arguing that the schizoid- that reflect a predisposition to psychosis. The pre-
paranoid personality represents a selective advan- ceding discussion on genetics presented two lines of
tage. According to Jarvik, these schizoid-paranoid argument: in one case, that there is a hypothetical
people, who carry the gene or genes predisposing to genetic basis to extraordinary creativity; in the other
schizophrenia, have an increased probability of case, that there is a genetic basis to-and possibly a
contributing to the gene pool since they, more than selective advantage for-a predisposition to psy-
their less paranoid cousins, are more likely to sur- chosis. If there is a common biological thread link-
vive and ensure the survival of their offspring. The ing creativity with forms of major mental illness, it
validity of Jarvik's thesis notwithstanding, we have is argued in this chapter that such a thread may be
little evidence to suggest that highly creative people cognition. Thus, we may make a leap of faith to
are disproportionately schizoid and/or paranoid. infer that certain biologically conceived cognitive
McConaghy (1960) did describe a mode of thinking styles that are peculiar to extraordinary creativity
with a predisposition to psychosis that was charac- possess common biological ancestry with certain
terized by paranoia, and Hasenfus and Magaro cognitive styles that are associated with a pre-
(1976) reported that paranoid schizophrenics, like disposition to major mental illness. Although it is
creative people, exhibit ideational fluency. not intuitively (or empirically) obvious that there
Jarvik's thesis is one possible response to the may be a selective advantage for certain forms of
problem of how an abnormal gene maintains itself psychosis, the concept itself helps to explain an-
given the low fertility of schizophrenics. Other in- other riddle presented at the beginning of this chap-
vestigators have also sought to identify some com- ter; namely, how it could be adaptive, from an evo-
pensating selective advantage for schizophrenia. lutionary standpoint, to wed the curse of mental
Huxley, Mayr, Osmond, and Hoffer (1964) noted, illness to the gift of creativity. The outcome of a
for example, that schizophrenics are highly re- major mental illness remains a curse and cannot be
sistant to certain physical stressors, such as infec- assumed, in any respect, to be adaptive. There are
tious diseases and severe bums. Carter and Watts those who have argued, however, that normal indi-
( 1971) also observed an increased resistance to viral viduals who have the abnormal gene without pene-
infections early in life among schizophrenics. Such trance for psychosis do in fact have some selective
resistance might give individuals with the abnormal advantage. Thus, even though it makes little sense
gene a selective advantage over normal individuals. to wed the outcome of mental illness to creativity, it
There would be no apparent advantage if one was may make sense to wed a predisposition to mental
actually schizophrenic; however, normal carriers, illness (i.e., "normal" carriers of the aberrant
those in whom the gene had no penetrance for the gene) to creativity.
disease, would have the selective advantage with- Several very interesting papers offer collateral
out the dire outcome of schizophrenia (Alland, evidence to support the feasibility of the hypothesis
1967). This assumes, of course, that the gene is of concerning a genetically transmitted cognitive style
low penetrance. It would seem that there are, in (or styles) common to highly creative people and
fact, at least two plausible theories to explain how a those manifesting a predisposition to psychosis
disadvantageous genetic condition is maintained in with subclinical features. In looking at eye-tracking
the population over time despite the reduced fitness dysfunctions among 30 sets of dizygotic twins dis-
of schizophrenics (Slater, Hare, & Price, 1971). cordant for schizophrenia, 5 of the schizophrenic
One theory is Jarvik's notion that there is a selective probands with good eye tracking had a healthy co-
advantage for normal (i.e., not psychotic) carriers twin with bad eye tracking (Holzman, Kringlen,
of the gene. Another theory is that of polygenic Levy, Proctor & Haberman, 1977; Holzman,
inheritance. In this case, a selective advantage for Kringlen, Levy, & Haberman, 1980). In a study
gene carriers would be unnecessary since the "re- involving schizophrenics, manics, and their par-
sponse to natural selection against a polygenic trait ents, it was found that a number of schizophrenic
associated with lowered marriage and fertility rates probands with good eye tracking had nonschizo-
would be very slow" (Gottesman & Shields, 1972, phrenic parents (one or both) with impaired eye
p. 329). tracking (Holzman, Solomon, Levin, & Waternaux,
1984). Although the aforementioned anomalous re- movement), 10 moved to the right. Only 5 broad
lation between psychiatric status and eye-tracking categorizers were consistent enough to be classified.
performance held for 30% of the schizophrenics, it Of those 5, 4 were left movers. The data did not
was present in only 8% of the manic patients. Re- support the hypothesis for this group.
cently, Matthysse and Holzman proposed a model to In summary, if we assume that eye tracking re-
explain these seemingly anomalous findings (Mat- flects, in part, a much more complex process of input
thysse, Holzman, & Lange, 1986; Matthysse & regulation, then it is logical to posit that the same
Holzman, 1987). In their genetic latent structure latent trait responsible for diminished eye-tracking
model, the manifest traits of schizophrenia and bad performance may also be responsible for other defi-
eye tracking are expressions of an inferred latent trait cits in the visual processing of information. In addi-
that is genetically transmitted. The latenttrait, which tion, the modes of information processing conferred
is not behaviorally observable, may be expressed as by that genetically transmitted latent trait may pro-
schizophrenia or aberrant eye tracking or both. Mat- vide an increased advantage for certain types of
thysse and Holzman (1987) further suggested that creativity.
the transmission pattern of the latent trait may con-
form more to Mendelian laws and be of higher pene-
A Taxonomic Consideration
trance than either of the manifest traits taken sep-
arately. Although all the questions and issues raised here
Research on eye movement laterality and provide fertile ground for future empirical investiga-
creativity lends further support to the aforemen- tion, itis virtually impossible, at present, to draw any
tioned genetic hypotheses. Hamad (1972) examined reliable conclusions that involve the elastic notion of
"right movers" and "left movers" (those who look creativity. Jarvik's hypothesis, like many others,
to the right or left) among mathematicians. In the may be supported, or not supported, simply as a
math department, 10 professors and 24 graduate function of the way that creativity is defined. Thus, it
students were questioned about their work. Among is imperative that we begin by bringing some defini-
those who displayed 80% to 90% unilateral eye tional order to the constructs we are examining.
movements, 19 were left movers and 13 were right Given the extraordinary heterogeneity of the domain
movers. The left movers reported using more imag- of behaviors (and products) described as "cre-
ery and engaging in more artistic activities than the ative," it is my impression that the first task must be
right movers. Because it is assumed that, during the development of a preliminary taxonomic system
introspection, there is a predominance of activity in that identifies coherent, homogeneous subgroups or
the hemisphere contralateral to the direction of eye types of creative activity. The proliferation of em-
movement, the results suggested that the right hemi- pirical and clinical data on creativity within the last
sphere of the left movers was involved in imagery decade has increased the urgency for an organized
(though evidence to support greater creativity among approach to this extraordinarily complex and diverse
the left movers was quite weak). Hines and Martin- behavioral domain. Indeed, the importance of tax-
dale (1974) also found some support for the associa- onomic differentiation is maximal when the area of
tion of "left looking" with creativity. In two sepa- inquiry is diverse and amorphous, as is the case with
rate studies, left-looking males scored significantly creati vi ty .
higher on the Remote Associates Test (RAT). These On a light note, the interaction of creative cogni-
differences did not hold up in a third study with tion with psychiatrically disturbed cognition calls to
females. Huang and Byrne (1978) used the same mind an insightful portent of Captain Ahab: "My
lateral eye-movement paradigm to test the hypoth- means are sane, my motive and my object mad"
esis that "narrow categorizers, " being more analy- (Melville, 1851). Surely, it was not the harpooner's
tic in the processing of information, would use the techniques that were mad but Ahab's fanatical de-
left hemisphere and "broad categorizers," being votion to killing Moby Dick. For present purposes,
more holistic, would use the right hemisphere. First- we may wish to adopt the obverse formula: mad
year female psychology students were administered means and sane motives.
Pettigrew's C- W Scale. Those who scored above 90
(n = 18) and below 47 (n = 15) were selected as ACKNOWLEDGMENTS. The author is grateful to
broad and narrow categorizers, respectively. The Dan Carter, Raymond Knight, Ruth Rosenberg, and
results supported the hypothesis with respect to nar- Peg Salamon for critical commentary on earlier ver-
row categorizers. Of the 11 subjects classified as sions of this chapter, as well as to Alison Martino for
consistent movers (greater than 75% unilateral editing and text entry.
References creativity. British Journal of Educational Psychology, 38,

197-201.
Alland, A. (1967). Evolution and human behavior. Garden City, Cropley, A. J., & Sikand, J. S. (1973). Creativity and schizo-
NY: The Natural History Press. phrenia. Journal of Consulting and Clinical Psychology, 40,
Alston, E. F. (1972). James Barrie's "M'Connachie"-His 462-468.
"writing half." American Imago, 29, 257-277. Cross,P. ,Cattell,R. B., &Butcher,H. J. (1967). The personality
Andreasen, N. J. C. (1980). Mania and creativity. In R. H. pattern of creative artists. British Journal of Educational Psy-
Belmaker and H. M. Van Praage (Eds.), Mania: An evolving chology, 37,292-299.
concept (pp. 377-386). New York: Spectrum. Dale, P. M. (1952). Medical biographies. The ailments of thirty-
Andreasen, N. J. c., & Canter, A. (1974). The creative writer: threefamous persons. Norman: University of Oklahoma Press.
Psychiatric symptoms and family history. Comprehensive Psy- Darlington, C. D. (1969). The evolution of man and society.
chiatry, 15, 123-131. London: George Allen & Unwin.
Andreasen, N. J. C., & Powers, P. S. (1975). Creativity and Dellas, M., & Gaier, E. L. (1970). Identification of creativity:
psychosis: An examination of conceptual style. Archives of The individual. Psychological Bulletin, 73, 55-73.
General Psychiatry, 32, 70-73. Drevdahl, J. E. (1954). An exploratory study ofcreativity in terms
Babcock, W. L. (1895). On the morbid heredity and predisposi- of its relationships to various personality and intellectual fac-
tion to insanity of the man of genius. Journal of Nervous and tors. Unpublished doctoral dissertation. University of Nebras-
MentalDisease, 20, 749-769. ka, Lincoln.
Barron, F. (1958). The psychology of imagination. Scientific Drevdahl, J. E. (1956). Factors of importance for creativity.
American,199,151-166. Journal of Clinical Psychology, 12, 21-26.
Barron, F. (1963). Creativity and psychological health. Drevdahl, J. E., & Cattell, R. B. (1958). Personality and
Princeton, NJ: D. VanNostrand. creativity in artists and writers. Journal ofClinical Psychology,
Barron, F., & Harrington, D. M. (1981). Creativity, intelligence, 14, 107-111.
and personality. Annual Review of Psychology, 32, 439-476. Dykes, M., & McGhie, A. (1976). A comparative study of
Becker, G. (1978). The mad genius controversy. Beverly Hills, attentional strategies of schizophrenics and highly creative
CA:Sage. normal subjects. British Journal of Psychiatry, 128, 50-56.
Behrens, R. R. (1975). Lunatics, lovers, and poets: On madness Edel, L. (1975). The madness of art. American Journal ofPsych i-
and creativity. Journal ofCreative Behavior, 9, 228-232. atry, 132,1005-1012.
Bell,Q. (1972). Virginia Woolf: A biography (Vol. 2). New York: Eiduson, B. T. (1958): Artist and non-artist: A comparative
Harcourt Brace Jovanovich. study. Journal ofPersonality, 26, 13-28.
Broad, W. J. (1981). Priority war: Discord in pursuit of glory. Ellis, H. (1904). A study of British genius. London: Hurst &
Science, 211,465-467. Blackett.
Bronowski, J. (1965). Science and human values. New York: Ellrich, R. J. (1974). Rousseau's account of a psychological
Harper & Row. crisis. American Imago, 31,80-94.
Bychowski, G. (1973). Marcel Proust and his mother. American Farmer, E. W. (1974). Psychoticism and person-orientation as
Imago, 30,8-25. general personality characteristics of importance for different
Callaway, E., & Naghdi, S. (1982). An information processing aspects of creative thinking. Unpublished Bachelor of Science
model for schizophrenia. Archives of General Psychiatry, 39, Thesis, University of Glasgow.
339-347. Gibson, C. (1889). The characteristics of genius. A popular
Carter, M., & Watts, C. A. H. (1971). Possible biologicaladvan- essay. London: Walter Scott.
tages among schizophrenics' relatives. British Journal ofPsy- Goertzel, M. G., Goertzel, V., & Goertzel, T. G. (1978). Three
chiatry, 118, 453-460. hundred eminent personalities. San Francisco: Jossey-Bass.
Cattell, R. B. (1966). The personality and motivation of the Gottesman, I. I., and Shields, J. (1972). Schizophrenia and
researcher from measurements of contemporaries and from genetics: A twin study vantage point. New York: Academic
biography. In C. W. Taylor & F. Barron (Eds.), Scientific Press.
creativity: Its recognition and development (pp. 119-131). Gough, H. G. (1979). A creative personality scale for the adjec-
New York: Wiley. tive check list. Journal of Personality and Social Psychology,
Cattell, R. B., & Butcher, H. J. (1968). Thepredictionofachieve- 37, 1398-1405.
ment and creativity. Indianapolis: Bobbs-Merrill. Gough, H. G., & Heilbrun, A. B. (1965). The adjective checklist
Cattell, R. B., & Drevdahl, J. E. (1955). A comparison of the manual. Palo Alto, CA: Consulting Psychologists Press.
personality profile of eminent researchers with that of eminent Grant, V. W. (1968). Great abnormals. New York: Hawthorn
teachers and administrators of the general population. British Books.
Journal ofPsychology, 46, 248-261. Gray, J. A. (1967). Strength ofthe nervous system, introversion-
Cattell, R. B., & Stice, G. F. (1957). The 16PFpersonalityfactor extraversion, and conditionability and arousal. Behavior Re-
questionnaire. Champaign, IL: Institute for Personality and search and Therapy, 5,151-169.
Ability Testing. Guilford, J. P. (1950). Creativity. American Psychologist, 5,
Claridge, G. (1972). The schizophrenias as nervous types. British 444-454.
Journal ofPsychiatry, 121,1-17. Guilford, J. P. (l956a). The relation of intellectual factors to
Claridge, G., Canter, S., & Hume, W. I. (1973). Personality creative thinking in science. In C. W. Taylor (Ed.), The identi-
differences and biological variations: A study oftwins. Oxford, fication ofcreative scientific talent (pp. 69-95). Salt Lake City:
England: Pergamon Press. University of Utah Press.
Cleckley, H. (1964). TheMaskofSanity(4thed.) St. Louis: C. V. Guilford, J. P. (1956b). The structure of intellect. Psychological
Mosby. Bulletin, 53, 267-293.
Cropley, A. J. (1968). A note on the Wallach-Kogan test of Guilford, J. P. (1959). Traits of creativity. In H. H. Anderson
(Ed.), Creativity and its cultivation (pp. 142-161). New York: ality: A survival advantage for carriers of genes predisposing
Harper. to schizophrenia? Neuropsychobiology, 3, 179-191.
Guilford, J. P. (1966). Intellectual resources and their values as Jones, E. (1958). The life and work of Sigmund Freud. New
seen by scientists. In C. W. Taylor & F. Barron (Eds.), York: Basic Books.
Scientific creativity: Its recognition and development (pp. 101- Karlsson, J. L. (1978). Inheritance of creative intelligence.
114). New York: Wiley. Chicago: Nelson-Hall.
Guilford, J. P. (1968). Intelligence, creativity, and theireduca- Keefe, J. A., & Magaro, P. (1980). Creativity and schizo-
tional implications. San Diego, CA: Robert R. Knapp. phrenia: An equivalence of cognitive processing. Journal of
Guilford, J. P. (1970). Creativity: Retrospect and prospect. Abnormal Psychology, 89, 390-398.
Journal ofCreative Behavior, 4, 149-168. Korb, R., & Frankiewicz, R. G. (1979, September). Aptitudes,
Haddon, F. A., & Lytton, H. (1968). Teaching approach and intellective styles, and personality characteristics as facili-
divergent thinking abilities. British Journal of Educational tators and differentiators of creativity. Paper presented at the
Psychology,38,171-180. Annual Meeting of the American Psychological Association,
Hamad, S. R. (1972). Creativity, lateral saccades and the non- New York.
dominant hemisphere. Perceptual and Motor Skills, 34, 653- Krebs, H. A., & Shelley, J. H. (Eds.). (1975). The creative
654. process in science and medicine. Proceedings of the C. H.
Harrow, M. , & Quinlan, D. (1977). Is disordered thinking unique Boehringer Sohn Symposium. Kronberg, Taunus, May 16-
to schizophrenia? Archives ofGeneral Psychiatry, 34, 15-21. 17, 1974. New York: American Elsevier.
Hasenfus, N., & Magaro, P. (1976). Creativity and schizo- Kuhn, T. S. (1970). The structure of scientific revolutions.
phrenia: An equality of empirical constructs. British Journal of Chicago: University of Chicago Press.
Psychiatry, 129, 346-349. Lange-Eichbaum, W. (1932). The problem of genius (E. Paul &
Hines, D., & Martindale, C. (1974). Induced lateral eye-move- C. Paul, trans.). New York: Macmillan.
ments and creative and intellectual performance. Perceptual Larkin, E. (1970). Beethoven's medical history. In M. Cooper
and Motor Skills, 39,153-154. (Ed.), Beethoven: The last decade (pp. 439-466). London:
Holden, C. (1987, April). Creativity and the troubled mind. Oxford University Press.
Psychology Today, pp. 9-10. Lelut, L. F. (1836). Du demon de Socrate, specimen d'une
Holton, G. (1978). The scientific imagination: Case studies. application de la science psychologique a celle de l' histoire.
Cambridge, England: Cambridge University Press. Paris: Trinquart.
Holzman, P. S., Kringlen, E., Levy, D. L., Proctor, L. R. & Lidz, T., Wild, C., Schafer, S., Rosman, B., & Fleck, S.
Haberman, S. J. (1977). Abnormal pursuit eye movements in (1963). Thought disorder in the parents of schizophrenic pa-
schizophrenia: Evidence for a genetic marker. Archives of tients: A study utilizing the object sorting test. Journal of
General Psychiatry, 34, 802-805. Psychiatric Research, I, 193-200.
Holzman, P. S., Kringlen, E., Levy, D. L., & Haberman, S. J. Lombroso, C. (1891). The man ofgenius. London: Walter Scott.
(1980). Deviant eye tracking in twins discordant for psycho- Lombroso, C. (1910). The man of genius. New York: Charles
sis: A replication. Archives of General Psychiatry, 37, 627- Scribner's Sons.
631. MacKinnon, D. W. (1960). What do we mean by talent and how
Holzman, P. S., Solomon, C. M., Levin, S., & Waternaux, C. do we test for it? In The search for talent (pp. 20-29). New
S. (1984). Pursuit eye movement dysfunctions in schizo- York: College Entrance Examination Board.
phrenia: Family evidence for specificity. Archives of General MacKinnon, D. W. (1962a). The nature and nurture of creative
Psychiatry, 41, 136-139. talent. American Psychologist, 17, 484-495.
Huang, M-s., & Byrne, B. (1978). Cognitive style and lateral MacKinnon, D. W. (I 962b). The personality correlates of
eye movements. British Journal of Psychology, 69, 85-90. creativity: A study of American architects. In G. S. Nielsen
Hudson, L. (1970). The question of creativity. In P. E. Vernon (Ed.), Proceedings of the XIV International Congress of Ap-
(Ed.), Creativity. Harmondsworth, Middlesex, England: Pen- plied Psychology, Copenhagen, 1961 (Vol. 2, pp. 11-39).
guin Books. Copenhagen: Munksgaard.
Huxley, J., Mayr, E., Osmond, H., & Hoffer, A. (1964). MacKinnon, D. W. (1965). Personality and the realization of
Schizophrenia as a genetic morphism. Nature, 204, 220- creative potential. American Psychologist, 20, 273-:281.
221. MacKinnon, D. W. (1967). The study of creative persons: A
Jamison, K. R. (in press). Manic-depressive illness and accom- method and some results. In J. Kagan (Ed.), Creativity and
plishment: Creativity, leadership, and social class. In F. K. learning (pp. 20-35). Boston: Beacon Press.
Goodwin & K. R. Jamison, Manic-depressive illness. MacKinnon, D. W. (1972). The role of personality traits in the
Oxford, England: Oxford University Press. development of scientific abilities. In A. Rosca (Chm.),
Jamison, K. R., Gerner, R. H., & Goodwin, F. K. (1979). Symposium 18: The detection and training of scientific talent
Patient and physician attitudes toward lithium. Archives of (Vol 1, pp. 515-518). Proceedings, 17th International Con-
General Psychiatry, 36, 866-869. gress of Applied Psychology. Liege, Belgium, July 25-30,
Jamison, K. R., Gerner, R. H., Hammen, C., & Padesky, C. 1971. Brussels: Editest.
(1980). Clouds and silver linings: Positive experiences associ- MacKinnon, D. W., Crutchfield, R. S., Barron, F., Block, J.,
ated with primary affective disorders. American Journal of Gough, H. G., & Harris, R. E. (1958, April). An assessment
Psychiatry, 137, 198-202. study of air force officers: Part I. Design of the study and
Jarvik, L. F., & Chadwick, S. B. (1973). Schizophrenia and description of variables. Lackland Air Force Base, TX: Air
survival. In M. Hammer, K. Salzinger, & S. Sutton (Eds.), Force Personnel and Training Research Center. (Technical
Psychopathology: Contributions from the social, behavioral Report WADC-TR-58-91 [I], ASTIA Document No. AD
sciences (pp. 57-73). New York: Wiley. 151040)
Jarvik, L. F., & Deckard, B. S. (1977). The odyssean person- Martindale, C. (1977-1978). Creativity, consciousness, and
cortical arousal. Journal of Altered States of Consciousness, participation of patients receiving lithium. Psychosomatics,
3, 69-87. 23, 81-87.
Martindale, C., & Greenough, J. (1973). The differential effect Pickering, G. (1974). Creative malady. New York: Oxford Uni-
of increased arousal on creative and intellectual perfonnance. versity Press.
Journal of Genetic Psychology, 123, 329-335. Prentky, R. A. (1979). Creativity and psychopathology: A neu-
Martindale, C., & Annstrong, J. (1974). The relationship of rocognitive perspective. In B. A. Maher (Ed.), Progress in
creativity to cortical activation and its operant control. Journal experimental personality research (Vol. 9, pp. 1-39). New
of General Psychology, 124, 311-320. York: Academic Press.
Martindale, C., & Hasenfus, N. (1978). EEG differences as a Prentky, R. A. (1980). Creativity and psychopathology: A neu-
function of creativity, stage of the creative process, and effort rocognitive perspective. New York: Praeger.
to be original. Biological Psychology, 6, 157-167. Prinzhorn, H. (1972). Artistry of the mentally ill. New York:
Martindale, C., & Hines, D. (1975). Creativity and cortical Springer-Verlag.
activation during creative, intellectual and EEG feedback Reichsman, F. (1981). Life experiences and creativity of great
tasks. Biological Psychology, 3, 91-100. composers: A psychosomaticist's view. Psychosomatic Medi-
Martindale, C., Hines, D., Mitchell, L., & Covello, E. (1984). cine, 43, 291-300.
EEG alpha asymmetry and creativity. Personality and Indi- Robinson, D. N., & Sabat, S. R. (1977). Neuroelectric aspects
vidual Differences, 5, 77-86. of infonn'ltion processing by the brain. Neuropsychologia,
Matthysse, S., & Holzman, P. S. (1987). Genetic latent struc- 15, 625-641.
ture models: Implication for research on schizophrenia. Robinson, L. (1977). Visionaries and madmen: Are creativity
Psychological Medicine, 17, 271-274. and schizophrenia linked? Lawrence Berkeley Laboratory
Matthysse, S., Holzman, P. S., & Lange, K. (1986). The genet- News-Magazine, 2, 7-10.
ic transmission of schizophrenia: Application of Mendelian Roe, A. (1951). A psychological study of physical scientists.
latent structure analysis to eye tracking dysfunctions in schizo- Genetic Psychology Monographs, 43, 121-235.
phrenia and affective disorder. Journal of Psychiatric Re- Roe, A. (1952). A psychologist examines 64 eminent scientists.
search, 20, 57-76. Scientific American, 187, 21-25.
Maupassant, G. de. (1900). Le horla. Paris: P. Ollendorff. Roe, A. (1953). The making of a scientist. New York: Dodd,
McConaghy, N. (1960). Modes of abstract thinking and psycho- Mead.
sis. American Journal of Psychiatry, 117, 106-110. Roe, A. (1956). The psychology of occupations. New York:
McConaghy, N. (1961). The measurement of an inhibitory pro- Wiley.
cess in human higher nervous activity: Its relation to allusive Rosen, G. (1969). Madness in society: Chapters in the historical
thinking and fatigue. American Journal of Psychiatry, 118, sociology of mental illness. New York: Harper & Row.
125-132. Rosman, B., Wild, C., Ricci, J., Fleck, S., & Lidz, T. (1964).
McConaghy, N., & Clancy, M. (1968). Familial relationships of Thought disorder in the parents of schizophrenic patients: A
allusive thinking in university stUdents and their parents. further study utilizing the object sorting test. Journal of Psy-
British Journal of Psychiatry, 114, 1079-1087. chiatric Research, 2, 211-221.
Mednick, S. A. (1962). The associative basis of the creative Rothenberg, A. (1976). Homospatial thinking in creativity.
process. Psychological Review, 69, 220-232. Archives of General Psychiatry, 33, 17-26.
Melton, A. W. (1964). The taxonomy of human learning: Over- Rothenberg, A. (1979). Creative contradictions. Psychology To-
view. In A. W. Melton (Ed.), Categories of human learning day, 13, 55-62.
(pp. 325-339). New York: Academic Press. Rothenberg, A. (1983). Psychopathology and creative cogni-
Melville, H. (1851). Moby Dick: or, The whale. New York: tion. Archives of General Psychiatry, 40,937-942.
Harper & Brothers. Schaffer, P. (1981). Amadeus. New York: Harper & Row.
Mitchell, A. R. K. (1972). Schizophrenia-The meanings of Scholander, P. F. (1978). Rhapsody in science. Annual Review
madness. New York: Taplinger. of Physiology, 40,1-17.
Murphy, G. (1979, December 2). Schumann: A short, troubled Shapiro, R. J. (1970). The criterion problem. In P. E. Vernon
life. Rochester Democrat and Chronicle, p. 5C. (Ed.), Creativity (pp. 257-269). Harmondsworth, Mid-
Nicholls, J. G. (1972). Creativity in the person who will never dlesex, England: Penguin books.
produce anything original and useful: The concept of Slater, E., Hare, E. H., & Price, J. S. (1971). Marriage and
creativity as a nonnally distributed trait. American Psychol- fertility of psychiatric patients compared with national data. In
ogist, 27, 717-727. 1. 1. Gottesman & L. Erlenmeyer-Kimling (Eds.), Differential
Nisbet, J. F. (1912). The insanity of genius: And the general reproduction in individuals with mental and physical disor-
inequality of human faculty physiologically considered. ders. Social Biology, 18, S60-S73.
London: Stanley Paul. Solomon, M. (1975). The dreams of Beethoven. American Im-
Nordau, M. (1900). Degeneration. New York: D. Appleton. ago, 32, 113-144.
Ogilvy, D. (1971). The creative chef. InG. A. Steiner (Ed.), The Stein, M. 1. (1971). Several findings of a transactional approach
creative organization (pp. 119-213). Chicago: University of to creativity. In G. A. Steiner (Ed.), The creative organiza-
Chicago Press. tion. Chicago: University of Chicago Press.
Pavlov, 1. P. (1955). [Selected works] (S. Belsky, trans.). Storr, A. (1971). Problems of creativity . Contemporary P sycho-
Moscow: Foreign Languages. analysis, 7, 115-133.
Pezzullo, T. R., Thorsen, E. E., & Madaus, G. F. (1972). The Storr, A. (1976). The dynamics of creation. Hannondsworth,
heritability of Jensen's level I and II and divergent thinking. Middlesex, England: Penguin Books.
American Educational Research Journal, 9, 539-546. Taylor, D. -W. (1960). Thinking and creativity. Annals of the
Phillips, R. H. (1982). Mood, creativity, and psychotherapeutic New York Academy of Sciences, 91, 108-127.
Terman, L. M. (1926). Genetic studies of genius (Vols. I, II). Vandenberg, S. (1968). The nature and nurture of intelligence.
Stanford: Stanford University Press. In D. Glass (Ed.), Genetics (pp. 3-58). New York: Rockefel-
Tollefson, D. (1961). Response to humor in relation to other ler University Press and the Russell Sage Foundation.
measures of personality. Unpublished Doctoral Dissertation. Winslow, J. H. (1971). Darwin's victorian malady. Philadel-
University of Illinois. phia: American Philosophical Society.
Tomlinson-Keasey, C., Warren, L. W., & Elliott, J. E. (1986). Wittkower, R. R. (1973). Genius: Individualism in art and art-
Suicide among gifted women: A prospective study. Journal of ists. In P. P. Wiener (Ed.), Dictionary of the history of ideas
Abnormal Psychology, 95, 123-130. (pp. 297-312). New York: Scribners.
Trilling, L. (1950). The liberal imagination: Essays on literature Woody, E., & Claridge, G. (1977). Psychoticism and thinking.
and society. New York: Viking Press. British Journal of Social and Clinical Psychology, 16, 241-
Tsanoff, R. A. (1949). The ways of genius. New York: Harper & 248.
Brothers. Wright, S. (1954). Some psychological and physiological corre-
Tucker, D. M., Stenslie, C. E., Roth, R. S., & Shearer, S. L. lates of certain academic underachievers. Unpublished Doc-
(1981). Right frontal lobe activation and right hemisphere toral Dissertation, University of Chicago.
performance. Archives of General Psychiatry, 38, 169-174.
CHAPTER 16
Examining Counselors' Creative

Processes in Counseling
P. Paul Heppner, Karen Fitzgerald, and Carolyn A. Jones
When working with clients, the process of successful therapy is tion there are underlying elements of spontaneity
akin to the process of a continuous weaving, of picking up vari- and intentionality that can be seen in creative coun-
ous colored and textured threads to create a new fabric. (Carolyn
Jones, personal communication, January, 1986)
seling behavior.
The purpose of this chapter is to examine the role
Creative counseling involves feeling connected with the client of counselors' creativity within counseling. The
in a mutual atmosphere of energy, trust and individualized com- subject of creativity has. been examined within a
munication. (Helen J. Roehlke, personal communication, Janu- variety of contexts, such as art, scientific in-
ary, 1986)
ventions, and other public arenas. Little attention,
Intense reactions to and awareness of the connections between however, has been given to creativity within the
present difficulties and events in the past will result in the ap- therapeutic process. Creativity is often studied in
plication of new and effective knowledge. (Richard Thoresen, relation to problem solving. Because a counselor's
personal communication, January, 1986)
role is to help other people solve problems, the
Therapists who have a broad range of knowledge and who are therapeutic process may well be a productive set-
capable of eliciting internal images that capture the essence of a ting within which to examine creativity. Creativity
client's situation are constructive and insightful. (Wayne P. An- in the counseling process could be examined from
derson, personal communication, January, 1986) several perspectives: (a) from the counselor's point
of view, (b) from the client's perspective, or (c)
These are statements from creative, effective from the interaction between the therapist and the
therapists who were describing the process which, client. In this chapter, we will focus solely on
for them, represents innovative and productive creativity from the counselor's perspective, al-
counseling interactions. Their words allude to though the other perspectives merit examination as
themes of flexibility, knowledge, confidence, and well. In so doing, we realize that we have ar-
uniqueness. They describe the process of under- tificially divided the counseling process, which has
standing clients and of intervening in ways that pro- as its foundation the interaction between counselor
duce constructive change. Although this process is and client.
clearly different for each therapist, upon examina- We will begin the chapter by briefly defining the
term counseling and discussing some of the essen-
P. Paul Heppner, Karen Fitzgerald, and Carolyn A. tial characteristics of the counseling process. Next,
Jones • Department of Psychology, University of Missouri, we will discuss the role of counselor creativity with-
Columbia, MO 65211. in counseling, suggesting that creativity occurs in at
271
least two major areas of counseling; namely, under- clients. Although the focus for change is on the
standing a client's problem and facilitating client client in a counseling situation, experienced coun-
change. In the third section of the chapter, we will selors acknowledge that they, too, may be affected
speculate about some of the variables that affect by the relationship in ways that initially may not be
counselor creativity within counseling. Finally, we known or not easily controlled.
will conclude by discussing research strategies and The counseling process is exceedingly complex,
how the formal educational training of therapists and client change is often difficult to evaluate
might provide a unique setting in which to investi- (Highlen & Hill, 1984). In addition, the possible
gate how creativity can be fostered or inhibited in routes to change are innumerable and many times
counselor trainees. . idiographic. Counseling is also unique in that it
involves interaction; this is a special human rela-
tionship in which the client and the counselor en-
The Counseling Process counter each other and develop a working alliance.
Each counselor-client dyad is unique and estab-
Counseling has been defined as a psychological lishes idiosyncratic communication patterns within
specialty in which "practitioners help others im- the therapeutic relationship.
prove their well-being, alleviate their distress, re- Counseling involves scientific and artistic ele-
solve their crises, and increase their ability to solve ments. The scientific elements of predictability and
problems and make decisions" (Fretz, 1982, p. rationality are necessary in order to facilitate client
15). Counseling is a process that involves an in- change. The artistic side of counseling allows for
teraction between client and counselor with the the uniqueness and vitality of human interaction
shared goal being client change or growth, often and includes mystery and evocation. But coun-
occurring within the context of a highly involving selors, unlike artists or scientists, have less control
and emotional professional relationship. This goal over the medium with which they work. Clients are
can be accomplished through various methods de- not frozen in time, waiting for the next intervention,
rived from the study of human behavior (see Cor- but rather are evolving in an environment of contin-
sini, 1984). uous external stimulation. Using the analogy of
Counseling can also be viewed as an information weaving a new fabric, it is here that new threads are
exchange between the client and the counselor added from the counselor, while threads from with-
(Heppner & Fitzgerald, 1987). The client attempts in the client are altered. In a sense, creativity within
to communicate to the counselor the nature of his or counseling can be the link between the predictable
her experiences and the difficulties and conflicts and the mysterious in human interaction, the known
that are associated with those experiences. The and the unknown.
counselor attempts to understand the client's expe-
riences in the context of this information and to gain
an awareness of the client's personality dynamics. Counselor Creativity within Counseling
In doing this, the counselor utilizes a broad range of
knowledge about personality development and dy- Creative thought and behavior are manifested in
namics, personal problem solving or coping pro- many ways in the counseling process. Our defini-
cesses, and communication processes. tion of counselor creativity involves the combina-
Given the counselor's understanding of the cli- tion of information, often in unique or novel ways,
ent, and the development of a professional working that is ultimately used to elucidate or solve a client's
alliance with the client, the focus shifts to client problem by extending the client's experiential
change. The counselor's goal is to help the client world in some way. A common characteristic of
change. This is typically done through a variety of creative events within counseling is "double mem-
interventions that are based on knowledge about bership" (Koestler, 1964); that is, creative thought
counseling techniques and the process of changing involves seeing events in two frames of reference,
human behavior. These interventions can be con- which heretofore were seen as unconnected, as part
ceptualized as imparting know ledge and skills as a of a theme which then provides new insight into the
way of altering maladaptive information-process- client's situation for both the client and the coun-
ing patterns that lead to maladaptive behavior. selor. Thus, some kind of transformation of infor-
Given their experiences in the world, counselors, mation occurs (Pickard, 1987). This transformation
too, are subject to change, especially with previous may involve thoughts, feelings, and behaviors that
CHAPTER 16 • EXAMINING COUNSELORS' CREATIVE PROCESSES 273
are new or different for the client. Thus, creative similar to the story of the blind men describing an
events within counseling may require combining elephant-it is strongly influenced by the perspec-
events or information in a novel way. This often tive of the therapist.
involves providing a new piece to the client's We make use of empathy, intuition, and the-
puzzle and is usually accompanied by a feeling of oretical and scientific knowledge to create a de-
surprise or a sense of the unexpected. scriptive image or working model of a client within
Another characteristic of creative events pertains ourselves. This image, constructed from our expe-
to the facilitation of the counseling process. Cre- rience with the client, is an aid in diagnosis and
ative counseling is not simply atypical interventions treatment planning. The challenge lies in creating a
or novel descriptions of a client. Rather, a critical balance between an empirically standardized de-
feature is that the outcome of creative events in scription and an image that does justice to the
counseling is an extension of the client's experien- unique, dynamic aspect of a client.
tial world and discernible client progress. In other Understanding the essence of a client's problem
words, the role of creative behavior in counseling is involves an accurate representation of the client's
purposeful, not merely interesting or aesthetic. difficulties in the context of his or her history and
Thus, to evaluate creative processes in a counselor, personality dynamics. This often requires perceiv-
it is necessary not only to examine how the coun- ing linkages between aspects of client experiences
selor is processing information but ultimately to that may seem unrelated. Many times, an experi-
estimate the impact on the client as well. (Because enced and competent counselor can accurately
the latter point is beyond the scope of this chapter, it comprehend the essence of a client's problem with-
will not be discussed much here.) Sometimes client out a great deal of trouble, and within a relatively
progress may be miniscule; other times the creative few number of counseling sessions. For example,
integration may result in a major turning point or the majority of clients with normal developmental
critical incident in therapy (see Kell & Mueller, problems pertaining to social relationships, such as
1966). Regardless of the quantity, the presence of dating, studying, career planning, and interper-
change or progress in the client is an essential char- sonal communication, typically present few prob-
acteristic in evaluating counselor creativity. lems for veteran counselors. Creative thought and
We propose that counselors utilize creative pro- behavior may not be required in client assessment in
cesses in at least two major areas: (a) to understand these situations because past experience suggests
the essence of the client's problem(s), and (b) to what the underlying issues are likely to be. Con-
develop interventions to facilitate desired client versely, there are other clients and client problems
change. For our immediate purpose, we will deline- that present a befuddled array of information, and
ate the different opportunities for creative ex- the counseling process may then reach an impasse.
pression by counselors in a somewhat temporal and Counselors often acknowledge such situations by
linear fashion. We recognize, however, that this noting that they are "feeling stuck," that "some-
approach does not fully represent the complex thing is missing," or that "the pieces don't hang
nature of creativity within the counseling process together." The client may appear uncharac-
where counselor, client, and counselor/client in- teristically "resistant"; the client-counselor in-
teraction variables are deeply embedded in a con- teraction may be strained and unproductive; and the
textual matrix. Moreover, the different processes counselor may continually discover what seems to
involved in understanding the client's problems and be new and unrelated information. In essence, the
developing interventions are oftentimes intertwined counselor and the client together are unable to make
with each other and are highly interdependent in a sense of the relevant information and to integrate
nonlinear fashion. that information to pinpoint and understand the
A counselor might initially engage in creative nature of the client's difficulties.
processing with a client during the process of case An impasse can be a serious obstacle in counsel-
conceptualization. In deriving personality theories, ing because progress is blocked. Experienced thera-
psychologists have used conceptual skills to form pists attempting to circumvent such obstacles often
broad impressions regarding the healthy develop- describe periods of intense concentration in which
ment of human beings as well as those attributes they attend very closely to one or more of their
that indicate some type of psychological dysfunc- auditory, kinesthetic, or visual senses. In essence,
tion or personality disorder. The procedure with they use whatever modality to obtain more informa-
individual clients is much the same. This process is tion and then use that information to understand the
client's difficulty. Extraneous and nonessential egies from memory and apply that knowledge to the
stimuli are blurred in this state of focused attention. client's problem at hand. As these plans are "tested
Counselors talk about "really hearing the client, " out, " the counselor can obtain feedback from the
and, with this focus, imperceptible nuances or sub- client about the effectiveness of the intervention.
tleties in the client's "story" may emerge with Many times an experienced and competent coun-
striking clarity. Counselors may "suddenly have a selor can effectively implement intervention strat-
feeling about the client," perhaps signaling that egies with a client that result in the desired client
some key element has yet to be resurrected despite change. Often these interventions are not especially
the appearance that everything of importance has unique, and there is minimal trial and error involved
been accounted for. They may "feel" affectively to in the design or the application of the intervention.
some degree what the client dares not feel as the For example, certain problems, such as phobias,
client tries to avoid painful emotional material and study skills deficits, and lack of assertiveness, can
attempts to maintain a sense of control. The coun- be addressed by relatively standard treatment meth-
selor's experience of the explicit or implicit feeling ods. In these circumstances, there is limited need
state of the client means that he or she "partakes of for more creative thought or behavior on the part of
the quality of the feelings and not the quantity" the counselor. However, there are other situations
(Greenson, 1967, p. 368). In working through an in which the therapist has tried more familiar and
impasse, the counselor may also utilize a descrip- logical interventions, all of which have failed. As
tive image of the client, and as the image "reacts with problems associated with case conceptualiza-
with ideas, feelings, memories or fantasies, etc." tion, counseling may reach another impasse, this
(Greenson, 1967, p. 369), the counselor may dis- time during the intervention phase of the therapeu-
cover some missing piece in the therapeutic puzzle. tic process. At such times, counselors often ac-
When considering a new piece of information, knowledge feeling frustrated, either with them-
the experienced therapist will often test his or her selves or with the client. This type of impasse
hunches or opinions either explicitly with the client during the intervention phase also effectively
(e.g., could it be this?) or implicitly by collecting blocks the therapeutic process and is another oppor-
confirmatory information. Thus, this creative coun- tunity in which creative processes in therapy can be
selor process may occur at once (voila!) or gradu- effective.
ally by extended interactions with the client. This It is important to note that creative counselor
additional information allows a more complete inte- behavior does not only occur during impasses in the
gration of the previous information and a more counseling process. One therapist mentioned the
useful picture of the client, so that the counseling times when she really felt happy and playful and
process can now proceed in a more productive man- when the counseling process was "really rolling";
ner. The composer Brahms described such a pro- subsequently, she felt inspired to behave very cre-
cess as feeling "vibrations which thrill my whole atively in session. Another spoke of the increased
being ... [which] assume the form of distinct confidence one can feel to take risks when the coun-
mental images . . . the ideas flow upon me . . . seling process is progressing smoothly. However, it
and not only do I see distinct themes . . . but they seems that some of the most striking creative events
are clothed in the right forms" (Abel, 1964, pp. occur at points of difficulty when the counselor
19-21). In like manner, a novel synthesis of mean- feels stuck. It may be that these events are simply
ingful material can result in an emotional rush for more meaningful or memorable due to the tension
the counselor as a result of the counselor's creative of the impasse and the greater need for a novel,
discovery of what was unfamiliar and previously purposeful perspective.
unknown to perhaps both client and counselor. One therapist we spoke to noted her unexpected
As the counselor and, in particular, the client awareness of feeling sad during a particular coun-
better understand the nature of the client's difficul- seling sequence. When the therapist reflected that
ties, and as counseling goals are established, then feeling, the client was then able to explore an aspect
either implicit or explicit intervention plans are then of herself that effectively opened the door for fur-
developed. Depending on the complexity of the cli- ther change. Other therapists report utilizing spon-
ent's difficulties and the comprehensiveness of the taneous images that develop in their awareness as a
goal(s), this can be a quick or a very time-consum- way of synthesizing bits of information obtained
ing process. The task for the counselor is to retrieve from the client. Another therapist described the pro-
a broad array of knowledge about intervention strat- cess of incubation; for example, as he thought about
a particular client at times throughout the week, his Table 1. Content and Process Variables
affective reactions to the client became more pro- Affecting Counselor Creative Processes
nounced and then served as a basis for an interven-
tion in the next counseling session. Similarly, an- Variables Content Process
other therapist described taking a coffee break Variables af- Breadth and Ability to re-
during a session, primarily to offer himself an op- fecting the depth of main open to
portunity to scan his sensory awareness and inte- utilization of knowledge new informa-
grate new information. In short, it seems that expe- knowledge base tion
rienced counselors engage in a variety of cognitive Counselor self- Ability to be re-
and affective processes that allow them to integrate efficacy ceptive to
information, sometimes very small pieces of infor- one's experi-
mati on , which either provides some hints for under- ences
standing the client and/ or subsequent interventions. Variables af- Counselor emo- Counselor effort
The interactive quality of creativity is essential in fecting the tional and or involve-
understanding its effectiveness. When we choose quality of the physical ment
one way or another to intervene in a client's system, counseling well-being
we are taking a step in the "dance" of therapy. relationship
Sometimes we may logically or intuitively decide
which intervention to use. But sometimes the ra-
tionale for an intervention may be difficult to identi-
fy, and the client's response may be difficult to roughly into two categories: (a) variables affecting
predict. Often we express a sense of doing some- the utilization of knowledge, and (b) variables af-
thing "on a hunch." In these cases, the reactions of fecting the quality of the counseling relationship.
our clients can give us clues as to why we chose a We have further divided these categories into con-
particular intervention or direction and can help us tent and process variables (see Table 1).
to deepen our understanding of the client. Content refers to variables that are finite and
Sometimes the counselor-client interaction quantifiable, and process refers to more fluid vari-
leads to unexpected, unintended, and even un- ables and involves the counselor's style of process-
wanted outcomes. Eileen Pickard (personal coming information. A critical content variable is the
munication, May 24, 1987) described a "relay of counselor's knowledge base, which seems to us to
insights" in creative processing, by which she be the essential foundation of being an effective
meant that one person's insights may be received therapist. A related content variable is a counselor's
quite differently by another and may lead to unin- sense of self-efficacy, which seems to affect how a
tended developments on the part of the initiator. counselor regulates or uses his or her cognitive,
What may begin as uncreative within the intentions affective, and behavioral responses. Two process
of one person may spark creative insights in the variables that are related to knowledge reflect on
receiver, and vice versa. Such "relays" depict the aspects of the counselor's style of processing infor-
creative processing within counselors and clients. mation: specifically, (a) remaining open to new in-
formation and possibilities, and (b) being receptive
to one's experience.
Variables Affecting Creativity within Because counseling is based on interactions be-
Counseling tween two people, our second group of variables
revolves around the interpersonal quality of the
The purpose of this section is to discuss some of therapeutic relationship. A critical content variable
the variables that affect a counselor's creativity within the relationship is the physical and emotional
within counseling. This section is speculative and is well-being of the therapist. The process variables
based on our own counseling experiences as well as pertain to the effort or involvement which the thera-
on applied and laboratory problem-solving re- pist invests in the client.
search. We believe that a number of different vari-
ables affect a counselor's creativity within counsel-
ing and most likely are interrelated; as a result, our Knowledge Bases
discussion is in no way meant to be exhaustive. A major variable affecting creativity within
The variables we have identified can be grouped counseling is the counselor's knowledge base, in-
cluding, in particular, knowledge about personality anxiety, and ability to tolerate confusion and ambi-
development and dynamics, the change process, guity (for related problem-solving research, see
and the counseling process. The counselor's knowl- Heppner, Hibel, Neal, Weinstein, & Rabinowitz,
edge bases directly affect the counselor's abilities in 1982; Heppner, Reeder, & Larson, 1983).
case conceptualization and developing interven-
tions to facilitate desired client change. Research
suggests that one's knowledge bases affect the Remaining Open to New Information
speed and accuracy of processing information (An- There is sometimes a danger in relying ex-
derson, 1983); we suspect that this is the case within clusively on one's knowledge bases and person-
counseling as well. The more relevant knowledge ality / counseling theories; the therapist may be
the counselor has available about counseling ac- unable to understand accurately the client's prob-
tivities, the more facile the counselor will be in lem or to facilitate client change because the thera-
processing information within the counseling expe- pist recognizes only information relevant to his or
rience. Conversely, the more the counselor labors her knowledge bases and theories, thereby restrict-
in the theoretical conceptualization of human being or limiting his or her vision. The counselor may
havior and the counseling process, less time, ener- be cognitively, perceptually, and emotionally posi-
gy, and space will be available to process informa- tioned to look for what is already "known" to be a
tion about a particular client and his or her unique possibility. In short, there sometimes is a risk in
situation. knowing, especially when the counselor makes er-
In a simplified way, knowledge bases reflect dif- roneous or biased assumptions about the client (see
ferences along a novice-expert continuum as dis- Havens, 1982). Thus, a critical variable is the man-
cussed in the information-processing literature (An- ner in which a therapist processes information with-
derson, 1983). The counselor's knowledge bases in counseling so as to remain open to new pos-
are like a foundation, a base from which to begin. sibilities. This involves a willingness to set aside
Several of the therapists we interviewed remarked one's know ledge, to surrender one's beliefs about a
that their experience and reading in the field of phenomenon, to "not know" in order to learn more
psychology (and also great literary works on the deeply about some aspect of a client's life. The
human condition) provided an excellent storehouse ability to tolerate ambiguity and anxiety is impor-
of information which they could tap into for insight. tant here as one relinquishes predictability. The
Large knowledge bases do not guarantee creative poet Keats labeled the process "negative ca-
behavior within counseling, but high levels of pability" and used such phrases to describe this
creativity most likely are not achieved without an process as removing "barriers to knowledge,"
adequate knowledge base. "removing the self," "negating or annihilating the
self" to permit one to experience reality without the
burden of consciousness of "self" (see Bate, 1939;
Therapist Self-Efficacy
Margulies, 1984). One therapist told of "spon-
The information in the counselor's knowledge taneous images" that arise when she resists the
bases is obviously a critical factor within creativity. temptation to assign prematurely categories or la-
Another and equally important factor is how the bels that she perceives about a client. "Not know-
counselor appraises or evaluates his or her counseling" means the counselor must resist the lure of
ing knowledge and skills. Such self-evaluation is a status afforded his or her position as an "expert."
metacognitive variable (Butler & Meichenbaum, Paradoxically, the approach of not knowing may
1981) and could be conceptualized as self-efficacy allow a counselor to obtain more information to
(Bandura, 1986). The extent to which a therapist understand more clearly the obstacles a client is
believes in him- or herself (and in the counseling experiencing.
process as well) is a function of previous counseling
experience and is most likely correlated with his or
Being Receptive to One's Experience
her know ledge bases about counseling. We concep-
tualize self-efficacy as a content variable (rather Clients bring problems into counseling that they
than a style variable), which is a measurable quan- have been unable to resolve themselves, in part,
tity of belief in oneself as a therapist. A strong belief because they are unaware of their experiences (and,
in oneself as a therapist most likely affects one's thus, are lacking information), or are unable to find
performance in a number of ways, such as per- a workable solution. Sometimes counselors can be
severance, risk-taking, involvement, performance helpful by "taking on" the client's problem (with
high levels of empathy) and using their own reac- (see Gelso & Carter, 1985). The quality of the rela-
tions to uncover missing information and psycho- tionship will affect the amount of effort expended
logical processes. Or sometimes counselors can and the level of involvement by both counselor and
obtain information by monitoring how they are client. With a high degree of effort, often there are
personally responding to the client. A therapist quantitative and qualitative differences in the
talked of the new understanding or the almost' 'un- amount of information that is processed. Likewise,
canny insights" that emerge when he allows him- the counselor's interest and emotional involvement
self to listen' 'very quietly" to all parts of his expe- in the client often elicits deep, elaborative cognitive
rience as he empathizes with the client's situation. processing within the counselor. In a way, the client
Another process in creativity involves counselors becomes a "current concern" (Klinger, 1971) of
being receptive to their experiences within and be- the counselor. Klinger has noted that creativity
tween counseling, particularly their respondent, af- often arises from "important" problems; problems
fective, and cognitive processes (for related re- that have goals that are meaningful in some way to a
search, see Klinger, 1971). More experienced person. Thus, a high level of counselor effort and
counselors talk about reacting to clients with a cer- involvement results in a high level of information
tain feeling (e.g., sadness, grief, pride) and then processing on the part of the counselor, thereby
using that awareness to facilitate the counseling increasing the probability that creative uses of in-
process in some way. Counselors vary widely in formation will occur. This high level of counselor
their reception of their own impulses, affects, and involvement often entails intermittent processing
ideas. The need to "move beyond words" was ex- between counseling sessions, or incubating with a
pressed by one therapist who makes great use of problem for a while, until ideas are combined in a
imagery in his work. Some counselors ignore or are unique fashion.
unaware of many of their reactions to clients, or However, sometimes a counselor tries too hard,
believe them to be inappropriate in that setting, and, becomes overinvolved with a client, and his or her
thus, are unable to use this information to solve the focus of attention narrows and results in missed
client's problem. Being receptive to one's inner ex- information. Several writers have noted that cre-
periences may be a function of individual difference ative insight occurs during a period of "relaxed
variables (e.g., introversion-extraversion: Myers & tension" (Ghiselin, 1952; Rugg, 1963) so that the
McCaulley, 1985) as well as a learned capacity to solution is not forced (Klinger, 1971). Or some-
use one's inner events. In short, sometimes being times counselors (especially trainees) may be over-
creative within counseling involves processing in- concerned about their counseling performance,
formation from several perspectives (counselor, which not only results in high levels of anxiety or of
client) as well as several modalities (cognitive, af- being overly focused on their emotional responses
fective) and senses (visual, auditory). but also reduces their ability to process information.
Sometimes more experienced counselors describe
their creative insights as occurring while their
Well-Being
senses are involved in a highly charged emotional
Because counseling is an interpersonal activity, counseling relationship; yet, concurrently, they are
which is often emotionally charged, the emotional able to remove themselves to cognitively process
and physical well-being of the therapist is an impor- what is transpiring in the counseling session and the
tant variable affecting creativity. An emotionally or therapeutic process. This phenomenon has been de-
physically tired therapist often has more difficulty scribed in other contexts as the capacity to be, at
processing information within counseling, and once, both participant and observer (Capra, 1975),
many times is less able to understand the psycho- and involves the ability to balance internal and ex-
logical events in the client's life. In short, the well- ternal messages.
being of the therapist is a condition that can inhibit
or facilitate how the counselor processes informa-
tion and, consequently, may affect creativity within Conclusion
counseling. We have discussed several variables that affect
counselors' creativity within counseling. From our
experience, there seems to be a curvilinear rela-
Therapist Effort and Involvement
tionship between most of the variables we have
A critical variable in counseling pertains to the discussed and the creative process; that is, a lack of
relationship between the counselor and the client knowledge, self-efficacy, openness to new infor-
mation, receptivity to one's experiences, and in- information processing (e.g., cognitive, affective,
volvement most likely inhibit creative responses. imaging) with problem identification and interven-
But high levels of these variables may also inhibit tions on the change process (attitudes, cognitions,
creativity. For example, a large knowledge base behaviors) within clients. Investigators could em-
about counseling coupled with an exaggerated ploy methodologies, such as Interpersonal Process
sense of self-efficacy as a therapist may reduce the Recall (Kagan, 1975), Thought-Listing (Cacioppo
therapist's openness to new information and inhibit & Petty, 1981), Hill's counselor intentions and cli-
novel approaches to counseling. ent reactions (Hill & O'Grady, 1985; Hill, Helms,
It seems that counselor creativity in counseling isSpiegel, & Tichenor, 1988), or even Martin's sche-
most often a function of some optimal level of infor- mata analysis (Martin, 1985). Such methodologies
mation processing by the therapist. Each of the vari- could be used to assess how clients or counselors
ables we discussed either facilitates or inhibits the are processing information, especially during im-
therapist in cognitively or affectively processing in-passes. Comparisons could be made along an ex-
formation relevant to the client's problem. Most pertise dimension (novice trainees, doctoral level
likely, counselor creativity within counseling is not interns, experienced and skillful therapists), per-
simply related to the amount of information pro- haps initially through intensive single-subject de-
cessed, but also how the information is processed signs and later between groups. Correlational de-
and combined in novel yet effective ways. It is our signs could also be utilized to examine the
hunch that more experienced, highly skilled, and relationships among some of the variables we sug-
creative therapists have well-differentiated road- gested earlier (e.g., counselor knowledge bases,
maps of the counseling process for different types self-efficacy, remaining open to new information)
of clients, and that they can process information and creative counselor processes.
more quickly, accurately, and in nonlinear ways. Another setting that may be quite fruitful to ex-
Their successful experiences may in tum bolster amine is the formal training of psychotherapists.
their therapeutic confidence, which in tum allows Much has been written about the impact of formal
them greater tolerance for risk taking and ambigu- education on creative behavior in general and on the
ity. Most likely, these counselors can also discrimi- cognitive processes that are involved in creative
nate between relevant and less relevant information production (Burgett, 1982; Simonton, 1983). As
and, subsequently, may have more time to concen- yet there has not been a systematic analysis of the
trate on finding the missing piece or to combine manner in which creativity and problem solving are
information in unique and helpful ways. addressed in counselor education programs. Re-
searchers might examine the effects of counselor
education on counselor creativity, preferably via a
longitudinal analyses, but also through a cross-sec-
Final Comments tional analysis. A useful research model may be one
employed in examining the effect of counselor edu-
This chapter discussed the role of counselor cation on trainees" research interests (see Royalty,
creativity within counseling and some of the vari- Gelso, Mallinckrodt, & Garrett, 1986). Another
ables that seem to affect creative counselor behav- approach may be to examine the supervisory pro-
ior. We believe that counseling would provide a cess of trainees in relationship to the development
productive setting in which to study creativity. In of creative processes in counselors. Supervision is
fact, counseling would provide a unique oppor- widely recognized as one of the most important
tunity for studying creativity within the context of teaching mechanisms of counselor trainees; within
an interpersonal relationship. It may prove to be the last five years, there has been a dramatic in-
particularly productive to examine creativity in crease in supervision research, most notably with
terms of how the counselor processes information regard to the developmental processes of trainees
and, subsequently, how the client processes (e.g., Heppner & Roehlke, 1984; O'Leary-Wiley
information. & Ray, 1986; Reising & Daniels, 1983). Re-
There are a number of ways that researchers searchers might examine the relationships between
might proceed. The interpersonal influence re- supervisor variables (e.g., supervisor style, super-
search paradigm (see Corrigan, Dell, Lewis, & visor behaviors related to teaching and emotional
Schmidt, 1980; Heppner & Dixon, 1981) might support, supervisor theoretical orientation and in-
also be useful to analyze the effect of a counselor's terpersonal values, level of creative processes mod-
eled by the supervisor) with various indices of train- programs were directed more toward such process
ee creativity (e.g., trainee self-report variables, objectives than were most American programs.
information processing as indicated above, type In short, a counselor's role is to help other people
and effectiveness of trainee interventions with cli- solve problems. Because creativity is often studied
ents, such as metaphors and novel fantasies) across in problem solving, the training of counseling psy-
different trainee developmental levels (see chologists might be a fertile area to investigate how
Rabinowitz, Heppner, & Roehlke, 1986, for a creativity can be fostered or inhibited in trainees.
model of the latter). Likewise, counseling psychologists might provide
It is our concern that formal education in ad- a population with which to study naturally occur-
vanced degree programs may inhibit or discourage ring creative events, which could form the basis for
the creative capabilities of trainees. For example, descriptive research. It is unfortunate that so little
trainees are often asked to justify their interven- attention has been given to how therapists process
tions. Although valuable, this challenge process information. We know very little how more experi-
may also discourage divergence and inhibit more enced or novice counselors use their cognitive and
creative thinking. In addition, therapists must be affective processes, or how they combine informa-
ethically responsible for their clients' welfare and tion to develop creative client conceptualizations or
safeguard them from harm. Thus, trainees may feel therapeutic interventions. Likewise, we know very
hesitant to consider any approach that is not cus- little about how clients' experiential world is im-
tomary, or they may be directly cautioned about pacted by creative counselor responses, and the dy-
being creative. Without any assurance of what the namic interplay between the two. Perhaps there
reaction to a creative (unique, surprising, nontradi- might be mutual benefits for creative collaborative
tional) intervention will be, many counselors (es- efforts between cognitive psychologists, who are
pecially novices) hesitate to risk what seems like interested in exploring creative processes, and
nonconventional acts. Even though learning the counseling psychologists, who desire to use cre-
fundamentals of counseling enhances the trainee's ative behavior to create healthy change in their
ability to be a therapist, the trainee's potential to be clients.
intuitive or creative can be inhibited by an over-
emphasis on the predictable, traditional methods. ACKNOWLEDGMENTS. The authors are grateful
Thus, from a research perspective, it may also be to the following therapists for their thoughts regard-
useful via some of the research models suggested ing creativity within counseling: Wayne Anderson,
above to assess if trainees' creative processing in- Helen Roehlke, Richard Thoreson. In addition, we
creases or decreases over time during graduate would like to thank Helen Roehlke and Eileen Pick-
training. ard for their editorial assistance in preparing this
It is important to note that facilitating creativity chapter.
in counselor trainees does not need to be done at the
expense of the more traditional educational objec-
tives. In fact, we believe that counselor educators
References
and supervisors must continue to insist on rigor and Abel, A. M. (1964). Talks with the great composers. Gannisch-
demonstrable competence in trainees' academic Partenkirchen, Germany: G. E. Schroeder-Verlag.
preparations. A curriculum that fosters the creative Anderson, J. R. (1983). The architecture of cognition.
capabilities in trainees does not need to be divided Cambridge: Harvard University Press.
Bandura, A. (1986). Socialfoundationsofthoughtandaction:A
along academic/nonacademic lines. Rather, train- social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall.
ing for creativity and problem solving could be add- Bate, W. J. (1939). Negative capability: The intuitive approach
ed to the curriculum, with the goal of enhancing the in Keats. Cambridge: Harvard University Press.
counselor's ability to process information from sev- Burgett, P. J. (1982) . . . . On creativity. Journal of Creative
eral modalities (e.g., thinking, feeling, imaging). Behavior, 16, 239-249.
Butler, L. , & Meichenbaum, D. (1981). The assessment of inter-
In this regard, however, it might be useful to exam- personal problem-solving skills. In P. Kendall & S. D. Hollon
ine the leaming objectives most commonly associ- (Eds.), Assessment strategies for cognitive-behavioral inter-
ated with graduate training, with a keen eye toward ventions (pp. 197-225). New York: Academic Press.
process versus content objectives, such as ini- Cacioppo, J. T., & Petty, R. E. (1981). Social psychological
procedures for cognitive response assessment: The thought-
tiative, discovery leaming, independence, and self- listing technique. In T. V. Merluzzi, C. R. Glass, & M.
evaluation skills. Anecdotally, it appeared to the Genest (Eds.), Cognitive assessment (pp. 309-342). New
first author that some English counselor-training York: Guilford Press.
Capra, F. (1975). The tao of physics. Berkeley: Shambala. tions illustrated in a case study and with therapists of varying
Corrigan, 1. D., Dell, D. M., Lewis, K. N., & Schmidt, L. D. theoretical orientations. Journal of Counseling Psychology,
(1980). Counseling as a social influence process: A review. 32,3-22.
Journal of Counseling Psychology, 27,395-441. Hill, C. E., Helms, 1. E., Spiegel, S. B., & Tichenor, V. (1988).
Corsini, R. (Ed.). (1984). Current psychotherapies (3rd ed.). Development of a system for categorizing client reactions to
Itasca, IL: F. E. Peacock. therapist interventions. Journal of Counseling Psychology,
Fretz, B. R. (1982). Perspectives and definitions. The Counsel- 35, 27-36.
ing Psychologist, 10(2), 15-19. Kagan, N. (1975). Influencing human interaction: Eleven years
Gelso, C.l., & Carter, 1. A. (1985). The relationship in counsel- with IPR. Canadian Counsellor, 9, 44-51.
ing and psychotherapy: Components, consequences, and the- Kell, B. L., & Mueller, W. 1. (1966). Impact and change: A
oretical antecedents. The Counseling Psychologist, 13(2), study of counseling relationships. New York: Appleton-Cen-
155-243. tury-Crofts.
Ghiselin, B. (1952). The creative process. New York: New Klinger, E. (1971). Structure and function of fantasy. New
American Library. York: Wiley-Interscience.
Greenson, R. R. (1967). The technique and practice ofpsycho- Koestler, A. (1964). The act of creation. New York: Macmillan.
analysis (Vol. 1). New York: International Universities Press. Margulies, A. (1984). Toward empathy: The uses of wonder.
Havens, L. (1982). The risks of knowing and not knowing. American Journal of Psychiatry, 141, 1025-1033.
Journal of Social Biological Structures, 5, 213-222. Martin, 1. (1985). Measuring clients' cognitive competence in
Heppner, P. P., & Dixon, D. N. (1981). A review of the inter- research on counseling. Journal of Counseling and Develop-
personal influence process in counseling. Personnel and ment, 63, 556-560.
Guidance Journal, 59, 542-550. Myers, l. B., & McCaulley, M. H. (1985). Manual: A guide to
Heppner, P. P., & Fitzgerald, K. M. (1987). Human intel- the development and use of the Myers-Briggs Type Indicator.
ligence: Implications for counseling. Journal of Counseling Palo Alto, CA: Consulting Psychologists Press.
and Development, 65, 266-267. O'Leary-Wiley, M., & Ray, P. B. (1986). Counseling supervi-
Heppner, P. P., & Roehlke, H. 1. (1984). Differences among sion by developmental level. Journal of Counseling Psychol-
supervisees at different levels of training: Implications for a ogy, 33, 439-445.
developmental model of supervision. Journal of Counseling Pickard, E. M. (1979). The development of creative ability.
Psychology, 31, 76-90. London: National Foundation for Educational Research.
Heppner, P. P., Hibel, 1. H., Neal, G. W., Weinstein, C. L., & Rabinowitz, F. E., Heppner, P. P., & Roehlke, H. 1. (1986).
Rabinowitz, F. E. (1982). Personal problem solving: A de- Descriptive study of process and outcome variables of super-
scriptive study of individual differences. Journal ofCounsel- vision overtime. Journal ofCounseling Psychology, 33,292-
ing Psychology, 24, 580-590. 300.
Heppner, P. P., Reeder, B. L., & Larson, L. M. (1983). Cog- Reising, G. N., & Daniels, M. H. (1983). A study of Hogan's
nitive variables associated with personal problem-solving ap- model of counselor development and supervision. Journal of
praisal: Implications for counseling. Journal of Counseling Counseling Psychology, 30, 235-244.
Psychology, 30, 537-545. Royalty, G. M., Gelso, C. 1., Mallinckrodt, B., & Garrett, K.
Highlen, P. S., & Hill, C. E. (1984). Factors affecting client D. (1986). The environment and the student in Counseling
change in individual counseling: Current status and theoretical Psychology. The Counseling Psychologist, /4, 9-30.
speculations. InS. D. Brown&R. W. Lent (Eds.), Handbook Rugg, H. (1963). Imagination. New York: Harper & Row.
of counseling psychology (pp. 334-396). New York: Wiley- Simonton, D. K. (1983). Formal education, eminence, and dog-
Interscience. matism: The curvilinear relationship. Journal of Creative Be-
Hill, C. E., & O'Grady, K. E. (1985). A list of therapist inten- havior, 17, 149-162.
PART IV
Applications
Part IV of the volume is, by far, the largest of our sections. Three chapters are devoted to various aspects of
increasing creativity in writing. Another three chapters focus on creativity in science and social science,
and the last two chapters examine procedures whereby creativity may be facilitated at different stages of
life.
CHAPTER 17
Foundations for Creativity in the

Writing Process
RHETORICAL REPRESENTATIONS OF ILL-DEFINED PROBLEMS
Linda J. Carey and Linda Flower
Introduction: An Approach to Studying presents an innovative solution to a significant rhe-

Creativity in Writing torical problem.
In this situational view of creativity, it becomes
A creative act is usually defined as one that has a important to ask not just how did the "genius"
valuable or interesting product and that is in some class of writers produce traditionally acknowledged
way original or surprising (Hayes, 1981). How- (i.e., literary) works, but how do people in that
ever, whether we characterize a particular act as larger class of "effective" writers produce a cre-
"creative" clearly depends on the context or cir- ative response to a given rhetorical problem? Cre-
cumstances in which it takes place. For example, ative responses to the rhetorical problems raised by
we evaluate the creativity of a child's drawing using school, professions, and public life are necessary
different criteria from those we would apply to a on a regular basis for these groups to sustain what
painting by Monet; a creative act may be enriching others see as quality work. This source of practical
to one individual or it may have earth-shaking con- creativity, in the face of significant but not infre-
sequences. Although creativity in writing is popu- quent rhetorical problems, is an asset for society
larly associated with literary genres, other genres, that schools, professions, and public groups have to
such as expository writing, also offer opportunities depend upon.
for creative products. For example, a research re- Against this backdrop of practical creativity,
port, a proposal, or a magazine article could be which is demonstrated by a wide range of exposito-
judged creative if it presents information in a new ry writers facing a variety of rhetorical problems,
and valuable way to meet the needs and constraints we want to look at the individual writer at work and
of its audience and purpose-that is, if the text ask: What are the cognitive processes in expository
writing that produce, or at least create an oppor-
Linda J. Carey and Linda Flower • Center for the Study of tunity for, a creative response?
Writing, Carnegie-Mellon University, Pittsburgh, PA 15238. The cognitive research we will discuss does not
283
284 PART IV • APPLICATIONS
try to predict or even account for creativity per se, gaps" of the problem with specialist knowledge;
which, we would argue, has multiple sources. Fur- each problem-solver's solution will be unique be-
thermore, the studies we review were not explicitly cause it reflects the solver's own unique knowledge
designed to measure creativity itself but rather to and values (Hayes, 1981). Thus, the very nature of
analyze features of expertise. What this research an ill-defined problem stimulates creativity in the
does show us is how certain features of the writing problem solver. Many problems in architecture, in
process itself (viz., working with ill-defined prob- design, and in music would certainly fall under this
lems, task representation, integrating topic and rhe- umbrella of an ill-defined problem (see Reitman,
torical knowledge, and strategies for global revi- 1964; Simon, 1973) and these are, of course, areas
sion) provide a cognitive mechanism that has the in which we would expect to find expressions of
potential to produce a uniquely adaptive response. creativity. Similarly, many writing tasks also pre-
These features operate as an opportunity or an invi- sent a solver with an ill-defined problem and,
tation for creativity-an invitation that writers hence, with a potential for creativity.
often decline. Although all writing tasks are ill-defined in com-
In this chapter, we examine the composing pro- parison with, for example, game problems (such as
cesses of expert writers working in expository gen- the Tower of Hanoi problems discussed by Newell
res. We take a problem-solving perspective (see & Simon, 1972), some tasks are more ill-defined
Newell & Simon, 1972; Simon, 1986) which postu- than others. Such tasks as a routine rejection letter
lates that creativity does not depend on "special" or a lab report for a physics class, for example, may
abilities or on unconscious processes and insights, require little active problem solving if the writer has
but rather on ordinary cognitive processes that are already written several of these before and, hence,
applied in powerful ways. We draw on research into is very familiar with the discourse conventions. The
writers' composing processes and particularly into writer can simply "fill in the blanks" in his or her
their planning and revision processes (Flower, standard outline (or "script") for the task with ap-
Schriver, Carey, Haas, & Hayes, 1987; Hayes, propriate content. Other tasks may only require
Flower, Schriver, Stratman, & Carey, 1988), writers to access their knowledge about a particular
which indicates that when expert writers tackle aca- topic and to reproduce it on paper. (Reporting the
demic or professional expository tasks, they engage minutes of a meeting would be one obvious exam-
in active and complex problem solving in order to ple.) Here we will discuss those expository tasks
define their task and solve their rhetorical problems that do require active problem solving. How do
in unique and interesting ways. good writers develop creative solutions to complex,
In the first section, we discuss how the ill-defined ill-defined writing tasks and why are student writ-
nature of many writing problems and the cognitive ers' texts often disappointingly routine?
processes experts use to solve these problems in- We will address these questions by looking at
teract to provide an opportunity for creative think- three processes which our research suggests are cru-
ing. In the second section, we examine how dif- cial to dealing with ill-defined writing tasks: (1)
ferences in writers' representation of their task can constructing an elaborated and flexible representa-
affect the originality and overall quality of their tion of the task; (2) integrating topic knowledge and
final products; and in the last two sections, we look rhetorical knowledge; and (3) applying and control-
at how writers' planning and revision processes can ling problem-solving strategies.
provide opportunities for working creatively in ex-
pository genres.
Constructing a Representation of the Task
By their very nature, ill-defined problems do not
Creativity and Ill-Defined Problems present solvers with a ready-made task representa-
tion. Rather, in response to vague task specifica-
How do expository writing tasks provide oppor- tions, such as "build a house" or "write a research
tunities for creativity? To address this question, we proposal," a problem solver has to take an active
will follow the lead of Newell, Shaw, and Simon role in defining the boundaries of the problem and
(1964) who propose that a creative act is the act of in specifying a set of goals and criteria for the task.
solving an ill-defined problem. These are problems Thus, when faced with an unfamiliar writing task, a
in which solvers have to define the problem for writer has to construct his or her own unique repre-
themselves and in which they have to "fill in the sentation of the rhetorical problem to be solved
CHAPTER 17 • FOUNDATIONS FOR CREATIVITY IN THE WRITING PROCESS 285
(Flower et at., 1987). Although a topic may be their goals and require them to re-represent the
assigned (as in many college writing tasks), or the problem or to modify a current representation. For
format may be prespecified (there are standard example, in writing a proposal, a writer may real-
guidelines for writing a proposal), the conceptual ize, on reading the introduction, that the purpose of
structure the writer creates around a topic, and the the research may be clear but not persuasive. The
function to which those format features are put, writer may then have to develop a set of unique
reflect the private goals of the writer. goals and subgoals for persuading the reader that
For instance, expert writers spend considerable this particular project is needed and feasible. This
time and attention elaborating a network of goals, process of constructing and reconstructing an image
constraints, and criteria as they compose (Flower et of the task in response to the growing text often
at., 1987). They may draw inferences about the leads writers to the discovery of some unique and
audience and set goals for dealing with it, or they valuable insights about their task (Flower & Hayes,
may translate format features, such as an obligatory 1980) and, hence, provides opportunities for cre-
introduction, into goals. These goals, of course, ative thinking.
could reflect a conventional plan (e.g., better start This phenomenon of building a dynamic task rep-
out with a topic sentence and some background) or a resentation seems to be important to creativity in
uniquely adaptive one (e.g., how about showing domains other than writing. For example, a study
them, in some subtle way, what they appear to as- by Getzels and Csikszentmihalyi (1976) found that
sume?). Setting goals does not guarantee unique artists whose work was rated highly for creativity
goals, but the planning process we have observed often discovered new dimensions to a problem as
goes well beyond a simple selection and transcrip- they worked on it; these creative artists were modi-
tion of topic knowledge. Even on a "normal" aca- fying their problem representation in response to
demic task, such as that faced by the contributors to their emerging drawings. Or, as Perkins (1981)
this volume, there is likely to be enormous variety points out, creative problem solvers remain ready to
in the way the task "write about creativity change their decisions in the light of new knowl-
in ... " is framed and in the top-level goals and edge gained from working on the problem and thus
constraints writers give themselves. These goals, remain open to new insights.
created in interaction with the writer's topic knowl- The process of constructing and modifying a rep-
edge, then determine how much of that knowledge resentation of the task presents inexperienced writ-
is used and how and why it is used. ers with several difficulties that limit their oppor-
Because complex writing tasks are quite literally tunities for creativity. First, novices may simply
constructed by the writers, people with the same "jump into" the problem without spending time
assignments give themselves significantly different and effort on re-representing and discovery. The
rhetorical problems. The very nature of these ill- result may be that they are then forced to work with
defined problems-which writers structure and de- an initial representation which is too abstract or
fine for themselves-then, is an important basis for which does not address important features of the
creativity in writing. Some writers exploit this po- task. Certainly, in our work on planning, we found
tential by giving themselves unique and valued that novice writers tended not to build the complex
problems to solve. In a study of writers' initial task networks of goals and subgoals that typified the
representations (Carey, Flower, Hayes, Schriver, plans of the expert writers (Flower et at., 1987),
& Haas, 1986), we found that the nature of the goals and, as consequence, their texts failed to address
that a writer develops in his or her task representa- important audience needs.
tion may well affect the quality of the final product Second, novices may become committed to one
and provide an opportunity for creative problem- representation of their task and not be open to re-
solving. In a later section, we will discuss how representing the task in light of new discoveries.
qualitative differences in writers' initial task repre- For example, Beach and Eaton (1984) found that
sentations are mirrored by differences in the quality many of their student writers typically relied on the
of their final texts. pattern of organization of the "five-paragraph
Building a representation of an ill-defined prob- theme" and were unable to adapt their texts to the
lem is also a dynamic process (Simon, 1973). Ex- constraints of audience and purpose. Similarly,
pert writers' representations often change as they Britton, Burgess, Martin, McLeod, and Rosen
progress through the task: new facets of the problem (1975) found that students' representations of rhe-
may occur to experts during writing that change torical problems were fixed by their perceptions of
the demands of the teacher and school context, when the leamerltextbook reader needs an orienta-
rather than by their sense of audience and purpose. tion to the field. In this "writer-based prose," writ-
As a result, their texts tended to be formulaic and ers take a cognitively less demanding route of talk-
routine and to lack interesting and new angles or ing to themselves (Flower, 1979). For some, this is
perspectives. an efficient first -draft strategy, and they transform
their text into "reader-based prose" later. The in-
teraction of rhetorical and content knowledge, then,
Integrating Topic Knowledge and Rhetorical
is both a problem that some writers do not solve and
Knowledge
a constraint that generates uniquely adaptive solu-
Another opening for creativity in writing is the tions for other writers.
way in which people manage the constraints of inte-
grating their topic knowledge and rhetorical knowl-
Developing and Applying Problem-Solving
edge. Simply put, writers must often manipulate or
Strategies
transform what they know to meet the constraints of
a unique rhetorical situation. If the subject is com- As we have seen, ill-defined problems require a
plex or new, writers will not have available pre- solver to manage both an evolving set of goals and
packaged or organized pieces of information that constraints and a large body of knowledge. To ac-
can simply be slotted into the text. Rather, they will complish this management task successfully, a
have to engage in difficult knowledge-transforming problem solver needs to use strategies for cutting
operations to adapt what they know to meet the down the search process (Simon, 1973). Our re-
rhetorical goals of, for example, involving and in- search indicates that expert writers have a wider
teresting a particular audience (Scardamalia & Be- repertoire of such strategies than do inexperienced
reiter, 1988). This "juggling act" between two dif- writers. For example, we found that our expert
ferent knowledge domains-content knowledge planners had several strategies for resolving goal
(i.e., about the substance of the text) and rhetorical conflicts and for consolidating information; our
knowledge (i.e., about the constraints of audience, novices, on the other hand, often simply sidestep-
genre, and purpose) presents an opportunity for ex- ped these problems and, as a result, missed oppor-
pert writers to restructure their knowledge in in- tunities for developing promising aspects of the text
sightful ways and thus provides a potential for cre- (Flower et at., 1987). Similarly, expert revisers
ative expression. seem better equipped to deal with difficult global
However, this juggling act appears to present dif- revision problems because their strategies are more
ficulties for developing writers. As Scardamalia efficient and more flexible; in contrast, novices fre-
and Bereiter's study indicates, inexperienced writ- quently rely on a time-consuming and risky trial-
ers often take a "knowledge-telling" approach to and-error procedure of rewriting the text until it
expository writing; they simply write all they know "sounds better" (Hayes et at., 1988). (We will
about a topic without considering such rhetorical discuss this work on revision in more detail in the
features as audience and purpose. Similarly, Lan- final section.)
ger (1984) found that having a lot of information on Strategies, like some of the other features we
a topic did not necessarily help students to write have discussed, seem to operate alternatively as a
coherently on that topic; they were unable to struc- path or a roadblock to creativity. Rose (1984), doc-
ture their knowledge into an appropriate rhetorical umented some of the ways in which the possibility
pattern to develop, for example, a cause/effect es- for creativity is shut down by the rigid rules and
say. In our studies of planning, we found that writ- strategies that students bring to writing. Student
ers whose plans mainly consisted of "content writers who regularly blocked or failed to do
plans" produced texts that were less well adapted assignments voiced absolutist assumptions about
for audience and purpose than writers who inte- how the writing process should operate (e.g., as a
grated content and rhetorical planning (Carey et at., spontaneous and elevated act in which formal pre-
1986). planning is inappropriate), whereas nonblockers
Adults who are capable of adapting what they recognized a variety of acceptable approaches
know to what a reader needs or wants often fail to do determined by the context. The high-blocking
so; they report information when the reader needs a students also invoked a variety of rigid "rules" to
recommendation; they define technical concepts which a writer must adhere (although they did not
agree on the rules). Some of these rules demanded writer creates can, of course, be simple and/ or con-
concentration on the surface features of text rather ventional, but it can also be an elaborated and
than on the conceptual structure, for instance, and unique construction. Hence, the process of writing
led to premature editing. As one student put it, "I itself opens the door for creative cognitive acts.
write with the thought . . . that this is going to be Also, the recursive nature of the writing process
it . . . so it had better be good the first time observed in studies of planning and revision allows,
through" (p. 73). Finally, some of the limited strat- and in fact calls for, re-representation. As writers
egies students brought to writing, such as depend- generate information through the activity of com-
ing on the five-paragraph theme, allowed only lim- posing itself, they also perceive new goals and con-
ited structures of thought, and they led to straints, and their image of their task is itself fre-
"incremental" planning in which students planned quently open to transformation. The process opens
and composed in small, unconnected segments. a door, but whether or not writers do indeed inte-
The strategies, assumptions, and rigid rules of grate this growing network of goals and topic
the high-blocking students seemed, then, to inter- knowledge into a coherent, much less unique con-
fere with their day-to-day functioning as writers in figuration is a question we will look at in the context
the university. What is interesting for our purpose is of specific parts of the process.
that this maladaptive approach to writing seems to In tune with our focus on practical creativity, we
operate by shutting down the processes that can are looking at the way the "normal" processes of
foster creativity-the processes of integrating and composing can lead to creative results. We are as-
reintegrating information and of taking a flexible, suming, as has Simon, that creative acts do not
context-dependent approach to managing strat- depend on "special" or extraordinary basic pro-
egies. For high-blockers, the constraints that lead cesses. In the research we are about to describe, we
creative writers to unique, adaptive plans have cannot assert that the written products of the sub-
ceased to be generative. jects studied would be judged as creative, because
Having a repertoire of strategies may not be suffi- public judgment of uniqueness plus value is our
cient to ensure success in solving ill-defined writing standard. The research did not set itself up to study
problems; a writer must also know when to use a creativity per se but expert performance. So the
particular strategy and be able to monitor and test its conclusions one can draw are necessarily limited.
effectiveness. This metacogitive ability-that is, What we do propose to do is to look more closely at
the ability to regulate one's own strategic action how writers tackle these difficult processes of
(see Brown, 1980)-allows writers to have more building a representation of a writing task, of inte-
control over their own processes. In our revision grating topic and rhetorical knowledge, and of de-
research (Hayes et al .. 1988), we found that expert veloping and applying problem-solving strategies.
revisers made conscious decisions about which We examine these processes in the context of three
strategies to use, based on such criteria as the nature different kinds of professional and college writing
and density of the problems in the text and the prag- assignments: a task that required interpretation and
matic constraints of the task (e.g., the amount of synthesis of reading materials; one that involved
time available.) The novice writers in our study, planning and writing a short expository text; and
however, often failed to deal adequately with text one that was essentially a revision task.
problems because they were inflexible and kept The processes we have chosen have three
with one strategy, even if it was obviously not qualities: they are generative and integrative pro-
working for them. Similarly, research on reading, cesses and have the potential to affect the top levels
summarizing, and learning (Brown & Smiley, of a writer's goal structure. These processes do not,
1978; Brown, Day, & Jones, 1983), suggests that in themselves, produce a creative response to a rhe-
novices' lack of conscious regulation restricts their torical problem, but they do provide, we will argue,
ability to use effective comprehension strategies, the operational foundation for responses that are
such as rereading, making inferences, and extract- both novel and highly adaptive to the context for
ing gists, even if they are aware of these strategies. writing. From an educational point of view, these
To sum up, from the perspective we are taking, seem to be aspects of the writing process that writ-
many writing tasks are best described as ill-defined ers learn to manage with some difficulty, that lead
problems in which writers construct their own net- to effective writing, and that have the potential, as
work of goals and plans. The initial representation a high-level generative and integrative processes, to
lead to that special effect we label creative. ic of time management, rariging from Alan Lakein
on how to "take control of your time and your
life," to advice from Cornell's study skills center,
The Effect of Task Representation on to William James's comments on working through
Originality fatigue. The assignment was designed to include all
the sacred words of college assignments: synthe-
The practical creativity on which we concentrate size, interpret, be comprehensive, and the essay
in this chapter requires an original but highly adap- was referred to at the end of the readings as "your
tive response to a rhetorical situation. A creative statement. " The intent was to create a Rorschach
solution has to function within a considerable set of assignment that would allow students to examine
constraints set by the context. For instance, writing their own process and interpretation of this college
in these situations is often in response to an assign- assignment. On the day the assignment was due,
ment that is posed by an organization, a manager, or students did a self-analysis of how they had viewed
a teacher. The study by Britton et al. (1975) of the task in terms of the information source they
2,000 samples of school writing in Great Britain, used, the format they thought appropriate, their
grades 6-12, suggests that school assignments may overall orgallizing plan, and some of their dominant
severely limit opportunities for both originality and strategies and goals. This analysis was supported
imaginative adaptation. By grade 12, 61 % of the by a checklist based on the results of three prior
scripts studied were categorized as papers written pilot studies that had shown us a wide range of
by "pupil to examiner," "as a demonstration of response to two parallel versions of this assign-
material mastered or as evidence of ability to take ment. Certain features of these responses are rele-
up a certain kind of style" (p. 122). The terms vant to creativity.
assigned or school writing are sometimes used as For this assignment, 43% of the freshmen stu-
synonyms for noncreative ahd purposeless exposi- dents organized their paper as either a summary of
tion. Sadly enough this may be an apt description the source texts or as a review and comment-a
for how some teachers use writing (as a test for plan in which the source texts provide a conceptual
content knowledge or correctness) and how their frame and substance to which the writer can add a
students learn to see assigned writing. However, a commentary at selected spots (Ackerman, in press).
look at the writing processes of college students Only 25% of the students attempted a synthesis in
suggests that assigned writing is, like any other which their essay was organized around a control-
constraint, subject to interpretation. The task the ling concept that gave structure and coherence to
writer gives him or herself may be a much better their discussion. Students who gave themselves the
predictor of creativity than whether it is assigned task of synthesis had to invent a concept that could
and or highly constrained. make sense of these data, even though, like normal
As a case in point, we can take a close look at the library research, the information did not fall into a
task representation process itself, as it occurred in a simple or obvious pattern; and they had, at least in
large group of college writers writing a paper on an theory, to deal with the contradictory claims of
assigned body of readings. these "authorities." An even smaller group, 11 %,
said that they treated the paper as an interpretation
Alternative Representations of a Standard of the source texts, organized around a purpose of
College Task their own. To use this plan, writers had to imagine a
reader who would find the information useful, or to
In this study, four sections of a freshman class see an issue or problem to be examined, and use that
were given a standard, open-ended college assign- purpose to organize an interpretation of the relevant
ment that read: material. This plan typically makes heavy use of the
Here are some notes, including research results and observa- writer's own ideas.
tions, on time management. Your task is to read and interpret this When asked to describe their own strategies, 2%
data in order to make a brief (1-2 page), comprehensive state- of the freshmen said that they used the strategy of
ment about this subject. Your statement should interpret and
"adapting to the reader"; 6% chose to "use the text
synthesize all of the relevant findings in the text.
for my own purpose." However, when they were
The text was two pages of notes and conflicting asked to predict what the experienced, masters level
claims taken from various "authorities" on the top- students had done when they did this task, the fig-
ures for these adaptive strategies jumped to 24% seemed very close to the limited "pupil to exam-
and 15%. It seems that the students themselves saw iner" recitation of knowledge that Britton's group
that the strategies one used were not a function of observed and deplored.
the assignment but of the writer. Some students In contrast to Martha, Kate, a beginning graduate
who commented on this difference attributed it to student in professional writing, concentrated on the
the greater freedom they perceive master's students way the audience determined the focus of the paper
to have. and defined what information was' 'relevant. " The
Finally, the checklist, though naturally in- task she gave herself required adapting her reading
complete, also listed a number of goals previous to what she imagined her potential readers would be
writers had mentioned. The freshmen students interested in. The following protocol shows us how
identified their goals with choices, such as "pre- introducing this additional constraint of audience
senting learning" (20%), "covering key points" solved the problem of generating enough to say:
(18%), and "do the minimum" (13%). Zero per-
cent checked the goal of "creativity." Taken to- Hmm. Let's see. Ok. I don't see how I'm going to get 1-2 pages
gether, these analyses show some students giving out of studies and experts that pretty much back each other up.
Damn. I wish I'd never sold my acoustics textbook [reference to
themselves tasks in which a creative response is
an interruption by a roommate.] Well, who are my readers?
unlikely to occur. These data come from the stu- About 15 people who mostly teach the freshmen course ....
dents' own self-analyses; the data based on judges' Yes, so why would anyone care to hear about this? ... So they
analyses of the texts showed even lower frequency want to know how to write better, but maybe more importantly
of syntheses and interpretations. However, these how to teach better. Oh yes, and Linda Flower, but there's no
doubt she's interested in that to. Ok. They're my primary au-
self-reports are interesting because there were clear dience, so ....
cases in which students attempted more ambitious
tasks, even though the judges were unable to per- As a result of these goals, Kate generates ideas
ceive the result of that effort in the written text. The not found in the source and begins her text by telling
thinking-aloud protocols and the in-class presenta- readers two ways in which her information might be
tions of three of these students can give us a more of use to them. This view of writing as an adaptive
explicit view of how these interpretations were cre- enterprise is apparently not new to Kate. In her own
ated (Flower, in press). presentation and in response to Martha's very dif-
Martha was a junior engineering major in one of ferent plan, Kate remarks that, as a past economics
the pilot studies. Students in the class had done a major, she even has a private formula for her pro-
thinking-aloud protocol of themselves as they read cedure: T = I(A). The topic or information she uses
the source texts and wrote their essays. The pro- is a function of the audience.
tocols let them look closely at their own cognition Both Kate and Martha appeared to bring well-
and prepare an in-class presentation on an "in- learned plans for interpreting writing tasks with
teresting feature" of their own writing process. them. And in one of the pilots, 50% of the students
Martha's presentation focused on her own very said that they paid no attention to the assignment
clear procedure for doing the task. She used what itself but invoked their standard paper-writing strat-
the class came to call the "gist-and-list" strategy: egies. In light of the clear variety among these rep-
you read through the text with some care, find the resentations, one wonders how often the "stan-
key words in each paragraph, and summarize it dard" strategy is the optimal one for all these
trying to capture the main idea. You then write a writers. Our final example shows a student actually
paper organized around this set of gists. The most negotiating a decision point in her task representa-
interesting feature of her plan was its caution rule: tion and considering the costs and benefits of alter-
Sometimes a new idea will occur to you as you are native interpretations.
writing-a new connection, an insight, or new way
to organize. If that happens, you must decisively "Interpret and synthesize" [student is rereading the
ignore this possibility for it will only confuse you assignment].
and your paper. Martha's representation of the task What the hell does that mean? Synthesize means to pull together,
placed priority on accurate summarization, effi- no, to make something up. Why should I want to make some-
thing up?
ciency, and coherence. Furthermore, this was a [She then rereads, commenting on the wording of the assign-
practiced strategy for her, "just like doing a re- ment. ... ]
search paper." The task Martha gave herself Synthesis sounds like I'm making a chemical compound. Hmm.
Put together. closing it in others. In the next section, we will look

[Re-reads1 •• All of the relevant findings in the text. ' , more closely at the content of the initial task repre-
How can I do this?
sentations of a group of expert and novice writers.
At this point the writer apparently decides that, in

fact, she does not want "to make something up"
and begins to summarize the readings.
It is not surprising that different representations Integrating Topic and Rhetorical
of a writing task could have the effect of fostering or Knowledge in Initial Planning
discouraging original thinking and an adaptive re-
sponse to the needs of readers. The striking feature A writer's task representation can, as we have
of these protocols and presentations is that the stu- seen, open up possibilities for originality, or con-
dents who held these different representations ap- fine a writer to producing a routine, standard text.
peared to assume that they were simply doing the To develop an effective representation of an ill-
task as it was assigned, rather than constructing an defined task, writers need to adapt their knowledge
interpretation, and were surprised to discover the about the topic to meet the constraints of audience
range of options others considered possible-if not and purpose. This integration of content and rhe-
actually mandated by the assignment. Students who torical knowledge can provide unique opportunities
were limiting the likelihood of creativity may have for experienced writers, whereas for others it can be
done so in the name of least effort, but they ap- prohibitively difficult. In this section, we will take a
peared to do so in the faith that this was also closer look at the process of building a task repre-
expected. sentation of a typical ill-defined professional writ-
Of course, it does not follow that even students ing task. What qualitative differences do we see
who set the goals of being creative will be able to do between experts' and novices' task representations
so. One student in the pilot class said that she habit- that might help account for experts' practical
ually set high standards for creativity in her own creativity and students' run-of-the-mill responses to
writing. However, as the protocols made clear, this many academic or professional writing tasks?
goal actually functioned as a test that she applied In the study we will discuss here (Flower et al.,
during composing to fledgling ideas and initial bits 1987), five experienced composition teachers and
of prose, which generally failed to pass the test of seven student writers with varying degrees of skill
this rather harsh and premature internal critic. Set- were given an expository assignment in which they
ting the goal did not seem to promote creativity were asked to write a short article describing their
itself. job for Seventeen magazine, that is, for an audience
To sum up, on a reasonably complex task, such of girls aged about thirteen or fourteen. The task
as these college assignments, task representation was ill-defined in that each writer could draw on
becomes a critical part of the process. The goals and and adapt a unique body of personal knowledge
constraints that students invoke may have sources about her job and develop a unique set of goals to
in unquestioned past experience or in active, in- reach the young readers of the article. Clear! y, there
ferential efforts at representing features of the task; was a great variety of ways of approaching this task
however, the qualitative differences in these repre- and, hence, a potential for interesting, adaptive so-
sentations point to their constructed nature. Some lutions to a complex rhetorical problem. Our writ-
of these representations shut down the possibility of ers on this task had to do what many professionals
creativity by giving low priority to the writer's own who write for lay audiences have to do:(l) deter-
ideas, by specifying a linear composing process un- mine their own goals, including representing the
deflected by those discoveries writing itself engen- audience to themselves and deciding how to meet
ders, and by eliminating the need for an adaptive the audience's needs, and (2) decide what knowl-
transformation of knowledge and treating writing as edge was relevant, given their goals. For many
knowledge-rehearsal rather than as a potentially writers, even the question of defining their job was
useful, rhetorical act. Because task representation open to debate. Their answer depended, in part, on
in writing is (1) a highly interpretive process but (2) their image of the audience and the goals they set for
often unrecognized and, hence, closed to critical this essay. And that decision, in tum, appeared to
examination, it appears to play an important role in depend partly on what information would be easy to
opening the door for creativity in some cases and access.
We looked at our writers' initial plans, taken teachers came up with a structure that mir-
from verbal protocols of our subjects' working on rored the important aspects of his job as a
our task. (By initial plans, we mean plans articu- writing teacher; "What about a sequence
lated before writing a first sentence, see Carey et ... (a writing teacher) reads papers, makes
al., 1986.) We found that even though all our writ- up assignments . . . I want to start with what
ers developed initial plans for content (i.e., for what they think is obvious."
information they should include), our most suc- 4. Other rhetorical goals: In this category, we
cessful writers also developed quite complex found diverse plans for addressing this partic-
rhetorical plans for their text (e.g., plans related to ular audience, for projecting the writer's per-
audience, to overall purpose, or to organizational sona, for the language and tone of the text.
structure). Although our writers' specific rhetorical For example, given the needs of his readers,
goals and plans for the task were diverse (some one writer decided that' 'the tone and style in
were personal or even idiosyncratic, e.g., I must Seventeen ... should be light and lively,
change their minds about what an English teacher filled with slang. "
does), we found that our writers' rhetorical plans
included information in these four major categories:
The writers in our study who produced the most
1. Overall purpose or "theme": This category successful and innovative texts (in general, but not
encompassed goals and plans for what the es- exclusively, the expert writing teachers) were those
say should accomplish, for an overall focus, who developed plans in all these categories before
or for a unifying idea around which other beginning to write. These writers were able to build
ideas could be developed. For example, one what we have termed a rhetorical representation of
writing teacher came up with a top-level goal their task that provided an overall framework for
to focus on "how a teacher differs from a generating and selecting information to include in
professor"; another, on using this essay to the text. They integrated topic knowledge with rhe-
"raise (students') horizons and help them to torical concerns and interrelated rhetorical goals so
examine their own future." that, for example, they saw purpose as closely con-
2. Audience: These were plans that developed a nected with audience needs. This rhetorical repre-
representation of the audience for the text. sentation provided a unifying and coherent "theo-
For example, one writer represented her au- ry" of the task as they defined it for themselves.
dience as "people like myself, or people like Let us take a look in more detail at how one of our
I was, but adjusted for twenty years"; another most successful writers built a rhetorical represen-
subject considered their current interests: tation of the Seventeen magazine task. This writer
, 'they're all in school, they're taking English, produced an innovative text that provided an in-
for many of them English will be a favorite teresting angle on his job as a college writing teach-
subject." Many of our experienced writers, er and that was particularly appropriate to the in-
in fact, actively struggled with alternative terests and needs of his readers. The excerpts, taken
views of their readers. from a verbal protocol ofthis writer as he was work-
3. Structure: In this category, writers developed ing on our task, include some of the subject's most
goals and plans for organizing or structuring important initial planning episodes and our catego-
the text. For example, one of our writing rization of the goals and plans he is developing.
Clause number Analysis

Episode 1
16 Job-English teacher rather than professor. Content
18 In fact that might be a useful thing to to focus Theme
on. Defines focus of text
19 ... how a teacher differs from a professor.
20 And I see myself as a teacher.
21 That might help my audience to reconsider their Theme/Goal
notion of what an English teacher does. Sets top-level goal
23 (Reads) "young female teenage audience."

24 They will all have had English. Audience
25 Audience-they're all in school. Represents relevant background
26 They're taking English. of readers
27 For many of them English may be a favorite
subject.
29 But for the wrong reason-some of them may have Audience
the wrong reasons in that English is good Represents attitudes
30 because it's tidy.
31 Can be a neat tidy little girl.
32 Others turned off because it seems too prim.
33 By God, I can change that notion for them. Theme/Goal
Episode 2
45 All right I'm an English teacher.
47 I know they are not going to be disposed to hear Audience
what I'm saying. Represents attitudes
48 Partly for that reason and partly to put them in the
right-the kind of frame of mind, Goal: Prepare audience
49 I want to open with an implied question or a direct
one, Structure
50 and put them in the middle of some situation, Develops skeleton structure
51 then expand from there to talk about my job more
generally,
52 and try to tie it in with their interest. Goal: Involve audience
53 So one question is where to begin. Structure
54 Start in the middle of-probably the first day of Plans introduction
class.
55 They'd be interested.
56 They'd probably clue into that easily Audience
57 because they would identify with the first days of Represents a shared reference
school,
58 and my first days are raucous affairs.
59 It would immediately shake 'em up Goal
60 and get them to think in a different context. Develops audience goals
What features do we see in these excerpts that Second, the goals that he generates provide an
would lend support to our hypothesis that experts integrated rhetorical framework for his planning.
build a rhetorical task representation? Instead of piecemeal idea generation or brainstorm-
First, this writer's task representation is rich in ing, we see this writer generating and organizing
rhetorical information. In these episodes, he devel- content to meet his particular persuasive purpose.
ops plans in all the rhetorical categories discussed For example, he starts out with a scenario that will
above. Not only does he develop plans for specific put the audience in the right frame of mind and will
content to include in his text, but he also develops a help them to think about English teaching in a dif-
theme, a partial structure, a quite detailed represen- ferent context. His text plans are thus adapted to fit
tation of his audience, and a set of task-specific with his guiding focus. In addition, many of his
goals. For example, he develops a representation of goals interact with each other, and two goals may
the audience that includes their background (they be instantiated by a single text plan. For example,
take English in school and may enjoy it) and their he comes up with the idea to talk about his first
attitudes (they are not going to be disposed to hear day of class because this would further his goal to
what I'm saying.) From this representation he de- shake up his audience and because this would be
velops a set of goals for his audience-to "shake something his readers could identify with. Thus,
them up" and make them think in a different this one text plan instantiates two of his important
context. goals.
In sum, this writer is generating a set of goals and What alternatives to a rhetorical representation
plans that make this task uniquely his own. Al- might writers use? We found in our study that several
though the task instructions provide him with some of our inexperienced writers focused their initial
loosely defined goals and constraints (the topic, the planning almost exclusively on generating ideas
genre, and the audience), he uses the umbrella of a about what content to include and that they ignored
rhetorical framework to adapt, elaborate, and in- many of the rhetorical features that we found in our
stantiate these goals. The task that he ends up doing experts' initial plans. Thus, they engaged in a type of
is very much a task that he himself has created. It is "knowledge-driven" planning without transform-
a task that results from a coherent theory of the task ing their knowledge to meet the constraints of au-
based on rhetorical principles and concepts. dience and purpose. Below are some early protocol
Clause number Analysis

Episode 1
6 I'm going to assume here I'm an engineer. Content
8 I guess a research-a research engineer. Defines subject matter of text
10 However, there's a graduate student Content (to end of Episode I)
II because I'm a graduate student. Explores different aspects of job
12 So ... really my job is I'm going to school.
13 Let's see-I'm a graduate student,
14 and I'm an engineer,
15 and I'm a research engineer.
16 So everything is there.
17 So maybe I should explain here instead is that I'm
a graduate student pursuing a Ph.D. in engineering.
20 I do research work.
22 I teach a course.
Episode 2
24 (Reads) "for a thirteen to fourteen teenage
audience"
25 So we have to address the fact that this girl is Audience
seventh or eighth grade.
27 Okay so these are all the things going through my Content
head relative to engineering-research engineering. Reviews possible
28 These are all the things I'm supposedly doing details to include
29 and I'm pursuing a Ph.D.
30 Working on a thesis.
32 My job here is wrong, the way I'm interpreting my
job.
34 The way I'm interpreting my job means what I'm Content
doing with my life at this moment in time. Redefines topic
37 The girl is approximately in seventh or eighth
grade.
38 ... the assignment has to appeal to a broad range Goals
in intellect. Sets very general audience
goals
39 It must explain simply what I am doing.
43 I have to generate an essay. Structure
44 We'll assume it's about two pages. Specifies genre and length
48 That's not really hard.
49 I really have my first line Process comments
50 so I'm going to rip off the page here.
52 I think I can write this out very quickly.
episodes from one student writer who was working purpose) than those who did not. In fact, our writers
on our Seventeen magazine task. This writer's final whose plans were judged to be rich in rhetorical
text was rated eighth out of twelve for quality. information were judged independently to have the
This student's task representation looks different best texts (see Carey et al., 1986). (There was a
from the expert's in several ways. First, the subject high positive correlation, r = .874, between the
generates few rhetorical goals in contrast to the ex- quality of initial planning, in terms of the rhetorical
pert's rich rhetorical planning. Rather, most of the features discussed earlier, and text quality.) In addi-
planning is related to features of the topic, for ex- tion, we found that:
ample, the main aspects of his job as a graduate
student in engineering and whether being a re- 1. Writers who developed plans in all the rhe-
searcher is in fact his job. He includes no goals for torical categories of audience, purpose,
an overall theme or focus to the text; little informa- theme, and other rhetorical features, as well
tion about organization beyond a very general sense as content plans, received significantly higher
of the genre and the length; almost no information scores for their final text (p = .0149, by the
on his audience's interests or characteristics, be- Mann-Whitney test) than writers whose plan-
yond that provided in the task instructions; and no ning failed to recognize one or more of these
goals for the text beyond a rather vague sense that categories; only one of the writers whose text
he should "appeal to a broad range in intellect" and received a score in the lower half of the dis-
"keep the text simple. " The very general rhetorical tribution of scores covered all these catego-
goals that he does come up with (e.g., address the ries in his initial plans. In addition, those writ-
fact that this girl is in the 7th or 8th grade) are not ers whose plans included goals for an overall
further instantiated with subgoals and text plans, focus or "theme" for the text received signif-
and he is unable to build upon them. icantly higher scores than those who did not
Second, we do not see the integrated overall (p = .0185). Our weaker writers demon-
framework to guide his planning process. His plan- strated a lack of concern for rhetorical fea-
ning does not seem to be organized by rhetorical tures either because they did not have the ap-
features but rather generated haphazardly by fea- propriate rhetorical knowledge, or because
tures of the content with which he is struggling. they did not realize the importance of rhe-
Thus, unable to transform his content to meet the torical planning.
constraints of the assignment and unable to build his 2. Conversely, the writers who produced the
own unique set of rhetorical goals for his task, this more successful texts had a lower percentage
writer has little option but to "knowledge-tell." of content plans than did writers who pro-
His text reflects this limitation as he produces a duced the poorer texts. (We found a negative
rather dry description of energy research, which our correlation, r = - .366, between the amount
raters judged to be low in meeting the needs of the of content planning and text quality.) This
audience. suggests that writers who focus largely on
In sum, without any unifying theory of his task, planning specific content before they begin to
this writer has little choice other than to plunge in write may be missing and/or ignoring impor-
and write, letting his topic information drive his tant rhetorical goals and constraints, such as
planning and generating process. Although he does adapting a text for the audience.
plan more extensively than many of our other sub-
jects, his plans do not help him to carve out a very We want to be cautious about drawing inferences
appropriate and effective representation of his task. about creativity from differences these studies have
The two writers we have looked at in detail were shown us between expert and novice writers. How-
typical in many ways of our writers in this study. ever, we have tried to isolate certain features of the
Several of our subjects built quite complex rhe- writing process that lay a foundation for original
torical representations of the task, analogous to the and useful solutions to rhetorical problems.
one in our first case study, whereas others focused Because creativity is a uniquely adaptive re-
almost exclusively on content plans, as did the writ- sponse to such a problem, it seems important that
er in the second case. Overall, we found that those some writers base their effort on first exploring
writers who engaged in rhetorical planning before multiple dimensions of the rhetorical problem they
writing produced texts that were more successful face. In response to that situation, they appear to
(particularly in terms of meeting the needs of the develop a unifying theory of the task that integrates
audience and in developing an effective rhetorical their knowledge with their goals. Insightful adap-
tive effort is an important part of creativity, but it selves ill-defined in that there are many different
does not, of course, necessarily involve originality. goals and sub goals that a writer might develop to
However, these writers also showed us one of the guide his or her revision and many possible solu-
foundations of such inventiveness. Instead of rely- tions to the problem. The process of solving global
ing on their prior knowledge and resorting to direct revision problems widens a writer's options and
knowledge-telling or relying on standard schemata thus increases his or her potential for producing an
or conventions, writers who develop a dynamic and effective and innovative final version of a text.
rhetorical representation of their task plunged into In our study of revision (Hayes et aI., 1988), we
extended constructive planning. They not only gen- found that one important revision process in the
erated fresh inferences, new goals, and adaptive repertoire of our expert writers (and a stumbling
plans but also created a unique configuration of block for many of our novices) was the ability to
such information adapted to the entire problem as diagnose successfully global text problems. We
they perceived it. Although the process will not characterize diagnosis as building a representation
ensure the product, and we must not discount other of a text problem that provides a writer with some
factors, including dolfhain knowledge, integrative specific information about the type or category of
rhetorical planning offers a strong cognitive foun- problem and some possible strategies for improving
dation for rhetorical c~eativity. the text (Flower, Carey, & Hayes, 1985). Some-
times when evaluating a text (either their own or
one written by another person) writers will sense
that" something does not sound right," that is, they
detect a potential trouble spot. If a writer goes on to
expand upon such an intuition and explain why the
Developing and Applying Problem- text is poor by coming up with a name or category
Solving Strategies for Revision for the problem, then he or she is making a fully
fledged diagnosis. A diagnosis differs from a detec-
Writers who come up with innovative solutions tion (e.g., "this is terrible' ') in much the same way
to ill-defined tasks rarely do so without making that a patient's heartfelt but weakly specified com-
substantial revisions to their text; in fact, the kind of plaint differs from a physician's more multidimen-
practical creativity we are discussing in this chapter sional, high-information diagnosis, involving a
often requires a writer to spend considerable time problem, symptoms, and solutions. The power of
and effort on improving. and sometimes rethinking, diagnosis resides in identifying the situation as a
a text before arriving at a final version. For experi- particular problem type, which, in tum, activates a
enced writers, revising may involve' 'reseeing" the body of tests and solution procedures. Thus, diag-
text on a global level and making major changes to nosis allows for strategic revision.
the meaning and to the overall structure (Faigley & We will now look at some case studies of writers
Witte, 1981; Sommers, 1980). Our research on the diagnosing and solving ill-defined text problems.
cognitive processes of revision (Hayes et at., 1988) We draw on data from a study of expert and novice
indicates that making these kinds of substantial text writers who were given the task of revising a poorly
changes involves complex problemcsolving ac- written rather pedestrian memo from one athletics
tivities that depend on strategies for pruning down coach to another about women's sports oppor-
the many options for the text into a productive set of tunities on campus. Our subjects were asked to pro-
alternatives, and for choosing between these alter- duce a revision of the text that could be given to
natives. In this section, we look in more depth at the freshman women students as a one-page handout to
kinds of strategies writers develop for revising an introduce them to campus sports (see Hayes et at.,
expository text. 1988). Again, this was a fairly typical professional
Text problems that are global in scope (i.e., en- writing task in that, as well as solving language and
compassing the whole text or a substantial propor- organization problems, our writers needed to
tion of it) or that relate to broader and less clear-cut change the overall focus of the text and adapt it for a
(i.e., global) issues, such as audience and purpose, new audience who might not be very motivated to
invariably have no easy answers. Both these types read it. We will examine in some depth our writers'
of global problems require more than the quick fix diagnoses in relation to a particularly problematic
procedures that are needed for correcting routine section of the text that was part of larger, whole-text
mechanical or grammatical errors; rather, they are coherence and audience problems, and that also
difficult for writers to solve because they are them- contained several fairly local problems, such as
faulty topic focus, negative tone and structure, nections between diagnoses, either at the same text
faulty parallelism, and wordiness. This section of level or across text levels, the lines on the figures
the original text read as follows: indicate these connections.
Let us take a closer look at our expert's diagnoses
I don't want to infer that the only chance women get for par- in Figure 1. In order to build a representation of the
ticipating in sports is on varsity teams. Intra-mural sports are not
the same as varsity sports in which the rules are better, equipment
major problems with this excerpt, the expert (BA)
is better, with the techniques of the players being more devel- relates her concept of the overall purpose of the
oped. Irregardless, 1M sports may be the choice for many wom- memo (to recruit for teams in order to lobby for
en-they can be just as much fun and take less time. funding-P44-46) with paragraph level diagnoses
that focus on: missing information (P43); il-
Our writers represented the problems with this logicality in the argument structure (if the memo is
excerpt of text in qualitatively different ways. First, recruiting for varsity teams, why talk about intra-
the experts tended to build integrated networks of mural teams (P47-48); and unhelpful and uncon-
diagnoses to deal with global text problems. They vincing information about intramural sports that has
noticed connections between independent problems no purpose (P54-55, 57, 59). She successfully di-
at all levels of the text: they saw individual choices agnoses some of the individual local problems with-
of diction, style, and content in terms of a larger in this paragraph; however, these diagnoses are
multi sentence context; they saw concrete evidence closely related to each other and to a larger whole-
of broad global problems in smaller problems. Con- text problem. By creating this hierarchically struc-
versely, the novices tended to see local discon- tured and integrated representation, BA is able to
nected problems that were not tied to the larger work at several text levels simultaneously. She can
context; they either did not see the larger problems thus analyze the text by working on local examples
or did not connect global problems to local ones. that contribute to the larger problem.
Second, because many of the experts in our study In contrast, the novice subject (ML) (Figure 2)
built elaborated representations of text problems, does not relate this paragraph to the text as a whole
they came up with a variety of solution procedures. in her diagnoses. Rather than seeing one large prob-
The novices, however, tended to build more simple lem, ML sees several small word-level problems
representations that did not allow them to come up (faulty use of an abbreviation-P50, and faulty
with very effective solutions. Because the experts word choice-P54, C 126) that are not related either
were able to locate these different facets of complex to each other or to the sentence-level problems of
problems in specific text features, they had a wider incoherence (ClOl) or "negativeness" (CI16),
range of strategies to help them solve both indi- which she diagnoses. However, we do see this
vidual local problems and larger global ones. novice beginning to work more like the expert,
though still on a paragraph rather than a whole-text
Case Study 1: Building an Integrated Problem level: she attempts to integrate a rather fuzzy para-
Representation graph-level diagnosis that this paragraph contains
Although we found interesting individual dif- "negative stuff" (P41) with her diagnosis that a
ferences within our subject groups, this case study particular sentence seems illogical (What's the
is especially pertinent to understanding creativity good of intramural sorts if they are so bad-
because the two writers typify differences in the C116).
amount of integration in our experts' and our These two writers were fairly typical of the ex-
novices' problem representations. Figure 1 shows pert and novice writers in our study. In Figure 3, we
all the diagnostic comments made by one of our represent schematically the diagnostic comments
teaching experts (BA) about this paragraph, and made by two other teaching experts and two other
Figure 2 shows all those made by a novice (ML). novices as they considered the problems in our sam-
For each subject, we distinguished the text level to ple excerpt of the original text. Figure 3 indicates
which the comment applied (i.e., whole textlmulti- the text levels of the problems that these writers
paragraph, paragraph/ intersentence, sentence, saw, and the extent to which their problem repre-
word level) and the general category of problem sentations were integrated. (The lines on the figure
mentioned in the comment. We looked specifically represent explicit connections between the writers'
at whether the revisers built integrated networks of diagnoses and/or detections of text problems.)
diagnoses that related text problems and spanned The problem representations of the experts and
text levels. Where the subjects made explicit con- novices (see Figure 3) were very different: first, the
experts' diagnoses were integrated, whereas the they often see only the isolated local problems, their
novices' were more fragmented and independent of solutions are more likely to be ad hoc and less goal
each other; second, the experts diagnosed problems driven than the experts' solutions.
at several levels of the text, particularly above the
sentence level, whereas the novices focused almost
Case Study 2: Using Diagnosis to Develop
exclusively on sentence and word-level problems;
Solution Strategies
finally, the experts diagnosed global problems
(such as audience and purpose), whereas the In this second case study, we look at how one of
novices often could detect that something global our experts and one of our novices used their diag-
might be wrong with the text (e.g., one subject noses of text problems to develop procedures for
commented that "this whole paragraph sounds improving the text. First, we examine some epi-
ridiculous") but were unable to come up with a sodes taken from the protocol (P) and cued recall
fulfledged diagnosis. Our analysis suggests that ex- (CR) of an expert writing teacher who effectively
perts' ability to integrate diagnoses of local prob- diagnosed and solved the problems in our example
lems with each other and with whole-text diagnoses excerpt. (The statements that are in quotation marks
allows them to see local problems in the context of in these episodes indicate where the writer was
more global ones. This interaction of the local and reading from the original text.) We have selected
the global provides experts with a rich conceptual some key episodes that illustrate how the subject
framework upon which to base their solution pro- dealt with one of the major problems in the' 'I don't
cedures. Novices, on the other hand, appear to be want to infer" paragraph, a problem he categorized
driven mainly by the text itself, an approach that as "negativeness." Notice how he starts by work-
does not allow them to import a representation of ing with the whole paragraph as a single movable
the unstated purpose of the whole text. Because unit.
Protocol
P73 What I guess I'd do now is to move this Cl90 and then there's all this information about
paragraph, or the contents of it varsity sports.
P74 the one about personal benefits, up here. Cl91 So what I tried to do was to substitute
P75 Continue with the positive things about it information about intramural sports.
rather than the negative. C228 . . . either the words or the structure itself,
connotation of the words or the structure itself
C163 If you look at these two sentences of which I
is negative.
guess I actually cut everything but about three
C229 It's negative, negative, negative.
words.
C230 I don't have a logical mind.
Cl64 They are both split in the negative.
C231 I can't follow negatives.
Cl65 "I don't want (to infer)"
C232 I go make them over.
Cl66 "Intramural sports are not"-which I think is
C233 And if I say that I do not want you to think
not the way a recruiting document should be.
that I don't care about your reluctance, I've
Cl67 I think you should put in a positive statement.
lost myself,
C168 And secondly, both of them are really lead-ins
C234 because I'm talking about a double negative
that don't get you anywhere.
about a negative mental attitude,
Cl74 But since we use in addition to varsity
C235 and I can't follow it.
sports . . . intramural sports attract these acts
C236 I don't think most people can.
and so on
C237 I just changed the structure.
Cl75 where you tell what it is rather than what it is
C240 I'm simplifying the sentence structure and
not.
making it more direct.
Cl76 which is just a delaying tactic.
C261 That negative approach thing ("feel that
Cl88 The paragraph is claiming from the very participating in sports is another pressure they
beginning that she is going to talk about don't need.")
intramural sports, C262 They don't need it.
Cl89 yet as I said, the first two sentences introduce C263 But it's also that I just changed the whole
intramural sports with a negative sentence structure.
C264 The sentence structure as the writer had it is C442 so I brought the-so I changed the order
that many new students feel that participating from-
in sports is another pressure they don't C443 the way this writer had it was the negative to
need. the positive.
C266 I set up the structure differently. C444 I brought the positive to the benefits, to the
negative,
C441 Then it seemed to me to emphasize the C445 which I close with another positive statement
benefits first as opposed to the reluctance or to de-emphasize the negatives or the
the problems students feel, complaints of the problems people feel.
This expert built an elaborated problem represen- rather difficult "mapping" of general and abstract
tation that provided him with a rich knowledge base problems onto the text appears to be an important
for solving the problems in the text. His multi- step in moving from diagnosis to action. This writ-
faceted diagnosis sees the "negativeness" problem er's diagnoses enable him to come up with a range
not only in the words but also in the syntax, in the of concrete and useful strategies for improving the
ordering of information, and in the rhetorical effect text that are closely tied to text features. Thus he
of the text. For example, he diagnoses that the con- decides to
notations of the words and the grammatical struc-
ture are both negative (C228); that the "lead-ins,"
1. move the paragraph to continue with the
which lead nowhere (C168), highlight inappropri- positive statements (P73-75): multipara-
ate and negative information; and that the sentences
graph level
shift the focus to the worst aspects of sports and
2. put in a positive statement (C167): paragraph
delay the presentation of the key ideas (C174-l76).
level
In addition, he sees these individual problems as
3. substitute information (in the paragraph)
symptoms of a larger, more global problem: a nega-
about intramural sports (C19l): paragraph
tive approach is inappropriate for the overall rhe-
level
torical purpose (as he represents it to himself) ofthe
4. change the order of the statements from nega-
handout, that is, for a recruiting document (C166),
tive/positive to positive/negative/positive
and he recognizes recurring patterns of negative
(C44l-445): paragraph level
structure and tone throughout the whole text (P73- 5. simplify the sentence structure to make it
75). In fact, he generalizes about the rhetorical ef-
more direct (C240): sentence level
fect of negative ness on readers, that is, that a "dou-
ble negative about a negative mental attitude" is too
difficult for most readers to follow (C234-236). The elaboration and location of problems in the
This writer was able to be very specific in text, which we noted in the expert's diagnoses, pro-
his diagnosis of this complex problem. He effec- vide a contrast to the performance of one of the
tively locates problems in actual text features at better and more fluent of the novices. The protocol
severallevels-multiparagraph (P73-78), para- comments represent this subject's diagnosis of this
graph (C188-191), and sentence level (C26l). This same problem paragraph.
Protocol
P41 "I don't want" ... This is all negative stuff. P50 This 1M There's no 1M before
"in which the rules are better, the equipment P51 so you have to go back
is better" P52 so this should be Intra-mural again.
P42 It can all be cut. P54 Irregardless ... it can't be irregardless.
P45 They (are) a little more casual I guess. P55 I guess I've got to change that.
P47 Urn this 1M down here, I'll go back to that P56 So we'll cut that out
too. P57 and it should be "1M sports ...
P48 I'll make basic corrections P58 They can be just as much fun as varsity
P49 and then go back and improve. sports and take less time."
Pl28 "I don't want to infer" ... keep varsity, Pl81 They tend to be a bit more casual.
Pl29 intramural sports are . . . tend to be a bit
C1l4 (Experimenter: You crossed out the whole ... )
more casual and less demanding.
CllS Well it's just so negative.
Pl29a 1M sports may be ... perhaps a more
C1l6 It's like, so what's the good of 1M sports
influential . . .
C1l7 if there so bad.
Pl30 Participating is an extra, urn let me see . . .
C1l8 They're just a bit more casual, not as
P132 is an addition.
disciplined.
P133 Oh well, I'll go back to that.
Cl2S (Experimenter: And just?")
Pl79 "I don't want to infer that the only chance Cl26 I don't like "just" either.
women get for participating in sports is on Cl27 "Even," "just," and "like," I don't like.
varsity teams." There is also ... there are (Had commented earlier: "I was always told
also 1M sports in any composition class I had that 'like' is
Pl80 which are not quite the same as varsity. just bad to use. So I always put 'such as. ')
Where the expert's comments show an elabo- simply may not know how to locate the source of
rated representation of the text problems, the their own responses as readers; they may not see the
novice's comments are sparse and undeveloped. features of the text that prompted their responses.
The expert saw an interplay of lexical, syntactic, These two case studies illustrate the key role of
semantic, genre-related, and rhetorical problems; diagnosis in developing effective representations
the novice sees only one- "negative stuff" (P4I). and solutions to ill-defined revision problems. The
Her diagnosis is based on a simulation of a reader's problem representations our experienced writers
response (C1l6-CIIS), a form of diagnosis that built were not just bigger but qualitatively different
normally provides only a very generalized and from the novices' representations. The experts were
open-ended definition of the problem. This sub- (1) seeing individual problems as part of a concep-
ject's response is also typical of novice diagnoses in tual whole and (2) developing a wider range of
that it depends primarily on the reviser's social or effective strategies based on several dimensions of
topic knowledge rather than on rhetorical knowl- the problem. Overall, we found that our expert writ-
edge; she imagines what someone might say about ers had several general strategies for dealing with
1M sports and rewrites the content of the text to fit ill-defined revision problems:
her own ideas (P45, CIIS)
Broad, intentional diagnoses, of the sort both ex- 1. In the act of building a representation of an ill-
pert and novice were making, pose a built-in prob- defined problem, expert writers are seeing an inte-
lem for the reviser who must map something like a grated set of more familiar subproblems. When a
general sense of negativeness or a lack of focus onto global, ill-defined problem is located in features of
specific features in the text. Generally speaking, the the text, the diagnosis becomes more operational
more global the diagnosis, the harder it is to locate because it now says something about where and
exactly the problem in the text and the fewer the how to proceed. It is the difference between being
strategies are available for acting on the diagnosis. told to "be more clear" and being told to "define
Here, while the expert has several procedures oper- your key terms earlier in the paragraph." Finally,
ating on different text levels, the novice has few. In once these abstract problems are given a local con-
fact, she demonstrates a typical novice strategy: text and a name in the form of subgoals and text
first she localizes the problem in a single offending features, they can lead to text changes, which,
clause (P4!); then she solves the problem by delet~ though the changes themselves are local, are con-
ing (P42) and substituting her own new information trolled by a larger, integrated problem and plan.
(P45). She returns to this section of the text on two 2. As these writers build their problem represen-
later passes and simply rewrites the sentence to fit tation, they create subgoals that function as tests for
her own ideas about intramural sports (PI2S-l33; success. The novices' criteria tended to be vague,
P17S-ISO). In addition, on some occasions in "well it's just so negative," whereas the experts'
which she is unable to map a general diagnosis to criteria were much more specific, "the connota-
the text (Pl33), she finally decides to leave the tions of the words and the structure itself is
problem alone. We suspect that novice revisers negative. "
TEXT LEVEL
(purpose is)
Recruiting for for funding
Whole text/
multiparagraph
57-59
Paragraph
convinc-
ing about
purpose intramural
sports
Sentence • Parallelism
B
Word
Figure 1. Expert's diagnostic comments.
3. Creating tests and subgoals could have the Experts try to maintain this part/whole balance by
result that a writer only deals with the isolated local invoking global tests. For instance, toward the end
problems and loses sight of the larger context. For of the protocol of one of our expert writers, after he
example, meeting the goal to introduce some new has made a number of local revisions, the writer
technical terminology could interact with a more decides to "read (his text) over and see what it
global goal to maintain a chatty, informal tone. sounds like" to test for a potential whole-text prob-
TEXT LEVEL
Whole text/
multiparagraph
Paragraph
li1
Sentence
116
Doesn't say "also" (It's) just so
(not just talking negative What's the
about varsity) good of intramural
sports if they are
so bad
Word .50 .5li .126

Faulty use It can't be I don't like"just"
of abbreviation irregardless and "even"
Figure 2. Novice's diagnostic comments.

EXPERTS' COMMENTS
TEXT LEVEL
EXPERT 1 EXPERT 2
Whole text
multiparagraph
..~
. =========
..
Paragraph ~
Sentence • .
Word
•
NOVICES' COMMENTS
NOVICE 1 NOVICE 2
Whole text
multiparagraph
...I
Paragraph
Sentence • • •• • •
Word
• -
Figure 3. Networks of experts' and novices' diagnostic comments (. = detection of a problem;· = diagnosis of a problem).
lem: "I have a feeling there is still a shift in tone. ning with goals and gists gives writers more flexi-
The first paragraph is more formal and it gets less bility than working with prose and allows them to
formal as it goes along." Thus, the problem solver tum vague, amorphous problems into more man-
is able to shift back and forth from representing ageable ones.
local problems to seeing their global effects. In this chapter, we have tried to use recent re-
4. Finally, a very important expert strategy was a search in cognition and the composing process to
planning strategy. Often, when experts are diag- describe some features of this process that help lay a
nosing and developing solutions to ill-defined prob- foundation for creativity. Creativity, whether it is
lems, they work on a more abstract level with goals practical creativity or genius-type work, clearly
and gists rather than with actual text. This allows rests on many capabilities, including rich knowl-
revisers to integrate various top-level goals before edge, social expectations, and the context for per-
getting into specifics and provides an overall frame- formance, in addition to cognitive actions. How-
work for their diagnoses and text changes. Local ever, our analysis would lead us to contribute two
problems can then be confronted only after a tenta- observations to this picture of creativity.
tive global plan is formulated. Because goals and The first observation is that the writing process
gists are an effective way to chunk lots of informa- itself shows us some powerful mechanisms for
tion, revisers can come up with several alternative creativity in the act of building a representation of a
plans for improving the text. On the other hand, a problem; in the process of exploring mUltiple di-
plan that looked promising in the abstract may not mensions of that problem and integrating that
work out at a local level, and the reviser may be knowledge into a unifying theory of the task and an
pushed back into re-representing the problem. For integrated plan; and, finally, in the act of using
dealing with complex revision problems where revision as an opportunity to re-envision a text and
there is not one easy answer, this strategy of plan- to deal with individual elements of a text as part of a
meaningful, purposeful whole. These powerful but creativity? How can they make the process of think-
demanding processes in writing are a potential ing creatively operational-an act student writers
source of both adaptive and original solutions to can work toward? Does instruction sometimes pre-
rich rhetorical problems. sent a romantic view of creative work that fails to
The second observation is that these features- recognize the role that energetic and enterprising
this potential inherent in the cognitive process of cognition can play?
experienced writers-is no guarantee of creativity.
As educators, on the other hand, it strikes us that 3. Planning and revision. If we tum our atten-
many of our students are shutting down this poten- tion to teaching, one important question we
tial in themselves. The strategies that lead them to would like to answer is: Does the more lim-
depend on knowledge-telling, to take limited and ited problem solving we see in students' writ-
tentative looks at their rhetorical problem, to build ing process and, in particular, in their ap-
simplistic images of their task, to invoke rigid rules proach to planning and revision simply reflect
in the face of a complex situation, and to use revi- what these writers do not do, or what they
sion as a limited tool for repairing local problems all cannot do? A line of research that we think
produce a process that does not rise to inventive and will be particularly fruitful will be work that
generative problem solving when it needs to. Al- focuses directly on the strategies writers use
though we cannot guarantee creativity, we can, we for open-ended tasks and on their own
believe, help writers develop the strategic, cog- awareness of or metacognition about their op-
nitive tools they need to meet situations that de- tions and process.
mand it.
The focus of this chapter has been on practical Currently, we are investigating whether students
creativity-on the variety of creative responses that may have untapped, hidden planning skills which
all sorts of writers bring to many kinds of writing they could draw on, if they could be prompted or
problems, in school and in professional work. It sensitized to do so. For example, if student writers
follows that one important line of research growing are explicitly asked to plan on a more expert level
out of the work we have discussed will pursue the (e.g., to focus on rhetorical planning, to elaborate
illusive trail of creativity in real-world situations, goals, or to deal with conflicts) would we find sharp
particularly those situations in which a creative gains in the sophistication of their planning? Or
(i.e., a valuable) response to a rhetorical situation is would we see that applying these strategies, es-
called for, or is likely to make a difference. pecially in the context of ill-defined tasks, requires
Education and writing in school is one place better, more direct instruction. The results of this
where creativity makes a difference. Our survey research will, we hope, provide a basis for helping
suggests some of the following research areas and student writers to develop their potential in the
questions: classroom. By building on students' hidden
strengths as well as focusing on their specific areas
1. Ill-defined tasks. What is the relation between of difficulty, we can develop their awareness of
the expert and novice strategies used in vari- powerful cognitive strategies and provide a spring-
ous kinds of writing and the power of those board for a more creative approach to the writing
strategies to foster creativity? How do writers process.
with different levels of expertise handle the
problem and the potential of ill-defined tasks?
2. Task representation. To what extent are stu- References
dents' performances conditioned, on standard
school tasks that involve writing, by the ways Ackerman, J. (in press). Students' self-analyses and judges'
perception: Where do they agree? In L. Flower (Ed.),
they represent the task to themselves? Are Reading-to-write: Exploring a cognitive and social process
students sometimes constructing representa- New York: Oxford University Press.
tions that diminish a task or that even oblit- Beach, R., & Eaton, S. (1984). Factors influencing self-assess-
erate the opportunity for creativity? ing and revising by college freshmen. In R. Beach & L. Brid-
well (Eds.), New directions in composition research (pp.
149-170). New York: Guilford Press.
What is the role of instruction in fostering a cre- Britton, J., Burgess, T., Martin, N., McLeod, A., & Rosen, H.
ative approach to ill-defined tasks? For example, (1975). The development of writing abilities (11-18).
how do teachers communicate their expectations of London: Macmillan.
Brown, A. L. (1980). Metacognitive development and reading. Hayes, J. R. (1981). The complete problem-solver. Philadel-
In R. Spiro, B. Bruce, & W. Brewer (Eds.), Theoretical phia, PA: The Franklin Institute Press.
issues in reading comprehension (pp. 453-481). Hillsdale, Hayes, J. R., Flower, L., Schriver, K., Stratman, J., & Carey,
N J: Erlbaum. L. J. (1988). Cognitive processes in revision. In S. Rosenberg
Brown, A. L., & Smiley, S. S. (1978). The development of (Ed.), Advances in applied linguistics: Vol. 2. Reading, writ-
strategies for studying texts. Child Development, 49, 1076- ing and language process (pp. \76-240). Cambridge, En-
1088. gland: Cambridge University Press.
Brown, A. L., Day, J. D., & Jones, R. S. (1983). The develop- Langer, J. A. (1984). Where problems start: The effect of avail-
ment of plans for summarizing text. Child Development, 54, able information on responses to school writing tasks. In A.
968-979. Applebee (Ed.), Contextsforlearning to write (pp. 135-148).
Carey, L. J., Flower, L., Hayes, J. R., Schriver, K., & Haas, C. Norwood, NJ: Ablex, 1984.
(1986). Differences in writers' initial task representations Newell, A., & Simon, H. A. (1972). Human problem-solving.
(ONR Technical Report, No.2). Pittsburgh, PA: Carnegie- Englewood Cliffs, NJ: Prentice-Hall.
Mellon University. Newell, A., Shaw, J. c., & Simon, H. A. (1962). The process of
Faigley, L., & Witte, S. (1981). Analyzing revision. College creative thinking. In H. E. Gruber, G. Terrell, & M.
Composition and Communication, 32, 400-414. Wertheimer (Eds.), in Contemporary approaches to creative
Flower, L. (1979). Writer-based prose: A cognitive basis for thinking (3rd ed., pp. 63-119). New York: Atherton Press.
problems in writing. College English, 41,19-37. Perkins, D. N. (1981). The mind's best work. Cambridge: Har-
Flower, L. (in press). Task representation. In L. Flower (Ed.), vard University Press.
Reading-to-write: Exploring a cognitive and social process Reitman, W. R. (1964). Heuristic decision procedures, open
(Technical Report, Center for the Study of Writing at Berkeley constraints, and the structure of ill-defined problems. In M.
and Carnegie Mellon, New York: Oxford University Press. W. Shelley & G. L. Bryan (Eds.), Human judgments and
Flower, L., & Hayes, J. R. (1980). The cognition of discovery: optimality (pp. 282-315). New York: Wiley.
Defining a rhetorical problem. College Composition and Rose, M. (1984). Writer's block: The Cognitive dimension.
Communication, 31,21-32. Carbondale, IL: Southern Illinois Press.
Flower, L., Carey, L. J., & Hayes, J. R. (1985). Diagnosis in Scardamalia, M., & Bereiter, C. (1988). Knowledge-telling and
revision: The expert's option (Communications Design Center knowledge transforming in written composition. In S. Rosen-
Technical Report, No. 27). Pittsburgh PA: Communications berg (Ed.), Advances in applied linguistics (Vol. 2., pp. 142-
Design Center. 175). New York: Cambridge University Press.
Flower, L., Schriver, K., Carey, L. J., Haas, c., & Hayes, J. R. Simon, H. A. (1973). The structure of ill-structured problems.
(1987). Planning in writing: A theory of the cognitive process Artificial Intelligence, 4,181-201.
(ONR Technical Report, No.1). Pittsburgh, PA: Carnegie- Simon, H. A. (1986). Some computer models of human learn-
Mellon, University. ing. In M. Shafto (Ed.), How we know. San Francisco, CA:
Getzels, J. W., & Csikszentmihalyi, M. (1976). The creative Harper & Row.
vision: A longitudinal study of problem finding in art. New Sommers, N. I. (1980). Revision strategies of student and expe-
York: Wiley. rienced writers. College Composition and Communication,
31, 378-387.
CHAPTER 18
Cognition and Writing

THE IDEA GENERATION PROCESS
John A. O'Looney, Shawn M. Glynn, Bruce K. Britton, and

Linda F. Mattocks
The interest in ideas and where they come from is an writing processes, such as organizing, goal setting,
old one, dating back at least to pre-Socratic philoso- translating, and reviewing.
phers and continuing to the present-day theories of Our purpose is to examine the relationship be-
creative thinking. The question of where ideas tween cognition and writing, giving particular em-
come from has been answered in a variety of ways: phasis to the process of idea generation. We will
ideas can come from the gods, from the imagina- consider each of the following topics in turn:
tion, from a mind trained in reasoning, from a hap-
hazard association of memories, from stimulating 1. A human information-processing model
images, from a subconscious bank of archetypes, 2. Some models of writing that relate idea gener-
from memories transformed by mental schemata, ation to other subprocesses.
and from attempts to solve problems. 3. Some methods of capturing and manipulating
Many of these theories of idea generation are still ideas when writing
popular, whereas others have been discarded. The 4. The cognitive constraints imposed on idea
search for a valid theory of idea generation remains generation when writing
important because new curricula for teaching writ-
ing, such as Linda Flower's (1981) Problem-Solv- We will critically review the literature relevant to
ing Strategies for Writing, are being based on par- idea generation when writing, attempt to pull this
ticular methods of idea generation. In these new literature together theoretically, and identify some
curricula, students are taught how to generate ideas remaining research issues that need to be addressed.
and how to integrate idea generation with other The writing we speak of here is considered by us to
be "creative writing" in that it requires of both
John A. O'Looney • Department of Language Educa- novices and experts the production of new and ef-
tion, University of Georgia, Athens, GA 30602. Shawn fective ideas, such as those found in good essays,
M. Glynn and Linda F. Mattocks • Department of Educa-
tional Psychology, University of Georgia, Athens, GA 30602.
articles, short stories, and novels. We are not
Bruce K. Britton • Department of Psychology, University speaking of relatively routine writing tasks, such as
of Georgia, Athens, GA 30602. producing a shopping list.
305
Human Information-Processing Model Kintsch's (1980, p. 12) notion of one's long-tenn

memory as a "propositional network that is orga-
The infonnation-processing model of human be- nized only in the sense that certain relationships
havior holds that a limited capacity "working exist among the propositional nodes" is essentially
memory" controls ongoing, conscious mental op- a conception of memory as a spatial entity; that is,
erations, such as those that enable us to read, write, ideas with strong relationships are stored together
and solve arithmetic problems (Baddeley, 1978; and tend to be retrieved together. The stronger the
Britton, Glynn, & Smith, 1985). In working memo- relationship, the more likely one propositional node
ry. new incoming infonnation is maintained tem- will be recalled when a node close to it is retrieved.
porarily through the process of rehearsal. While Polson (personal communication to Kintsch,
maintaining this new infonnation, we can operate 1980) found that many writers systematically
on it (e.g., comprehend, infer, or compute), we can search their memory along one associative path to a
integrate it with our existing related knowledge given depth before exploring other branches to the
(stored in our long-tenn memory system), and we same depth. We have illustrated this strategy in
can return the product of this integration to our Figure 1. For example, when a mystery writer cre-
long-tenn memory system for pennanent storage. ates a motive for a murderer, the writer may first
Working memory is like a cognitive workbench on consider jealousy in depth before moving on to re-
which we can fashion new intellectual products. venge or greed. This depth-first strategy, Polson
The processing capacity of working memory is speculated, places the least strain on one's working
limited in tenns of (1) the amount of infonnation memory. However, though this strategy reduces the
that can be maintained at one time and (2) the demands imposed on the writer, it is possible that
number of cognitive operations that can be per- following it will lead the writer to neglect other idea
fonned simultaneously. For present purposes, pro- paths.
cessing capacity will be defined as "the limited What is particularly important for memory is that
pool of energy, resources, or fuel by which some the retrieval cues are, to a large extent, under the
cognitive operations or processes are mobilized and control of the writer. Manipulation of these cues,
maintained" (Johnston & Heinz, 1978, p. 422). Kintsch (1980) suggested, detennines a search set
The processing capacity of a writer's working that the writer can then explore: "The level of con-
memory has been found to be related to his or her straint imposed by the retrieval cue and the density
writing capability (Benton, Kraft, Glover, & Plake, of the knowledge space in the area of the search
1984; Tetroe, 1984). With maturation and the ac- jointly detennine how many nodes will be included
quisition of strategic knowledge on how to manage in the search set" (p. 14). Once the retrieval cues
cognitive resources, the effective capacity of the and knowledge base are detennined, the search set
writer's working memory increases. is automatically constituted. We have illustrated
this in Figure 2.
Once an item has been recovered from the search
Memory Probes
set, the recovered item is added to one's retrieval
Much of the raw material for our ideas comes cues, and the association between the sampled item
from our long-tenn memory, or our store of referen- and the original retrieval cue is heightened. Thus,
tial knowledge. The mode of operation of one's retrieval cues are continually redirected by the
long-tenn memory has been the subject of numer- sampled items; however, the direction the search
ous experiments by Kintsch and his colleagues process takes sometimes may not be well suited to
(e.g., Kintsch, 1986; van Dijk & Kintsch, 1983). the goal of the search.
Kintsch (1980), in a discussion of component writ- In an empirical test of Kintsch's model, Cac-
ing processes, has suggested that the idea genera- camise (1981) found a significant difference in idea
tion process begins with a probe of long-tenn mem- generation between writers who wrote for an au-
ory. This probe is constrained by infonnation about dience of adults (a less constraining task) and those
the topic and the intended audience. Kintsch hy- who wrote for children (a more constraining task).
pothesized that different topic and audience cues Writers who wrote for children had a great deal
would place different constraints on the search for more difficulty generating and developing ideas. It
ideas in one's long-tenn memory. In particular, a took them twice as long to write the same number of
more specific topic or audience would help a writer ideas. Caccamise suggested that in order to max-
narrow the search for ideas. imize idea generation, it may be necessary to first
CHAPTER 18 • THE IDEA GENERATION PROCESS 307
BREADTH OF SEARCH
SEARCH
SET 1
•
SEARCH
SET 2
•
SEARCH
SET 3 SET N
+
SEARCH
DEPTH
OF
SEARCH
Figure 1. Breadth and depth of memory search.
generate ideas without regard to an audience, and Caccamise suggested that generating ideas for
only later reshape ideas to reflect the audience's writing is a special case of the concept of the "ill-
needs. structured problem." Ill-structured problems are
not easy to formalize and often demand more from
the problem solvers than their limited computa-
Problem Solving
tional capacities can achieve. An ill-structured
Caccamise (1981) maintained that Kintsch's problem in writing is one in which a writer must
memory model is best applied to writing when it is juggle several constraints simultaneously. Consid-
combined with a model of a problem-solving sys- er, for example, when a writer is asked to develop
tem. The resulting general model can be viewed as a an argument appropriate for an audience whose
serial processor that can only address a small members vary widely in age, beliefs, and cultural
number of inputs and outputs at a time. backgrounds. The writer must give extra considera-
KNOWLEDGE BASE
RETRIEVAL CUE #1
RETRIEVAL CUE #2
RETRIEVAL CUE #N
o 00
000
00 0
000
~------+-----~
Figure 2. Retrieval cues and knowledge base determine the search set.
tion to these constraints, plus the usual ones, such board and then step back and ask their class ques-
as idea generation, sentence structure, punctuation, tions about how the composition could be extended;
and spelling. the teachers could then implement the students'
Cooper and Matsuhashi (1983), Flower and suggestions on the board. Gagne has suggested that
Hayes (1980), Young (1981), Newell, Shaw, and eventually a child could act as his or her own "sur-
Simon (1963) and others also have written of idea rogate partner" in generating conversational cues
generation as a problem-solving process. Young, in for continuing to write.
particular, has stressed the importance of a task
environment that forces a person to formulate the
problem, not simply solve already posed problems. Purpose
Young has suggested that some of the strongest In the human information-processing model,
incentives for communication are found in the pro- purpose is viewed as driving behavior. Because of
cess of discovering a problem and exploring reason- the work of Moffett (1968), Britton, Burgess, Mar-
able solutions to that problem. He also has sug- tin, McLeod, and Rosen (1975), and Kinneavy
gested that discovering real problems and their (1971), ways of classifying writing into discourse
solutions can be intrinsically rewarding activities types (e.g., narration and exposition) based on pur-
for children who are beginning to write. pose (e.g., description and explanation) are avail-
Beginning writers often encounter the difficult able. Knowledge of discourse types could be con-
problem of finding sufficient information on a topic sidered a heuristic for writers. Understanding the
in their long-term memory both to begin and sustain connection between purpose and discourse type
a piece of writing. Unlike spoken discourse, in could be especially useful to the student writer who
which new ideas can be generated out of a dialec- is aware of a purpose, but who is in need of a mode
tical exchange, writers often have to develop search in which to pitch his or her voice.
strategies in order to gain access to information After a writer has determined a purpose and se-
stored in their long-term memories. lected the appropriate discourse type for that pur-
Scardamalia, Bereiter, and Goelman (1982) pose, specific objectives within the discourse type
found that fourth-grade writers produced shorter es- can be identified. Collins and Gentner (1980) hy-
says than sixth-grade writers. This difference could pothesized four basic objectives that writers have:
be explained simply by pointing to a difference in (1) making the text enticing, (2) making the text
the knowledge base (Voss, Vesonder, & Spilich, comprehensible, (3) making the text memorable,
1980) or to differences in the children's ability to and (4) making the text persuasive. The identifica-
manage the demands made on their working memo- tion of a clear objective will help a writer narrow the
ry (Glynn, Britton, Muth, & Dogan, 1982). Scar- focus of his or her idea generation process.
damalia et al., however, pointed to a third explana-
tion. They suggested that younger children, whose
communication experience is predominantly in the
spoken mode, rely more on the conversational cues Models of Writing: Relationship of Idea
to prompt a memory search for relevant ideas than Generation to Other Writing
do older children. To test their hypothesis, Scar- Subprocesses
damalia et al. asked fourth- and sixth-grade chil-
dren to continue to write after they appeared to have Models of writing help us to conceptualize the
stopped writing. Even though both groups wrote writing process as a set of component, interrelated
more after being cued, only the younger writers' subprocesses. We will discuss several models and
compositions showed an increase in the coherence the role idea generation plays in each. The models
of what they wrote. These results suggested that the are those of Hayes and Flower (1980), Rose (1984),
younger writers had stopped before they had used Cooper and Matsuhashi (1983), and Wason (1978).
all the relevant ideas in their memory set. The older
children apparently possessed a set of internal cues
The Hayes and Flower Model
or prompts that the younger children lacked.
Gagne (1985) has recommended that teachers Hayes and Flower (1980) have described a model
model self-questioning strategies to help young (see Figure 3) of writing that has received wide-
children acquire internal cues. For example, teach- spread acceptance among writing theorists. In de-
ers could write part of a short composition on the vising their model, Hayes and Flower posited the
TASK ENVIRONMENT
WRITING ASSIGNMENT TEXT

Topic PRODUCED
Audience SO FAR
Motivati I)g Cues
THE WRITER'S LONG-TERM PLANNING TRANSLATING REVIEWING
,
MEMORY
~ f-!ORGANIZINGI I READING
Knowledge of Topic
~ ~
I EDITING
j1 I
Knowledge of Audience GOAL
Stored Writing Plans SETTING
I I
L MONITOR I
Figure 3. The Hayes and Flower (1980) model of writing. From Cognitive processes in writing (p. 11) by L. W. Gregg and E. R.
Steinberg, 1980, Hillsdale, NJ: Erlbaum. Copyright 1980 by Lawrence Erlbaum Associates, Inc. Reprinted by permission.
task environment and the writer's long-term memo- rhetorical situation (i.e., audience, topic, and text).
ry as the context in which the writing process takes Poor writers tend to limit their solutions (and thus
place. The writing process itself consists of plan- their ideas) to problems in text production, such as
ning, translating, and reviewing. Idea generation, format or spelling. In their verbal report protocols,
organization, and goal-setting are seen as sub- good writers usually make reference to their au-
processes of planning. The translating process pro- dience or assignment about twice as often as the
duces language corresponding to the ideas that are poor writers.
generated and the goals that are set in the planning
process; however, the recursive nature of these pro-
Writer's Block and Rose's Model
cesses makes it difficult to say where idea genera-
tion ends and translating begins. The study of writer's block by Rose (1984) tends
In the Hayes and Flower model, the function of to support the conclusions of Flower and Hayes
the generating process is to retrieve from long-term (1980). Rose found that nonfluent writers, or
memory items that are relevant to the writing task. "high-blockers," tended to plan incrementally;
Memory probes originate from the task environ- that is, they did not think through the ideas they
ment, specifically from cues about the topic and the wanted to express, nor did they write out any orga-
audience. From the first probe, items tend to be nizational approach before beginning a first draft.
retrieved in associative chains. Hayes and Flower Rose also found that high-blockers assumed a
contend that most writers will break a retrieval model of writing as spontaneous, and thus were in a
chain when they access one irrelevant item. The bind when ideas simply did not flow. Furthermore,
most persistent memory searches they observed in high-blockers tended to use confusing or conflict-
verbal report protocols never extended more than ing rules, plans, and strategies. Because the con-
three retrievals beyond relevant material. flicts tended to occur at the global level, the poor
According to Flower and Hayes (1980), ideas are writers spent much of their composing time on fruit-
generated as a response to a rhetorical problem that less cogitation rather than on directed idea genera-
includes an assignment, an audience, and the writ- tion. Rose found that high-blockers paused, on the
er's personal goals for the reader and text. When average, 9.1 more minutes (out of 60) than low-
generating ideas to solve the rhetorical problem, blockers. He concluded that much of the efficiency
more successful writers attend to all aspects of the of low-blockers is rooted in their ability to use flex i-
ble rules, appropriate plans, goals to produce first as summanzmg and concluding; (c) sequencing
and define later, and in their general absence of roles, such as adding one assertion to the next, re-
conflict. High-blockers, on the other hand, used placing a statement with an alternative, and narrat-
rigid rules, such as "You must always develop at ing; (d) relationship roles, such as comparison/
least three points in your paper." Such a rule, stated contrast, inferring, evaluating, cause/result, and
as an imperative, can frustrate a novice writer who qualifying; and (e) development roles, such as
might only have one argument; and instead of de- exemplifying, defining, and describing. Under-
veloping the one point he or she does have, the standing model patterns of these sentence roles
writer is likely to give up entirely (Kean, Glynn, & can help a writer to better express his or her
Britton, 1987). ideas.
In his model, Rose (1984) has posited three types In order to produce the content of a particular
of composing subprocesses: rules, interpretive sentence, Cooper and Matsuhashi (1983) have sug-
plans, and discourse frames. Rules refer to lin- gested that the writer moves through nine idealized
guistic, stylistic, rhetorical, and social linguistic stages of sentence planning:
processes that are responsive to context and pur- 1. Formulating a proposition. In this stage, a
pose. Good rules flex to match the context; rigid writer, guided by a schema for the discourse type,
composing rules are enacted without a context. retrieves information from long-term memory and
Blocked writers usually have limited, rigid rules. uses the information to form the basis of a message
Interpretive plans influence decisions about how or proposition.
much to write, what to emphasize, what to classify, 2. Framing the proposition. Framing involves
and what to define. These plans are most fruitfully choosing an action or state that can be appropriately
made in the context of the audience and task. Block- hung on the objects or people in one's proposition.
ing occurs, Rose has contended, when the writer's Chafe (1977) has described such frames as case
interpretive activities are inappropriate to the task frames that explicitly determine the role-agen-
or audience. tive, affective, recipient, attribute, etc.-each
A third composing subprocess concerns the use noun will play in the sentence. By placing a noun in
of discourse frames. Rose discovered that "incre- various case frames, a writer can change the seman-
mental planners" planned in that fashion because tic thrust of the sentence and in doing so generate
they knew of no other way to plan. They did not new ideas.
possess control over an array of discourse frames 3. Placing the proposition. In this stage, a writer
(e.g., description and persuasion). Instead they must decide upon what he or she wants to accom-
tended to adapt a few discourse frames to all sorts of plish with respect to his or her audience. Each sen-
content and plans. If a student's knowledge of dis- tence is generated to accomplish an act (called a
course frames is limited, the complexity and per- "speech act"). Although there are many possible
haps the fluency of his or her writing will be speech acts, a writer's task is facilitated by the
limited. knowledge that most nonfiction discourse consists
of a relatively small number of representative
forms, for example, asserting, commanding, ques-
The Cooper and Matsuhashi Model tioning, suggesting, stating, and hinting. Cooper
Cooper and Matsuhashi's (1983) model of the (1984) has argued that the principles writers use to
composing process is more intricate than that of organize content come not from the content, but
Rose. These researchers have asserted that there are from writers' intentions expressed as speech acts.
two main subprocesses of composing: (1) develop- 4. Directing the proposition. When directing the
ing global discourse plans, and (2) developing sen- proposition, the writer must generate an idea as to
tence plans. They have argued that a writer's global how his or her theme will be continued in the next
plans (for a piece of discourse) are influenced by his sentence. The decision to place certain people or
or her conception of purpose, audience, discourse objects in the subject or major theme position in a
type, and discourse structure. sentence essentially will decide the direction of the
Global discourse plans are supported by sentence piece of writing. Deciding on a thematic direction
plans. The basic roles that sentences can play in- also will have consequences for further idea genera-
clude (a) generalizing roles, such as stating a thesis tion: some ideas will be eliminated in the process,
or restating a proposition; (b) rhetorical roles, such whereas others logically will come to the forefront.
5. Connecting the proposition. In this stage, the from prior clauses that fade during recoding.
writer takes into account what is given or what has 9. Transcribing the proposition. The compo-
preceded and what is new. In addition to deciding nents of the psychomotor act itself often will domi-
what is given and what is new, at this stage the nate the writer's consciousness until they have be-
writer must develop ideas on how to make his or her come somewhat automatic. As Cooper and
new ideas cohere with the old. Knowing that cohe- Matsuhashi (1983) have argued,
sion is created through the use offive major types of
cohesive ties-reference, substitution, ellipsis, a major goal of the development of writing ability is for transcrip-
conjunction and lexical reiteration, and colloca- tion increasingly to occupy subsidiary awareness, enabling the
writer to use focal awareness for other plans and decisions. (p.
tion-will enable a writer to generate alternative 33)
cohesive structures to suit various purposes and
modes of discourse (Halliday & Hasan, 1976).
6. Wording the proposition. This is the stage
Wason's Models
whereby the writer codes his or her thoughts into
specific lexical sequences. The difficulties writers Wason (1978), among others, has noted two
encounter at this stage have been discussed by rather contradictory models for idea generation
Chafe (1977) in terms of the problem of codability. when writing. In one model, the writer generates
Some objects or events are easy to code because ideas in note or outline form with minimal attention
they are part of our common experience (e.g., the to syntax and style. The writer then translates these
colors of a traffic signal). Others are more difficult notes or outlines into a draft of written discourse. In
because they demand specialized knowledge or a the other model, the writer composes the first draft
unique perspective (e.g., the meaning of an icon of using complete sentences and paragraphs, as ideas
an extinct race). Our choice of words will, as we are generated.
attempt to write coherently, lay a path for the fur- One advantage of the first model, which we will
ther generation of ideas. However, this path is cir- call the proposition-generating model, lies in the
cumscribed by (1) our previous choices with regard fact that processing capacity for generating content
to discourse type, purpose, functional, and struc- more likely would be available if that capacity were
tural sentence roles; (2) our knowledge of the sub- not being used to generate plans for coordinating
ject; and (3) our vocabulary. patterns of sentence functions, roles, speech acts,
7. Presenting the proposition. In this stage, the and coherence. A second advantage relates to the
writer is concerned with issues of usage and gram- notion that ideas flow at a rate faster than the writing
mar; that is, the writer transforms the already com- of prose can capture. Taking notes, then, would
posed sentence into proper standard English. have the distinct advantage of keeping pace with the
8. Storing the proposition. Atthis stage, the writ- flow of ideas.
er's major concern is storing the proposition in his One advantage of the second model, which we
or her memory. A proposition that is longer than a will call the sentence-generating model, is found in
few words must be rehearsed and re-rehearsed to its ability to explain how writers often generate
insure it is not lost in the process of transcribing it ideas by expanding on a point. This process can be
on paper. seen as the cumulative effect of following a dis-
If writers, like speakers, form sentences with course pattern or creating a unique frame, place,
clause-size components (Daiute, 1980), one would direction, connection or wording form a propo-
expect more errors would be found in longer claus- sition.
es, because producing these clauses would burden a A second advantage of the sentence-generating
writer's working memory. And, indeed, Daiute model can be seen among writers who decide not to
(1981) found that the erroneous sentences produced focus the bulk of their attention on idea production
by secondary school students are eight words long- and thereby leave little processing capacity for de-
er, on the average, than correct sentences. Daiute veloping the vehicles that carry ideas to readers
(1984) suggested that writers who do not make syn- (e.g., figurative language). Thus, although propo-
tax errors have a variety of complex sentence struc- sition generating has the advantage in generating a
tures stored in long-term memory. These stored lot of ideas in rough form, sentence generating has
schemata for sentence structures help good writers the advantage of generating ideas in the form of
to remember important grammatical information good prose.
Cognitive Constraints Imposed on Idea formats that made possible the effective allocation
Generation When Writing of processing capacity. The more verbal writers
typically used proposition-based formats (e.g.,
Graesser (1984) has identified two sources of proposition lists, outlines, and diagrams) when
writing difficulty: (1) writing makes demands on constructing a preliminary draft. In each of these
our limited processing capacity because it requires formats, a large pool of preliminary draft arguments
us to juggle a variety of subprocesses; (2) writing was displayed in a form that lent itself to rapid
makes demands on our topic knowledge because it identification and comparison of raw ideas. The use
calls on us to generate and develop sound ideas. We of such formats enabled the more verbal writers to
will discuss these two sources of difficulty in the reallocate processing capacity and thereby to pro-
following sections. duce additional arguments. Writers with low verbal
ability, on the other hand, used proposition-based
formats infrequently.
Processing Capacity Limitations
To produce a text, writers must do more than Freewriting. To cope with their processing ca-
simply generate ideas; they must sequence them, pacity limitations, students of writing should be dis-
place them into sentences, and comply with the couraged from attempting to draft finished products
rules of sentence mechanics. According to a cog- on their first try. Advocates of the "freewriting"
nitive workbench interpretation of text production, technique have long discouraged premature editing:
each operation imposes a distinct demand on writ-
ers' limited processing capacities. Many people are constantly thinking about spelling and grammar
The results of Glynn et at. (1982) indicate that, as they try to write. Editing, in itself, is not the problem. Editing
when planning and translating operations are de- is usually necessary if we want to end up with something satisfac-
tory. The problem is that editing goes on at the same time as
manded simultaneously from writers, processing producing. The editor is, as it were, constantly fiddling with
capacities are taxed severely. These researchers what he's doing while he's in the middle of trying to do it.
conducted two experiments on persuasive writing. (Elbow, 1973, p. 5)
In their first experiment, production was lowest
when preliminary-draft arguments were generated The Glynn et al. (1982) experiments provide em-
in the form of sentences. Production increased sig- pirical justification for the use of the freewriting
nificantly when arguments were generated in the technique. If idea generation is the most important
form of ideally sequenced propositions (idea units concern of the writer, then perhaps mechanics, sen-
summarized in several words). Production was tence formation, and sequence operations should be
highest when arguments were generated in the form postponed until later drafts of a document. If these
of unsequenced propositions. In the latter circum- latter operations are postponed, processing capacity
stance, when both the sentence-formation operation can be focused on the production of ideas in simple,
and the sequence operation were eliminated, writ- proposition-like forms.
ers were able to allocate more of their capacity to It should be kept in mind that good writers are
the task of argument production. These findings distinguished from poor writers by superior knowl-
were later replicated by Hayes and his colleagues edge as well as superior technique. Thus, the free-
(J. R. Hayes, personal communication, April 4, writing technique alone should not be expected to
1985). remediate deficiencies in the idea generation of
In the second experiment by Glynn et al., the poor writers. The value of the technique resides
availability of additional capacity was found to solely in its ability to focus processing capacity
have no effect on the argument production of writ- temporarily on idea generation.
ers with low verbal ability. They interpreted their
findings as suggesting that additional capacity can Automaticity. Theorists differ on how much of
provide writers only with the opportunity to draw the writing process can become automatic with
from their pool of persuasive arguments; the provi- practice. Flower and Hayes (1980) have argued that
sion of additional capacity cannot by itself remedi- a great deal of writing process, including the pro-
ate deficiencies in the knowledge stores of low- duction of grammatical sentences, can become au-
ability writers. tomatic and demand very little conscious attention.
In these experiments, writers' metacognitive Cooper and Matsuhashi (1983) have proposed a
knowledge was reflected in a tendency to adopt contrasting view; they argued that even though
planning to produce written discourse can be guided by underly- the essays they wrote. As a result, the essays were
ing syntactic and discourse competence ... it is nearly totally a more coherent and cohesive.
conscious, non-automatic planning process .... [and that] cer- Perhaps the most sophisticated study of the rela-
tain phrase and sentence schema may provide local constraints,
and a discourse schema may guide the process from a lofty tionship between topic-specific knowledge and
abstraction level; but each clause is still new ... even the most quality of writing was conducted by Langer (1984),
effortless writing requires close attention to planning every who analyzed two sets of papers written for a high
clause and sentence. (p. 34) school history class. Langer compared the ratings
of holistic scores, syntactic complexity, coherence,
It is unlikely that writers can achieve a high level audience, and function that these papers earned
of automaticity in the generation, coordination, and with measures of topic knowledge taken prior to the
integration of novel, complicated ideas; however, writing task. Not surprisingly, Langer discovered a
they can achieve it with sentence formation, punc- strong and consistent relationship between topic-
tuation, and spelling. The extent to which the latter specific background knowledge and the quality of
processes are automated will determine the extent student writing and, moreover, found that the orga-
to which additional processing capacity can be ded- nization of the students' knowledge influenced the
icated to idea generation while writing. The de- quality of their writing:
mands of the writing task, especially in situations in
which novice writers have been given tasks requir- When the assignment calls for a simple reiteration of facts, or
ing more resources than they currently possess, elaboration of a given idea, a large amount of unintegrated (or
have been known to obstruct the learning of writing loosely linked) information will suffice. However, when the
student is required to present a thesis, analyze, and defend it, the
skills (Scardamalia, 1981). degree of organization of knowledge, as opposed to simple flu-
We recommend that novice writers be encour- ency, will determine success. (p. 41)
aged to cope with their limited processing capaci-
ties in two ways: first, by concentrating on idea The findings of several other studies suggest that
production and temporarily setting aside the consid- poor writers may have difficulty bringing relevant
eration of other subprocesses; and, second, by auto- knowledge to bear when writing because they are
matizing sentence formation and mechanics opera- preoccupied with the more basic skills of spelling,
tions to whatever extent is possible. The latter can punctuation, and word choice (Atwell, 1981;
be achieved by systematic practice and such exer- Birnbaum, 1982; Kellogg, 1987; Pianko, 1979).
cises as "sentence combining." Good writers, in contrast, are more concerned
with their purpose and the representation of their
ideas.
Topic Knowledge Limitations
Does the amount and organization of a writer's Idea Revision. According to Gagne (1985, p.
knowledge about a topic influence the quantity and 217-218), revision involves two stages: (1) eval-
quality of the ideas he or she produces when writ- uating what has been written to determine how well
ing? Relevant to this question are several "bio- it meets one's goals, and then (2) remediating un-
graphical" studies in the area of reading com- successful parts of the product. There is a develop-
prehension that examined the effect of readers' mental sequence for these two stages. Studies of
personal history and life experiences on their con- elementary school children conducted by Bartlett
struction of meaning (e.g., Anderson, 1977; Good- (1982) and by Scardamalia and Bereiter (1983)
man & Goodman, 1978; Harste, Burke, & Wood- show that fourth graders are capable of evaluating
ward, 1982). These studies pointed out the ways in their writing and recognizing the presence of prob-
which readers' organization of their past experi- lems; however, only the children in the higher
ences affected their responses to new experiences. grades are able to tackle successfully these prob-
Also relevant to the question of topic knowledge lems during revision. But even in high school and
are studies that have measured or manipulated stu- college, there are some students who fail to evaluate
dents' knowledge of particular domains, and that their writing carefully enough to detect problems;
then asked the students to write about topics in those and the problems they do detect tend to be related to
domains. Voss et al. (1980) demonstrated that col- mechanics rather than ideas (Bridwell, 1980;
lege students highly knowledgeable about a particu- Faigley & Witte, 1981; Glynn et al., 1982; Stallard,
lar topic, such as baseball, included more elab- 1974).
orative detail and fewer irrelevant shifts of topic in Because beginning and remedial writers do not
revise their ideas sufficiently, Cohen and Scar- used to capture and develop ideas when writing. It
damalia (1983) developed a program of self-diag- is not possible here to provide in-depth instructions
nostic statements (e.g., "needs an example to ex- for using each particular method; readers are re-
plain idea") to prompt revision behavior. The ferred to the original sources cited for that purpose.
program was successful, with the most frequent re- Teachers often recommended the following ways
vision being the addition of more ideas; it was of developing ideas:
prompted by the statement "too few ideas."
Besides prompts, there are other things a teacher 1. Writing down any ideas related to an assigned
can do to encourage students to review their written topic
ideas. Gagne (1985) has recommended that the 2. Keeping a journal of interesting ideas and
teacher provide students with practice and feed- events
back, and a wide variety of examples and nonexam- 3. Using source materials
pIes of writing problems. In addition, we recom- 4. Thinking in terms of analogies
mend that students be encouraged to revise other 5. Discussing ideas with a peer or teacher who
students' writings. Because there is less ego in- asks questions (i.e., "conferencing"; Freed-
volvement, some students may find it easier to ex- man, 1982).
amine critically the work of others. Eventually,
they will be able to apply a degree of objectivity to Some beginning writers discover these methods
their own work. We further recommend that stu- on their own, whereas others rely upon their teach-
dents be encouraged to give a formal presentation in ers for an introduction. Bereiter and Scardamalia
class based on their written ideas. The questions (1979) have endorsed the teaching of such methods
that listeners raise may signal biases and ambigu- because, they say, beginning writers often are not
ities in the speaker's ideas. Later, when the speaker even aware of which topics they know a little or a
is revising, he or she will be more sensitive to these great deal about. Relatively simple, informal meth-
biases and ambiguities, realizing that it is not possi- ods such as these often evolve into complex, formal
ble for readers to ask questions. methods.
We will now discuss some of these formal, com-
Incubation. Although the effects of an incuba- plex methods of idea development that are taught to
tion period are well documented in the problem- writing students. We propose a framework for
solving literature, there has been very little direct viewing these methods: some are derived from a
study of the effects of incubation on idea generation general problem-solving method, based largely on
and revision during writing. Quite possibly, an in- cognitive psychology, whereas others are derived
cubation period would allow a writer's biases and from a general neoclassical invention method,
idiosyncratic references to become more apparent based on traditional schools of rhetoric.
to him or her.
Bereiter and Scardamalia (1979) have spoken of
Problem Solving
the supposed benefit of a "cooling-off" period for
obtaining a psychological distance between one's Problem solving, which is a general method that
initial draft and any subsequent revisions. How- is rooted in cognitive psychology, has been articu-
ever, in a study that Bracewell, Bereiter, and Scar- lated by Young (1976), Flower and Hayes (1980),
damalia (1979) conducted, they did not find that a Flower (1981), and many others. It revolves around
one-week incubation period for revision led to any planning processes in writing.
significant improvement in papers written by Problem solving has been defined as a movement
novice writers (over and above what would have from a perceived need or sense of difficulty,
been generated if the papers had been revised im- through the exploration of relevant information and
mediately after being written). More research, em- ideas, to the intuition of possible solutions (Young,
ploying longer intervals and complicated writing 1976). After identifying a problem, the writer's
problems, needs to be carried out. next step in his or her problem-solving planning is
to define the writing goals (Anderson, 1985). Not
A Framework for Methods of Idea all the goals in one's writing relate to content. Be-
reiter (1980) has identified three basic levels of goal
Development
orientation: (1) structure, consisting of text mode,
The purpose of this section is to propose a frame- paragraph, and sentence; (2) content, consisting of
work to view some of the many methods that are ideas and their relationships; and (3) purpose, con-
sisting of the goals of the writer and the relationship Generating ideas in a simple associative manner
with the audience. may be the most effective planning procedure for
Idea generation using the problem-solving meth- children, but Martlew (1983) believes that this type
od involves generating ideas or solutions to prob- of facilitation overlooks the need to plan and
lems on syntactic, semantic, and pragmatic levels. organize text in a cohesive manner. Simple asso-
Because the problem-solving method is global in ciative memory searches must be integrated with
scope, Young (1981) has suggested that we regard other planning procedures if children are to learn to
the more specific "invention" techniques as part of use appropriate semantic and syntactic structures,
this generalized process of inquiry, that is, the prob- economy in organization, and originality of thought
lem-solving method. Young argues that, apart from (Kroll, 1978; Ochs, 1979). Indeed, unless the
the problem-solving method, invention techniques teacher is careful, associative listing could rein-
often function merely as retrieval aids. Within the force what Bereiter (1980) calls "associative
general problem-solving approach, however, spe- dumping," an immature strategy used by the begin-
cific invention techniques can contribute to the dis- ning writer who believes that writing simply in-
covery of new truths and to eventual solutions. volves dumping on paper everything he or she
Three specific invention techniques particularly knows about a particular topic.
useful within the context of the general problem-
solving method are associative listing, brainstorm- Brainstorming. A variation of associative list-
ing, and synectics, which will now be discussed. ing, the brainstorming technique, involves generat-
ing as many ideas as possible, with no immediate
Associative Listing. Bereiter and Scardamalia judgments about their quality (Osborn, 1953).
(1979) found that simply asking a child to write Johnson, Parrott, and Stratten (1968) demonstrated
more could be effective in improving fluency and the effectiveness of brainstorming in a variety of
idea production. Similarly, they found that it was creativity tasks, such as the generation of plot titles.
more productive to ask children to list single words Participants who brainstormed generated a greater
associated with a topic than it was to ask them to list number of ideas, including a greater number of
ideas. Listing words, it seems, made it possible for high-quality ideas, than participants who were
the children to access their memories easily while asked to generate only high-quality ideas. Brain-
avoiding the syntactic demands that are associated storming does include a judgment component, but it
with generating ideas in the form of sentences. A is reserved until later when all possible ideas have
listed word strongly associated with an entire idea been generated. Group and individual brainstorm-
can act as a retrieval cue for the idea when it comes ing procedures can be easily implemented in class-
time to write. Such listing also allows a freer flow of room writing situations by teachers who wish to
ideas because writing sentences takes longer than stimulate their students' idea production (Glover,
thought and can slow or block idea production. 1979; Glover & Gary, 1976).
Similarly, Caccamise (1981) has suggested that
writers often believe they are out of ideas when Synectics. The technique of synectics cap-
actually they have simply exhausted the contents of italizes on analogical reasoning. It requires writers
one of many possible search sets that hold the acti- to compare an assigned topic to topics they might
vated idea nodes (see Figures 1 and 2). Because not typically consider to be related. For instance, a
working memory limits the number of activated writer might be asked to describe the similarities
nodes, writers might think that they have come to between the topic he or she has chosen (e.g., social
the end of their ideas, when actually they have only life in the business office) and a distantly related
come to the end of one cluster. Perhaps because topic (e.g., the African jungle). In this case, an
they have considered the associations incorporated analogy might be drawn between the office water
in that cluster in depth, elaborated them, and possi- cooler and the jungle water hole. Even though the
bly summarized them, writers will acquire a pre- analogies often are farfetched and not used in the
mature sense of closure. To avoid this sense of actual composition, they may help writers to flex or
closure, Caccamise suggests that writers may want warm up their minds in a way that will help them to
to distribute their idea generation searches over sev- perceive a striking and relevant analogy when it
eral separate times rather than on one occasion. In does present itself.
this manner, writers will conduct searches that are The rationale behind synectics is based in part on
not limited by previously generated items or by studies of creative individuals. For cognitive theo-
search strategies that have been prematurely closed. rists, creativity is not just a result of an individual's
unique system of associational bonds but is also an model under critical settings of different input vari-
outcome of a person's handling of information and ables), (6) taxonomize (by finding a set of constants
combining of data to develop effective solutions. that provides an underlying explanation for the vari-
According to Cropley (1970), creative thinkers tend ation in the critical cases generated), and (7) dimen-
to take greater risks in their thinking and are more sionalize (by placing a structure on the taxonomy).
willing to take in large quantities of information.
Furthermore, creative individuals also are more
Neoclassical Invention
likely to code their information (i.e., connect it to
prior knowledge so as to make it meaningful) in The method of neoclassical invention and the two
unusual or novel ways. Cropley believes that the that follow were identified by Young (1976) in his
willingness to treat apparently unrelated sets of data topographical survey of major invention methods.
as if they were related is a disposition characteristic Neoclassical invention is based on the rhetoric of
of creative individuals. This disposition is one as- Aristotle, Cicero, and Quintillian; it involves the art
pect of "divergent thinking." The opposite dis- of composing persuasive arguments in support of a
position is known as "convergent thinking," proposition. Once a thesis has been determined, the
which is characteristic of individuals who tend to writer makes appeals based on moral character
see a narrower range of connections among data or (ethos), based on the audiences' emotions (pathos),
ideas. and based on logic (logos). Arguments or ideas are
Synectics, then, is a way of stimulating the work- discovered by use of topics, or strategic prompts
ings of a creative mind. It is suggested that synec- that involve a writer in developing his or her argu-
tics can help convergent individuals begin to see ments by using definitions, comparison, contrasts,
data equivalences that are not readily apparent to antecedents, consequences, and contraindications
them otherwise. (see Table 1). The continuing usefulness of clas-
The importance of invention techniques, such as sical invention techniques is attested to by the de-
synectics, has a basis in developmental as well as in velopment of a computer program based on Aristo-
cognitive psychology. Growth in the ability to coor- tle's "invention topics" (Bums, 1984).
dinate physical concepts, such as height and
weight, has come to be considered a benchmark in
cognitive development (Piaget & Inhelder, 1969). The Dramatist Method
As Scardamalia (1981) has pointed out, cognitive Whereas the previous method prompted the gen-
development is construed as taking more and more eration of persuasive arguments, the dramatist
variables into account during single judgmental method (Burke, 1969) prompts the generation of
acts. Training in invention techniques may help be- motives. It relies on a pentad of strategic prompts-
ginning writers to learn how to develop their ideas.
It is this progressive ability to develop ideas, Scar-
damalia asserts, that underlies advances in writing, Table 1. Catalogues of Techniques for Idea
such as the movement from unelaborated to elabo- Generation
rated discourse and the movement from a single- Examples of topics
argument thesis defense to a multifaceted defense
that anticipates counterarguments. Argument from opposites
A fortiori-argue from less to more (e.g.,
Techniques in Combination. Specific prob- "care for your family, care for your
lem-solving techniques, such as associative listing, country")
brainstorming, and synectics, can be combined into
formal, multitechnique methods such as the one Dissociation
proposed by Collins and Gentner (1980), which in- appearance/ reality
cludes aspects of techniques we have already dis- means/ends
cussed, plus several others. Altogether, Collins and individual! group
Gentner recommend seven ways of manipulating a spirit/letter
writing topic: (1) identify dependent variables, (2) Methods of paragraph development
generate critical cases, (3) compare to similar cases narrative
(analogize), (4) compare to dissimilar cases (con- classification
trast or differentiate), (5) simulate (by running a
Table 2. Dramatist Pentad a events; contrast a chunk of experience with other
units; say how much a thing can change and still be
1. Act: What exactly was done? the same; and locate events in a larger class of
2. Scene: Where did it happen? things, in a temporal sequence, in a physical con-
3. Agent: Who did it? text, and in a class system. Odell found that the
4. Agency: By what means was it done? students did learn to examine data more thoroughly
5. Purpose: Why was it done? and to perform a greater number of intellectual op-
aEach of the five heuristic probes can be seen as an aspect of the erations on these data. They also learned to present
other; for example, war can be an act, a scene, an agency (as a more evidence to support their statements. Odell,
means to an end), a purpose (in schemes proclaiming a cult of however, cautioned that his conclusions are tenta-
war), and as a collective act.
tive and asked whether other idea development
methods might not have had an even greater influ-
act, scene, agent, agency, and purpose-to evoke ence upon students' writing.
statements about the motive implicit in rhetorical
acts (see Table 2).
Questions to Prompt Future Research
Tagmemics The idea generation process in writing represents
The method of tagmemics is based on language an exciting and still largely unexplored area of re-
theory and was specifically developed from axioms search. These six questions about the idea genera-
of language structure. The method shares a goal tion process can guide the formulation of compre-
common to classical invention, that is, finding ar- hensive research plans or programs:
guments that are likely to produce changes in the
audience. Moreover, the method assists in the re- 1. How does a student become competent at
trieval of information and the analysis of prob- generating ideas during writing?
lematic data. Unlike classical invention, however, 2. Why do some students become competent in
it is not designed for use with a particular mode of idea generation faster than other students?
discourse. It involves responding to a series of 3. How do students become, not just competent,
questions organized under the special organiza- but creative in their idea generation?
tional concepts called "particle," "wave," and 4. What can a teacher do to help a novice writer
"field" (see Table 3). become an expert in generating ideas?
Odell (1974) taught a core group of Pike's tag- 5. How should a teacher assess a student's com-
memic discovery procedures to writing students in petency and creativity in the skill of idea gen-
order to evaluate the procedures' effectiveness. eration during writing?
Specifically, he taught students how to segment 6. How should a teacher's assessment pro-
Table 3. Tagmemic Invention

Knowledge of a thing (a) Particle (b) Wave (c) Field
1. How it differs from ev- Describe its features What are the features of How do the parts of the
erything else (contrast) the novel writing pro- novel fit together?
cess?
2. How much it can change At what point would it How does this process How does this novel as
and still be itself become a poem or a differ from other simi- a system work in con-
report? lar processes? trast to other novels?
3. How it fits into larger How does it relate to a How does the process of How does the novel as a
systems system of literature? the novel fit into the system relate to the
process of literature? larger system of
which it is a part, that
is, literature?
cedures change over time as the student be- mation-processing model emphasized the impor-
comes increasingly proficient at generating tant role of our limited-capacity working memory,
ideas? the cognitive workbench on which we fashion our
ideas while writing. Much of the raw material for
Thinking about the preceding fundamental re- our ideas comes from long-term memory, the
search questions can help in the formulation of knowledge base or propositional storehouse that we
plans for future research on the idea generation pro- strategically probe by means of retrieval cues and
cess. To illustrate, we have formulated a research search sets.
plan that we present in the next section. We found that idea generation could be better
understood by combining the information-process-
An Illustrational Plan for Future ing model with a model of a problem-solving sys-
tem. In this combined model, purpose plays an es-
Research sential role in helping the writer to narrow the focus
of his or her idea generation.
Good writers usually have more ideas to write The models of writing we discussed helped us to
about than poor writers; however, one should not conceptualize the writing process as a set of compo-
conclude that good writers are good simply because nent, interrelated subprocesses of which idea gener-
they may have more ideas. They may be good writ- ation is one. There were some general similarities
ers because they can strategically manipulate their among the models discussed: for instance, the
ideas and coordinate their manipulations with other Hayes and Flower model and the Cooper and Mat-
writing subprocesses. suhashi model are hierarchical in nature, with idea
To properly test this notion, one could control generation seen as a subprocess of a higher level
topic knowledge before examining writers' use of subprocess. And all models defined proficient writ-
idea manipulation strategies. For example, a class ers as those who could access and coordinate sub-
of writers could be taught a body of unfamiliar but processes in a flexible manner in response to differ-
related information which they would learn to a ent demands of the writing task, such as those posed
given level of mastery. Ideally, this topical infor- by the "assignment," the audience, and the writ-
mation should form a conceptual system (e.g., ers' personal goals. All models also stressed the
"photosynthesis" if a science topic was called for). importance of possessing a repertoire of sub-
The writers would then be given a series of writ- processes that generate not just content ideas~ ?ut
ing assignments that would require them to tap pri- metacognitive ideas that relate to the overall wntmg
marily this body of information for ideas. During process.
the writing and afterward, writers would be asked to The commonalities of the models are over-
report verbally their thoughts. These verbal reports
shadowed, however, by their unique aspects. For
and associated written products would be analyzed example, the Hayes and Flower model assigns spe-
to gain insight into how writers manipulate their
cial importance to problem solving, viewing the
ideas. Differences in the quality of the written prod- rhetorical situation as a problem that must be solved
ucts would be due largely to variations in the writ-
by juggling rhetorical demands. Rose's model is
ers' ability to manipulate their ideas and express unique because it stresses the idea-generating be-
them in writing. havior of deficient writers, or high-blockers, as
This plan would also lend itself to an examination much or more than the behavior of proficient writ-
of the ways in which the writing process could be ers. The Cooper and Matsuhashi model, with its
used to help students better understand the compli-
stress on sentence roles within discourse types, is
cated conceptual systems that they encounter in unique in the weight it gives to the medium as part
content areas, such as science, mathematics, and
of the message. And, finally, Wason's two contra-
social studies. In other words, this plan could pro-
dictory models, though generic and oversimplified,
vide additional empirical support for curricula that
remind us that more than one model may be neces-
stress the notions of "writing to learn" and "writ-
sary to conceptualize the idea generation processes
ing across the content areas."
of different writers, or of the same writer in differ-
ent developmental stages.
Summary and Conclusions In this chapter, we identified two general sources
of writing difficulty. One was a writer's processing
This chapter has examined idea generation in the capacity limitations. We argued that writing is a
writing process. Our discussion of the human infor- cognitively demanding process because it can call
for a variety of complex subprocesses to be carried methods which we have discussed. On the other
out concurrently. Although these subprocesses can hand, if a student seems to have good content ideas
severely tax novice writers' limited processing ca- that are poorly developed and organized, the in-
pacities, such procedures as freewriting can help structor might want to show this student writer how
writers to cope with their limitations. to use discourse forms, sentence functions, roles,
With experience and practice, writers can make and plans to develop ideas for organizing and elab-
some processes highly automatic and, therefore, orating on existing ideas.
less demanding on their processing capacities. And Finally, we posed some fundamental questions
even though it may not be possible to achieve auto- for future research. We also presented an illustra-
maticity in idea generation, the relative automation tional plan: to examine how ideas are manipulated
of punctuation, spelling, and sentence construction and expressed in writing, students would first mas-
makes it possible to dedicate more processing ca- ter a conceptual system that they would then "tap"
pacity to idea generation. Thus, we recommend that for a series of writing assignments. This plan could
novice writers cope with their limited processing provide additional support for curricula that empha-
capacities in two ways: first, by concentrating on size "writing to learn" and "writing across the
idea generation, while temporarily setting aside content areas."
other subprocesses, and, second, by automatizing
mechanics and sentence-formation operations to
whatever extent is possible with the aid of appropri- References
ate practice exercises.
Anderson, J. R. (1985). Cognitive psychology and its implica-
The second general source of writing difficulty tions (2nd ed.). San Francisco, CA: W. H. Freeman.
we identified was a writer's topic-knowledge lim- Anderson, R. C. (1977). The notion of schemata and the educa-
itations. We suggested that the amount and organi- tional enterprise. In R. C. Anderson, R. J. Spiro, & W. E.
zation of a writer's topic knowledge influences both Montague (Eds.), Schooling and the acquisition of knowl-
edge. Hillsdale, NJ: Erlbaum.
the quantity and quality of the ideas produced when Atwell, M. (1981, November). The evolution of text: The inter-
writing. We further suggested that poor writers tend relationship of reading and writing in the composing process.
to downslide by concentrating on mechanics, Paper presented at the annual meeting of the National Council
whereas good writers tend to upslide by concentrat- of Teachers of English, Boston, MA.
ing on ideas. Baddeley, A. D. (1978). The trouble with levels: A reexamina-
tion of Craik and Lockart's framework for memory research.
It was pointed out that revision involves two Psychological Review, 85, 139-152.
stages, evaluation and remediation. Beginning and Bartlett, E. J. (1982). Learning to revise: Some component pro-
remedial writers may need intensive instruction in cesses. In M. Nystrand (Ed.), What writers know: The lan-
one or both stages. The instruction should include guage process, and structure of written discourse (pp. 345-
363). New York: Academic Press.
self-diagnostic prompts, practice, feedback, exam- Benton, S. L., Kraft, R. G., Glover, J. A., & Plake, B. S.
ples and nonexamples, revising the writing of oth- (1984). Cognitive capacity differences among writers.
ers, and speaking in formal settings about the ideas Journal of Educational Psychology, 76, 820-834.
one intends to revise. In addition, allowing ideas to Bereiter, C. (1980). Development in writing. In W. Gregg & E.
incubate prior to revision may help some writers to R. Steinberg (Eds.), Cognitive processes in writing (pp. 73-
93). Hillsdale, NJ: Erlbaum.
better detect biases and ambiguities in their ideas. Bereiter, c., & Scardamalia, M. (1982). From conversation to
We next proposed a framework for viewing the composition: The role of instruction in a developmental pro-
formal methods for capturing and developing ideas cess. In R. Glaser (Ed.), Advances in instructional psychology
when writing. A relatively recent development is (pp. 1-64). Hillsdale, NJ: Erlbaum.
Birnbaum, J. C. (1982). The reading and composing behaviors
the general problem-solving method that incorpo- of selected fourth- and seventh-grade students. Research in
rates more specific techniques, such as associative the Teaching of English, 16, 241-260.
listing, brainstorming, and synectics. Bracewell, R. J., Bereiter, c., & Scardamalia, M. (1979,
Although a complete method of idea generation April). A test of two myths about revision. Paper presented at
in the writing process is still to be developed, the the annual meeting of the American Educational Research
Association, San Francisco.
methods that are available suggest several ways in Bridwell, L. S. (1980). Revising strategies in twelfth-grade stu-
which instruction in writing can be improved for dents' transactional writing. Research in the Teaching of En-
individual beginning writers with specific prob- glish. 14. 197-222.
lems. For example, if a student writer has shown Britton. B. K .• Glynn, S. M., & Smith, J. W. (1985). Cognitive
demands of processing expository text: A cognitive work-
thoughtfulness in class discussions but displays a bench model. In B. K. Britton & J. Black (Eds.),
lack of content ideas in a paper, an instructor might Understanding expository text (pp. 227-248). Hillsdale, NJ:
recommend one or more of the idea development Erlbaum.
Britton, J., Burgess, T., Martin, N., McLeod, A., & Rosen, H. nitive demands. Journal ofEducational Psychology, 74, 557-
(1975). The development of writing abilities (11-18). 567.
London: Macmillan. Goodman, D., & Goodman, Y. (1978). Reading of American
Burke, K. (1969). A grammar of motives. Berkeley, CA: Uni- children whose language is a stable rural dialect or a lan-
versity of California Press. guage other than English. (Final Report, Contract No.
Burns, M. H. (1984). Recollections of first-generation comput- 100-30087). Washington, DC: National Institute of
er-assisted prewriting. In W. Wresch (Ed.), The computer in Education.
composition instruction (pp. 15-33). Urbana, IL: National Graesser, A. (1984). The impact of different information sources
Council of Teachers of English. on idea generation: Writing off the top of our heads. Written
Caccamise, D. J. (1981). Cognitive processes in writing: Idea Communications, 1, 341-365.
generation and integration. Unpublished doctoral disserta- Gregg, L. W., & Steinberg, E. R. (Eds.). (1980). Cognitive
tion. University of Colorado, Boulder, Colorado. processes in writing. Hillsdale, NJ: Erlbaum.
Chafe, W. (1977). The recall and verbalization of past experi- Halliday, M. A. K., & Hasan, R. (1976). Cohesion in English.
ence. In R. W. Cole (Ed.), Current issues in linguistic theory London: Longman.
(pp. 215-246). Bloomington, IN: Indiana University Press. Harste, J. C., Burke, C. L., & Woodward, V. A. (1982). Chil-
Cohen, E., & Scardamalia, M. (1983, April). The effects of dren's language and the world: Initial encounters with print. In
instructional intervention in the revision of essays by grade six J. A. Langer & M. Smith-Burke (Eds.), Reader meets author:
children. Paper presented at the annual meeting of the Ameri- Bridging the gap (pp. 105-131). Newark, DE: International
can Educational Research Association, Montreal. Reading Association.
Collins, A., & Gentner, D. (1980). A framework for a cognitive Hayes, J. R., & Flower, L. S. (1980). Identifying the organiza-
theory of writing. In E. R. Steinberg & L. W. Gregg (Eds.), tion of writing processes. In L. W. Gregg & E. R. Steinberg
Cognitive processes in writing. Hillsdale, NJ: Erlbaum. (Eds.), Cognitive processes in writing. Hillsdale, NJ:
Cooper, C. & Matsuhashi, A. (1983). A theory of the writing Erlbaum.
process. In M. Martlew (Ed.), The psychology of written lan- Johnson, D. M., Parrott, G. L., & Stratten, R. P. (1968). Pro-
guage: Developmental and educational perspectives (pp. 3- duction and judgment of solutions to five problems. Journal of
39). New York: Wiley. Educational Psychology, 59 (Monograph Suppl. No.6).
Cooper, M. (1984). The pragmatics of form: How do writers Johnston, W. A., & Heinz, S. P. (1978). Flexibility and capacity
discover what to do when? In R. Beach & L. Bridwell (Eds.), demands of attention. Journal of Experimental Psychology:
New directions in composition research (pp. 109-126). Lon- General, 107, 420-435.
don: Guilford. Kean, D. K., Glynn, S. M., & Britton, B. K. (1987). Writing
Cropley, A. J. (1970). S-R psychology and cognitive psychol- persuasive documents: The role of students' verbal aptitude
ogy. In P. E. Vernon (Ed.), Creativity (pp. 116-125). Har- and evaluation anxiety. Journal of Experimental Education,
mondsworth, Middlesex, England: Penguin Books. 55, 95-102.
Daiute, C. A. (1980). A psycholinguistic study of writing. Kellogg, R. T. (1987). Effects of topic knowledge on the alloca-
Unpublished doctoral dissertation. Teachers College, Colum- tion of processing time and cognitive effort to writing pro-
bia University, New York. cesses. Memory and Cognition, 15, 256-266.
Daiute, C. A. (1981). Psycholinguistic foundations of the writ- Kinneavy, J. L. (1971). A theory of discourse. Englewood
ing process. Research in the Teaching of English, 15, 5-22. Cliffs, NJ: Prentice-Hall.
Daiute, C. (1984). Performance limits on writers. In R. Beach & Kintsch, W. (1980, July). Psychological processes in discourse
L. Bridwell (Eds.), New directions in composition research production. Paper presented at the workshop on "Psycho-
(pp. 205-224). London: Guilford. linguistic Models of Production," University of Kassel, West
Elbow, P. (1973). Writing without teachers. New York: Oxford Germany.
University Press. Kintsch, W. (1986). Learning from text. Cognition and Instruc-
Faigley, L., & Witte, S. (1981). Analyzing revision. College tion, 3(2), 87-108.
Composition and Communication, 32, 400-414. Kroll, B. M. (1978). Cognitive egocentrism and the problem of
Flower, L. S. (1981). Problem-solving strategies for writing. audience awareness in written discourse. Research in the
New York: Harcourt Brace Jovanovich. Teaching of English, 12, 269-281.
Flower, L. S., & Hayes, J. R. (1980). The cognition of discov- Langer, J. (1984). The effects of available information on re-
ery: Defining a rhetorical problem. College Composition and sponses to school writing tasks. Research in the Teaching of
Communication, 31, 21-32. English, 18, 27-44.
Freedman, S. (1982). The student teacher writer conference: Martlew, M. (1983). The development of writing: Communica-
Key techniques. Journal ofEnglish Teaching Techniques,12, tion and cognition. In F. Coulmas & K. Ehlich (Eds.), Writing
38-45. infoeus (pp. 257-275). New York: Mouton Publishers.
Gagne, E. D. (1985). The cognitive psychology of school/earn- Moffett, J. (1968). Teaching the universe of discourse. Boston:
ing. Boston: Little, Brown. Houghton Mifflin.
Glover, J. A. (1979). The effectiveness of reinforcement and Newell, A., Shaw, J. C., & Simon, H. (1963). The processes of
practice for enhancing the creative writing of elementary creative thinking. In H. E. Gruber, G. Terrell, & M.
school children. Journal of Applied Behavior Analysis, 12, Wertheimer (Eds.), Contemporary approaches to creative
487. thinking: A symposium held at the University of Colorado (pp.
Glover, J. A., & Gary, A. L. (1976). Procedures to increase 63-119). New York: Atherton Press.
some aspects of creativity. Journal of Applied Behavior Anal- Ochs, E. (1979). Planned and unplanned discourse. In T. Giron
ysis, 9, 79-84. (Ed.), Syntax and semantics, Vol. 12: Discourse and syntax.
Glynn, S. M., Britton, B. K., Muth, K. D., & Dogan, N. New York: Academic Press.
(1982). Writing and revising persuasive documents: Cog- Odell, L. (1974). Measuring the effect of instruction in prewrit-
ing. Research in the Teaching of English, 8, 228-240. Stallard, C. K. (1974). An analysis of the writing behavior of
Osborn, A. F. (1953). Applied imagination. New York: good student writers. Research in the Teaching of English, 8,
Scribners. 206-218.
Piaget, 1., & Inhelder, B. (1969). The psychology of the child. Tetroe, 1. (1984, April). Information processing demand ofplot
New York: Basic Books. construction in story writing. Paper presented at the meeting
Pianko, S. (1979). A description of the composing processes of of the American Educational Research Association, New
college freshmen writers. Research in the Teaching of En- Orleans.
glish, 13, 5-22. van Dijk, T. A., & Kintsch, W. (1983). Strategies of discourse
Rose, M. (1984). Writer's block: The cognitive dimension. comprehension. New York: Academic Press.
Carbondale, IL: Southern Illinois University Press. Voss, 1. F., Vesonder, G. T. V., & Spilich, G. 1. (1980). Text
Scardamalia, M. (1981). How children cope with the cognitive generation and recall by high-knowledge and low-knowledge
demands of writing. In C. H. Frederiksen & 1. F. Dominic individuals. Journal of Verbal Learning and VerbaLBehavior,
(Eds.), Writing: The nature, development and teaching of 19, 651-667.
written communication (Vol. 2, pp. 81-103). Hillsdale, Nl: Wason, P. C. (1978). Specific thoughts on the writing process.
Erlbaum. Paper presented at the Cognitive Processes in Writing-Inter-
Scardamalia, M., & Bereiter, C. (1983). The development of disciplinary Symposium on Cognition, Carnegie-Mellon
evaluative, diagnostic, and remedial capabilities in children's University.
composing. In M. Martlew (Ed.), The psychology of written Young, R. (1976). Invention: A topographical survey. In G.
language: Developmental and educational perspectives (pp. Tate (Ed.), Teaching composition: 10 bibliographical essays
67-95). New York: Wiley. (pp. 1-43). Fort Worth, TX: Texas Christian University.
Scardamalia, M., Bereiter, c., & Goelman, H. (1982). The role Young, R. (1981). Problems and the composing process. In C.
of productive factors in writing ability. In M. Nystrand (Ed.), H. Frederiksen & 1. F. Dominic (Eds.), Writing: The nature,
What writers know: The language process and structure of development, and teaching of written communication (Vol. 2,
writers discourse (pp. 173-210). New York: Academic Press. pp. 59-66). Hillsdale, Nl: Erlbaum.
CHAPTER 19
Creating the Conditions for Creativity

in Reader Response to Literature
Carolyn A. Colvin and Roger Bruning
Introduction any other activity that can be offered in an educa-

tional setting, the opportunity for students to write
In many classrooms today, the major task of stu- can give voice to the self and can channel self-
dents is to acquire and demonstrate a set of skills expression to productive, generative, and creative
that are consistent with instructional objectives. ways. To this end, we focus in this chapter on a
Many educators would agree wholeheartedly with particular kind of writing-reader response-in
this focus and its implications of goal directedness, which writing is conceived of as an act of respon-
objectivity of measurement, and accountability. At sive self-expression, of composition, and of cre-
the same time, however, an objectives-oriented ativity. The writing we will describe is not done for
classroom prescribes a particular role for the stu- evaluation by a teacher, as most school writing cur-
dent, that of "achiever of objectives." Certainly, rently is, but instead serves as a tool for construct-
the real lives of students are vastly richer in their ing and negotiating meaning with others. In re-
many dimensions than this narrow role as achiever sponse writing, students "write to learn" and write
of objectives can give expression to. To the extent to discover what they know. Through the act of
that this role is emphasized and valued over others, writing and by sharing with others their comprehen-
students' experiences in schools correspondingly sion of what they have read, students begin to un-
have become circumscribed, even impoverished in derstand themselves and to express themselves in
their scope. Lying untouched are students' in- creative ways.
terests, their values, and, indeed, the full array of In this chapter, we will attempt to show that the
their capabilities as human beings. writing of reader response not only can create
We believe that self-expression is fundamental to the basic conditions for meaningful learning, but
all meaningful learning and that the interests and also can playa fundamental role in creativity by
values underlying students' lives must be ex- helping learners construct new meanings. To reach
pressed.1t is our contention that, perhaps more than these goals, however, the process of writing itself
must become the focal point-writing not as prod-
uct but as a creative act of writing to learn, of
Carolyn A. Colvin • Department of Teacher Education, San
Diego State University, San Diego, CA 92182. Roger "making meaning," and of "constructing one's
Bruning • Department of Educational Psychology, Univer- world through writing" (Mayher, Lester, & Prahl,
sity of Nebraska, Lincoln, NE 68588-0641. 1983). When the proper conditions are set for stu-
323
dents to write in response to what they have read, study in instruction, a position held most strongly
the elements of creativity become genuine pos- by advocates of the so-called New Criticism, with
sibilities. These conditions, we will argue, include its focus on objectivity and scientific analysis. Most
evoking the subjective reactions of the individual notable in arguing for the need to shift attention
and activating a wide range of cognitive processes. toward what readers do in their responses to literary
The uniqueness of the individual, his or her experi- works was Louise Rosenblatt, whose book Liter-
ences, and the materials at hand set the stage for the ature as Exploration was published in 1938. Al-
novel expression of creativity through the writing of though the book received relatively little attention
reader response. when first published, by the 1960s it had generated
a devoted following (Purves, 1976). Currently, in
its fourth edition, it is widely regarded as the classic
exposition of the role of the reader coming to under-
Reader Response: Responding to stand or transact with text.
Reading by Writing Within the perspective advocated by Rosenblatt,
the term "response to literature" (Cooper, 1985) or
In the following sections, we will attempt to de- simply "reader response" has come to refer to the
scribe the nature of reader response and outline the transactions of readers with literature. Cooper's
features that make it uniquely suited for self-ex- (1985) and others' use of the term (e.g., Bleich,
pression and cognitive activation. We will argue 1985) suggests the importance of the reader's role,
that activation of memory for life-events-episodic personality, and culture in the reading experience,
memory-and generation of affective responses but without displacing the centrality of the text.
based on one's episodic memory are critical to cre- Rejected is the idea that a text can be "translated"
ative reader response. Further, we will describe the by the reader; instead, reading is seen as a highly
role of the teacher in creating a context in which subjective process in which readers understand
reader response can be used to foster students' what they read through the filters of their own expe-
creativity. We will indicate steps by which diversity riences and personalities. Emotional reactions are
of reader response through writing can be encour- an essential part of reader response.
aged; that is, describe ways teachers may be able to Our use here of the term reader response thus
create "settings for spontaneity," where students grows out of this general usage (See Tompkins,
are able to discover, transform, and express the 1980) and the advocacy of Bleich (1978) and Pe-
knowledge they have (Rosenblatt, 1976). Finally, trosky (1982) of the position that comprehension is
we outline an agenda for research in which some of best demonstrated by students' extended written re-
the issues we raise can become areas of inquiry sponses to text, followed by discussions and nego-
likely to yield information useful to theorists and tiations with others of the subjective meanings of
practitioners. that text. As used in this chapter, then, reader re-
sponse refers to writing in response to reading with-
in the following general parameters: The immediate
What Is Reader Response?
stimulus for writing is a written text of some kind;
At roughly the same time but with seemingly the writing must be extended, usually beyond a
little awareness of the others' work, individuals paragraph or more; the writing must be personal,
working in literary theory and in psychology began involving the subjective perceptions of the indi-
to reject predominant notions of the objectivity of vidual; and the writing should evolve within the
knowledge and to voice views that emphasized the context of a "community of readers and writers" in
role of learners' associations, inferences, and in- which it is created, discussed, and revised. Al-
terpretations in determining meaning (Cooper, though writing and reading are the key elements,
1985). In psychology, this perspective was ad- other forms of verbal expression, such as speaking
vanced against the prevailing behaviorism by indi- and listening, may be equally important. In our
viduals such as Bartlett (1932) and Dewey (1938), discussion in this chapter, however, we will focus
and later by cognitive theorists, such as Neisser on writing and reading and their relationship to one
(1967), Bransford (Bransford, Barclay, & Franks, another and to the creative process.
1972), Rumelhart (1977), and Jenkins (1974).
At issue in literary theory was the emphasis on Why Writing in Response to Reading? When
the dominance of the objective text as a core of one examines the relationship of writing to reading
CHAPTER 19 • CREATIVITY IN READER RESPONSE 325
from a theoretical perspective, it is evident that both Researchers who have examined the relationship
reading and writing are based on the construction of of reading and writing perspectives have suggested
meaning (Birnbaum 1986; Birnbaum & Emig, that to have readers write in conjunction with their
1983; Harste, Woodward, & Burke, 1984; Pe- reading may help them become more like Iser's
trosky, 1982; Tierney & Pearson, 1983; Salvatori, ideal reader (Bleich, 1978; Comprone, 1986; Pe-
1983; Shanahan & Lomax, 1986). Writing and trosky, 1982, 1985; Tierney & Leys, 1986). As
reading are natural companions in the learning pro- Petrosky has stated:
cess. Readers draw on prior knowledge and experi-
ences, reading strategies, and the extent of their Writing about reading is one of the best ways to get students to
literary sophistication to create meaning from text. unravel their transactions so that we can see how they understand
They also must be sensitive to the devices and cues and, in the process, help them elaborate, clarify, and illustrate
their responses by reference to the associations and prior knowl-
used by the author in order to integrate these ele- edge that inform them. (Petrosky, 1982, p. 24)
ments into the meanings they construct from the
text.
Texts, in tum, have been formed by authors with Petrosky has suggested that when students are
a sense of audience, specific and nonspecific inten- asked to write about what they have read, they en-
tions, and both directly expressed and tacit mes- gage in a creative act of composing. They make
sages. In creating text, authors have at their dis- meaning from the information available to them,
posal numerous tools of literary expression- their personal knowledge, and the cultural and con-
choice of vocabulary, sentence structure, cohesive textual frames in which they find themselves (Hor-
devices, and the like-to give form and meaning to ton, 1983; Moffett, 1984). Instead of the kind of
their writing. Ultimately, however, it is in the in- reading that demands only recall or very little more,
teraction between reader and the text that com- Petrosky feels students should be encouraged to
prehension occurs and meaning is created (Langer provide a structured written response calling on sig-
& Smith-Burke, 1982). nificant prior knowledge and experiences.
Readers come more or less ably prepared for their Taking up Petrosky's theme, the kind of creative
interactions with a text (Adams & Bruce, 1982; reader response we advocate is one that
Harste et al., 1984). For instance, the preparation
that readers have had for the reading of a given text teaches us how to think, lets us read without the pressures of
recall and then, when we are finished, it begs us to speak our
(e.g., prior knowledge or experiences) is one deter- minds about what we have read and, in the process, it asks us to
miner of the quality of their transactions with text substantiate our interpretations and opinions-our reading-
(Beach & Liebman-Kleine, 1986; Holland, 1975, with evidence from our lives and the texts. (Petrosky, 1982,
1985; Petrosky, 1982). Beyond preparation, the p. 21)
purposes that direct readers in their quest for mean-
ing from the text also will have a powerful influence We believe that virtually all teachers at all educa-
on their ability to respond actively to text materials. tionallevels and in all subject areas can use writing
If readers come to a text with expectations that they in this generative sense. By creating a "community
will be successful in their task, that the reading of readers and writers" (Tierney & Leys, 1986),
process will be enjoyable, and that they will create they can nurture a new identity in students as indi-
something that will enable them to leave with more viduals whose experiences are valuable in new
than when they began, then the quality of their read- learning and can create a context, a "setting for
ing is likely to be far different from that of those spontaneity" (Rosenblatt, 1976) in which students
who anticipate problems and expect to find the can discover, transform, and express the knowl-
reading tedious. Similarly, readers who are guided edge they have.
by objectives or expect to be quizzed over their
reading will read with far more narrow purpose than Two Examples ofReader Response. To give our
those who are aiming for personal satisfaction and readers a fuller flavor of the kind of activities that
self-expression as outcomes of the reading task. might be subsumed under the rubric of reader re-
Most readers fall short ofIser's (1974, 1978) depic- sponse, we would like you to consider the following
tion of the "ideal reader" -ready to participate two examples. Both are drawn from the general
fully in the reading and to integrate the experiences area of literature. In the first case, envision yourself
of reading with the rest of his or her knowledge responding to the following instructions in an En-
(Comprone, 1986). glish class for high school seniors.
Please read the following poem carefully. Then, when you have the reading of the Pound poem and for negotiating
finished, take 15 or 20 minutes to respond to the following three meanings in their "community" of learners. The
questions in writing. The questions (Bleich, 1978) are: (I) What
initial composing of meaning is critical not only
do you see in the poem? (2) How do you feel about what you see?
and (3) What associations-thoughts and feelings-does the because it is a creative expression of each student's
poem create in you? notion of meaning, but also because it forms the
basis for further negotiations of meaning. Further
The River Merchant's Wife: A Letter l elaboration and expression of thought are at the
heart of the creative expression described earlier.
While my hair was still cut straight across my forehead
I played about the front gate, pulling flowers. The initial creative expression enables students to
You came by on bamboo stilts, playing horse, take the next step of making their writing' 'public"
You walked about my seat, playing with blue plums and to carry on further dialogue with their peers.
And we went on living in the village of Chokan: Only after the process of writing, interaction, and
Two small people, without dislike or suspicion.
rewriting has beea played out, we believe, is their
At fourteen, I married My Lord you. writing (which represents the accumulation of their
I never laughed, being bashful.
creative expression) ready for evaluation by the
Lowering my head, I looked at the wall.
Called to, a thousand times, I never looked back. teacher.
Our second example also is appropriate to the
At fifteen I stopped scowling,
I desired my dust to be mingled with yours
study of literature, but may be used in a variety of
For ever and for ever and for ever. content areas. We have used it, for instance, in
Why should I climb the look out? demonstrations of reader response for preservice
At sixteen you departed, teachers to help them begin to form their philoso-
You went into far Ku-to-yet, by the river of swirling eddies phies of teaching. It may be used in a similar man-
And you have been gone five months. ner to help students examine their notions of heroes
The monkeys make sorrowful noise overhead.
or of influential people in their lives.
You dragged your feet when you went out. We use a children's story as a stimulus, reading it
By the gate now, the moss is grown, the different mosses aloud and asking students to access their personal,
Too deep to clear them away!
The leaves fall early this autumn, in wind.
SUbjective memories in order to provide an initial
The paired butterflies are already yellow with August written response. Then, once students have com-
Over the grass in the West garden; posed this response, they are asked to distance
They hurt me. I grow older. themselves from their writing and view it more as
If you are coming down through the narrows of the river Kiang,
Please let me know beforehand,
an informed reader might. What evolves from this
And I will come out to meet you exercise is the kind of writing that acts as a mediator
As far as Cho-fu-Sa. between the subjective, personal response created
earlier and the eventual formal statement that ex-
Now further imagine yourself taking your writing, presses the student's philosophy of teaching. The
perhaps the next day, to a small group of 2 or 3 of sequence as we have used it is as follows:
your classmates. There you read what you have
written, listen to what others have written; and af- The teacher reads aloud from Ben's Trumpet by Rachel Isadora
(1979). The book is extensively illustrated; the text is relatively
firm those things in your own and others' writing brief, describing a boy, Ben, who wants to be a trumpeter, but
that you have found interesting. Finally, imagine has only an imaginary instrument.
again reading the poem and what you have written
in response to your reading, then revising your writ- Ben's Trumpet2
ing, adding any new perceptions and insights In the evening, Ben sits on the fire escape and
gained from your interactions with the group. listens to the music from the Zig Zag Jazz Club.
This sequence of writing-response activities uses He joins in, playing his trumpet.
Sometimes he plays until very late and falls
writing to allow students to create an initial mean- asleep in the hot night air.
ing for the poem they have just read. Using this Every day on the way home from school, Ben stops
composition, the students then have a basis for join- by the Zig Zag Jazz Club.
ing groups of their peers to hear others' responses to He watches the musicians practice.
The pianist, the saxophonist, the trombonist, and the drummer.
1From Personaeby Ezra Pound. Copyright 1926 by Ezra Pound. 2From Ben's Trumpet by Rachel Isadora. Copyright 1979 by
Reprinted by permission of New Directions Publishing Rachel Isadora Maiorano. Reprinted by permission of Green-
Corporation. willow Books and the author.
But most of all Ben thinks the trumpeter is the eat's meow. ing life to these thoughts through the act of writing,
Ben feels the rhythm of the music all the way home. students are able to better understand what it is they
He plays for his mama, grandmother and baby brother. know and feel. Selecting characteristics of their
And for his papa and his friends.
One day, Ben is sitting on the stoop and playing his trumpet. "teacher" for use in their philosophy statements
"I like your horn," someone says. allows them the opportunity to distance themselves
It is the trumpeter from the Zig Zag Club! from the purely subjective and to consider such
Ben smiles and watches him walk to the Club. things as audience, format, and style. The frame-
The next day, after school, Ben stops and listens
to the musicians practicing a red hot piece.
work for future writing and revision has been
He starts blasting away with his trumpet. formed by their intensely personal memories that
Some kids in front of the candy store watch him. were activated in the initial writing phase.
"Hey, what ya doing?" they yell. In both examples, the major elements of reader
Ben stops and turns around. response are present-a text, reading, writing a re-
"What you think ya doing?" they ask again.
''I'm playing my trumpet," Ben answers. sponse to it, sharing that response with others, and
"Man, you're crazy! You got no trumpet!" elaborating or modifying that written response.
They laugh and laugh. These elements can be replicated in any subject area
Ben puts his hands in his pockets and walks slowly and with students of almost any age and level of
home.
He sits on the stoop and watches the blinking
sophistication.
lights of the Zig Zag Club. He sits there a long In this basic sequence of events and in its multi-
time, just watching. tude of variations lie the potential for cognitive ac-
Down the street the band comes out for a break. tivation, the construction of new meanings by the
The trumpeter comes over to Ben. learner, and the opportunity for self-discovery and
"Where's your horn?" he asks.
"I don't have one," Ben says. creative self-expression. The written responses
The trumpeter puts his hand on Ben's shoulder. composed by the reader come out of reading and
"Come on over to the club," he says, rereading both the text and the writing done in re-
"and we'll see what we can do." sponse to the text. The response itself may be elabo-
rated or transformed by the reactions of others. In
Once the book has been read, the students are asked the forward-backward motion of the comprehen-
to write about the person who has been their' 'trum- sion process is the creation of meaning. Readers,
peter." In other words, they are to reflect on the acting as writers, push forward to create meaning
person who has served in a similar capacity as the from text.
trumpeter did for Ben and then describe the quali-
ties of that person. The students are asked to think
of a particular incident that will exemplify to a read-
er the importance of this single individual. Creative Dimensions of Reader
Following this composing session and small Response
group discussions, the students are asked to reread
their compositions and underline all of the words or Writing in response to reading has features few
phrases that describe this person. To form a bridge other modes of expression can match for tapping
between this personal, subjective response and the students' potential for creative expression. Some of
more abstract philosophy of teaching they are being these lie in its form, others in the cognitive opera-
asked to formulate, students can take these power- tions that either contribute to writing or can be ap-
ful descriptors and use them as a basis for formulat- plied to it. The following section examines some of
ing their philosophy statements. Our experience these features.
here is that preservice (and inservice) teachers often
will write about those "teachers" who have been
Student Self-Expression
instrumental in both positive and negative ways in
their teaching careers. Those individuals and their A self-evident aspect of any extended writing,
qualities also seem to be at the heart of how both self-expression is a fundamental condition for
novice and mature teachers want to define them- creativity. Although writing, particularly extended
selves as teachers in the teaching profession. Once writing, currently is a relatively rare component of
their initial composing is complete, students feel most classroom interactions (Applebee, 1981, 1984;
more confident in their reflections, examinations, Birnbaum & Emig, 1983; Bruning, 1984; Good,
and reconsiderations of their own thoughts. In giv- 1983; Graves, 1983), it has great potential for broad-
ening the role of the student from a passive recipient nal work of Minsky (1975), cognitive theorists and
of information to that of an active learner with an researchers (e.g., Anderson, 1977; Anderson &
essential role in learning. When directed effectively Pearson, 1984; Rumelhart, 1977, 1980; Rumelhart
by teachers, writing requires both production and & Ortony, 1977; Schank & Abelson, 1977; Siefert,
originality-key building blocks of creative McKoon, Abelson, & Ratcliff, 1986) have argued
expression. persuasively for the centrality of framelike struc-
tural features in human cognitive functioning. In
general, their argument is that frames provide a
Activation of Cognitive Processes
means for information assimilation and direct the
Writing activates a wide variety of the cognitive cognitive processes of the individual. Readers'
processes of the individual, including perceptual, frames for key passage concepts and for text struc-
memorial, and planning and decision-making di- ture, for instance, dictate the kind and amount of
mensions (see Carey & Flower, Chapter 17 in this attention that is devoted to different categories of
volume; Frase, 1982), and plays an active role in information in a passage. These structures also un-
the construction of knowledge and knowledge derlie the inferences that readers make while read-
structures. We will explore each of these in tum. ing and assist them in filling in gaps in memory by
aiding in the necessary reconstruction of implied
Dimensions of Cognitive Activation. Flower information (Anderson & Pearson, 1984). Johnston
and Hayes and others (Bracewell, Frederiksen, & and Pearson (1982), for instance, have suggested
Frederiksen, 1982; Flower & Hayes, 1980, 1981, that the influence of such frames is so powerful in
1984) have described the cognitive processes that reading that prior knowledge about a topic may be a
are involved in writing as falling into three broad better predictor of reading comprehension than ei-
categories: (a) prewriting or planning, the stage of ther an intelligence test or a reading test scores.
generating a conceptual structure that will allow Just as frames are an integral part of reading and
expression of an idea for a specific reader; (b) other comprehension activities, so also do they play
translating, the transformation of the conceptual a powerful role in writing, as writers plan, translate
structure into the actual writing through such ac- propositions into expression, and edit and revise
tivities as vocabulary selection, choosing the lan- (Meyer, 1982; Scardamalia & Bereiter, 1982).
guage forms through which ideas are expressed, Among the several kinds of frames that have been
and ordering the expression of ideas into a mean- suggested as applying to writing (see Bracewell et
ingful sequence; and (c) revising, the operations al., 1982) are story grammars or narrative frames
that the writer undertakes to evaluate whether the for sequences of events, expository frames for ex-
text is adequate and to make changes to assure its pository text, conversational frames for conversa-
adequacy. Even a cursory evaluation of these three tions among characters, and procedural frames for
categories of activities makes plain that mUltiple explaining operations and activities. For instance,
dimensions of cognitive operations almost certainly where one writer may proceed with little attention
are invoked in any extended writing activity. to reader characteristics and simply concentrate on
To write at all, writers minimally must have ac- writing down information recalled, another may
cess not only to the content of what is to be written, take identical information and produce writing
but also to the means by which the content might be crafted toward a specific audience with a specific
expressed. Writing must draw on memory, both for purpose in mind. The difference between the two
semantic knowledge, in the form of specific recall writers lies in "framing skills" (Bracewell et al.,
of vocabulary, concepts, and principles, and for 1982) that the latter writer brings to bear on the
knowledge that might more properly be called pro- writing task.
cedural (Anderson, 1983, 1987), such as mor- From the standpoint of the creative process, then,
phological, syntactic, and pragmatic aspects of the not only the activation of memory but the nature of
language expression. that activation is critically important. A large
Beyond simple recall, however, is the necessity number of studies of creative processes (e.g., see
to recall and to use information in an organized Glover, 1980; Torrance, 1983) have shown the util-
fashion, ordered within schemata or frames. The ity for creative activity of stimulating the produc-
role of such cognitive structures as a critical dimen- tion of ideas. Indeed, the keystone of most indices
sion in guiding comprehension and recall is well of creativity, aside from originality, is simple vol-
established in the literature. Building on the semi- ume of idea production. What reader response af-
fords is the chance to bring information from mem- textual frames we happen to find ourselves in. (Petrosky, 1982,
ory to consciousness, where it potentially is p. 26)
available for use in the creative act. Then, as the
reader/writer struggles with the challenges of writ- In the planning, composing, and revising sequence
ing and with the task of combining information with required by writing, writers somehow must com-
the structural aspects of what is recalled, the fuller bine the ways in which they have framed their
potential of creative responding arises. To write is world, the structures of the topic and their knowl-
to activate knowledge about the pragmatic uses of edge about it, and the demands of the writing task,
language, understandings of the implied commu- and must express all these within a framework that
nication between writer and reader, and perceptions can be comprehended by a reader. These are inher-
of the purposes for writing. To the extent that writ- ently creative acts, and virtually all but the most
ing necessitates these cognitive functions, the po- formulaic of writing tasks require them. Thus, the
tential will increase for their creative expression. writing task in reader response can be construed as
one of accommodation of knowledge to knowledge
Constructive Processes in the Act of Writing. and structure to structure. Frameworks of knowl-
Beyond its ability to mobilize the organized knowledge must be combined with or subsumed within
edge of the student, reader response stimulates new frames, hierarchically organized knowledge
constructive aspects of the cognitive processing frames must be reconciled with the sequential di-
system-the creation of new meanings. The view mension of writing, and cognitive content and
of human cognition as actively constructive is not structure must be expressed within the conventions
new. The concept of constructive memory is pre- of language.
sent in the seminal work of Bartlett (1932) and The act of reader response also has another con-
clearly articulated in more recent research (e.g., structive dimension. Planning, composing, and re-
Bransford et al., 1972; Paris & Lindauer, 1976, vision in writing of any length take time and gener-
1977). Unless individuals are under explicit in- ate a stream of cognitive activity. Although the
structions to produce it, exact reproduction in recall cognitive processes of writing often have a cyclical
of events is quite rare. Instead, memory almost al- quality, in the physical act of writing word must
ways involves transformation of the input by the follow word, sentence sentence, and paragraph
individual (Loftus & Loftus, 1976). paragraph. As these activities move ahead, students
Constructive models (see Paris & Lindauer, must relate what they are reading and writing to
1977, for a detailed discussion) stress that new in- what is in their memory. Because of the extended
formation is transformed by affective, social, and time and attention required by writing, discovery of
cognitive processes of the individual. What is "forgotten" information is likely to be enhanced.
learned is created by each individual; it is as much a With each new activation of information, yet other
product of that persons' existing cognitive struc- information is likely to be activated (Anderson,
tures and affective tones as it is of the material to be 1983, 1987; Stanovich & West, 1983) and the prob-
learned. ability of other recall enhanced, in much the same
The individual learner in this view is not passive manner as lively conversation stimulates rich recol-
but active. What is learned and subsequently re- lections and associations. Further, as writers strug-
membered is not reproduced so much as it is cre- gle to reconcile the information encountered in
ated. Information undergoes' 'blending, condensa- reading with recollections of their experiences in
tion, omission, invention, and similar constructive order to express a judgment, partly formed ideas
transformations" (Paris & Lindauer, 1977, p. 35). begin to take shape and evolve into new, more com-
Not only content is transformed, however; equally plete forms. In order to respond to their reading,
important are new organizational structures that the writers must make tacit understandings explicit (Pe-
individual generates. trosky, 1982).
The constructivist viewpoint obviously is com- The degree to which such constructive processes
patible with that of theorists who see the function of occur is no doubt a function of many factors, in-
writing, at least in the sense of reader response, as cluding the nature of the writing task the readers
one of creatively constructing new meanings (Tier- must complete, the cognitive abilities and the affec-
ney & Leys, 1986; Tierney & Pearson, 1983). tive resources of readers, and the particular social
When we write, we compose by making meaning from available and cultural contexts in which our readers-as-writ-
information, our personal knowledge, and the cultural and con- ers live. What we would argue, however, is that stu-
dents' resources for responding creatively through meanings they have constructed. Each person in the
writing are vast, far greater than is supposed or- group ordinarily has a copy of what the group mem-
dinarily. To tap these resources requires that they be bers have written; the leader is the person whose
permitted to approach text without artificial rewriting is being discussed. In such groups, each
strictions and to respond, especially initially, in individual can obtain assistance in developing
their own terms (Rosenblatt, 1976). Only if a teach- ideas, determine if the writing is clear and easy to
er has made expression of this kind possible, that is, understand, and, if desired, get advice on such writ-
has created a "setting for spontaneity," can the ing mechanics as spelling, grammar, and usage. In
processes be nurtured in which students can clarify Murphy's view, group members should be encour-
and enlarge their reactions to text. aged to respond to peer writing by complimenting
what is effective, asking clarifying questions about
the content of the paper, and, as appropriate, mak-
Creative Interaction with Others through
ing suggestions for additional clarity. The general
Reader Response
focus is constructive but is intended to force the
Writing provides a "window into the mind." students as readers and writers to examine and ex-
The self-expressive nature of writing that we are pand upon thoughts, feelings, and associations that
proposing is a part of reader response that provides have driven their response. Because the writers'
teachers with an unusual opportunity to enhance affective reactions and personal knowledge will
students' creative processes. Each event of reader have been evoked in the writing, such discussion
response, beginning with the readers' responding in will reveal a variety of perspectives because each
writing to their reading, followed by their sharing group member has seen things differently (Murphy,
that writing with their peers, and finally concluding 1987). Discussion, in tum, creates new perspec-
with their rewriting what they have written, is an tives. As Murray (1986) has stated:
occasion for creative expression and offers teachers
the chance to encourage it. When one says something, lays it out in the open, gathers it up
for the speaker and listener, a presence is effected .... What-
ever the object, once it is said, ... it now enjoys its own pres-
Reader Response as Cognitive Record. Earlier ence in the discussion, has its own meaning, its perspectivalness,
we argued together with Petrosky (1982) that its possibilities, and its unfolding history. (p. 81)
"comprehension arises from an immersion in the
particulars of texts, readers' knowledge, and con-
texts" (p. 22) and that readers' writing reflects that Rewriting
comprehension. In our view, creative reader re- Rewriting is a vital part of creative reader re-
sponse must evoke the subjective reactions of the sponse. As fresh perspectives grow out of the in-
individual in order to activate the widest possible terplay among individuals' reading, writing, re-
range of cognitive processes. In writing, the pro- sponding, and rereading and from new possibilities
cesses are moved from inside the head to an out- contributed by others, these perspectives can be
ward public reflection of the self. captured in revisions. In rewriting early drafts stu-
Writing provides at least a partial record of one's dents are not asked necessarily to alter their original
cognition at the time of its creation, a record about perceptions, but only to have listened actively to
which judgments can be made thoughtfully and ob- others' perceptions and to have made an honest
jectively both by the writer and by others (Bereiter, attempt at describing their own writing and reacting
1984). To the extent that either students or teachers to that of others (Murphy, 1987).
desire, criteria for judgment of creativity can be Just as in the original writing of reader response,
invoked in both one's own and others' reactions to the goal in revision is to express one's own personal
the writing. reaction to the text, a reaction with a bias toward
one's personal experience and the effect that it has
Reader Response Discussion Groups. In reader generated. Ideally, the written response will have in
response, an important mechanism for providing it components of a broadened critical perspective
reactions is that of peer response groups, groups of that includes the reactions of others both to the text
three or four students who discuss what has been and the student's writing. Others will have shared
written by each. As typically used in reader re- their personal viewpoints with the reader/writer and
sponse (Murphy, 1987), their function is to provide thus will have enhanced the potential of what he or
the readers/writers with an opportunity to share the she now might write. In practice, of course, helping
students to expand and broaden their perspectives standard classroom assessment procedures help in-
may not be a simple matter. As Ritchie (1988) has sure this unifonnity. Many teachers also seem to be
pointed out, having once struck upon a powerful quite strongly committed to the idea that, even
and personal connection with what they have read, though instructional methods might be indivi-
some students are absolutely unwilling and unable dualized, learning objectives should be the same for
to move away from their original writing. Particu- all. But common objectives for all necessarily fail
larly with adolescents and late adolescents, this to be able to account for the interplay of the stu-
would not be unexpected for developmental rea- dents' personalities and experiential qualities with
sons. Nonetheless, they will have been presented the content. To the extent that one embraces a con-
with the opportunity to modify their perspective, an structive, creative model of learning, the notion of
opportunity not often available in most current, unifonn outcome becomes less and less attractive.
more traditional approaches to writing. Focusing on student expression, particularly a
fonn of student expression that draws upon the fun-
damental human relationships and emotions known
to the student, also is likely to conflict with strongly
Enabling Conditions for Creativity in held student attitudes toward "how education is. "
Reader Response Initially, most students are likely to resist, perhaps
even strongly, approaches that seem not to offer a
In the foregoing discussion, little has been said clear sense of "what is required" and "whatto do"
about the role of the teacher in reader response (Murphy, 1987). For many students, too, the op-
classrooms. In one sense, this is appropriate. The portunity to express one's honest reactions to what
emphasis of reader response is on the students' has been read will create a great deal of insecurity
creating their own meaning through their transac- and uncertainty. As Rosenblatt (1976) has indi-
tions with text. Student experiences and their use of cated, with the opportunity to express the self
those experiences in responses to text are para- comes responsibility. For many students, it is far
mount. Obviously, however, the teacher's actions more comfortable to give this over to a teacher-
are critical in creating a "setting for spontaneity." authority than to take this responsibility on them-
selves. Teachers also may be well socialized in their
roles as infonnation source, critic, and ultimate au-
Creating a Supportive Classroom Culture
thority. Consequently they, too, like the students,
The teacher's role in guiding reader response is a may unconsciously resist the shift in role necessi-
complex and exacting one. The first task is con- tated by free student expression in reader response.
textual-creating a classroom climate in which per- Teachers who use the methods of reader response
sonal, emotional statements are encouraged and must recognize that students who write with a sense
honored. Developing a classroom culture in which of their own identity, hear the understandings of
personal experiences and self-expression are valued their work by their peers, and revise-in other
and protected requires a considerable shift in goals words, who participate in a "community of readers
from those emphasizing content to those stressing and writers" -are less subject to and dependent
self-expression and individual development. Con- upon the authority of the teacher. No matter what its
tent need not be abandoned, but the teacher must potential for student learning and development,
focus equally on the experiences and identities of change such as this initially may cause great dis-
the students through which the content is experi- comfort to many teachers. The rewards are great,
enced and transfonned. As Rosenblatt (1976) has however, as both teachers and students begin to
stated, turning attention to the students' back- value the experiences of their lives and to trust in
grounds and temperaments may cause the teacher to their own individual expression. Through sharing
fail to do justice to the text, but the students' re- this expression with others, each can develop great-
sponses are, as she states with great insight, "part er insights into human life and values and, ultimate-
of our teaching materials" (p. 51). ly, express them in acts of creativity.
As we indicated at the outset of this chapter,
however, making personal expression the center-
piece of classroom process comes into conflict with Drawing on Students' Episodic Memory
a strong emphasis in our educational system on uni- Portraying abstract verbalization and analysis as
fonnity of outcome. Instructional objectives and the "correct" response to literature quickly will
destroy responsiveness in virtually all students. Ad- ment, experience, and background. She contended
olescents often are untouched by the appeal of liter- that, at the outset, a literary work will have meaning
ature because for a reader only in personal terms. These personal
responses may be immature, incomplete, or even
for them words do not represent keen sensuous, emotional, and mistaken, but they nonetheless are the initial mean-
intellectual perceptions. This indicates that throughout the entire
course of their education, the element of personal insight and
ing of the work to the reader.
experience has been neglected for verbal abstractions. (Rosen-
blatt, 1976, p. 50) Only on the basis of such direct emotional elements, immature
though they may sometimes be, can [the reader1 be helped to
build any sounder understanding of the work. The nature of the
The strong implication is that, for there to be a student's rudimentary response is, perforce, part of our teaching
potential for creativity, readers who respond must materials. (Rosenblatt, 1976, p. 51)
be allowed to take as full advantage as possible of
their personal experiences and the emotional re- The Role of Emotion in Cognition. Current
sponses they generate. These resources lie gener- models of cognition lend support to Rosenblatt's
ally within the realm of episodic memory. assertion. Although by and large the element of
For some time now, memory theorists have pro- emotion had figured only sporadically in earlier the-
posed that general information, on one hand, and ories and models of cognition (see Read, 1984),
information about specific events, on the other, emotion now has begun to appear as an element in
may be stored in separate memory systems. Tulving current models of cognitive processes.
(1972, 1983), for instance, has proposed a distinc- One such model is that of Bower and his associ-
tion between semantic memory and episodic memo- ates (Bower, 1981; Bower & Cohen, 1982; Gilligan
ry that corresponds to these categories. In the se- & Bower, 1984). Their perspective, consonant with
mantic representational system is abstract, decon- those of most current models of memory (e.g., An-
textualized information that has been developed derson, 1983, 1987; Collins & Loftus, 1975) is that
across multiple experiences. Episodic memory, in memory can be represented as a rich associative
contrast, is autobiographical and unique, consisting network. This network is composed of concepts,
of representations of particular events. Recent for- schemata, and events, which form the nodes of the
mulations of the semantic-episodic distinction network. These nodes are linked by multiple path-
(e.g., Whittlesea, 1987) have indicated that it may ways to one another. Central to the model is the
derive not so much from different stores for the process of spreading activation, in which activation
semantic and episodic types of information, but of one or more nodes leads automatically to activa-
from the encoding and retrieval activities that are tion of associated nodes. In such network models,
used by the individual. In other words, demands of spreading activation is seen as the primary mecha-
some tasks, contexts, and purposes may lead to the nism by which memory is enhanced (e.g., through
encoding of information with more of the decontex- creating additional associative links through elab-
tualized features of semantic memory, while other oration) and whereby specific information is re-
tasks, contexts, and purposes may lead to encoding trieved from storage and brought into conscious-
with strong episodic features (Hintzman, 1986; ness.
Whittlesea, 1987). In Bower's model, however, emotions also are
central cognitive units, linked like other units to
The Necessary Role of Emotion. Our judgment, related events and concepts. In the words of
for which the rationale is presented below, is that Gilligan and Bower (1984), the network contains
reader response through writing is most likely to "emotion nodes" (p. 571) that, once stimulated,
create meaning and to enhance creativity when its activate autonomic response patterns, subjective
starting point is in the stimulation of the episodic experiences, and interpretive rules. These units re-
memory of the reader/writer. Further, it is our verberate back to the activated emotion units and
strong belief that emotional and not analytical or cause the activation to persist for some time.
critical responses on the part of the reader are most Two features of the emotion nodes are important
likely to stimulate critical expression. In Literature for our present discussion. First, activation pro-
as Exploration, Louise Rosenblatt argued per- duced by emotion is hypothesized to be unusually
suasively that the reader's response to literature in- persistent in the memory network. Whereas ordi-
evitably will be in terms of the student's tempera- nary concepts disappear nearly instantaneously
from consciousness when, say, the topic of conver- engender full learner involvement and creativity are
sation shifts, activation from emotion continues and likely to be in episodic, not semantic memory. In
only dies out gradually through a natural damping the former are the directly experienced events of
process. That is, the complexity of neurophysiolo- student's lives. These memories, with the full
gical, motor, and subjective-experiential reactions, richness of emotion that they can evoke, will pro-
once activated, tend to persist much longer than vide the best starting point for students' creative
those stemming from "purely cognitive" units. transactions with text. In contrast, for the teacher to
Second, Bower contends that emotions are call for an "understanding" of text based on the
stronger sources of activation than conceptual or abstracted knowledge in semantic memory seems
schematic nodes. As Gilligan and Bower (1984) much less likely to evoke from the learner either the
have stated: vividness of recollection or the motivation to drive a
vital interaction between the learner and the text.
One might think of nodes in the network as small voltage sources
(or signal boosters), and that, once aroused, emotion nodes sim- To value reading as a highly subjective experi-
ply send more "voltage" than concept nodes. This would cause ence with a strong emotional component, teachers
an emotion to gain greater control than a thought over the direc- must acknowledge, we believe, the "mine-ness"
tion and content of subsequent thoughts. (p. 5) (Heidegger, 1927/1962) of every students' exis-
tence; that each experience, including the hoped-for
In this conception, then, stimulation of affective
transaction with text, is comprehended based on
dimensions of students' episodic memory would be
self-reference. Often, however, we as teachers be-
expected to produce activation that is both more
have as if materials we ask students to read have
durable and more extensive than activation that did
about them a completely objective character-a
not tap this affective component. Thus, when stu-
"truth" to be uncovered by the student or pointed
dents are asked to link what they have read-a
out by us to the student. The origin of meaning in
poem, newspaper article, a letter-to their own ex-
reading, however, arises first from the reader and
perience, particularly elements of their experience
not from any particular text being read by the reader
evocative of emotion, we would argue that engage-
(Applebee, 1985; Holland, 1968, 1975, 1985;
ment and motivation to continue will be high be-
Tierney & Pearson, 1983). Personal experience
cause of the nature of the activation produced.
frames all reading; understanding is created out of
Moreover, not only will it be likely that a relatively
who readers are and the responses the text evokes
larger number of related concepts will be activated,
(Beach, 1983; Bleich, 1975a,b).
but that their activation will persist longer. To the
Another factor, perhaps more speculative, also is
extent that multiple concepts and schemata con-
likely to make the evocation of episodic memory as
tinue to be activated and the reader struggles to
a starting point for the reader's transactions with
express the meaning they carry, the fundamental
text more productive than attempting to stimulate
elements are put into place for creative responding.
semantic memory. For younger readers and writers,
Although emotion is proposed as an appropriate
in particular, we would argue that the contents of
starting point for student response, reason is not
episodic memory are much more likely to be large,
excluded. Reason, to return to Rosenblatt's (1976)
stably organized, and accessible compared to the
view, is equally vital to informed response but
contents of semantic memory. Whereas the former
should "arise in a matrix of feeling" (p. 227). In
reflects the whole body of the student's life experi-
her view, rationality is not a force arrayed against
ences, the latter is likely to contain a relatively
emotion, but instead a force applied to emotion in
smaller amount of abstracted qualities of experi-
an attempt to understand it. Murray (1986) echoes
ence organized around selected concepts and top-
this same thought:
ics. The poetry example that was provided earlier
Provocation, discontinuity, delayed and suspended judgment are illustrates this point; to generate images, to associ-
ingenious and creative procedures, methods, employed in the ate feelings, to recall life events when reading this
process of imaginative thought, but they gain their respect and poem is relatively easy to do, no matter what an
prove efficacious only within the confines of the structure pro-
vided by sound, rational, logical thinking. Neither is successful
individual's training. In contrast, however, to
without the other; both are fruitful when together. (p. 35) "analyze" the poem by applying techniques oflit-
erary criticism, or by relating it to other literary
Our argument, then, is that those memories most works, or somehow to "explain" its meaning is
likely to stimulate the quality of response needed to likely to be much more difficult for most readers.
Teaching Methods in Reader Response Writing for New Audiences

One of the most straightforward ways that writ-
An underlying assumption of this chapter is that ing can be moved into new forms, yet still derive
all reader response is inherently creative from the from student experience, is to focus that writing on
standpoint of the individual. For a student to re- a new audience. For instance, students in a science
spond to reading with extended writing, multiple class might be asked to use information from a read-
dimensions of cognition and language must be acti- ing in ecology in a letter to a company president, a
vated and combined in a creative process. At the state senator, or the President (or, for that matter, in
same time, however, we would acknowledge that letters to pygmys, astronauts, surfers, or real-estate
not all student expression in reader response will be developers). At least implicitly, most writing in the
judged creative or even divergent from that of oth- schools is done for the teacher as reader, with little
ers. A "setting for spontaneity" and self-ex- consequent attention to the pragmatic use of lan-
pression through writing are necessary for creative guage as purposeful and persuasive activity. Sim-
student expression, but not sufficient. When apply being forced to express a body of information to
plied to reader response activities, the following a new audience will require transformations of that
dimensions may add to the possibility that the writ- information and greater personal involvement of
ten expression may take new and creative forms. the reader/writer. Further variation can be intro-
duced by asking readers to convey their reactions to
more than one audience, varying in age, political
Arranging Priming Activities outlook, or some other significant dimension.
The central act of reader response is readers'
transactions with text, transformed through writ- Writing in New Language Forms
ing. The efficacy of these transactions is based,
however, on the assumptions that readers (1) have When students do write, they ordinarily are
sufficient knowledge available to have meaningful taught to value writing as a product. Their writing
transactions with the text, (2) have the motivation to must be neat, mechanically correct, and well orga-
engage in meaningful transactions with the text, nized. As a consequence, the writing they do is
and (3) have some degree of ease with the act of often formulaic. For example, one commonly
self-expression. Although some students will have taught structure for writing essays that is well
no difficulty on any of these dimensions, manyoth- learned by many students is to present an introduc-
ers initially may see no connection between their tory paragraph containing a thesis sentence, then to
own experiences and the text and must be helped to give reasons to support that thesis, and then to sum-
see that connection. Still others will have strong marize the argument in a final section. Even though
predispositions to react "objectively," and not this structure may be useful for some purposes, it is
sUbjectively. only one of a myriad of possibilities. Instead, the
We nonetheless see a nut1J.ber of approaches that teacher might ask students to create a sequence of
teachers can use to prepare their students for their their recollections, or imaginary conversations,
reading. For instance, prior to reading, the teacher speeches by historical figures, sets of instructions,
may wish to initiate a discussion on the general product warnings, codes, telegrams, signs, and so
topic of the reading or writing to establish a pre- on. By letting the students explore new forms of
paratory set or, through using analogy or metaphor, writing, but still requiring them to use their direct
link the topic to one familiar to the students. To experience in responding to the text, new combina-
evoke subjective responses, the teacher may ~ish to tions of idea, language, and structure will emerge.
ask students to recollect experiences they have had,
images they can visualize, and sensations they have
Taking New Perspectives as a Writer
felt. The teacher may ask students to talk to peers,
to observe actions of others, to write for informa- Most writing is from "I" or "we" to "you."
tion, or to interview individuals. The purpose of all Obviously, however, the perspective taken by the
such activities, of course, is to activate the direct writer can be altered. Students can be asked to write
sensory experience that will form the basis for the as if they were a second person writing to a third, or
reader response. even as if they were another person, for example, a
parent, writing to themselves. They may wish to what they read. What they comprehend grows not
read and write as visitors from outer space, as a only out of the content and structure of what is read,
hyperperceptive newborn baby, as the world's but out of the readers themselves and their tempera-
oldest person, or as "Star Trek's" Dr. Spock. Nor ments, intentions, and cultural perspectives (Ro-
need their perspective be limited to that of hu- nald, 1986).
mans-readers can write with their minds' eyes in The general acceptance of a constructive view of
the roles of a family pet, a radish growing in the learning, however, is not matched by specific
garden, a zoo animal being traded to a new zoo, an knowledge about the impact of writing on reading
ancient redwood tree, or an amoeba. and on the role each plays in the creation of new
Each change of perspective, of course, will force knowledge. There seems to be little doubt about the
students to grapple with a new array of cognitive general efficacy for learning and recall of active
structures and is likely to help stimulate additional processing of information (e.g., Glover, Benton, &
points of contact between what they already know Bruning, 1983), and the following section is pre-
and their reading. As they write from a perspective mised on the reasonable assumptions that writing is
that is new to them, unfamiliar associations and a way of activating cognitive processes and that
unusual cognitive structures are likely to be gener- such activation is likely to facilitate learning. The
ated that will make learning both meaningful and present chapter, however, has proposed that a spe-
memorable. cific pedagogical technique-reader response-is
likely to have positive effects on both learning and
the probability of creative responding. Whether this
Writing with Humor and on Unusual Topics is so is, of course, an important issue for research
Creativity, we know, often has a strong compo- and within that issue are a number of questions of
nent of humor. For most teachers and students, both theoretical and applied interest.
however, learning is seen as serious business and,
in school, almost all reading and writing are done 1. Are there distinct differences, both cognitive
with serious purpose on sober-minded topics. and affective, in the quality of writing result-
Important goals for learning, however, can be ing from prereading and prewriting activities
reached in many different ways. When process be- presumed to stimulate semantic memory as
gins to be valued over product, consistent with the opposed to those that stimulate episodic
constructivist view we have advocated, the specific memory?
content and methods of learning are less critical
than the processes they engender. For instance, the As we have argued throughout this chapter, the
ability to think freely and well about economic or "setting for spontaneity" created by the teacher for
historical or social issues and about the values in- reader response is likely to be a critical variable.
herent in decisions about them may best be devel- Not only must the teacher create an atmosphere of
oped by whimsy, by humor, through use of meta- openness and acceptance, but students must learn
phor, and by satire. For the typical student, the how to interact with their peers in constructive and
chance to write a satirical response to an issue may pleasurable ways. Furthermore, the activities that
prove to be far more motivating, intellectually stim- precede reading and writing will determine what
ulating, and ultimately creative than a more "se- students will do. Instructions to recall and write
rious" response to that same topic. from personal experience in response to a poem, as
we have contended, are likely to engender vastly
different conditions than instructions that direct
Issues for Research in Reader Response readers to their semantic knowledge, for instance,
instructions to "read and analyze the poem."
In both literary theory (e.g., Moffett, 1984; Pe- Whether, however, the creative consequences of
trosky, 1982; Rosenblatt, 1976, 1978, 1985) and these setting conditions in fact differ is a matter that
psychological theory (Anderson, 1983; Neisser, should be addressed empirically.
1982; Pearson, 1985), a constructive view of the
reader is well accepted. The research consistently 2. How does the nature of what is read affect the
has shown that readers' worlds are projected into creative qualities of what is written?
Some materials are inherently more evocative impact do the reactions of other readers/writers
than others. Those written from a "personal" have on what the student does in rewriting and how
standpoint to which students can readily relate, for does this relate to level of cognitive development?
instance, may evoke better and more complete writ-
ing and discussion than those with more gener- 6. What conditions for rewriting will move stu-
alized, abstract, textlike structures. Other dimen- dents in the direction of divergence and what
sions of the reading materials that may also greatly consequences, both cognitive and behavioral,
affect the potential for creative responding are the will the divergence have on comprehension of
extent of imagery used by the author, the strength of the original text?
point of view taken, the readers interest in the mate-
rials, and the familiarity of the topic to the reader. For many, perhaps most students, rewriting con-
notes doing minor editing, and sometimes little
3. What are the cognitive consequences for more than recopying a messy first draft. What tech-
creativity of being forced to take different niques can assist students in making decisions about
perspectives in writing in response to given the need to rewrite and help them transform what
text materials? they initially have written into something more
creative?
At present, we know little about the effects on
creative expression of having to write from an "un- 7. How does one bridge the gaps between stu-
ordinary" perspective. Does having to take differ- dents' experiences, the text, and the task of
ent perspectives in writing lead to new insights, writing?
new cognitive structures, new forms of expression?
Does the unusualness of the perspective make a In setting conditions for reader response, the
difference? Does the degree with which the read- teacher frames the task of reading or writing within
er/writer can identify with the perspective have an a particular perspective. What is the impact of that
impact on the creativity of expression? perspective, particularly as metaphors are used to
constrain or widen possible reactions to the literary
4. What are the cognitive consequences for work?
creativity of writing for different audiences?
Development of pragmatic awareness in writers Summary

and speakers is an important goal for many teach-
ers. Our interest here, however, is in the creative Writing is a powerful means of expressing the
consequences of writing for varied audiences. Are self. Through writing, students can begin to under-
students, for example, more willing to write cre- stand their emerging identities and to construct new
atively for an audience of their peers than for, say, meanings. When the proper conditions are set for
the teacher or for a hypothetical "reader?" Em- students to respond with writing to what they have
pirical research in which the conditions for writing read, the elements of creativity become genuine
are systematically varied along this dimension possibilities.
would assist in answering this question. The term reader response has come to refer to the
transactions of readers with literature, to a point of
5. How does what the student initially writes view in which reading is seen as a highly subjective
affect subsequent rereadings and rewritings? process whereby readers understand what they read
through the filters of their own experiences and
Reader response creates an unusually rich in- personalities, and to a set of related classroom ac-
teraction among reading, writing, and subsequent tivities in which readers write in response to what
rereading and rewriting, and raises interesting ques- they have read, discuss their writing with their
tions about how points of view might undergo trans- peers, and rewrite. The view advanced in this chap-
formation. Does the substance and structure gener- ter is that reader response provides unique pos-
ated initially and reported in the initial writing sibilities for creativity. Among these are the oppor-
persist? That is, are the initial cognitive structures tunity for students to express their own perspective,
by which the student transacts with the text immuta- the extent of cognitive activation that occurs in the
ble or can they be transformed by later inputs? What kind of writing required in reader response, the in-
teractions with peers about what has been written, literature. Journal of Research and Development in Educa-
and the processes generated by rewriting. tion, 16(1),47-54.
Beach, R., & Liebman-Kleine, J. (1986). The writinglreading
The necessary conditions for creative reader re-
relationship: Becoming one's own best reader. In B. Petersen
sponse, we argue, include a supportive classroom (Ed.), Convergences: Transactions in reading and writing
culture in which self-expression is valued. It also (pp. 64-81). Urbana, IL: National Council of Teachers of
requires the activation of affective responses from English.
episodic memory, and specific teaching techniques Bereiter, C. (1984). Learning about reading from writing.
Written Communication, 1(2), 163-188.
that include appropriate priming activities and Birnbaum, J., & Emig, J. (1983). Creating minds, created texts:
planned variations in the assigned writing. Writing and reading. In R. P. Parker & F. A. Davis (Eds.),
The major research issues raised by this chapter Developing literacy: Young children's use of language (pp.
are those relating to the role of emotion in creative 87-104). Newark, DL: International Reading Association.
Birnbaum, J. C. (1986). Reflective thought: The connection
reader response and the functional roles of episodic between reading and writing. In B. T. Peterson (Ed.),
and semantic memory. Other questions include the Convergences: Transactions in reading and writing (pp. 30-
nature of the interaction among writing, reading 45). Urbana, IL: National Council of Teachers of English.
comprehension, and rewriting, and the effects of Bleich, D. (l975a). Readings and feelings: An introduction to
imposing variations in the writing task, such as subjective criticism. Urbana, IL: National Council of Teach-
ers of English.
writing for different audiences and taking different Bleich, D. (I 975b). The subjective character of critical in-
perspectives in writing. terpretation. College English, 36(7), 739-755.
Bleich, D. (1978). Subjective criticism. Baltimore: Johns
ACKNOWLEDGMENTS. The authors express their Hopkins University Press.
sincere gratitude to Robert Brooke and Joy Ritchie Bleich, D. (1985). The identity of pedagogy and research in the
study of response of literature. In C. Cooper (Ed.),
of the Department of English at the University of Researching response to literature and the teaching of liter-
Nebraska-Lincoln for their insightful comments on ature: Points of departure (pp. 253-272). Norwood, NJ:
a preliminary version of this chapter. Ablex.
Bower, G. H. (1981). Mood and memory. American Psychol-
ogist,36, 129-148.
Bower, G. H., & Cohen, P. R. (1982). Emotional influences in
References memory and thinking: Data and theory. In M. S. Clark & S. T.
Fiske (Eds.), Affect and cognition (pp. 291-331). Hillsdale,
Adams, M., & Bruce, B. (1982). Background knowledge and NJ: Erlbaum.
reading comprehension. InJ. A. Langer&M. T. Smith-Burke Bracewell, R. J., Frederiksen, C. H., & Frederiksen, J. D.
(Eds.), Reader meets author/bridging the Gap. A psycho- (1982). Cognitive processes in cOmposing and comprehend-
linguistic and sociolinguistic perspective (pp. 2-25). New- ing discourse. Educational Psychologist, 17(3), 146-164.
ark, DE: International Reading Association. Bransford, J. D., Barclay, J. R., & Franks, J. 1. (1972). Sen-
Anderson, J. R. (1983). The architecture of cognition. tence memory: A constructive versus interpretive approach.
Cambridge: Harvard University Press. Cognitive Psychology, 2, 193-209.
Anderson, J. R. (1987). Skill acquisition: Compilation of weak- Bruning, R. H. (1984). Key elements of effective teaching in the
method problem solutions. Psychological Review, 94, 192- direct teaching model. In R. L. Egbert & M. Kluender (Eds.),
210. Using research to improve teacher education: The Nebraska
Anderson, R. C. (1977). The notion of schemata and the educa- Consortium (Teacher Education Monograph No.1, pp. 75-
tional enterprise. In R. C. Anderson, R. J. Spiro, & W. E. 88). Washington, DC: ERIC Clearinghouse on Teacher
Montague (Eds.), Schooling and the acquisition of knowledge Education.
(pp. 415-431). Hillsdale, NJ: Erlbaum. Collins, A. M., & Loftus, E. F. (1975). A spreading-activation
Anderson, R. c., & Pearson, P. D. (1984). A schema-theoretic theory of semantic processing. Psychological Review, 82,
view of basic processes in reading comprehension. In P. D. 407-428.
Pearson (Ed.), Handbook of reading research (pp. 225-293). Comprone, J. J. (1986). Integrating the acts of reading and writ-
New York: Longman. ing about literature: A sequence of assignments based on
Applebee, A. N. (1981). Writing in the secondary school. James Joyce's "Counterparts." In B. Petersen (Ed.),
Urbana, IL: National Council of Teachers of English. Convergences: Transactions in reading and writing (pp. 215-
Applebee, A. N. (1984). Contexts for learning to write. 230). Urbana, IL: National Council of Teachers of English.
Norwood, NJ: Ablex. Cooper, C. R. (1985). Introduction. In C. R. Cooper (Ed.),
Applebee, A. N. (1985). Studies in the spectator role: An ap- Researching response to literature and the teaching of liter-
proach to response to literature. In C. R. Cooper (Ed.), ature: Points of departure(pp. ix-xix). Norwood, NJ: Ablex.
Researching response to literature and the teaching of liter- Dewey, J. (1938). Experience and education. New York: Collier
ature: Points of departure (pp. 87-102). Norwood, NJ: Books.
Ablex. Flower, L., & Hayes, J: R. (1980). A cognitive process of theory
Bartlett, F. C. (1932). Remembering. London: Cambridge Uni- of writing. College Composition and Communication, 32,
versity Press. 365-387.
Beach, R. (1983). Attitudes, social conventions and response to Flower, L., & Hayes, J. R. (1981). Plans that guide the compos-
ing process. In C. H. Frederickson & J. F. Dominic (Eds.), Minsky, M. (1975). A framework for representing knowledge.
Writing: The nature, development and teaching of written In P. H. Winston (Ed.), The psychology of computer vision
communication (2nd ed., pp. 39-58). Hillsdale, NJ: Erlbaum. (pp. 211-277). New York: McGraw-Hili.
Flower, L., & Hayes, J. R. (1984). Images, plans and prose: The Moffett, J. (1984). Reading and writing as meditation. In J.
representation of meaning in writing. Written Communica- Jensen (Ed.), Composing and comprehending. Urbana, IL:
tion, 1(1), 120-160. ERIC Clearinghouse on Reading and Communication Skills.
Frase, L. T. (1982). Introduction to special issue on writing. Murphy, C. c. (1987). Enhancing eleventh graders critical
Educational Psychologist, 17, 129-130. comprehension of literary texts. Unpublished doctoral disser-
Gilligan, S. G., & Bower, G. H. (1984). Cognitive conse- tation, University of Nebraska, Lincoln, NE.
quences of emotional arousal. In C. E. Izard, J. Kagan, & R. Murray, E. L. (1986). Imaginative thinking and human exis-
B. Zajonc (Eds.), Emotions, cognition, and behavior (pp. tence. Pittsburgh, PA: Duquesne University Press.
547-588). Cambridge, England: Cambridge University Neisser, V. (1967). Cognitive psychology. New York:
Press. Appleton-Century-Crofts.
Glover, J. A. (1980). Becoming a more creative person. Neisser, V. (1982). Memory observed: Remembering in natural
Englewood Cliffs, NJ: Prentice-Hall. contexts. San Francisco: W. H. Freeman and Company.
Glover, J. A., Benton, S., & Bruning, R. H. (1983). Levels of Paris, S. G., & Lindauer, B. K. (1976). The role of inference in
processing: Effects of number of decision on prose recall. children's comprehension and memory for sentences.
Journal of Educational Psychology, 75, 382-390. Cognitive Psychology, 8, 217-227.
Good, T. L. (1983). Classroom research: A decade of progress. Paris, S. G., & Lindauer, B. K. (1977). Constructive aspects of
Educational Psychologist, 18, 127-144. children's comprehension and memory. In R. V. Kail, Jr., &
Graves, D. (1983). Writing: Teachers and children at work. J. W. Hagen (Eds.), Perspectives on the development of mem-
Exeter, NH: Heinemann. ory and cognition (pp. 35-58). Hillsdale, NJ: Erlbaum.
Harste, J. C., Woodward, V. A., & Burke, C. L. (1984). Exam- Pearson, P. D. (1985). Changing the face of reading comprehen-
ining our assumptions: A transactional view of literacy and sion instruction. Reading Teacher, 38(8), 724-738.
learning. Research in the Teaching of English, 18, 18. Petrosky, A. (1982). From story to essay: Reading and writing.
Heidegger, M. (1962). Being and time (J. Macquarrie & E. College Composition and Communication, 33, 19-37.
Robinson, trans.). New York: Harper & Row. (Original work Petrosky, A. (1985). Response: A way of knowing. In C. R.
published 1927) Cooper (Ed.), Researching response to literature and the
Hintzman, D. L. (1986). "Schema activation" in a multiple- teaching of literature: Points of departure (pp. 70-83). Nor-
trace memory model. Psychological Review, 93, 411-427. wood, NJ: Ablex.
Holland, N. H. (1968). The dynamics of literary response. New Pound, E. (1983). The river merchant's wife: A letter. In A. W.
York: Oxford University Press. Allison, H. Barrows, C. R. Blake, A. J. Carr, A. M. Eastman,
Holland, N. H. (1975). Five readers reading. New Haven: Yale & H. M. English (Eds.), The Norton Anthology of Poetry (3rd
University Press. ed., pp. 963-964). New York: W. W. Norton.
Holland, N. H. (1985). Reading readers reading. In C. R. Purves, A. C. (I 976). Foreword to L. M. Rosenblatt, Literature
Cooper (Ed.), Researching response to literature and the as Exploration. New York: Modern Language Association of
teaching of literature: Points of departure (pp. 3-21). Nor- America.
wood, NJ: Ablex. Read, P. B. (1984). Forward to C. E. Izard, J. Kagan, & R. B.
Horton, S. R. (1983). Thinking through writing. Baltimore, Zajonc (Eds.), Emotions, cognition, and behavior. Cam-
MD: Johns Hopkins University Press. bridge: Cambridge University Press.
Isadora, R. (1979). Ben's trumpet. New York, NY: Green- Ritchie, J. (1988, March). Reading, writing, and identity-Con-
willow Books. nections and misconnections. Paper presented at the 1988
Iser, W. (1974). The implied reader. Baltimore, MD: Johns Conference on College Composition and Communication, St.
Hopkins University Press. Louis, MO.
Iser, W. (1978). The act of reading. A theory of aesthetic re- Ronald, K. (1986). The self and the other in the process of
sponse. Baltimore, MD: Johns Hopkins University Press. composing: Implications for integrating the acts of reading
Jenkins, J. J. (1974). Remember that old theory of memory? and writing. In B. Peterson (Ed.), Convergences: Transac-
Well, forget it! American Psychologist, 29, 785-795. tions in reading and writing (pp. 231-245). Urbana, IL: Na-
Johnston, P., & Pearson, P. D. (1982). Prior knowledge, con- tional Council of Teachers of English.
nectivity, and the assessment of reading comprehension Rosenblatt, L. M. (1976). Literature as exploration (3rd ed.).
(Tech. Rep. No. 245). Urbana: University of Illinois, Center New York: Noble & Noble.
for the Study of Reading. Rosenblatt, L. M. (1978). The reader, the text, the poem: The
Langer, J. A., & Smith-Burke, M. T. (1982). Reader meets transactional theory of the literary work. Carbondale, IL:
author/bridging the Gap: A psycholinguistic and so- Southern Illinois University Press.
ciolinguistic Perspective. Newark, DE: International Reading Rosenblatt, L. M. (1985). The transactional theory of the literary
Association. work: Implications for research. In C. R. Cooper (Ed.),
Loftus, G. R., & Loftus, E. F. (1976). Human memory: The Researching response to literature and the teaching of liter-
processing of information. New York: Wiley. ature: Points of departure (pp. 33-53). Norwood, NJ: Ablex.
Mayher, J. S., Lester, N., & Prahl, G. M. (1983). Learning to Rumelhart, D. E. (1977). Understanding and summarizing brief
write/writing to learn. Upper Montclair, NJ: Boynton/Cook. stories. In E. LaBerge & S. J. Samuels (Eds.), Basic pro-
Meyer, B. F. (1982). Reading research and the composition cesses in reading: Perception and comprehension (pp. 265-
teacher: The importance of plans. College Composition and 303). Hillsdale, NJ: Erlbaum.
Communication, 33, 27-49. Rumelhart, D. E. (1980). Schemata: The building blocks of
cogmtlOn. In R. Spiro, B. Bruce, & W. Brewer (Eds.), context. Journal of Experimental Psychology: General. 112.
Theoretical issues in reading and comprehension (pp. 33- 1-36.
58). Hillsdale, NJ: Erlbaum. Tierney, R., & Leys, M. (1986). What is the value of connecting
Rumelhart, D. E., & Ortony, A. (1977). The representation of reading and writing? In B. T. Peterson (Ed.), Convergences:
knowledge and memory. In R. C. Anderson, R. J. Spiro, & Transactions in reading and writing (pp. 15-29). Urbana. IL:
W. E. Montague (Eds.), Schooling and the acquisition of National Council of Teachers of English.
knowledge (pp. 99-135). Hillsdale, NJ: Erlbaum. Tierney, R., & Pearson, P. D. (1983). Toward.a composing
Salvatori, M. (1983). Reading and writing a text: Correlations model of reading. Language Arts. 60(5), 568-579.
between reading and writing patterns. College English, 45(7), Tompkins, J. P. (1980). An introduction to reader-response crit-
657-667. icism. In J. Tompkins (Ed.), Reader-response criticism:
Scardamalia, M., & Bereiter, C. (1982). Assimilative processes From formalism to post-structuralism (pp. ix-xxvi). Bal-
in composition planning. Educational Psychologist, 17(3), timore, MD: Johns Hopkins University Press.
165-171. Torrance, E. P. (1983). Creativity in the classroom.
Schank, R. c., & Abelson, R. P. (1977). Scripts. plans. goals Washington, DC: National Education Association.
and understanding. Hillsdale, NJ: Erlbaum. Tulving, E. (1972). Episodic and semantic memory. In E. Tul-
Shanahan, J., & Lomax, R. G. (1986). An analysis and com- ving & W. Donaldson (Eds.), Organization of memory (pp.
parison of theoretical models of the reading-writing rela- 382-402). New York: Academic Press.
tionship. Journal of Educational Psychology. 78(2), 116- Tulving, E. (1983). Elements of episodic memory. Oxford, En-
123. gland: Oxford University Press.
Siefert, C. M., McKoon, G., Abelson, R. P., & Ratcliff, R. Whittlesea, B. W. A. (1987). Preservation of specific experi-
(1986). Memory connections between thematically similar ences in the representation of general knowledge. Journal of
episodes. Journal of Experimental Psychology: Learning. Experimental Psychology: Learning. Memory. and Cogni-
Memory. and Cognition, 12. 220-231. tion. 13. 3-17.
Stanovich, K., & West, R. (1983). On priming by a sentence
CHAPTER 20
Learning via Model Construction

and Criticism
PROTOCOL EVIDENCE ON SOURCES OF
CREATIVITY IN SCIENCE
John Clement
Introduction gies and other divergent processes are also exam-

ined. It is argued that these insights can involve
There is growing recognition that mental models fairly sudden reorganizations in the structure of a
play a fundamental role in the comprehension of mental model but do not necessarily involve ex-
science concepts. The process of learning via model traordinary or unconscious thought processes. This
construction appears to be central to theory forma- introduction and a summary of findings at the end
tion in science and central for science instruction constitute an overview of the chapter.
but is still very poorly understood. This chapter
uses evidence from case studies, in which a scientist
is asked to think out loud, to argue that nonformal Questions about the Nature of Scientific Theory
reasoning processes that are neither deductive nor Formation
inductive can play an important role in scientific
model construction. The construction process is Galileo's theory of motion, Faraday's concept of
complex and involves repeated passes through a the magnetic field, Darwin's theory of natural se-
cycle of hypothesis generation, evaluation, and lection, and Einstein's theory of relativity are com-
modification. monly cited examples of creative achievements in
"Aha" episodes that show that a scientist can science. Each is a major event in the history of
generate creative insights via spontaneous analo- scientific ideas, and, in each case, something very
new emerged that affected the entire scientific com-
munity and subsequently affected civilization as a
John Clement • Scientific Reasoning Research Institute, whole. Analyzing how such achievements take
University of Massachusetts, Amherst, MA 01003. The research place is a worthwhile goal, but achieving this goal
reported in this study was supported by the National Science has unfortunately proven to be surprisingly diffi-
Foundation under Grants MDR-8470579 and MDR-8751398.
Any opinions, findings, and conclusions or recommendations
cult. The universally recognized importance of ad-
expressed in this chapter are those of the author and do not vances in science has not made the problem of de-
necessarily reflect the views of the National Science Foundation. scribing the processes by which they were created
341
any .easier. In Darwin's case, for example, it is Background Questions from

possible to argue that the theory of natural selection Philosophy of Science
was built up gradually through a large number of
detailed empirical observations. But, on the other The Source and Pace of Theory Change
hand, it also is possible to argue that the theory was
the result of a mental breakthrough well after the Eurekaism versus Accretionism. It is useful to
B:agle's voyage in the form of an insight that con- separate out two major issues that are involved in
stituted a sudden reorganization of Darwin's ideas. the controversy over hypothesis formation, the
Thus, even with respect to specific historical exam- pace of scientific theory change and the source of
ples, disagreement emerges as to the basic sources ~ew theories (represented, respectively, in ques-
and pace of theory change in science. tIons 1 and 2, above). With respect to the pace of
At issue here is an important question concerning theory change, one can contrast eurekaist and ac-
the nature of science. Cast in its most global and cretionist positions. A eurekaist claims that a theory
~xtre~e form, the question is: Does science change
can be changed at a very fast pace by an insight that
In an Incr~~ental manner, with a series of many reorganizes its structure. In its strongest form, eu-
small empmcal observations inching it forward, or rekaism is associated with sudden flashes of in-
do occasional large breakthroughs occur in the spiration, possibly following a period of incubation
mind of the scientist in the absence of new data or nonconscious mental activity. Thus, some ideas
each causing a great leap forward in the field? On~ may form in and arrive suddenly from the uncon-
purpose of this chapter is to determine whether the scious mind.
methodology of protocol analysis has the potential An accretionist or incremental view of the pace
to illuminate some aspects of this question by using of scientific theory change holds that a scientist
data from transcripts of scientists solving problems gains knowledge in small pieces and puts them to-
aloud. I will concentrate most on an example of a gether deliberately at a slow and even pace. This
break~hrough episode in a thinking aloud case study
process should lead to a smooth progression in the
and diSCUSS the senses in which it is or is not an attainment of knowledge-an incremental' 'march
example of a scientific insight or "eureka" event. of progress" without large-scale reorganizations.
In particular, the case study is used to address ele-
ments of the following more specific questions: Rationalism versus Inductivism. A second
major issue is the source of new theoretical knowl-
1. What is a scientific insight? Can one identify
edge. The question of the sources of and justifica-
"insight" events or "eureka" events in
tion for new knowledge is a central point of contro-
thinking aloud protocols? Why do insights
versy between the rationalist and empiricist
occur? Why do periods of slow and fast pro-
traditions in Western thought. The rationalist tradi-
gress occur in scientific thinking?
tion emphasizes the power of reasoning from prior
2. What processes are involved in the generation
knowledge and greatly values the consistency and
of a scientific hypothesis? In particular, are
beauty of the resulting theories. Reasoning power,
hypotheses always generated as inductions
coupled with the prior beliefs of the learner, are
from data? What role do analogies and
emphasized as sources of knowledge. On the other
thought experiments play in creative scien-
hand, the empiricist tradition emphasizes the im-
tific thinking? What is the role of explanatory
portance of careful observation and greatly values
models?
the reliability of repeatable experimental pro-
3. Are there any parallels between the tensions
cedures. Here the term induction will denote a pro-
observed in an individual scientist thinking
cess by which a more general principle is abstracted
aloud and the tensions Kuhn describes be-
from a set of empirical observations as the source. I
tween an anomaly and a scientific paradigm?
will use the term inductivism to refer to the belief
4. What impact do findings relevant to the above
that induction is the primary, if not exclusive,
questions have on the concepts of "knowl-
source of hypotheses in science. Stated most sim-
edge construction" and "discovery learn-
ply, in this view scientists gradually gather facts,
ing" in a theory of instruction?
use inductive reasoning to organize them into gen-
I will attempt to show that empirical evidence eral statements, and finally build up a pyramid of
can be collected which speaks to certain aspects general empirical laws that summarize all the
of these questions. gathered data. Theory-driven and data-driven ap-
CHAPTER 20 • SOURCES OF CREATIVITY IN SCIENCE 343
proaches in artificial intelligence can, to some ex- of the scientific enterprise. 1 Prior to this century,
tent, be thought of as modem inheritors of the ra- empiricists focused on observation as the primary
tionalist and inductivist viewpoints. source of knowledge in science, and the twentieth-
Although they refer to different issues, the eu- century logical positivists built on their tradition by
rekaist versus accretionist and the rationalist versus attempting to show that scientific knowledge could
inductivist controversies are not independent his- be grounded firmly in sense experience. In their
torically but tend to interact. Eurekaism tends to be view, careful observations, and the assumptions of
associated with rationalism, whereas accretionism a common scientific observation language and the
tends to be associated with inductivism. Thus, it is applicability of the laws of logic and probability,
sometimes useful to refer to an individual position could provide science with knowledge of the utmost
as "rationalist-eurekaist" or "inductivist-accre- reliability, if not certainty. Although the logical
tionist." A rationalist-eurekaist view of theory positivists concentrated on issues surrounding the
change is associated with the idea that scientists at justification of theories rather than their origin,
times must be very creative, whereas the induc- their empiricism also affected views of the origins
tivist-accretionist view suggests that scientists can of scientific knowledge. Science was described in
make progress by relying on small changes without an accretionist manner as building and extending
large creative breakthroughs. This simplified pic- theories incrementally, approaching truth in a
ture of two opposing camps can be used as a starting monotonic way. For example, Rudolph Carnap
point for introducing some important issues con- held the inductivist belief that science advances up-
cerning the nature of science. ward from particular empirical facts to generaliza-
Gould (1980) noted that writers on both sides of tions that summarize or provide an abbreviation for
this controversy have tried to claim Darwin's theory a body of such facts (Suppe, 1974, p. 15n). Cer-
of evolution as an example. Historically, induc- tainly positivism has influenced the methodology of
tivist-accretionists claimed that it was a prime ex- other disciplines (e.g., behaviorism in psychology)
ample of the power of induction, as facts gathered in this direction.
by Darwin during the voyage of the Beagle were Important attacks on the positivist position, such
slowly pieced together into a grand theory. Ra- as Popper's (1959) success in showing that induc-
tionalist-eurekaists claimed that Darwin had a sud- tion cannot confirm the truth of theories, Hanson's
den, crucial insight upon reading Malthus's theory (1958) claim that observations are "theory laden,"
of human population constraints. and Kuhn's (1962) claim that theoretical advances
But both of these positions runs the risk of being often precede the empirical findings used to support
oversimplified. As Gould (1980) put it: "Induc- them in science, have raised serious problems by
tivism reduces genius to dull, rote operations. Eu- arguing against the empiricist emphasis on sense
rekaism grants it an inaccessible status more in the experience as the pre-eminent basis for knowledge.
domain of intrinsic mystery than in a realm where Popper (1959) held that the proper role for data is in
we might understand and learn from it" (pp. 60- the criticism rather than the confirmation of hypoth-
61). The implied challenge here is to find a less eses. Hypotheses are conjectures made by scientists
simplistic view that helps to explain creative behav- rather than certainties abstracted from data. But
ior in a nontrivial way. In this chapter, accounting these conjectures can be reliably criticized and fal-
for the data from the case study leads to a more sified by collecting data. This allows science to
complex view of scientific discovery than either the make progress via a series of conjectural hypoth-
extreme eurekaist, accretionist, rationalist, or in- eses and reliable criticisms. Popper's work pro-
ductivist positions can provide. vided support for the model shown in Figure 1, the
Toward the end of the chapter, I also review hypothetico-deductive method. There are three
some recent historical studies of Darwin's insights
that point to the same conclusion.
1Placing different scholars on these two broad spectra ignores
many differences between them and requires a number of sim-
Philosophical Positions plifications. For example, some scholars (e.g., positivists and
I give a brief outline here of how these two broad Popper, 1959) tend to concern themselves with the formaljusti-
fication of theories, whereas others (e.g., Hanson, 1958, Kuhn,
questions concerning the source and pace of scien- 1962) also focus on their psychological origin; arguments also
tific theories interact with some of the major twen- vary as to whether they refer to science as a whole or to the
tieth century philosophical positions on the nature individual scientist.
-
tionist view of theory change. In this view, normal
(A) Conjecture science may be accretionist in character, but revolu-
hypothesis tionary periods in science involve crisis and recon-
struction, implying that science progresses at an
uneven pace with periods of slow and fast change.
On the other hand, critics of Kuhn, such as Toulmin
(1972), have, in tum, questioned the reality of sci-
entific revolutions, arguing for a more continuous
(8) Deduce predictions
~
view of theory change. 2
and test empirically In summary, an inductivist-accretionist view of
science sees it as compiling facts and generaliza-
Fail Pass tions in a piece-by-piece fashion. Induction is the
primary process of hypothesis generation, with a
1 '1 one-directional flow of knowledge from data up-
ward to theories. In a rationalist-eurekaist view, on
Figure 1. Basic conjecture and empirical evaluation cycle in the other hand, significant theoretical develop-
hypothetico-deductive method. ments can occur when a scientist formulates mental
constructions at some distance from existing data
and can actually develop new ways oflooking at old
main stages shown here: (1) a hypothesis is formed data. Thus, knowledge can flow downward from a
by conjecture; (2) predictions deduced from the hy- newly invented, general theory to influence the for-
pothesis are tested empirically; (3) if the prediction mation of new specific theories, to reorganize one's
is incorrect, the hypothesis is rejected and the cycle view of existing data, and to suggest new places to
restarted. Popper maintained, contrary to the log- collect important data. Such reorganizations pre-
ical positivists, that a successful empirical test does sumably would require a large degree of creativity,
little to confirm a hypothesis, but that failing such a perhaps even extraordinary eureka episodes of
test is grounds for rejecting a hypothesis. Those insight.
hypotheses that survive the gauntlet of repeated These two views have been the subject of con-
testing become accepted laws. Favored laws tinuing controversy. Philosophers have taken vari-
emerge through the survival-of-the-fittest conjec- ous positions between these extremes, and some
tures, so to speak. However, Popper's emphasis on have attempted to point to examples from the histo-
conjecture also opens up the possibility that a non- ry of science supporting their position. However, in
inductive, nonaccretionist process, or even a eureka historical studies, it is always difficult to find data
event could be involved in hypothesis formation. saying much in detail about the actual process of
Popper's views have in tum been criticized in a hypothesis formation in the individual scientist. In
number of ways. The most relevant shortcoming for the next section, I consider several descriptions of
the purposes of the present study is that his classic this process as proposed by philosophers, after
work does not specify mechanisms for generating which I analyze a thinking-aloud case study to ex-
hypotheses; he relegates this task to psychology and amine these issues from an empirical base at a more
says only that hypotheses must be conjectural in detailed level. In this case study, examples of non-
nature. Also, Hanson's (1958) notion that observa- inductive reasoning in the formation of hypotheses
tions can be "theory-laden" implies that empirical will be examined in order to determine whether
testing in the hypothetico-deductive method may these types of data can challenge the inductivist
not be fully reliable and sufficient on its own as a position; and an identified "insight episode" will
means of hypothesis evaluation. (Other means of
hypothesis evaluation that are more rationalist in
character will be discussed in the next section.) 2Because this chapter focuses on thinking in the individual scien-
With regard to the pace of theory change, Kuhn's tist, I will not discuss here important work that emphasizes
(1962) ideas of revolution within a scientific disci- social factors in the development of scientific ideas. Although
these factors are undoubtedly significant, I believe that studying
pline and the creative "gestalt switch" that is re- hypothesis generation processes in the individual scientist is an
quired for an individual scientist to move outside of effective heuristic strategy for investigating a crucial part of the
his or her own paradigm argue against an accre- problem.
be examined to determine whether it can provide tion models of data-driven inductive processes
support for or against a eurekaist position. for generating certain scientific laws.
Some Possible Views of Hypothesis Answer 2: Creative Intuition

Formation Processes in the Individual Is some form of induction or guessing the only
Scientist source of scientific hypotheses? A number of recent
authors have answered no to this question by point-
How Are Hypotheses Formed? ing to the role of creativity, intuition, and the un-
In this section, it will be useful to concentrate on conscious in generating hypotheses (Koestler,
the more specific question, "What are the mental 1964, Polanyi, 1966; Rothenberg, 1979). Unfortu-
processes by which hypotheses are formed in an nately, I can only make the briefest mention of these
individual?" The answer to this question should long and detailed works here. Their views can be
involve some sort of model of the mental processes roughly characterized as replacing the "Hypothesis
being used. Discussion of this narrower question Formation by Induction" step in Box A of Figure 2
about individuals may be of some interest to those with a process labeled "Hypothesis Formation by
investigating the broader question about science as Creative Intuition." For example, Polanyi empha-
a whole, even though the latter issue is more com- sized the role of intuition and tacit knowledge in
plex. In fact, surprisingly little work has been ad- science. Rothenberg proposed a process of "Janu-
dressed to this question, especially in comparison to sian thinking," whereby a person is able to jux-
the complementary question, "How are scientific tapose seemingly contradictory ideas, as a common
hypotheses tested?" Here I give a brief overview of element in creative thinking. Koestler pointed to
several possible positions that can be taken on the "bisociative thought" -the ability to connect nor-
first question concerning formation. mally independent frames of reference-and to the
Popper's (1959) position and the hypothetico- role of the unconscious in accounting for creativity.
deductive method shown in Figure 1 can be taken as An interesting controversy emerged in this area.
a starting point here in the form of a non answer . The Perkins (1981) argued that all these descriptions
method shows one way in which hypotheses might attempt to point to extraordinary thinking pro-
be tested but does not show how they are generated. cesses; they attempt to supplement ordinary reason-
ing with something more powerful. He countered
this idea with the claim that most creative acts can
Answer 1: Hypothetico-Deductive Method Plus be explained plausibly by a model in which a person
Induction uses certain ordinary thinking processes more in-
Popper argued convincingly that induction can-
not be used to confirm the truth of scientific theo-
ries. However, some modem scholars retain some
form of induction in their model of scientific meth-
od as a way to suggest hypotheses. This can be
,.... (A) Hypothesis formation
by induction
represented by the model shown in Figure 2 that
combines the hypothetico-deductive method with
induction as a source of hypotheses. Here there is
no claim for a "logic" of discovery but only for a
fallible method for generating hypotheses. Further
experiments are performed in order to evaluate the ( B) Deduce predictions
~
inductions. Such a diagram is commonly implied in and test empirically
everyday characterizations of scientific method as a
combination of induction and deduction. Scholars Fail Pass
such as Harre (1983), Achinstein (1970), and Gre-
gory (1981) argue that induction can playa role in
hypothesis formation. However, they believe that
I I
other processes can be involved as well. Recently, Figure 2. Hypothetico-deductive method with the addition of
Langley (1979) has attempted to develop simula- induction as a source of hypotheses.
tensively, or with special goals in mind. In his cited example is that merely being able to make
view, the difference between a creative and an un- predictions from the empirical gas law, stating that
creative person is a difference of degree and pur- PV is proportional to RT, is not equivalent to under-
pose, not a difference of kind. Perkins also de- standing the explanation for gas behavior in terms
scribed authors like Koestler as contributing mainly of an imagable model of billiard-balI-like mole-
to the description of the products of creative think- cules in motion. Unlike the empirical law, the
ing; a remaining problem is to specify the processes model provides a description of a hidden process
of creative thinking in more detail. that explains how the gas works and answers
"why" questions about where observable changes
in temperature and pressure come from. Causal re-
Answer 3: Analogies and Successive lationships are often central in such models. The
Refinement Cycles as Sources of Explanatory model can not only add significant explanatory
Model Hypotheses power to one's knowledge but can also suggest
questions that stimulate the future growth of the
The work of another group of scholars in philoso- theory. In this view, the visualizable model is a
phy of science, including Campbell (1920), Harre major locus of meaning for a scientific theory.
(1961), Nagel (1961), and Hesse (1966), suggests (Brief summaries of these views are given in Harre,
that analogies may be a source of hypotheses. They 1967 and Hesse, 1967.)
argue that scientists often think in terms of the- The above authors, as well as Black (1979), ar-
oretical explanatory models, such as molecules, gue that models involve analogies to familiar situa-
waves, and fields, that are a separate kind of hy- tions (e.g., gases are analogous to a collection of
pothesis from empirical laws. Such models are not colliding balls). In Nagel's (1961) terms, such visu-
simply condensed summaries of empirical observa- alizable analogue models help scientists "make the
tions but rather are inventions that contribute new unfamiliar familiar. " This suggests that analogical
theoretical terms and images that are part of the reasoning may be an important noninductive source
scientist's view of the world, and that are not for generating such hypothetical models. More re-
" gi ven" in the data. cently, theory formation and assessment cycles
As shown in Figure 3, they see a distinction be- using analogies have been discussed by Clement
tween an empirical law hypothesis summarizing an (1981), Nersessian (1984), Holland, Holyoak,
observed regularity and what I will call an Nisbett, and Thagard (1986), and Darden and Rada
explanatory model hypothesis. Campbell's often (1988).
Most of the above authors also emphasize a ra-
tional (nonempirical) contribution to hypothesis
evaluation, holding that explanatory models are
evaluated also with respect to the criteria of sim-
MORE THEORETICAL plicity, aesthetic appeal, and consistency with other
accepted models.
o Formal principles
o Explanatory model hypotheses The Model Construction Cycle. Figure 4 repre-

sents an attempt to bring together several of these
features in a single idealized model of the hypoth-
esis development process for constructing scientific
o Empirical law hypotheses: models. Typically, such a process would be used to
mathematical or verbal descriptions develop an explanation for a newly recognized phe-
of patterns in observations nomenon. Essentially, the diagram depicts a
cyclical process of hypothesis generation, rational
o Observations and empirical testing, and modification or rejec-
tion. It is difficult to describe so complex a process
MORE EMPIRICAL
in a single diagram, but a simplified model will aid
in the present analysis. In contrast to Figure 2, in
Figure 3. Types of knowledge used in science. Explanatory Figure 4, when a hypothesis is evaluated nega-
models are distinguished from empirical laws. tively, it can sometimes be improved through modi-
( B) Activate possible
(A) Make initial analogies and rela-
observations I------
ted model elements
(c) Construct initial
,
model
~
(E) -I~
Reject or modify
( D) Rational (non -em-
piricail evaluation ~
model
(e.g. for consistency)
Fail
'--I p~r~
( F) Construct and per-

form empirical tests
Fail Pass
I -1-
Figure 4. Hypothesis development (model construction) cycle of conjecture, evaluation, and modification or rejection.
fication, instead of being completely rejected. ble on those occasions when one has the prior ad-
Thus, it may undergo a series of successive vantage of possessing the right variables, or compo-
refinements. nents of compound variables, to look for. But the
The double-ended arrows, in Figure 4, between explanatory model hypotheses being considered
"Make Initial Observations," and "Construct Ini- here would be formed by a less data-driven abduc-
tial Model" represent the idea that not only does tive process, possibly for just a single instance of
model construction respond to observation but that the phenomenon. Such a process might "pla-
one's focus of attention during observation can be giarize" the knowledge structure from an analo-
guided by one's initial model. This and other dou- gous case in memory to form the starting point or
ble-ended arrows indicate that the initial model gen- core of a new model. Or it might integrate several
eration process can be highly interactive and com- related model elements-constructing a new model
plex. It is still poorly understood. by combining several existing knowledge struc-
Essentially, the scientist must construct or piece tures previously known to the subject.
together a conjectured picture of a hidden structure Hypothesis evaluation can take place in two ma-
or process that explains why the phenomenon oc- jor ways. Empirical testing can add support to or
curred. Peirce (1958) and Hanson (1958) used the disconfirm a hypothesis. Rational evaluation can
term "abduction" (or retroduction) to describe the also support or disconfirm a hypothesis, depending,
process of formulating a hypothesis which, if it for example, on whether it is found to be consistent
were true, would account for the phenomenon in or inconsistent with other established theories.
question. The hypothesis can be a guess as long as it Evaluation processes cannot provide full confirma-
accounts for (predicts after the fact) the observation, but can lead a scientist to have increased or
tions collected so far. Empirical law hypotheses reduced confidence in a theory. Once generated, a
that consist only of a recognized regularity or re- hypothesis undergoes repeated cycles of rational
peated pattern in the variables, such as those dis- and empirical testing and modifications as needed.
cussed by Langley (1979), might be formed via a A limitation of the diagram that is not intended to be
more data-driven inductive process. This is possi- part of the model is the order in which rational and
empirical evaluations occur; tests can occur in dif- hypothesis formation, then it is a promising candi-
ferent orders on different cycles. date for a cognitive process underlying insight or
The endless loops in Figure 4 indicate that, ide- eureka events. This issue will be examined closely in
ally, theories in science are always open to new the section following the next one.
criticisms. However, as Kuhn (1962) pointed out,
scientists will sometimes ignore or discount some
criticisms in order to protect a favored theory. In Evidence from Thinking Aloud
practice, research groups may adopt a "protected Protocols on Model Construction Cycles
core" of theories that they take as givens (Lakatos, Using Analogies
1978).
A missing element in Figure 4 is the influence of
the subject's prior theoretical framework. This ele- Recently, cogmhve psychology has begun to
ment is difficult to depict, as it could affect so many study complex human cognition through the use of
of the processes shown. Because the scientist oper- protocol analysis. This section uses this method to
ates from a background of broader theoretical as- examine the process of hypothesis generation in
sumptions, these may have an early influence on the thinking-aloud protocols. Several examples of
model elements and analogies that come to mind, spontaneous analogies will be examined, as well as
and even (according to Hanson, 1958, and to Kuhn, a breakthrough episode that appears to be an exam-
1962) on what is observed. 3 ple of insight behavior. Instead of working back-
ward from historical records and outcomes, a more
direct analysis of the processes operating in the
Summary
thinking scientist will be attempted here.
The great difficulty, of course, is to have a video
camera trained on the scientist at one of the rare
In summary, little empirical work has been done moments when he or she formulates a hypothesis.
on the question of hypothesis formation processes One way to overcome this difficulty is to pose to the
in science, but philosophers have proposed several subject conceptually challenging but not over-
possibilities, including guessing, abduction, induc- whelming problems, which allow for the formula-
tion, and creative leaps. In addition, Campbell tion of hypotheses and explanations. The data dis-
(1920) and others have introduced the interesting cussed here were taken from interviews in which
distinction between empirical law hypotheses that advanced doctoral candidates or professors in tech-
are summaries of perceived patterns in observations nical fields were asked to think aloud as they solved
and explanatory model hypotheses that introduce such problems. Although the problems do not con-
visualizable models at a theoretical level and that cern issues on the frontier of science, in many cases
often contain currently unobservable entities. They they ask subjects to give a scientific explanation of a
suggest that analogies may be an important means phenomenon with which they are unfamiliar (i.e., a
of constructing the latter type of hypothesis. A pos- problem on the frontier of their own personal
sible synthesis of these ideas was proposed in Fig- knowledge). Thus, it is plausible that the thought
ure 4. It allows for the possibility that the hypo- processes analyzed will share some characteristics
thetico-deductive method, induction, abduction, with hypothesis formation and model construction
analogy, rational evaluation, and hypothesis modi- processes used on the frontiers of science.
fication may all play important roles at different
times in scientific thought.
The idea that analogies can be involved in hypoth- Use of Analogies and Models in Expert Problem
esis formation is often used to support a eurekaist Solutions
view of scientific discovery. If analogy generation is
a fast, creative process, and if it is important in In this section, evidence will be presented indi-
cating that analogies can be involved in a signifi-
cant way in generating the solution to a scientific
3The form of Figure 4 was developed via an extended successive
refinement process and was also designed to account for em- problem, and, more specifically, that they can
pirical data from protocols (like the ones to be discussed), and sometimes lead to a new model of the problem
not just as a summary of prior literature. situation.
YOU ARE GIVEN THE TASK OF ROLLING A HEAVY WHEEL UP A Use 0/ the Terms Analogy and Model. This ini-
HILL. OOES IT TAKE MORE. LESS. OR THE SAME AMOUNT OF
FORCE TO ROLL THE WHEEL WHEN YOU PUSH AT X. RATHER
tial example motivates the following uses the terms
THAN AT Y? analogy and model. I will refer to the occurrence of
a spontaneous analogy when the subject spon-
taneously shifts his attention to a different situation
ASSUME THAT YOU APPLY A FORCE PARALLEL TO THE SLOPE
AT ONE OF THE TWO POINTS SHOWN. AND THAT THERE ARE NO (called the analogous case) that he believes may
PROBLEMS WITH POSITIONING OR GRIPPING THE WHEEL. have relevant structural similarities to the original
ASSUME THAT THE WHEEL CAN BE ROLLED WITHOUT SLIPPING problem (also referred to here as the target). When
BY PUSHING IT AT EITHER POINT. this is true, the subject's cognitive structures repre-
senting the target and the analogous case will have
at least one structural relationship in common. In
the discussion that follows, I will refer to the lever
situation as the analogous case and to the structural
Y~ similarity relationship between the lever and the
wheel as the analogy relation.
Figure 5. Wheel problem.
Some analogies are used casually for "deco-
rative" purposes only. By contrast, the following
definition of a scientific model as a predictive anal-
ogy is intended to identify analogies that are used
for serious scientific purposes. Here, in the broad
Wheel Problem. I will first present a very brief sense of the term, a scientific model will refer to a
description of a solution to the "wheel problem" cognitive structure, where the subject believes that
illustrated in Figure 5. The question asks whether the model situation is analogous to the target situa-
one can exert a more effective uphill force parallel tion and believes that one may be able to use the
to the slope at the top of a wheel or at the level of the model to predict or account for observations made
axle (as in pushing on the wheel of a covered in the target.
wagon, for example). One way in which models are distinguished the-
Subject S2 compared the wheel to the analogous oretically from rote facts or procedures is by virtue
case of pushing on a heavy lever hinged to the hill of having a richer set of relational interconnections
(Figure 6B). He reasoned that pushing at the point within their structure as opposed to being a collec-
higher up on the lever would require less force. He tion of independent facts. A model M gives the
then made an inference by analogy that the wheel scientist a way of thinking about a target situation T
would be easier to push at the top (the correct an- that can predict how T behaves under certain condi-
swer). Here the lever is used as a model in some tions (whether this happens before or after the be-
sense for the wheel. havior is observed is not important for the defini-
/ ~~
Y~ SL---J
A B
Figure 6. Models used in wheel problem. (A) Original problem, (B) Lever model. (C) Spoked wheel without a rim model.
tion). The lever analogy for the wheel is a scientific A WEIGHT IS HUNG ON A SPRING. THE ORIGINAL SPRING IS
REPLACEO WITH A SPRING
model in this sense. Well-developed and successful
scientific models are also valued for being precise, - MADE OF THE SAME KIND OF WIRE.
- WITH THE SAME NUMBER OF COILS.
unambiguous, general, and predictive (Gentner, - BUT WITH COILS THAT ARE TWICE AS WIDE IN
1982). In addition, scientists often prefer models DIAMETER.
that are visualizable, causal, simple, and that con- WILL THE SPRING STRETCH FROM ITS NATURAL LENGTH.
tain familiar entities. (In a later section, the nar- MORE. LESS. OR THE SAME AMOUNT UNDER THE SAME WEIGHT?
rower category of an explanatory model will be (ASSUME THE MASS OF THE SPRING IS NEGLIGIBLE COMPARED
TO THE MASS OF THE WEIGHT). WHY DO YOU THINK SO?
defined as one that posits a material similarity in
which elements of M are assumed to actually exist
as hidden or nonobvious elements in T.)
t
(I) (2 )
~.
Improving the Model for the Wheel. The sub-
ject was confident that it would be easiest to move
the heavy lever in Figure 6B by pushing at point X,
but he was not so confident that it was a good
model for the wheel; he criticized the model by
i STRETCH
questioning whether there was a valid analogy
relationship between it and the case of the wheel. Figure 7. Spring problem.
Can one actually view the wheel as a lever, given
that the "fulcrum" at the bottom of the wheel
is always moving and never fixed? A second im-
proved analogue model described by this subject tion via analogies is not the most common method
was the spoked wheel without a rim shown in Fig- for solving problems.
ure 6C. The spokes allow one to view the wheel as Spring Problem. A second example of a prob-
a collection of many levers, thereby reducing any lem illustrating analogy use and model construction
worries about the moving fulcrum. This is a useful is the "spring problem" shown in Figure 7. That
model of the wheel for many purposes, including the wide spring will stretch farther seems to corre-
the present problem. spond to most people's initial intuition about this
In summary, after criticizing the "lever" model, problem. However, carefully answering the ques-
the subject was able to produce a second, more tion about why the wide spring stretches more (and
elaborate analogous case that provided an improved explaining exactly where the restoring force of the
model. This provides an initial example of a hy- spring comes from) is a much more difficult task.
pothesis generation, evaluation, and modification Because it asks why, this is largely an explanation
process leading to the formation and improvement question rather than a question with a single, well-
of a mental model via an analogy. defined answer. Thus, it is less like an everyday
"puzzle" problem and more like a theoretical
Creative Aspect of Analogies. As mentioned "why" question in science in which the answer is
above, an analogy is a related case that the subject an explanation.
believes is structurally similar to an original case. In a study of expert qualitative reasoning, I re-
However, the case also must differ in a significant corded ten professors and advanced graduate stu-
way from the original problem to be counted as an dents in technical fields while they were solving the
analogy. By this I mean that one or more features spring problem out loud (Clement, 1988). They
commonly assumed to be fixed in the original prob- were told that the purpose of the interview was to
lem are different in the related case. In order to study problem-solving methods and were given in-
generate an analogy like the lever analogy, the sub- structions to solve the problem "in any way you
ject must break away from the original problem can." After they reached an answer, subjects were
context. This "breaking the set" of the original asked to give an estimate of their confidence in their
problem appears to be one of the main reasons that answer. They were then asked if there was any way
generating an analogy is considered a creative act they could increase their confidence, and this often
and is most likely one reason that model construc- led to further work on the problem. Probing by the
ble and consistent. For simpler types of behavior,

such modeling can be fairly detailed and, in some
cases, can be expressed as a computer simulation.
For more complex or poorly understood phe-
Figure 8. Bending rod model. Longer rod bends more. nomena, an initial step in modeling can be achieved
by formulating a general description of structure
and process features-the basic units or cognitive
interviewer was kept to a minimum, usually con- objects to be used, the outline of a model, and a set
sisting of a reminder to keep talking. Occasionally, of "design criteria" that a more detailed model
the interviewer would ask for clarification of an would need to fulfill. The analysis of the case study
ambiguous statement. discussed in the remainder of this chapter will be
Some of the solutions were quite complex and aimed at the latter level.
took up to 50 minutes to complete. All subjects
favored the (correct) answer that the wide spring
S2's Protocol. In the spring problem, subject S2
would stretch farther, but the subjects varied con-
first generated the model of comparing a long hori-
siderably in the types of explanations they gave for
zontal bending rod with a short one (a weight is
their prediction. A number of subjects considered
attached to the end of each rod), inferring that seg-
the analogous case of a horizontal bending rod
ments of the wider spring would bend more and
(shown in Figure 8) or variations thereof. Most sub-
therefore stretch more. However, he was concerned
jects had a strong intuition that a longer rod would
about the appropriateness of this model because of
bend more than a shorter rod under the same weight;
the apparent lack of a match between seeing bend-
this analogy suggested to them that the wider spring
ing in the rod and not seeing bending in the wire in a
would stretch more. A number of other analogies
stretched spring. One can visualize this discrepancy
attempted in this problem are discussed in Clement
here by thinking of the increasing slope a bug would
(l988b), including two foam rubber blocks, one
experience walking down a bending rod and the
with large and one with small air holes in the foam,
constant slope the bug would experience walking
springs in series, springs in parallel, series circuits,
down the helix of a stretched spring. This discrep-
parallel circuits, and molecules in different
ancy led him to question whether the bending rod
crystals. Altogether, 31 significant analogies were
was an appropriate model for the spring. He then
observed, and they were generated by seven of the
constructed the analogous case of the "zigzag
ten subjects. Thus, a large number of spontaneous
spring" shown in Figure 9, apparently in order to
analogies were generated for this problem.
attempt to evaluate the analogy relation between the
spring and the bending rod and to attempt to con-
A Case Study of Hypothesis Generation struct an improved model. Because the full tran-
script is quite long, only verbatim excerpts are pre-
In the remainder of this section, I will focus on sented here. (My interpretive comments in the
the case study of subject S2 who appears to devel- transcript are indicated in brackets.)
op, criticize, and modify analogous cases for the
spring problem until he produces a new hypothesis
in the form of an explanatory model for how springs
work.
Purpose of Case Study. One of the main reasons

for doing an in-depth case study is to develop and
refine a basic vocabulary of concepts for describing
psychological observations and theories. The initial
challenge of such a study is to develop and describe
the "units" of behavior to be used in observation
and to propose an initial cognitive model in the form
of a set of cognitive structures and processes that
can account for the behavior and that is both plausi- Figure 9. Zigzag spring model. Wider spring stretches more.
5 S2: I have one good idea to start with; it find myself being tempted back to the
occurs to me that a spring is nothing but a straightened spring [rod] model again ...
rod wound up uh, and therefore maybe I I still don't see why coiling the spring
could answer the question for a rod. But should make any difference .... Surely
then it occurs to me that there's something you could coil a spring in squares, let's
clearly wrong with that metaphor because say, and it . . . would still behave more or
if I actually took spring wire and it was less the same. Ah! from squares, visually I
straight instead, it certainly wouldn't hang suddenly get the idea of a zigzag spring
down like a spring does .... It would rather than a coiled spring; that strikes me
droop . . . and its slope would steadily as an interesting idea (draws Figure
increase as you . . . went away from the 9). . . . Might there be something in that
point of attachment, whereas in a spring, idea? ...
the slope of the spiral is constant. I see a problem with this idea. The
7 S2: Why does a spring stretch? .. I'm still led problem . . . is that . . . the
back to this notion . . . of the spring stretch . . . has to do with . . . the joint.
straightened out [a bending rod] ... I'm But the springiness of the ... real spring
bothered by the fact that the slope doesn't is a distributed springiness; ... So ... I
remain constant as you go along it. It wonder if I can make the [zigzag]
seems as though it ought to be a good spring ... where the action ... isn't at
analogy, but somehow, somehow it doesn't the angles . . . it's distributed along the
seem to hold up .... length .... And I'm going to do that; I
23 S2: I feel I want to reject the straightened have a visualization .... Here's a
spring model-as a bad model of what a stretchable bar; (draws modified zigzag
spring is like. I feel I need to understand spring in Figure 10) a bendable bar, and
the nature of a spring in order to answer then we have a rigid connector. . . . And
the question. Here's a good idea. It occurs when we do this what bends . . . is the
to me that a single coil of a spring bendable bars . . . and that would behave
wrapped once around is the same as a like a spring. I can imagine that it would.
whole spring .... In the one-coil case, I
Here there is evidence that the subject is generat- zigzag spring with the lack of change in slope in the
ing a series of analogue models for the spring- original helical spring. However, these attempts do
from the rod to the angular zigzag spring to the provide evidence for another thought pattern in the
rectangular zigzag spring with stiff joints. The form of a repeated dialectic process of model con-
zigzag spring is eventually dropped, presumably struction, criticism, and modification.
because he was still critical of this model and could Next, S2 considers the analogy of a double-
not reconcile the bending going on in sections of the length spring instead of the double-width spring
appearing in the original problem.
37 S2: This rod here: as the weight moves along,

it bends more and more the further out the
weight is .... Hmmrn, what if I imagined
moving the weight along the
spring . . . would that tell me anything?
Would that? I don't know. I don't see why
it should. What if the spring were twice as
long ... instead of twice as wide? ... It
seems to me pretty clear that the spring
that's twice as long is going to stretch
Figure 10. Modified zigzag spring model. more .... Now if this is the same as a
spring that's twice as wide, then that

should stretch more .... Uhh, but is it the
same as a spring that's twice as wide?
Again, I just don't see why ... the coiling
[vs. a rod] should make any difference. It
just seems geometrically irrelevant to me
somehow .... But I ... can't-I have
trouble . . . bring that into consonance
with the behavior of an actually stretched Figure 11. Hexagonal coil model.
out spring . . . the slope problem anomaly
[increasing slope in the rod, but not in the
spring]-if I could resolve that depends on the depth of understanding sought by
anomaly . . . then I would feel confident the subject and on the subject's prior knowledge.
of my answer. . . but this anomaly The first level of depth in understanding is simply to
bothers me a lot. state an intuition that the wide spring will stretch
more; a second level is to give some plausible justi-
fication for this. For subjects who have previously
Again, he seems critical of the appropriateness of learned that there is twisting in the spring wire dur-
an analogy in the case of the double-length spring. ing stretching, they can, with some effort, achieve a
third quantitative level in identifying three causal,
57 S2: I feel as though I'm reasoning in circles. I linear factors leading to the result that the stretch is
think I'll make a deliberate effort to break proportional to the cube of the coil diameter. Proba-
out of the circle somehow. What else could bly the most difficult achievement occurs when the
I use that stretches . . . like rubber subject does not know about the invisible twisting
bands . . . what else in the wire, but is somehow able to construct that
stretches . . . molecules, polyesters, car hypothesis. S2 achieves this in the next section of
springs [leaf springs] ... what about the protocol to be discussed.
a ... two-dimensional spiral [watch] To see why this square coil model is helpful, note
spring? That doesn't seem to help. that it can, in tum, be understood in terms of two
simpler cases, the twisting rod and the bending rod,
as shown in Figure 12. That is, pulling the end of
At this point, the bending rod, double-length the lever" 1" down not only bends rod 1, but it also
spring, and zigzag spring analogies have each twists rod 2. (One way to comprehend this idea is to
pointed him to the correct answer to the problem, view rod I as a wrench that is twisting rod 2.) The
yet he remains unsatisfied with his understanding. same is true for all other adjacent rod pairs. Thus,
In line 57, he continues to search unsuccessfully for twisting is an important type of deformation in the
a more satisfactory analogous case. spring wire in this model.
This part of the protocol is reported in sections as
follows:
Insight Section. Subsequently, subject S2 pro-
duces an extremely productive analogy when he
generates the idea of the hexagonally shaped coil in :3
Figure II and moves from there to the idea of the
square-shaped coil in Figure 12. Imagining the
stretching of these polygonal coils apparently al-
lowed him to recognize that some of the restoring
forces in the spring come from twisting in the wire
instead of bending-a major breakthrough in the
/
solution that corresponds to the way in which en-
gineering specialists view springs. Much of the re-
X Y
mainder of this chapter will focus on this insight.
The impressiveness of the reasoning displayed
by different subjects in solving the spring problem Figure 12. Square coil model.
1. Subject is still in conflict about whether

spring wire is bending .
2. Subject generates a series of polygonal cOlI
analogies
3. Torsion discovery
4. Subject evaluates and adapts square coil as a
preferred model of the spring
5. Subject comments on his increased under-
standing
Figure 13. Single spring coil.
Section I: Subject Is Still in Conflict about different way, somehow. Let me just
Whether Spring Wire Is Bending generate ideas about circularity. What
could the circularity [in contrast to the
57 S2: I just . . . have the intuition that
rod] do? Why should it matter? How
a . . . straight rod ought to in some sense
would it change the way the force is
be a good model for a spring. But there
transmitted from increment to increment
are these anomalies that won't go away.
of the spring? Aha! Now let me think
And yet I can't see ... a better model.
about; Aha! Now this is interesting. I
79 S2: ... I'm just trying to imagine the
imagined; I recalled my idea of the square
coil . . . (traces circle about 7 inches in
spring and the square is sort of like a
diameter in air in front of self) a circle
circle and I wonder . . . what if I start
with a break in it. ...
with a rod and bend it once (places hands
81 S2: (Has just drawn Figure 13) ... you
at each end of rod in Figure 8 and
could just hold it there . . . and apply a
motions as if bending a wire) and then I
force there, and the spring stretches ....
bend it again?
I'll be damned if I see why it [the coil]
119 S2: What if I produce a series of successive
should be any different from that case
approximations to . . . the circle by
[the rod] ....
producing a series of polygons? Maybe
87 S2: ... if you start with a [stretched] helix
that would clarify because maybe that,
and unwind it . . . you should get a bow
that's constructing a continuous bridge, or
[bend], but you don't. I mean visually
sort of a continuous bridge, between the
imagining it, you don't. I don't see how
two cases [the rod and the coil]. Clearly
you could make the bow go away-just
there can't be a hell of a lot of difference
to wind it up-damn it!
between the circle, and, say, a
III S2: Darn it, darn it, darn it ... why should
hexagon ....
that [the difference between a rod and a
121 S2: ... or even a triangle ... square ...
coil] matter? ... I'm visualizing what
(draws hexagon in Figure 11) .... Now
will happen when you just take this single
that, a [hexagon] is essentially a circle. I
coil and pull down on it and it stretches;
mean, surely springwise that [hexagon]
and it stretches. . . .
would behave pretty much like a circle
(The subject spends a considerable amount of time try- does.
ing to resolve this issue without making progress.)
Section 3: Torsion Discovery
Section 2: Subject Generates a Series
121 S2: Now that's interesting. Just looking at
of Polygonal Coil Analogies
this it occurs to me that when force is
117 S2: (40 minutes into the protocol) I keep applied here, you not only get a bend on
circling back to these same issues without this segment, but because there's a pivot
getting anywhere with them. . . . I need here (points to X in Figure 11), you get a
to . . . think about it in some radically torsion effect. . . .
122 S2: Aha! Maybe the behavior of the spring model of the spring. What does it mean,
has something to do with twist (moves in terms of the square model, to increase
hands as if twisting an object) forces as the diameter of the spring? ... Now
well as bend forces (moves hands as if making the sides longer certainly would
bending an object). That's a real make the [square) spring stretch more.
interesting idea .... That might be the 135 I: How can you tell?
key difference between this [bending rod) 136 S2: (a) Physical intuition ... and also
which involves no torsion forces, and this recollection . . . the longer the segment
[hexagon). Let me accentuate the torsion (moves hands apart) the more the
force by making a square where there's a bendability (moves hands as if bending a
right angle. (Draws Figure 12). I like rod) .... (b) Now the same thing would
that. A right angle . . . that unmixes the happen to the torsion I think, because if I
bend from the torsion. have a longer rod (moves hands apart),
123 S2: Now . . . I have two forces introducing a and I put a twist on it (moves hands as if
stretch. I have the force that bends twisting a rod), it seems to me-again
this ... segment [1) and in addition I physical intuition-that it will twist
have a torsion force which twists more ....
[segment 2) at vertex, urn, X ... [in 143 S2: . . . So . . . doubling the length of the
Figure 12) (makes motion like turning a sides ... it will clearly stretch more.
doorknob with one hand). Both for reasons of torsion and for
reasons of the segment [bending).
Section 4: Subject Evaluates and Adopts Square
Coil as a Preferred Model of the Spring Section 5: Subject Comments on His
Increased Understanding
129 S2: (b) ... Does this (points to square-
shaped coil) gain in slope-toward the 144 S2: And my confidence is now 99% ... I
bottom? ... now feel pretty good about my
130 S2: (c) ... Indeed we have a structure here understanding about the way a spring
which does not have this increasing slope works although I realize at the same time
as you get to the bottom ... (e) it's only I could be quite wrong. Still, there seems
if one looks at the fine structure; the rod to be something to this torsion business; I
between the Y and the X, that one sees feel a lot better about it.
the flop effect [downward curvature). 178 S2: Before this torsion insight, my confidence
132 S2: (b) ... Now I feel I have a good model in the answer was 95% but my
of sp- of a spring .... Now I realize the confidence in my understanding of the
reason the spring doesn't flop is because situation was way way down, zero. I felt
a lot of the springiness of the spring that I did not really understand what was
comes from torsion effects rather than happening; now my confidence in the
from bendy [sic) effects .... answer is near 100% and my confidence
133 S2: And now I think I can answer the stretch in my understanding is like 80%.
question firmly by using this . . . square
Analysis of Transcript shown in dotted boxes with well-understood situa-

tions shown in solid boxes.
Models Used by S2. A hypothesized outline of Figure 14A (Line 81): S2 has already reduced the
the cognitive events producing S2's new under- spring situation to the equivalent single circular coil
standing in this last section is shown in Figure 14. situation as shown by the solid line labeled (1) in the
The figure shows hypothesized "snapshots" of a diagram. Also there is a tentative analogy relation
series of S2's final models as they develop over shown as a dotted line labeled (2), from the single
time, with solid lines showing confirmed analogy coil to the well-understood bending rod model.
relations, and dotted lines showing tentative analo- Figure 14B (Line 117): S2 then recalls his idea of
gy relations. Poorly understood situations are a square spring and generates the model of a hex-
a :- £i.-:~: -6:-
: ~: +:
----------~? 1_ ------ --,--." I rod
bending
____ L ____
/""-
~ ~ ~
spring single
loop kinesthetic very short
intuition rod
O:. . -----~:O:
:"-----, (3) r-----,
(~;'~gJo
(3 ) ;q;~;~~'\r I
b r- S-~JiJJo:-- ___ ~~) bending
t:
n ____ _____ n ____ \ _____ \
I ~:
, _ _ _ _ _ .J
:
\ . ____ '"
rod
spring single ./ ~
loop kinesthetic very short
intuition rod
Figure 14. Changes in S2's understanding as he constructs a mental model using analogies.
. /~ h
kmesthetlc very sort
d intuition rod
spring square
kine;;he~rv short
i nlu ilion rod
Figure 14. (Coot.)
agonal coil. In his words, this is "constructing a tension between an existing paradigm in the
continuous bridge, or sort of a continuous bridge, face of anomalies in science.
between the two cases [the rod and the circular
coil]. "4
Figure 14C (Line 121): While analyzing the hex- Insight Behavior. The short transcript excerpts
agon in terms of bending effects, it occurs to him displayed here do not convey the fact that the sub-
("Aha!") that there will also be twisting effects. At ject spent a considerable period of time (about 25
this point, he shifts to the simpler square model. minutes) alternately questioning and trying to justi-
Figure 14D (Line 123): By the final stage, S2's fy the initial bending rod model of the spring. Fol-
understanding of the underlying structure that lowing this frustrating struggle, the invention of the
makes a spring work has changed significantly. He polygonal coil with the subsequent torsion discov-
now appears to have a mental model of the spring as ery is a candidate for being termed a significant
working like a square coil that contains elements scientific insight for several reasons. 5 First, the idea
that bend and twist. His physical intuitions about is productive in the sense that it leads immediately
the difficulty of (1) bending and (2) twisting a long to a considerable amount of cognitive activity. In
rod versus a short one seem to playa role similar to fact, one is given the impression of a "flood" of
axioms; they are basic assumptions on which the ideas occurring immediately afterward. Progress is
rest of his conclusions are founded. made rapidly, as if the polygonal coil idea were a
In the discussion that follows, I will refer to the "trigger" that stimulates a series of further ideas.
square, the hexagonal, and the many sided coil Second, the torsion idea appears fairly quickly,
models collectively as polygonal coil models. To with little warning. Third, the subject changes his
anticipate, some of the conclusions I wish to draw hypothesized model of stretching-by considering
from this example, in the remainder of this chapter, torsion the subject introduces a new causal factor
are the following: into the system. Torsion constitutes a very different
mechanism from bending for explaining how the
spring resists stretching. (S2 is the only subject out
1. The recognition of torsion in the polygonal of 10 studied who clearly progressed from no
coil is a significant scientific insight in S2' s awareness of torsion in the spring to an understand-
attempt to understand the spring. ing of it as a factor.) Fourth, the subject says that he
2. S2 uses analogies to invent a model for how is now able to resolve the paradox of the apparent
the spring works in the form of the polygonal lack of bending in a helical spring and states that he
coil. feels he has achieved an increase in his understand-
3. This model can be classified as an explanato- ing of the system. Of course, his "theory of
ry model as opposed to an expedient model springs" could be developed further beyond the
because it proposes torsion as a causal factor polygonal coil idea, but the fact remains that this
actually operating in the spring. model is a significant advance over the bending rod
4. S2 produces models and insights via a suc- models. 6 Fifth, the subject reacts emotionally to his
cessive refinement process of hypothesis gen-
eration, evaluation, and modification or re-
jection. S2's process is noninductive. SThere are actually two parts to this insight: the construction of
5. The model generation process here is neither the polygonal coil, and the recognition of torsion in the coil.
The first part makes possible the second part, and both are
a pure eureka phenomenon nor a simple, accompanied by "aha's." The first part constitutes the genera-
smooth, methodical buildup of information. tion of a new representation for the target problem; the second is
6. Several divergent processes are used in gener- the new activation of a principle that can be applied to the new
ating hypotheses. representation. In much of the discussion that follows, it will be
convenient to treat these together as a single insight.
7. The recognition of an anomaly sets up a ten-
6In fact, twisting is the predominant source of stretching in a
sion condition that "drives" the dialectic helical spring. The idea that the spring wire bends is also par-
process, and that is partially analogous to the tially correct. By imagining the extreme case of a single circular
coil of a spring stretched out into an almost straight wire, one
can see that stretching produces some unbending as it removes
4The idea of "bridging" between analogous cases with a new the circular curvature originally put into the wire when it was
intermediate analogous case is an interesting nonempirical coiled. However, there is no bending in a vertical plane. Twist-
strategy in itself for evaluating the validity of the analogy rela- ing in the square coil can also be used to predict that the stretch
tion between two cases and is discussed in Clement (1986). varies with the cube of the coil diameter.
ideas, calling them "interesting" and exposing a ent analogy for generating· the answer to the
"key difference," as well as producing some em- problem. (Karmiloff-Smith and Inhelder, 1975,
phatic "aha" expressions with a raised tone of have documented a related distinction in children's
voice. Later in this chapter, I will attempt to formu- thinking.)
late a more careful definition for the term insight Hesse (1967) and Harre (1972) identified two
that is motivated by these factors. types of scientific analogue model: (1) a model that
shares only its abstract form with the target (Hesse
cites hydraulic models of economic systems as one
The Formation of an Explanatory Model via
example); I call this an expedient model; and (2) a
Analogies
model that has become in Harre's terms a "candi-
Explanatory versus Nonexplanatory Models. date for reality," in which a set of material features,
As discussed earlier, philosophers of science have instead of only the abstract form, is also hypoth-
developed an important distinction between explan- esized to be the same in the model and in the target
atory models and either empirical law hypotheses or situations. I will refer to the latter type of model M
formal quantitative principles, as shown in Figure as an explanatory model (or structural hypothesis),
3. It will now be useful to specify a more precise Me' if some of the basic objects, attributes, and
definition for the term explanatory model in order to concrete relations in M are hypothesized by the sub-
say whether S2 has developed one. Recall the project to be part of T and to underlie the behavior of
posal to use the term model to refer to a cognitive interest in T.
structure M, where the subject believes there is a This ordinarily means that the subject can attain
predictive analogy between some important rela- some degree of ontological commitment to (belief
tional aspects of the model M and some aspects of in the reality of) Me if empirical and rational support
the target situation T. One kind of model then is are obtained for it. Me is thought of as a hidden
merely an expedient and often temporary analogy structure within T that provides an explanation for
that predicts some aspects of the target's behavior. T's behavior. Usually Me contains some entities
M may happen to behave like T, and therefore pro- that are initially not directly observable or obvious
vide a way of predicting what T will do. Such an in T at that point in time.
expedient model may not provide a satisfying ex- This concept helps to account for the remarkable
planation for why T behaves as it does. M may say ability of scientists to formulate and propose hidden
nothing about the underlying process that explains structure and processes in nature before they are
T's behavior. An explanatory model, on the other observed more directly, such as atoms, black holes,
hand, should explain how T works, leading to a and the "bending" of light rays. An explanatory
feeling of "understanding" T. model can allow the scientist to see a phenomenon
S2 makes a clear distinction between confidence in a new way via an analogy to a hypothesized
in his answer to the problem and confidence in his visualizable structure that is considered to be hid-
understanding of the spring: den in the target situation to be explained. This is
something that empirical law hypotheses cannot do.
144 S2: ... There seems to be something to this In the case of the present protocol, the polygonal
torsion business; I feel a lot better about coil qualifies as explanatory, because the subject
it. ... believes that twisting and bending effects may actu-
178 S2: Before this torsion insight, my confidence ally be operating in the spring wire to produce its
in the answer was 95% but my behavior. Twisting and bending are concrete fea-
confidence in my understanding of the tures, but they are not ordinarily observed in
situation was way way down, zero. I felt springs. In this sense, the model expresses for the
that I did not really understand what was subject a hypothesis concerning the hidden struc-
happening; now my confidence in the ture underlying the way stretching produces defor-
answer is near 100% and my confidence mation and restoring forces in the spring wire. Fur-
in my understanding is like 80%. thermore, the square coil model removes the
anomaly of a potentially critical dissimilarity in the
original bending rod model-that of the lack of
This perceived increase in understanding is one in- cumulative bending in the spring. All these factors
dication that the polygonal coil has become an ex- presumably increase S2' s feeling of understanding
planatory model for the subject, not just an expedi- and of having a satisfying explanation for the be-
More similar
to target
, .... ---, ,
\, , I
... ---'
a b c
Simpler
Figure 15. Sequence of models generated by S2.
havior of a spring, as expressed in lines 144 and 178 elements could be something like the hidden ele-
of the protocol quoted above. ments in the target or not. Its explanatory status may
For these reasons, the polygonal coil with torsion grow gradually rather than in one decisive jump.
model qualifies as an explanatory model that pro- Improvements in the model may also raise its ex-
vides a hypothesis about the nature of springs. S2's planatory status. Indeed, this seems to be what oc-
statements lead one to believe it has become a pre- curred in S2's case. He used the bending rod early
ferred model of the spring that he will retain in on as a model, which gve him a prediction in which
memory. In this sense, S2's protocol is an example he was highly confident. However, he said his re-
of learning via the construction of an explanatory sulting understanding was very low. The recogni-
model. 7 tion of the lack of cumulative bending anomaly
appeared to prevent him from accepting it as an
Development of an Explanatory Model from an explanatory model. Cumulative bending is an
Initial Nonexplanatory Analogy. A further hy- important material property that is present in the
model but not in the target. A successive refinement
pothesis is suggested by S2's problem solution: an
process then led to a number of alternative models,
expert can develop an explanatory model via the
culminating with the polygonal coil model. The
modification and refinement of an initial model that
identification of torsion in the polygonal coil raised
is merely expedient or has low explanatory status.
S2's feeling of understanding significantly. This is
In this view, whether a model is explanatory is a
consistent with the interpretation that he had then
matter of degree. The explanatory status of a model
acquired some confidence that torsion is a real but
depends on the degree to which one believes that the
hidden mechanism operating in the spring. Thus,
model contains elements that are like elements hid-
S2 appears to take an initial, nonexplanatory analo-
den in the target to be explained.
gy (the bending rod model) and develop it, via crit-
It is reasonable that when an analogous case is
icisms and modifications, into a model that in fact
first proposed, it will often be unclear whether it
does have explanatory status for him.
has potential as an explanatory model-whether its
Simplifying Function of Models. In Figure 15,
7In one sense, I am appropriating the term explanatory here S2 considers a multisided coil but is unable to make
because, as Kuhn (1977) points out, what counts as explanatory further progress in his analysis before quitting. The
is different for Aristotle, Newton, and quantum physics. I am figure shows the set of polygonal models referred to
proposing that what counts for S2 in this problem fits the defini-
tion given-an analogue model that has material elements that by S2, placed in order of increasing simplicity or
are hypothesized as "candidates for reality." The sharing of analyzability from left to right. Note that these
material elements between model and target can be termed models attain a higher degree of perceptual resem-
material correspondence, and this assumption seems to be a blance to the spring in the opposite direction from
minimal requirement for something to have potential as an ex-
planation. Whether a satisfying explanation is actually attained,
right to left. Of the models shown, the bending and
however, will also depend on other factors, such as the support twisting rod models on the right are the simplest to
for and comprehensibility of the model. understand, but appear to be least like the spring
coil. One might be tempted to call the multigon in vations or new explanatory features to attend to; (2)
(b) the only "really" explanatory model in the an analogous case can serve as a rough initial model
sense that it is seen as actually present in the spring, of the target situation that is later developed and
whereas the others are not. But even in the multi- refined; (3) a developed explanatory model, what-
gon, there are material elements that are not present ever its origin, should in the end be linked by an
in the spring, such as fulcrum points and straight- analogy relation to the target situation.
line segments. Apparently, even the multigon
model is not a full candidate for the mechanism in
the spring. Summary of Evidence for a Model Construction
Hesse (1967) and Harre (1972) described some Cycle as a Noninductive Source for Hypotheses
models in science as simplifying models, in which The growth in S2' s ideas appears to have oc-
the scientist intentionally uses a model with features curred via the cyclical process of analogy genera-
that are different from those in T in order to make M tion, criticism, and modification (or rejection) that
simple enough to analyze. S2's polygonal spring is shown in Figure 16. This is a more general rea-
models appear to be simplifying models that are soning pattern that can help account for the transi-
partially explanatory; he sees the spring as probably tions between the states shown in Figure 14. Table 1
really twisting, as in the square coil, but not as summarizes evidence from the protocol that S2's
really square. The square provides a simplifying progress is a result of this kind of cyclical process
geometry-but S2 recognizes that ordinary spring rather than a result of either a convergent series of
coils are not square or pol ygonal. In summary, this deductions or an induction from observations. Fig-
appears to be a case in which modifications of an ure 16 is therefore a model of the processes that
initial analogy with low explanatory status led to the produce the observed behaviors shown in Table 1.
development of a model with considerably higher Here I assume that the bending rod and zigzag
explanatory status. However, the polygonal coil spring models are simplifying models, that the ex-
model is still a simplifying model, because some of tent to which they are explanatory is unknown to S2
its elements are recognized as not being present in at the time he proposes them, and that they are part
the helical spring. 8 of his attempts to develop an explanatory model.
Note that the cycle in Figure 16 corresponds to
Three Roles for Analogy. Even the most suc- the nonempirical processes B, C, D, and E in Figure
cessful models with no recognizable simplifying 4, the model construction cycle discussed earlier.
assumptions can be questioned as to their on- Process B is also implicated in the rapid search for
tological status-whether they are "really true" of analogies, such as "molecules, polyesters, and car
T. It is reasonable to take the point of view that a springs [leaf springs l" in line 57 of the transcript.
model can never be fully confirmed as true in a Thus, there is evidence in this case study that sup-
universal sense and should always be open to ques- ports the existence of the non empirical processes
tion. Another way to say this is that even in well- proposed in the model of hypothesis development
established scientific models, the relation between shown in Figure 4. We therefore appear to be in a
the idealized model and a real-life example is one of position in which real-time protocol evidence can
analogy, or partial resemblance, leaving open the be gathered to evaluate the plausibility of such mod-
possibility that other more refined or useful alter- els of scientific reasoning.
native useful models may be developed in the
future. Noninductive Hypothesis Generation. I will
This means that analogy can play a role in the now examine more carefully the claim that S2's
generation of new hypotheses in at least the follow- final model is neither the result of a convergent
ing different ways: (1) an analogous case can playa series of deductions nor an induction from observa-
provocative, heuristic role in suggesting new obser- tions. When S2 generates analogue model hypoth-
eses, they appear not to be deduced logically from
8M any sequences of mathematical models, especially in applica- prior principles; they are essentially reasoned con-
tions of the calculus, have the form shown in Figure 15. In this jectures as to what might be a fruitful representation
view, mathematical limit arguments, which examine properties for analyzing how a spring coil works. The reason-
as one passes from an analyzable simpler model and approaches ing involved does not carry the certainty associated
the limit of the target situation, are sophisticated attempts to
justify the intuitive validity of the analogy between the model
with deduction.
and the target situation. The role of analogies and models in Nor, apparently are the hypotheses built up in-
mathematical understanding is discussed by Fischbein (1987). ductively as abstract generalizations from observa-
Activate possible analogies

and related model elements
. Construct initial
model
-I
Reject or modify
model
• r
Rational ( non - empirical)
evaluation and criticism ~
~
Fail Pass
-I
I 1
Figure 16. Elements of model construction cycle which are supported by evidence in S2' s protocol (see also Table 1).
Table 1. Location of Evidence for Model Construction Cycle of Hypothesis Generation,

Criticism, and Modification or Rejection
Line Process a Hypothetical model Comments
5 G Horizontal bending rod Initial analogy

5 C Horizontal bending rod Bending in rod, but not in helix
23 G Square coil
Modifies square to produce zigzag model
23 M Zigzag no. 1
23 C Zigzag no. 1 Joints confounding
23 M Zigzag no. 2 with stiff joints Modifies zigzag no. 1 to produce no. 2
[Cl b Bending in zigzag, but not in helix
D Drops zigzag models
57 R Rod model
87 C Rod model Bending in rod, but not in helix
117 R Square coil
119 M Hexagonal coil
121 Makes torsion discovery in hexagon
122 C Hexagonal coil Hexagon geometry too complex
122 R Square coil Leads to successful prediction of restor-
ing forces without cumulative bending
in spring wire
aKey: G = Generates hypothesized model; C = Criticizes model; M = Modifies model; R = Reconsiders model; D = Drops or
rejects model.
blnferred in absence of direct evidence in protocol.
tions. S2 is unable to collect new data during the empirical law hypothesis that (other factors being
interview, and, consequently, his reasoning is inde- equal) wide springs will stretch more than narrow
pendent of new empirical processes. One can also springs. Apparently, this is a case in which the
consider whether he might be making new induc- development of a convincing explanatory model
tions on perceptual memories of prior observations, hypothesis can establish high confidence in an em-
but he does not appear to recall observing bending, pirical law hypothesis in the absence of new em-
twisting, zigzags, or squares in springs; instead pirical information. Kuhn (1962) discusses exam-
these appear to be newly imagined models. The ples of this pattern in the history of science.
novelty and nonobservability of the polygonal coil
with torsion model, and its evolution from crit- Argument for Not Separating the Context of Dis-
icisms of the earlier horizontal rod model, argue covery and thl! Context ofJustification. Finally, I
that the hypothesis generation process in this case want to consider a potential criticism of the model
was an imaginative construction and criticism pro- construction cycle shown in Figure 4. It is tradi-
cess rather than one of induction from observations. tional in philosophy to separate the contexts of hy-
Quite possibly, S2 would have made some new pothesis formation and hypothesis testing and eval-
observations of springs as well, had they been avail- uation in science. If I claim to be portraying
able (although it is doubtful that he would have hypothesis/ormation process in Figure 4, then why
observed torsion effects). But the present case study does it include hypothesis evaluation processes as
demonstrates the possibility that impressive pro- well? The answer concerns the observation that
gress in explanatory model construction can be loops in the cycle can at times be traversed ex-
made via noninductive processes. tremely rapidly. For example, S2's criticism of the
Of course, it is highly likely that empirical infor- bending rod model in line 5 of Protocol indicates
mation was involved in the original development of that the time interval between model generation and
the prior knowledge S2 uses. In attempting to speak criticism can be as small as 15 seconds. In addition,
to the rationalism versus inductivism issue, it is his modification of the zigzag spring model in line
important to identify the time period of focus. For 23 indicates that an entire generation, criticism, and
the purposes of this analysis, the focus is on the new modification cycle can take place within 90 sec-
knowledge developed during the hour or so of rea- onds. Although an evaluation in the form of a care-
soning in the interview rather than on the origins of fully designed laboratory experiment can take days
the prior knowledge S2 uses. For example, he uses or even years, other evaluation processes, such as
prior knowledge in the form of the concept of twist- certain nonempirical checks for consistency, can
ing. One assumes his earlier learning of the concept take place much more rapidly. In this sense, evalua-
of twisting involved empirical experiences with tion is an inherent part of the hypothesis formation
wrenches, cranks, and knobs. His new model of the process. Stated in traditional terms, it may often be
polygonal spring with torsion uses his old concept impossible to separate the "context of justifica-
of twisting as one of its elements, but the total struction" from the" context of discovery. " The history
ture of the model is a larger new construction. The of science tends to look at developments over a time
point is that the new knowledge developed by S2- scale of years or weeks. From this perspective it
the construction of a new explanatory model hy- may be impossible to separate these two contexts in
pothesis for how a spring works-was apparently the early stages of hypothesis formation when the
formed by processes during the protocol that were grain size of the time scale that one is using is
nonempirical. greater than fractions of an hour.
I do not wish to say here that some form of sug- In addition, generative techniques appear to be
gestion from patterns perceived in data cannot be used in the service of evaluative goals in this pro-
involved in some types of scientific hypothesis for- tocol. The initial generation of the zigzag and
mation. Rather, this case study acts more like an square spring ideas, for example, appear to be at-
"existence proof" in showing the possibility that tempts to evaluate the appropriateness of the bend-
noninductive construction processes can be very ing rod model. Subsequently, these become candi-
important in the formation of explanatory model dates for replacing the rod model.
hypotheses. In sum, the reason that evaluation processes ap-
pear in the model of hypothesis formation is that
An Explanatory Hypothesis Can Lead to an Em- they appear to be an inherent part of hypothesis
pirical Hypothesis. The final model of the poly- formation down to time periods, on occasion, of
gonal coil with torsion raises S2's confidence in the less than a single minute. In such cases, one sees a
rapid, dialectic interplay between generation and also prove useful for analyzing structural change in
evaluation processes. theoretical models in real scientific research.
Is There a Sudden Reorganizing Change in S2's

Eureka or Accretion? The Presence of Understanding?
Insight in S2's Protocol This subquestion itself can be broken down into
two parts: Is there a significant structural change?
I can now move to the second issue outlined in and Is it a sudden change? For the latter part, a
the introduction-the pace of theory change: Does pertinent time period must be identified over which
S2's reasoning contain eureka events that involve the change takes place, and a pertinent concept of
sudden reorganizations, or does he make progress "rate of hypothesis formation or modification"
smoothly in an incremental manner? The answer to must be defined. I will conclude that, although the
this question is not obvious. It seems possible to torsion discovery was not a "blinding insight"-
argue in either direction from this protocol. One can an instantaneous reorganization of his ideas-it
point to what appear to be sudden insights, but, on certainly was an impressive and relatively sudden
the other hand, sections precede these insights in breakthrough. The problem is to develop a rela-
which the subject prepares the context and ground- tively precise language for saying this.
work for having them. Sometimes his methods ap-
pear to be systematic, but, at other times, ideas Is There a Significant Structural Change? One
arrive in a rush, as if they are outside of his control. first needs to ask about the size of the change in
Thus, there seems to be mixed signals in the pro- representation or understanding produced by the
tocol on this issue. torsion insight. Does it simply add on a small new
fact or is it a complete reorganization? The type of
Defining a Pure Eureka Event change in understanding to be discussed here is a
structural change (change in relational structure as
In order to say something useful about the eureka opposed to surface features) in a currently assumed
question, one needs to become more precise about mental model.
the meaning of a eureka event. Here I will propose It is clear that the polygonal coil with torsion
an initial definition of the extreme case of a pure insight does not constitute a reorganization in S2' s
eureka event as an extremely sudden, reorganizing, understanding of any domain larger than the' 'theo-
extraordinary break away from the subject's pre- ry of springs" (such as the "theory of elastic mate-
vious ideas. I use the term extraordinary here to rials"). However, the insight does appear to add
refer to processes, such as unconscious or supernor- more than a simple fact; it appears to constitute the
mal reasoning, that are different from those used in addition of a significant set of structural relations to
ordinary thinking. If the appearance of a new hy- the subject's hypothesized model of the spring sys-
pothesis constitutes a break in the train of thought- tem, including the new causal chain of weight caus-
if the hypothesis comes "out of the blue" and ap- ing twisting and torsion, which, in tum, causes re-
pears unconnected to the subjects' previous ideas in sistance to stretching; and the new global effect of
the protocol-this would constitute one kind of evi- finding no cumulative effect of bending throughout
dence for an extraordinary and probably uncon- the square spring.
scious thought process. The accretion versus eureka Can the insight be characterized as a reorganiza-
question in extreme form then becomes: Is the sub- tion of the subject's mental model? In some senses
ject's accomplishment the result of a smooth, incre- it can, although the shift could have been larger.
mental, controlled, buildup from previous ideas? Torsion is a completely different geometric defor-
Or is it a sudden, reorganizing, extraordinary break mation than bending and constitutes a significantly
with his previous ideas? I will consider two sub- different hypothesis. The case here would be clear-
issues of this question expressed by the two pairs of er, though, if the subject had switched more com-
key words in it: sudden reorganizing and extraordi- pletely from the view of spring forces coming from
nary break. In this section, I would like to use the bending to the engineer's view that spring forces
analysis of the protocol as an initial test bed for come primarily from torsion. S2 did not go this far;
concepts that are developed to describe the quality instead he switched from using bending alone to
and pace of structural change in creative hypothesis using bending and torsion together in his explana-
generation activities. Some of these concepts may tions. But he did raise the question of which of these
two effects predominates. Although he was unable of which led nowhere. Can one identify a shorter
to reach an answer to this last question, when asked period of insight within the protocol? The bending
at the end of the interview, about whether the rod model was proposed within 1 minute after read-
stretch could be due completely to torsion, he felt ing the problem. Then a long period without lasting
that it was perfectly plausible. Given more time, the progress in model development of about 40 minutes
fuller transition might very well have taken place. ensued as the rod model was questioned, the zigzag
Clearly, the potential for a complete replacement of models were proposed and rejected, and other anal-
the deformation mechanism in the spring was ogies were tried. Finally, there was a breakthrough
created. in a 4-minute period during which the subject re-
What one can say then is that the subject achieved ferred to the square hexagonal coils, made the tor-
a major breakthrough in adding a major chain of sion discovery, and incorporated it into his final
casual factors to his model of the spring. This can be square coil model of the spring. When the subject
considered to be a reorganization in the sense that a finally generated the hexagonal coil toward the end
new system of relationships was created. Thus, the of the protocol, ittook less than 80 seconds for him
structural change in this subject's model of the to recognize the torsion effect, and less than another
spring appears to be of intermediate size. The 2 minutes to settle on the square coil as his final
change process was characterized by imaginative model of the spring. This 4-minute period is there-
attempts to switch to different problem representa- fore a candidate for the period of insight.
tions, most of which failed. When a productive However, the square coil idea was considered
representation was found (the polygonal coil), it led very briefly, only about 6 minutes into the protocol.
to the recognition of a system of new relationships But it was quickly dropped in order to consider the
involving force, torsion, and twisting. But it was zigzag spring. Thirty-four minutes later, it was
not a "revolutionary" change in the sense of reject- taken up again and led to the torsion insight. Should
ing and replacing a large, previously assumed body this 34 minutes between the dropping and re-
of established knowledge structures. However, it emergence of an idea be counted as part of the
does allow us to imagine the possibility that such a period of insight? I will assume not, because the
rejection and replacement could occur in science subject was following separate ideas during this
via a similar process. time which turned out to be blind alleys. If one
makes this assumption, one can point to this 4-
The Pace of Change in Understanding. I have minute segment as a relatively sudden "period of
taken a high rate of change in the currently assumed insight. " But the difficulties involved in defining
model as one defining characteristic of a pure eu- the period of insight here are clear. The benefit of
reka event. This rate of change could be defined as this exercise, however, is that it forces one to devel-
the ratio of the size of the change in the model's op some useful distinctions between concepts, such
structure to the time interval over which the change as structural change in a model, the period of in-
takes place. The latter concept may not be easy to sight, and the rate of structural change in a model.
operationalize as an observable variable, depending On the other hand, the insight was not in-
on the comparisons being made and the complexity stantaneous, and criticism and modification pro-
of the protocol, but it should at least playa clarify- cesses did occur during this 4-minute period as
ing role at the theoretical level. shown in Figure 16 and Table 1. This means that
It is a challenging task to point to a specific time from a microscopic perspective, which looks at the
interval in the protocol representing the' 'period of fine grain in the data, the insight appears to be
insight" because of the difficulty in defining the "unpackable" into potentially understandable sub-
latter. As an upper limit, the time for the subject's processes. This leads me to describe it as "fairly
total solution was 52 minutes. Thus, it is certain that sudden," rather than as an extremely sudden' 'bolt
the subject changed from the rod model of the from the blue. " This is the first sense in which the
spring to the square coil with torsion and bending in insight fails to qualify as a "pure eureka event."
a period smaller than this time. Viewed on a large In summary, there appears to be periods in the
time scale appropriate to the history of science, this protocol in which progress is made slowly or not at
would certainly be considered a tiny interval that all and others in which progress is quite rapid.
indicates a relatively sudden structural change. Those periods during which little progress is made
But much of this time was spent testing the sim- are frustrating to the subject but they, in fact, may
pler rod model and trying out other analogies, most provide necessary preparation for the later insight.
The pace of structural change is uneven rather than train of thought. Clearly, S2 breaks away from his
consistent, and progress comes intermittently. initial model of the problem. The torsion insight
When it does come, it is in the form of a relatively represents a real break (in the sense of "break-
sudden breakthrough that involves a significant through") with his previous bending rod model for
structural change in the subjects' hypothesized understanding the problem.
model. On the other hand, S2's work does not contain an
obvious break in the train of thought. It does seem
possible to construct a believable psychological ac-
Does the Subject Use Extraordinary Reasoning count of his thought process as a series of connected
Processes? conscious ideas. The growing series may actually
The second major subquestion to the main ques- look more like a branching tree or network than a
tion of whether there is a pure eureka event in the single chain, and there may be jumps of attention
protocol is whether S2 used extraordinary thought from the end of one branch to the end of another, but
processes during his breakthrough. If the processes the essential point is that a new idea does not appear
are found not to be extraordinary, one can go on to from nowhere; it is always plausible that it was an
ask the opposite question of whether the subject's outgrowth of S2's previous conscious ideas.
thinking is highly controlled in the sense that he Two major parts of S2's insight in the solution
always pursues a series of well-defined, conscious are the generation of the square coil analogy and the
plans and procedures. I will conclude that the tor- discovery of torsion. A plausible explanation for
sion-polygon insight was neither due to an unex- the torsion discovery can be given as follows. As S2
plainable, extraordinary process, nor due solely to a was examining adjacent sides in the newly con-
planned, methodological procedure. Rather, it was structed hexagonal coil model, an existing mental
the result of a dialectic process of conjecture, crit- schema for dealing with twisting situations was ac-
icism, and rejection or modification, involving rel- tivated. Such a recognition process is a common
atively uncontrolled divergent associations and event in everyday problem solving and should not
playful transformations on the one hand, as well as be considered extraordinary. It does happen to be a
relatively controlled strategies for mounting attacks key event in the solution to this problem. S2 was not
on the problem on the other. But these are all ordi- certain about this conjectured recognition at first
nary reasoning processes. S2's association and rec- and needed to examine it critically, which led him
ognition processes in particular can be viewed as to consider a square coil as an easier case.
divergent and creative, but these processes are nei- In the case of the original square coil analogy, it
ther conscious plans nor extraordinary. was generated while S2 was thinking about whether
there was a difference between a bending rod and a
Extraordinary Thinking. By extraordinary single spring coil:
thinking, I mean the use of special processes that
are outside of the set of normal reasoning processes 23 S2: Why should the coil have anything to do
used in everyday learning and problem solving. with-? It's just so arbitrary. Why does it
From a psychological point of view, this means I have to be a [circular coil]? Surely you
cannot imagine a plausible explanation for a partic- could coil a spring in squares, let's say,
ular thought process that is based on an ordinary and it . . . would still behave more or less
sequence of inferences, associations, guesses, esti- the same.
mates and criticisms. Two ways extraordinary
thinking could occur during a problem solution are:
if the subject performs some supernormal feat· of This is a highly creative idea but not one that neces-
synthesis without preparation; or, more generally, sarily involves extraordinary reasoning. Here S2
if there is a break in the train of thought, a jump to a appears to be imagining ways to bend a piece of
new train of thought that has no apparent connec- wire into a spring. The plausible ordinary process is
tion to any previous thought. This last kind of event one of imagining a simple transformation one could
might be evidence for unconscious processing. perform with one's hands.
The worth of this idea was not recognized imme-
Two Types o/Breaks. It is important, however, diately. Only after thinking hard about and confirm-
to distinguish between a break away from the sub- ing the lack-of-bending anomaly in the spring does
ject's currently assumed model and a break in the S2 return to the square coil idea in line 117 and use it
productively. Here there is a branch in the train of 1. The breakthrough is an important idea.
thought, but the return to the square coil idea can be (a) It is a key idea-an important component
seen as connected to its earlier appearance. of a solution.
In some cases, the connection to an idea in mem- (b) It overcomes a barrier that blocked pro-
ory may be a weak one-a loose association or gress; it comes after a frustrating series of
conjectured recognition or playful transformation false leads and blind alleys, after a period
rather than a deductive inference or a precise sub- in which little progress has taken place; it
question. Associations, transformations, and rec- resolves an anomaly.
ognitions in this light are divergent, unpredictable, 2. The breakthrough adds significantly to the
and sometimes highly creative processes, but not subject's knowledge. It produces a large
extraordinary ones in the sense of being uncon- structural change in the subject's model in
nected to the network of current representations. I which he
consider S2's overall achievement-the marshal- (a) Identifies new variables or causal factors
ing and orchestration of a large number of reasoning in the system
processes to produce the invention of a new explan- (b) Identifies a new hypothesized mechanism
atory model-to be extraordinary in the sense of in the form of an explanatory model; and
being unusually productive and creative. However, (c) States that it increased his understanding.
I can see no evidence that the reasoning processes 3. The subject's ideas are generated fairly quick-
he uses, taken individually, are extraordinary. The ly during the breakthrough, and he achieves
train of thoughts S2 reports weaves a "coherent rapid subsequent progress toward a solution.
story" in the sense that each new idea appears con- 4. The breakthrough is accompanied by more
nected to previous ideas and is therefore at least complex phenomena.
weakly constrained by previous ideas. (a) It is accompanied by indicators of emo-
S2' s ideas are also connected by the specific relational response-surprise, joy, and satis-
tionships implied in Figure 4, in which new ideas faction.
can grow out of modifications of or reactions to past (b) The subject realizes immediately that
ideas. This is an even more specific sense in which something important has been discovered
his insight did not emerge from out of the blue, and in the torsion idea.
it will be discussed below in the section on creative
processes. The following features are senses in which S2's
It should be noted that Tweney (1985) cited evi- breakthrough was not a pure eureka event.
dence to discredit the idea that Faraday's discovery
of induction was a "bolt from the blue," as some 1. The breakthrough idea was not generated ex-
have thought; and Perkins (1981) came to the con- tremely suddenly without preparation.
clusion, after reviewing the literature on insight in 2. The breakthrough did not involve the total
creative thinking, that there is no convincing body replacement of one hypothesized model with
of evidence that insights occur via special or ex- another.
traordinary processes. This does not eliminate the 3. The breakthrough is explainable via ordinary
possibility that such special processes might exist, reasoning processes; there is little evidence
but it does indicate that it is difficult to find convinc- that it was
ing evidence for them. (a) An extraordinary thought process,
(b) An unconscious process, or
(c) A break with all previous trains of
Defining Insight
thought.
I have discussed some senses in which S2's pro-
tocol does not provide evidence for a pure eureka One can now use the criteria developed in the
event. In this section, I will propose some criteria above list to define three categories of insight be-
for a less extreme kind of event, which I will term a havior. These definitions are, of course, to some
scientific insight. In order to sort out the different extent arbitrary; the goal is to try to define some
senses in which S2's solution is and is not an exam- useful categories that will help to make finer dis-
ple of insight behavior, it will be useful to refer to tinctions that can aid in analysis. The categories
the following list of the features of his polygon with (designed to refer to hypothesis development ac-
torsion breakthrough that are insight-like. tivities) are breakthrough, scientific insight, and
pure eureka event. They are defined in increasing not the expenditure of energy but the construction
order of specificity and unusualness so that the of new knowledge. With respect to the formation of
breakthrough category includes scientific insight, an explanatory model, progress did not take place
and the scientific insight category includes pure eu- as a smooth, incremental evolution of new knowl-
reka events. edge. Progress appears to be blocked when the sub-
A breakthrough is a process that produces a key ject is "locked into" his current conceptualization
idea-an important component of a solution-and of the problem for long and sometimes frustrating
that overcomes a barrier that can block progress periods. Most of the approaches he tries during this
toward a solution. period must be thrown away; they are not used later
A scientific insight is a breakthrough occurring as pieces of the final model. One analogy generated
over a reasonably short period of time that leads to a by the subject then led to a fairly sudden insight,
significant structural improvement in one's model which led to the formation of a new hypothesized
of a phenomenon; that is, it constitutes a shift from model. Thus, insight processes were found that are
the subject's previous way of representing the phe- not accretionist in character and that support a view
nomenon and leads to an increase in understanding of scientists as being capable of significant re-
of the phenomenon, as determined by the evalua- organizations in a relatively short period of time.
tion process in Figure 4. This is the descriptor that On the other hand, the major case against a pure
appears most appropriate for S2's breakthrough. eureka event is that these processes do not appear to
A pure eureka event is a scientific insight in be supernormal or unconscious ones. It was con-
which (1) there is an extremely fast emergence of a cluded that S2' s breakthrough can be considered a
new idea with little evidence of preparation; (2) the relatively sudden and structure-changing event that
new idea is a whole structure replacing the subject's includes relatively divergent and creative pro-
previous model or understanding of a situation; (3) cesses, but that there is not evidence for extraordi-
the process is not explainable via ordinary reason- nary processes. The upshot of the present analysis,
ing processes; extraordinary thought processes or then, is that rather than being an example of an
unconscious thought processes are involved. accretion or eureka process, the pace of progress is
This definition recasts the earlier initial defini- uneven, with' 'more revolutionary" and' 'less rev-
tion of a pure eureka event (an extremely sudden, olutionary" periods of work. S2's breakthrough
reorganizing, extraordinary break from the sub- can be characterized in the above terms as a scien-
jects' previous ideas) in a way that relates it to other tific insight but not as a pure eureka event.
types of insight behavior. For some purposes, re-
ducing everything to these three categories may be
less important than having something like the above
list of features for describing different ways in Creative Mental Processes
which an idea can be insightful. But the three terms
may provide a useful shorthand for some purposes. The various processes in the model construction
cycle can be divided into two main categories: the
productive processes of generation and modifica-
Summary
tion and the evaluative processes of empirical test-
This section has attempted to answer the ques- ing and rational evaluation. In this section, I exam-
tion, Was the polygonal coil with torsion break- ine questions about these individual processes and
through more like a sudden eureka event or an how they interact. Evaluative processes will be dis-
example of steady accretion? The case against ac- cussed first with respect to the role of anomalies,
cretion is the following: When one examines the leading to the view that a tension condition indi-
thinking-aloud case study microscopically over cated in the protocol is partially analogous to the
tens of minutes on a small time scale, one sees an motivating tension between an anomaly and a per-
arduous dialectic process of conjecture, evaluation, sistent paradigm in science. In a second section, I
and rejection or modification of hypotheses that discuss the role of transformation and invention in
precedes the breakthrough, as opposed to an event analogue hypothesis generation, processes that
that takes place instantaneously and effortlessly. create the possibility of provoking the recognition
Thus, in terms of effort alone, there is certainly a of a new principle in a novel construction. In a final
long and steady expenditure of energy on the part of section, I discuss the role of divergence and con-
S2. However, the issue of central concern here is straint in productive processes, leading to the view
that these processes are less constrained and con- Dialectic Tension. There is a palpable tension
vergent than-established procedures, but more con- obvious in the first section of the videotape that is
strained and "intelligent" than a blind selection conveyed only to a limited extent by the transcript: a
and variation process. frustration with not being able to resolve the anoma-
ly of the lack of bending in a helical spring. For
example in lines 87 and 111, S2 says:
Anomalies and Persistence in Protocols and
Paradigms 87 S2: ... if you start with a [stretched] helix
and unwind it . . . you should get a bow
In this section, I attempt to provide a deeper level [bend], but you don't. I mean visually
of explanation for the phenomenon of extended pe- imagining it, you don't. I don't see how
riods of little progress between insights in the pro- you could make the bow go away-just
tocol in terms of the dialectic view of model con- to wind it up-Damn it!
struction as a cyclical process of generation, 111 S2: Darn it, darn it, darn it ... why should
evaluation, and modification. Table 1 outlines evi- that [the difference between a rod and a
dence in the protocol for the presence of such a coil] matter?
dialectic process. One of the more subjective obser-
vations one can make of S2' s overall behavior in the
tape is to point to the impressive amount of stren- The tension apparently occurs between the rod
uous activity that he poured into this process. Even model and the lack-of-bending anomaly. This ten-
for those who admit that analogies can playa role in sion or disequilibrium condition appears to provide
scientific discovery, a common view is that a sub- a driving force that keeps S2 actively attacking the
ject may be passively reminded of an analogous problem even though he claims he is already 90%
situation C, and be able to transfer a prediction from sure of his answer. It bothers him enough to drive
C back to the problem. The image is of the insight him to search for a way to modify the rod model or
"coming to the subject" as a passive receiver. In replace it. This search takes up the better part of the
the present case, S2 is much more active and ag- 52-minute interview, and it is peppered with ex-
gressive: inventing tentative analogies, rejecting a pressions of frustration. Line 178 provides evi-
number of them, pursuing those that have promise dence that the reason for his dissatisfaction has to do
by criticizing them aggressively, and modifying with an important difference between having a con-
them in a series of thought experiments until he is fident prediction and having a feeling of under-
satisfied he has a valid model. A more apt informal standing. He speaks of having low confidence in his
image here than the passive receiver is a construc- understanding because the rod model predicts a
tivist one of the subject "aggressively constructing property that he feels should not occur in real
and testing different models in an effort to capture springs, even though he has high confidence in his
an understanding of the phenomenon. " predicted answer. I take this as an interesting exam-
What drives all this strenuous activity? In partic- ple of a situation in which good performance is not
ular, why does S2 persist in criticizing his under- equivalent to deep understanding, and, because of
standing when he is already 90% sure that the wider the subsequent events that raise his confidence, I
spring will stretch more? What drives the hypoth- take the important difference to be the lack of a
esis formulation process and keeps it working in the satisfying explanatory model.
face of little progress? Why is there a period of very
little progress followed by a period of insight in this Persistence of the Initial Model. Line 87 of the
protocol? For the last question, one could simply transcript is indicative of the fact that S2 finds it
say that there are a large number of possible paths to very difficult to give up the bending rod model. The
consider and that it is just a matter of luck that persistence of this model appears to be an example
determines when one will find a successful path. of an Einstellung effect; a problem space dominates
But there may be a deeper reason connected in at his thinking and prevents him from generating nec-
least one way with Kuhn's (1962) idea of intermit- essary new ideas. In order to make progress, S2
tent progress in science (periods of normal science must redescribe the problem using new descriptors;
and revolution.) In this section, I attempt to speak to he needs a new problem representation. But the rod
these questions in terms of conflict between a per- model keeps reappearing in the transcript. Even
sistent model and a perceived anomaly. though he proposes rejecting the model several
times, he is repeatedly tempted to return to it. It is as ative activity can be seen largely as a response to the
if the idea has an autonomous "life of its own." tension urging him to find a more satisfying model.
Thus, this example suggests that the tension be-
A Powerful Anomaly. Pitted against this per- tween a previously established model and a promi-
sistent model is a powerful anomaly. Bending in the nent anomaly can be a major driving force behind
vertical plane is central to the rod model, but S2 hypothesis generation.
cannot imagine a way for bending to take place in a Here it appears to require something as divergent
helical spring. (Here I am using the term anomaly in as analogy generation to break out of the Einstel-
the broad sense of a new finding that conflicts with lung effect that is formed by a persistent inadequate
previous ideas, whereas in some narrower usages, model. This provides an important connection be-
its referent is limited to a new nonconforming ob- tween the two previous major sections of this chap-
servation.) In summary, the symptoms of tension ter on model construction via analogy and the pres-
observed in the subject appear to be the result of a ence of insight in S2's protocol. The process of
conflict between a persistent initial model and a analogy generation, motivated by the recognition of
powerful anomaly. an anomaly, appears at times to be powerful enough
to break away from a persistent but inadequate
Analogy to the Persistence ofa Paradigm. When model or view. This is one way in which a scientific
the polygonal coil with torsion model is found, it insight can occur. Thus, the phenomenon of inter-
appears to finally break the tension. There may be a mittent progress that involves periods of little pro-
partial analogy here to Kuhn's (1962) idea of the gress punctuated by occasional insights can be seen
persistence of a paradigm in science. Even when as a natural outcome of psychological processes.
anomalies are known to exist, it is difficult to reject
a paradigm until something better is found to re-
Transformations, Invention, and Memory
place it. But this is very difficult to do because it
Provocation
requires breaking out of the current, stable point of
view. Here the bending rod model is hard to reject Transformations as a Source of Creativity. In
until the better model is found, and this requires a this section, I move to a discussion of hypothesis
great deal of imaginative effort. Compared to a generation processes and of analogy generation via
problem on the frontier of science, the scale here is, transformations in particular. It should be noted that
of course, very much smaller and easier. For exam- association apparently is not the only source of cre-
ple, there are no social forces to reinforce the sta- ative or divergent ideas in this protocol. For exam-
bility of the subject's initial model. Nevertheless, ple, after considering the bending rod case, S2 says
this tension between a persistent initial model and a in line 23, "Surely you could coil a spring in
recognized anomaly, which helps to explain the squares, let's say, and it . . . would still. . . ."
long period of slow progress followed by a period of Here he seems to be constructing a new case by
scientific insight in the protocol, is reminiscent of transforming the rod into a square coil rather than
Kuhn's descriptions. making an association to an existing idea in memo-
ry. Also, in line 37 the double-length spring analo-
Tension from an Anomaly as a Source of Moti- gy originates from the transformation of sliding a
vation. The tension associated with S2's dissatis- weight along a wire. A transformation occurs when
faction with his understanding apparently drives S2 alters features previously assumed to be fixed in
him to keep reattacking the problem repeatedly un- an existing problem representation to create a new
til he makes a breakthrough. In the present situa- representation. In a previous study, it was found
tion, the generation of a new or sharply modified that of the analogies generated by 10 subjects in
model is required in order to break the deadlock; solving the spring problem, more were generated
and it is in such cases that analogies should prove to via a transformation than were generated via an
be particularly useful, because they help the subject association (Clement, 1988). In that study, the term
break away from his current model. When they are transformation was used to refer to a general type of
successful, they apparently can lead to fairly large cognitive operation in the form of the alteration of a
and rapid changes in a mental model. S2 considers representation for any situation in working memo-
no less than 12 analogous cases during the protocol, ry, including an original target situation. Thus, the
including some that do not appear in the transcript modification process referred to in Figure 4 is a
excerpts given here, and this high degree of gener- transformation applied to the previously hypoth-
esized scientific model. Although association often to examine the extent to which these processes are
is cited as a primary source of creativity, it may be random or constrained. In fact, much of the pro-
that transformations are just as important, if not tocol preceding the torsion insight can be viewed as
more important, in scientific problem solving. divergent exploration to find clues for a new direc-
tion for analysis. Some relatively unconstrained di-
Invention of Analogous Cases. The novelty of vergent processes tnat occur in the protocol are as-
the zigzag and polygonal springs supports the claim sociations, transformations of the problem space,
that these are invented cases. For example, the the activation of analogous cases in memory, and
square coil was apparently constructed via a trans- the invention of new analogous cases. These pro-
formation, not recalled from memory. Although cesses can lead to multiple suggestions with no
analogous cases typically are thought of as sche- guarantee of success or even relevance. They are
mata already in long-term memory that are much less constrained and systematic than an estab-
activated or retrieved during problem solving, it lished, convergent procedure for solving a prob-
can also happen that the analogous case is invented lem. This leads to the following question: Are S2's
along with the analogy relation. Models generated processes so divergent as to constitute a random
by inventing an analogous case are in this sense trial and error process?
even more creative than those generated by being Certainly, S2' s divergent thinking seems to be
reminded of an analogous case. less systematic or formal than either logical deduc-
The polygonal coil is a new problem representation or methodical procedures of induction. And yet
tion amenable to a new method of analysis (tor- this less formal method of conjecture, criticism,
sion). In such an instance, the knowledge that one and modification allows him to make impressive
gains from an analogous case C need not be "stored progress in his understanding. In this process, it
in" C. Thinking about C may activate a useful sche- does not matter so much if one makes a faulty con-
ma (such as torsion) that has not previously been jecture; it may still be possible to transform it into a
applied either to the original situation to be ex- successful conjecture by carrying out a series of
plained or to C. This instance provides some sup- criticisms and modifications. In this section, I dis-
port for Black's view that the interaction between cuss the sense in which this successive refinement
the original and analogous cases can produce process goes beyond a random trial-and-error
knowledge in the form of an insight that was not strategy.
residing beforehand in either the original or the In its weakest form, the cycle in Figure 4 can be
analogous cases: "It would be more illuminating in described as a random trial-and-error process. By
some of these cases to say that the metaphor creates this I mean that the old hypothesis is discarded and a
the similarity than to say that it formulates some totally new random hypothesis is tried on each cy-
similarity antecedently existing" (Black, 1979, p. cle, without any learning or attempts at modifica-
37). In the present case study, in contrast to the tion between cycles. A less divergent strategy
usual view of analogy generation, the recognition would be to modify randomly part of the previous
of the key relationship (torsion) in the analogous hypothesis and keep the remainder in each cycle.
case occurs well after the generation of the analogy. This is analogous to a random variation theory of
The analogy plays a provocative role in activating a evolution. (See Campbell, 1960, for an exposition
principle whose applicability was previously unre- of this analogy.) However, there is evidence that
cognized, rather than a "direct source of trans- the generation and modification processes are not
ferred information" role. This issue is discussed random ones in the case of S2, and that they are
further in Clement (1988b). more powerful than the above two processes.
Thus, some analogies are invented rather than The first type of evidence is the general observa-
recalled, and some playa "provocative" role in tion of spontaneous analogy generation as a hypoth-
accessing new information rather than a "direct esis generation strategy. Analogous cases are gen-
transfer" role. erated by association or transformation processes
which means that they are connected in some way to
the target. The connection may not be a strong one,
Constrained Successive Refinement versus
but this is better than no connection at all.
Blind Variation
The second type of evidence indicates that, at
In this section, I tum to hypothesis generation times, a conscious constraint is held in mind when
and modification processes as sources of creativity generating a new association or transformation. For
within the model construction cycle. I want to begin example in line 57, S2 appears to focus on the idea
of stretching as a constraint as he generates several apparently because the square is simpler to analyze.
tentative analogies by association after asking him- This is a classic use of rational assessment criterion.
self, "What else stretches"? In a second example
in line 117, he generates polygonal coils while at- Less Constrained Methods. Not all generation
tempting to "generate ideas about circularity. . .. methods are highly systematic or constrained. The
Why should it matter? How would it change the generation of the double-length spring analogy in
way the force is transmitted" in the spring? The use line 37 provides an interesting example. Here the
of conscious constraints during generation is one analogy originates from the idea of sliding a weight
sense in which the model construction cycle can go along a rod. S2 then imagines this transformation
beyond a random variation and selection process. happening on the spring itself, as if it were simply
A further type of evidence is the observation of an "interesting thing to try." There is some evi-
an intelligent modification process in the cycle. dence here that he is exploring new and uncertain
Most of the analogies generated by S2 were rejected directions rather than trying to achieve a specific
in the end. But several did clearly serve as stepping goal using a conscious strategy of generation under
stones by preparing the way for suggesting better constraints. Although the analogy in this case does
ideas later on. This gives the cycle the property of not lead to a breakthrough, one cannot rule out the
successive refinement, in which one can learn from possibility that the ability to think playfully in a
the mistakes of the past. For example, the first relatively unconstrained manner would, at times,
zigzag spring in line 23 is criticized as a model be a powerful method.
because of the contaminating effect of bending at
the joints. This is then modified into a second Summary. Thus, I arrive at an intermediate
zigzag model with stiff joints, which is aimed at position concerning the nature of the subject's hy-
removing the criticism. As a second example, the pothesis generation and modification processes.
bending rod model is criticized because of an as- Compared to a pure eureka event, they form a more
sumed lack of cumulative bending in the spring. ordinary and connected train of thoughts. Com-
The introduction of the square coil model solved pared to a problem-solving process governed by
this problem by eliminating the cumulative bending established procedures, they are divergent pro-
effect. In these instances, the subject seems to gen- cesses that are relatively unconstrained. They can
erate or search for modifications that remove partic- produce novel inventions like the polygonal coil as
ular difficulties that the evaluation process has iden- well as a presumably infinite variety of other repre-
tified in an existing model. Thus, the cycle involves sentations. As they occur here within the model
intelligent modification based on information about construction process however, they often appear as
prior difficulties. This is a particularly powerful part of an intelligent successive refinement process
way in which generation and modification pro- rather than a blind variation and selection process.
cesses can be constrained. (See Darden, 1983,
Rada, 1985, and Darden and Rada, 1988, for fur-
ther discussion of nonblind hypothesis generation,
including the use of interrelations between scien- Darwin's Theory of Natural Selection
tific fields as a heuristic. Also, Holland et aI., 1986,
discuss goal-weighted summation of activation as a Having reviewed some philosophical views of
possible mechanism for guiding retrieval of rele- hypothesis formation processes in science and hav-
vant information, and Lenat, 1977, 1983, discusses ing presented some current findings from expert
heuristics for learning by discovery in mathematics. protocols, I will consider briefly a third approach to
From a broader perspective, in case studies of Fara- the study of creativity in science: the analysis of
day's and Darwin's thought, respectively, Tweney notebooks and other historical documents produced
(in press) and Gruber (1974) have proposed that by innovative scientists. I return to the example of
breakthroughs which appear to result from a fortu- Darwin's theory of natural selection mentioned at
nate "chance interaction" of several ideas were in the beginning of this chapter. Gould (1980) noted
fact significantly favored by a network of prior ac- that earlier writers had described the origin of this
tivities in the scientist's life.) discovery as the net result of a gradual buildup of
Finally, it should be noted that comparison and information-a process of accretion that occurred
selection between previously generated models can during Darwin's voyage on the Beagle, principally
also occur. For example, S2 settles on using the in South America. However, Gruber (1974) de-
square coil as a model over the hexagonal coil, bunked this view by pointing to evidence in Dar-
win's notebooks indicating that after the Beagle's that small individual forces acting over long periods
voyage, he, like a number of other naturalists, be- of time could effect vast changes in nature.
lieved in the existence of evolution (gradual change Thus, historical evidence in Darwin's case now
in species) but still had no model to explain it. He supports a more complex view that either induc-
lacked the theory of natural selection. It was only tivism or eurekaism. Both the fertile empirical
after a year and a half of conceptual struggle after ground of careful observations and the nonem-
his return to England that Darwin was able to for- pirical insights formed by key analogies to other
mulate a satisfactory theory. A particularly famous fields were apparently crucial in Darwin's case.
piece of evidence arguing against the accretion This analysis suggests that a more realistic hallmark
view is the important role of an analogy that oc- of genius than pure eureka episodes is the ability to
curred to Darwin when he read Malthus. In his generate a variety of tentative analogue models as a
autobiography (written much later) he wrote: starting point and then to carry out the long struggle
of repeated conjectures, criticisms, rejections, and
I happened to read for amusement Malthus on population. and modifications necessary to produce a successful
being well prepared to appreciate the struggle for existence which new theory. Although the time scale is much longer
everywhere goes on from long-continued observation of animals in Darwin's case, it is interesting that these are the
and plants, it at once struck me that under these circumstances same distinguishing criteria that emerge from the
favorable variations would tend to be preserved, and unfavorable
ones to be destroyed. (Darwin 189211958, pp. 42-43)
cases of model construction in the protocols dis-
cussed earlier. This suggests that perhaps the most
viable powerful form of scientific reasoning lies not
Darwin saw that factors similar to those that limited in the ability to "hit" on a perfect model at the
human population growth (such as a limited food outset, but in the ability to engage in such a dialec-
supply) might be the source of a selection factor in a tic, successive refinement cycle.
survival of the fittest model for animals. Thus, the
accretion by induction view is hard to maintain in
Darwin's case.
Does the Malthus episode then provide evidence Features of Creative Thinking and
for a eurekaist view of Darwin's achievement? The Implications for Future Research
recent analyses of Darwin's private notebooks car-
ried out by Gruber (1974) and Schweber (1977)
Creative Thinking
argue against this conclusion as well. They show
that Darwin struggled long and hard, considering To the extent that an extended analysis can re-
several hypotheses and gradually modifying and fit- move the initial subjective impressiveness of an
ting a number of pieces together into the theory of event, perhaps I am in danger here of seeming to
natural selection. The notebooks indicate the analo- trivialize the processes of analogy generation,
gy from Malthus was only one event in a compli- model construction, and insight as hypothesis de-
cated process of generation, evaluation, and velopment activities, and I would therefore like to
modification. avoid giving that impression. Clearly, once one has
Darwin read widely in fields outside of biology thought through the answer to a problem, the solu-
and apparently drew analogies from these fields in tion process can appear to be less impressive or
constructing his theory, including the ideas of vari- even obvious from hindsight. While one is actually
ation and selection (from breeding in domestic hus- solving a problem, however, creative reasoning,
bandry) and the idea of natural competition (from such as that exhibited by S2, is impressive in a
Malthus) (Darden, 1983). Gould believed Darwin number of ways.
also was influenced by the laissez-faire economics 1. First of all, there is the insight in the protocol
of Adam Smith, which showed that an ordered and that seems to lead to a "flood" of ideas. The speed
efficient economy could emerge from free competi- of progress during this episode is impressive.
tion. An analogy can be made to evolution here via 2. S2' s central achievement is the generation of a
the common idea of positive group change coming new structural hypothesis-the invention of a new
out of individual struggle. In addition, Gruber model of hidden mechanisms in the spring that he
(1974) cited Darwin's early geological theories on has never observed. This involves the identification
the growth of the Pacific barrier reefs over tens of of new causal variables in the system (such as tor-
thousands of years as fertile preparation for the idea sion) and new causal chains, as well as the identifi-
cation of a new global effect (lack of cumulative some strong restraints that can, in fact, act to con-
bending). trol the enterprise of model construction. Thus, al-
3. An important factor in producing this achieve- ternating between generative and evaluative modes
ment is the subject's desire to ask' 'why" questions in scientific thinking is seen as a powerful method,
and to seek a deep level of understanding beyond even when the generative methods are divergent in
what is required for the solution of the immediate character and new empirical tests cannot be
problem. Presumably, this urge to penetrate surface performed.
features and to conceptualize an underlying explan- 8. Perhaps S2's awareness of his own ability to
atory model at the core of a phenomenon is a basic criticize ideas, and the resulting faith in himself as a
motive underlying creative theory formation in self-correcting system, allows him a freer hand-
science. allows him to be more uninhibited in generating
4. S2 exhibits a remarkable persistence in this conjectures and in considering directions to pursue.
quest in the face of recognized internal inconsisten- It may be that generative ability and critical ability
cies and repeated failures. There is something of an are mutually supporting. Critical ability gives one
existential twist here: although the problem has no the freedom to be unusually associative or in-
practical significance for him, he puts enormous ventive. Generative inventiveness, or the ability to
energy into the problem of understanding as a chal- replace and repair what one removes, gives one the
lenge for its own sake. confidence or assurance to be critical of and to tear
5. S2' s playful and uninhibited inventiveness in down, at times, existing ideas. S2 seems willing to
producing conjectures and modifications of the consider "risky" analogies, such as the double-
problem is impressive. The analogous cases he gen- length spring and the bending rod, that appear to be
erated in searching for a better way to represent the very different from the original problem. However,
problem included the bending rod, polyester mole- it has been shown that, even when a risky initial
cules, leaf springs, watch springs, two types of analogy does not tum out to be explanatory, modifi-
zigzag springs, two or more types of polygonal cations of it may lead to an explanatory model.
springs, and double-length springs. He displays an Realization of this potential for debugging or re-
ability to think divergently and the flexibility to designing via criticism and modification may allow
modify thought forms in novel ways. This kind of one to feel freer to explore more imaginative mod-
flexibility is commonly recognized as a prominent els or a wider range of models. This freedom, in
characteristic of creative physicists and inventors. tum, would appear to be an important tool in the
6. There is a willingness in S2 to criticize vig- difficult job of breaking out of previous conceptions
orously and attack the validity of his own conjec- ofthe target situation. Again, rather than the ability
tures. He is able to engage in a dialectic conversa- to hit on the best possible idea in one stroke, it may
tion with himself, proposing new ideas on the one be that it is the ability to engage in a cycle of hypoth-
hand and criticizing them on the other. This seems esis construction and improvement that isthe most
to require viewing the failure of any single idea as powerful form of scientific thinking.
not very important; although, as has been shown, These eight qualities are some of the most im-
the apparent failure of seven or eight ideas to pro- pressive characteristics of creative thinking visible
duce a breakthrough does lead to some degree of in the case study.
frustration for S2.
7. With respect to Figure 4, one can contrast the
productive function of the generation and modifica- Implications for Future Research
tion processes with the evaluative function of the The conclusions reached here suggest that cre-
rational and empirical testing processes. The diver- ative hypothesis formation processes are still poorly
gent and creative generation processes (such as the understood but are not outside the realm of possible
use of analogies) represent a significant departure study. Mansfield and Brusse (1981) gave examples
from the more systematic, rule-governed processes of five aspects of the creative process in science: (1)
of theory growth that are envisioned by induc- problem selection; (2) extended effort; (3) setting
tionists, who would tend to see them as much too theoretical, empirical, and methodological con-
unrestrained to be part of the disciplined scientific straints; (4) changing constraints; and (5) verifica-
enterprise. However, the generation processes are tion and elaboration, including a process of for-
not entirely unconstrained, as has been discussed, mulating new constraints and testing them. Two
and the evaluative processes in Figure 4 provide areas that the present case study does not address
are problem formulation and empirical and meth- The above phenomena are not well understood
odological constraints. 9 These are important prob- and indicate that we are far from formulating ade-
lems for future research. quate explanations for many aspects of creative pro-
There are also other areas in which observations cesses. They still inspire awe, pointing to areas in
were made in the case study that the present analysis which the science of psychology remains quite
has said very little about. The first is the overall weak and in which further research is needed.
complexity in the details of S2' s thought processes,
including the presence of multiple goals, returns to
previously attempted solution paths, the balancing Educational Implications
of divergent and convergent processes, and the res-
olution of competing influences. In addition, each Learning via model construction is an area of
of the subprocesses shown in Figure 4 is in need of utmost importance to mathematics and science edu-
much more detailed study. Second, S2 can exhibit cation, an area that is very poorly understood. The
an "aha" reaction that something important has present study is essentially a study of learning via
been discovered, even before its implications have model construction in scientists. Thus, the study
been developed and articulated. For example, the will have interesting educational implications if it
aha episode upon considering the square coil in line can tell us more about processes that need to be
117 is of this form. Third, subjective observations fostered when students are learning scientific
from the videotape that are hard to capture in print models.
are the exuberance present in his aha episodes and Essentially, I will propose that the model con-
the tone of frustration present during his periods of struction cycle in Figure 4 may be useful as a de-
failure. This adds an emotional dimension to the scription of processes that need to take place in
process that is distinctly human. Fourth, I have only students who are learning to comprehend scientific
touched on the problem of how "guided" conjec- models. Because of space limitations, I can only
ture is guided-why one person's initial conjec- present a brief sketch of this idea here. The cycle is
tures are much more fruitful in the long run than relevant to three major educational goals: the con-
those of others. Fifth, S2's strong drive to ask tent goal of comprehending established scientific
"why" questions mentioned earlier (a kind of curi- models; the process goal of learning to solve ill-
osity) certainly has not been explained. Sixth, very structured problems; and the even more ambitious
little has been said about rational evaluation. Of process goal of learning scientific method or scien-
particular importance is the problem of how one tific inquiry skills. By attending to these different
evaluates the validity or appropriateness of a pro- content and process goals, educators may be able to
posed analogy. Matching "important" features is design instructional activities more effectively.
one method, as has been illustrated, but there may
be others as well (Clement, 1986). Finally, S2's Content Goals: Comprehending Scientific
flexibility in inventing new problem representa- Models
tions is hard to model. His image of the spring
appears to be malleable; he appears able to modify With respect to the first goal of comprehending
it into an infinite number of forms and variations. In established scientific models, several points can be
fact, there are a number of spontaneous imagery made. First, as has been discussed, many modem
reports in the protocol that suggest that certain scholars have argued that explanatory models are an
forms of spatial reasoning on spatial representations essential part of scientific understanding. As shown
may be central to S2's thinking here. Although the in Figure 3, explanatory models are a separate type
discussion of these reports is beyond the scope of of knowledge from either empirical laws or formal
this study, it opens up a large and important ques- quantitative principles. Easley (1978) and others
tion for future research on the nature of these pro- have noted the unfortunate tendency of educators to
cesses and the role they play in scientific thinking associate "real" scientific thinking with only the
(Clement, in press a,b). latter two types of knowledge.
A second point is that students learn complex
models via an internal construction process, not via
9Getzels and Csikszentmihalyi (1976) have described processes
of problem finding in art. Lenat (1983) described an attempt to
a direct transmission process during lecture. I can-
develop a simulation model for the process of selecting "in- not support this assumption fully here, but current
teresting" problems (theorems for analysis) in mathematics. research in science education is providing an in-
creasing amount of evidence in this direction. The those that are aimed at forming empirical law hy-
complex, tacit, nonobservable, and sometimes potheses or formal quantitative principles, because
counterintuitive nature of scientific models means the cognitive processes involved may be quite dif-
that misconceptions or "bugs" will be the rule ferent. If this is correct, students are unlikely to
rather than the exception during instruction, requir- learn explanatory models from laboratories that are
ing critical feedback and correction processes. This aimed at inductive reasoning; nor are they likely to
means that the learning of complex, unfamiliar, or learn them from the study of formal quantitative
counterintuitive models in science requires a kind principles.
of learning by doing and by construction and crit-
icism rather than by listening alone.
Problem Solving and Inquiry Skills
In this light, Figure 4 is seen as a potential model
for the learning of scientific concepts by construc- Figure 4 can also be thought of as a model of the
tion in the classroom. Educators inspired by Jean process of constructing a representation for an ill-
Piaget and others have advocated approaches that structured problem. (Here memories of prior expe-
are based on the processes of disequilibration, ac- riences can playa role in empirical testing if no new
commodation, and the construction of knowledge. empirical information is available.) In the case of
Unfortunately, the application of these concepts to content goals as discussed above, considerable sup-
instructional design suffers from a lack of precision port might be given by the teacher in guiding stu-
and consistency. The present approach may lead to dents through such a cycle. However, in order to
a more explicit model of the process of learning learn problem-solving skills, students eventually
scientific models. The findings from this study lend need to be able to generate problem representations
support to an educational strategy in which, rather by going through construction cycles without teach-
than "swallowing packaged ideas as a whole" in er support. Despite this difference, Figure 4 pro-
lecture, students are seen as developing partial vides the basis for seeing some significant overlap-
models, questioning them in face of anomalies, and ping in the strategies for achieving content and
working from their initial model to construct a more process (problem solving and inquiry) goals in sci-
adequate model. The term knowledge construction ence education.
has been much used in discussions of education. Finally, the most ambitious goal in science edu-
Perhaps the concepts of using prior knowledge cation is that of teaching scientific investigation or
(e.g., analogies), modifying models, and the suc- inquiry skills. In fact, it is extremely rare to find a
cessive refinement cycle can provide a more ex- class in which students are asked to propose and test
plicit picture of construction processes. scientific hypotheses for phenomena. Here again, it
A third point is that explanatory model construc- seems important not to assume that "discovery
tion takes place in the scientist via a set of processes learning" by induction from data is the predomi-
that are different from those used either in formal nant process in the scientific method. Model crit-
deductive proof, in the manipulation of quantitative icism and modification processes would seem to be
expressions, or in inductions from data. The present of crucial importance in the design of inquiry
study speaks perhaps most strongly to this last acti viti es .
point. It underscores the importance of processes Thus, the most general point to be made is that
that aid abduction, such as using analogue models the cycle outlined in Figure 4 may prove useful as
for developing understanding; fostering diseq- an outline of relevant learning processes for guiding
uilibrium in order to motivate efforts toward model educators in designing and evaluating instructional
construction; and fostering criticism and modifica- activities that are concerned with the learning of
tion (or rejection) processes for overcoming diffi- scientific models. Here I have only been able to
culties occurring in students' models that contain sketch some possible implications along these lines;
misconceptions. Very valuable nonempirical crit- further educational research and development
icism and modification processes can take place efforts are very much needed.
when students attempt to give explanations and ar-
gue about them in large or small group discussions
(Clement et al., 1987). This simple implication is Summary
probably greatly underemphasized in instruction.
Educators need to distinguish between activities This chapter began by posing questions concern-
that are aimed at forming explanatory models and ing the origins of hypotheses in science and the role
of insight or eureka events in creative scientific tion question, creative mental processes, and edu-
thi.nking. I have attempted to show that protocol cational implications.
~vlde~ce .can be used to argue against an overly
mductlOmst view of the source of hypotheses. It can
~lso. be used to argue against either a pure accre- Sources of Hypotheses
~lOm~t or.~'pure eurekaist view of the pace of change
m sClenuhc hypothesis formation. Instead, it has 1. A new scientific hypothesis in the form of an
led me to take a less simplistic view of hypothesis explanatory model can be developed via noninduc-
tive means in the absence of new empirical informa-
developI?~~t as illustrated in Figure 4, emphasizing
the posslblhty of both empirical and nonempirical tion. This lends support to the importance of a non-
sources of hypotheses and multiple passes through a inductive component in the hypothesis generation
cycle of generation, evaluation, and modification process.
2. The model construction process proposed was
(or rejection). In this cycle, hypothesis evaluation
one of successive refinement, involving repeated
can also originate from empirical and nonempirical
cycles of generation, evaluation, and modification
s?urces. In such a system, powerful scientific in-
or rejection. Table 1 summarized evidence from the
slghts can occur when a new model is developed
protocol that S2' s progress is a result of this kind of
that leads to a "flood" of new ideas. But this can
cyclical refinement process rather than a result of
happen without necessarily involving the extraordi-
either a convergent series of deductions or induc-
nary or unconscious reasoning processes that are
tions from observations.
associated with the term eureka event. The present
data support the view that the methods that are used 3. Such a cycle can be more powerful than a blind
trial and error or a blind variation and evolution
by scientists are varied and complex, and that the
process. For example, when difficulties have been
hypoth~tico-deductive method, rational evaluation,
i~entified in an existing model, subsequent genera-
~bductlOn, analogy, and induction may all play
tion and modification processes can serve to re-
lmportant roles at different times in scientific
thought. move the difficulties.
4. Hypothesis evaluation processes appear to be
Recent work in the philosophy of science was
an inherent part of hypothesis formation down to
drawn on to make several useful distinctions. The
resolution intervals of a single minute on occasion.
t~rm scientific model was used to refer to a predic-
The history of science tends to look at develop-
tive analogy. The term explanatory model was used
ments over years or weeks. From this perspective,
to distinguish those scientific models that are
the case study observation of very small cycle times
intended to represent nonobvious entities that are
for the non-empirical criticism and modification
present in the situation to be explained. The latter
loop in Figure 4 (as small as 90 seconds here) makes
te~ ~ll?ws one to distinguish between two types of
it very difficult to separate the "context of discov-
sClentIflc hypothesis: a hypothesis in the form of a ery" from the "context of justification" in the ear-
predictive, explanatory model that introduces new
ly stages of hypothesis formation. In such cases it is
entities that have not previously been (and may nev-
possible to have a rapid, dialectic interplay between
er be) ?bserved directly, and an empirical law hy-
generation and evaluation processes.
pothesls that summarizes patterns in observations.
5. The development of a convincing explanatory
These distinctions helped to describe creative
model hypothesis can also lead to the formation of
proc~sses in the case study of subject S2, who was
an empirical law hypothesis in the absence of new
workmg on the problem of whether a wide spring
empirical information: in this case, the final model
stre~ches more than a narrow spring. S2's central
of the polygonal coil with torsion supports the em-
achlevement was the generation of an explanatory
pirical law hypothesis that (other factors being
model-the invention of a new model of hidden
equal) wide springs will stretch more than narrow
mechanisms in the spring that he had not observed.
springs.
This involved the identification of new causal vari-
ables in the system (such as torsion) and new causal
chains, as well as the identification and explanation
Spontaneous Analogies
of a global e~fect (lack of cumulative bending).
. Th~ concluslOns of the study are organized below 1. Subjects were observed to generate and use
mto five categories: sources of hypotheses, the role spontaneous analogies as predictive models.
of spontaneous analogies, the eureka versus accre- 2. Many of the observed analogies apparently
were generated via a transformation of another sit- reasoning operations; they do not involve a sudden
uation. Although association is often cited as a pri- break in the train of thought that would indicate a
mary source of creativity, it may be that transforma- pure eureka event. The train of thoughts that S2
tions are just as important, if not more important. reports weaves a "coherent story" in the sense that
3. In a successful model construction cycle, an each new idea is connected to previous ideas and is
initial analogy with low-explanatory status can be therefore at least weakly constrained by previous
developed and modified to become an explanatory ideas. That he appeared to use the processes in Fig-
model that proposes the presence of a hidden struc- ure 4, in which new ideas can grow out of modifica-
ture operating in the target situation. tions of or reactions to past ideas, is an even more
4. This means that analogy can playa role in the specific sense in which his insight did not just
generation of new hypotheses in at least the follow- emerge from out of the blue. This does not prove
ing two different ways: (a) an analogous case can that important unconscious or non ordinary pro-
serve as a rough initial model of the target situation cesses cannot occur-Poincaire's famous insight
that is then developed and refined; (b) a developed upon entering a bus may have been one example-
model, whatever its origin, is linked by an analogy but it does indicate that important insights can be
relation to the target situation because it posits that generated when there is no evidence for such
elements and relations in the model are like ele- processing.
ments and relations in the target. This appears to 3. S2's insight occurred after a long struggle re-
contribute to a feeling of understanding when ele- sulting from the conflict between a first-order
ments in the model are familiar. model and a recognized anomaly. The conflict or
5. Rather than always being stored as cases that disequilibrium condition between a persistent
are activated in memory, some analogies (e.g., re- model and an anomaly appears to provide a moti-
vised mental models) are novel, invented cases. vating force for a more intense level of activity for
6. In some instances, the knowledge that one hypothesis development that is not dependent on
gains from an analogy does not come directly from other external motives. The persistence of the sub-
the analogous case. The analogy can play a ject's initial model and the tension between it and
provocative role by triggering the application of a the perceived anomaly may be partially analogous
principle that has never before been applied to ei- to the persistence of a paradigm in the face of anom-
ther the target or analogous cases. In these in- alies in science. An important function of the strat-
stances, the most important relationships in the egy of searching for analogous cases is that it may
analogous case is recognized after the generation of help the subject to break away from such a stable
the analogous case. persistent model. This helps to explain the presence
of intermittent periods of negligible progress and
rapid change or insight as a natural outcome of
The Eureka versus Accretion or Pace of
psychological processes.
Conceptual Change Question
1. Three possible levels of insight were defined: a
breakthrough, which overcomes a barrier that has Creative Mental Processes
blocked progress; a scientific insight, which is a
relatively sudden breakthrough leading to a signifi- 1. A subject can use relatively unconstrained,
cant improvement in a model; and a pure eureka divergent, hypothesis generation processes that can
event, which is an insight that is not explainable via lead to insights, including analogy, association,
ordinary reasoning processes. transformation, and invention processes.
2. Aha episodes were observed in association 2. Divergent and creative processes represent a
with a scientific insight involving the formation of a significant departure from the more systematic pro-
new explanatory model. Such an insight can be cesses of hypotheses generation envisioned by in-
quite powerful and impressive and can lead to a ductionists, who would tend to see the former as
rapid improvement in conceptual understanding. much too unrestrained to be part of the scientific
However, although insights can involve creative enterprise. However, the evaluation processes in
thinking, (a) they can involve preparation and con- Figure 4 can provide some strong restraints. Thus,
firmation efforts, and (b) they do not necessarily alternating between generative and evaluative
involve unconscious or other extraordinary thought modes in scientific thinking is seen as a powerful
processes that are outside the domain of normal method, even when the generative methods are di-
vergent in character and new empirical tests cannot ries that are aimed at inductive reasoning, or from
be performed. lectures on formal quantitative principles.
3. Divergent processes are relatively uncon- 4. The learning process outlined in Figure 4 at-
strained compared to other processes, but there is tempts to give an explicit cognitive meaning for the
evidence that generation and modification pro- term knowledge construction. As such, it may
cesses can be guided by some constraints. This prove useful as a model of relevant learning pro-
makes the model construction cycle more powerful cesses for guiding curriculum planners and practi-
than a blind selection and variation process. tioners in designing and evaluating instructional ac-
4. Recent analyses of Darwin's notebooks have tivities in science education.
suggested that a more indicative hallmark of genius
than pure eureka episodes is the ability to generate In conclusion, it appears to be possible to devel-
tentative analogue models as a starting point and op models of creative hypothesis formation pro-
then to carry out the long struggle of a cycle of cesses that are tied to empirical information from
repeated generation, criticism, and modification or thinking-aloud protocols. Many aspects of creative
rejection that is necessary to construct a successful reasoning processes remain poorly understood:
new theory. In fact, these are the same prominent "guided" conjecture, anticipation in the aha phe-
features that emerged from an analysis of model nomenon, the apparent malleability of the spatial
construction in the thinking-aloud case study. It imagination, emotional factors, question asking,
was conjectured that the most viable powerful form and sources of curiosity, to name just a few. They
of scientific reasoning may lie in the ability to en- still inspire awe. Nevertheless, creativity is a more
gage in such a dialectic cycle, rather than in the accessible object of study than some would claim; it
ability to invent a completed model in one stroke. is not always an "instantaneous crystallization
5. Thus, the examples discussed here motivate a transmitted from the unconscious." Current tech-
conception of advanced scientific thinking that in- niques make the process of studying creativity a
cludes nondeductive, noninductive, and divergent productive and exciting one: by using protocol anal-
processes. These processes can play an important ysis and other methods, significant progress can be
role in producing predictive, explanatory models made in increasing our understanding of it. Exactly
that are novel inventions. how much we will be able to understand and explain
in this complex domain-how far our model con-
struction cycles will take us-remains to be seen.
Educational Implications
1. The findings mentioned thus far suggest an ACKNOWLEDGMENTS. I would like to thank
educational strategy in which, rather than "swal- Ryan Tweney, David Perkins, Ernst von Glasers-
lowing packaged ideas as a whole" in lecture, stu- feld, John Lochhead, and David Brown for their
dents are helped to develop partial models, to crit- very helpful comments, and, in particular, Lindley
icize them, and to work from their initial model to Darden for her many insightful criticisms and
construct a more adequate model. suggestions.
2. The findings also underscore the importance
of: using analogue models for developing under-
standing; fostering disequilibrium in order to moti- References
vate efforts toward model construction; and crit-
Achinstein, P. (1970). Inference to scientific laws. In R.
icism and modification (or rejection) processes for Stuewer (Ed.), Minnesota studies in philosophy of science,
overcoming difficulties occurring in students' mod- vol. 5: Historical and philosophical perspectives of science
els that contain misconceptions. (pp. 87-104). Minneapolis: University of Minnesota Press.
3. Explanatory models are an essential part of Ackennan, R. (1965). Theories of knowledge: A critical intro-
scientific understanding that is a separate type of duction. New York: McGraw Hill.
Black, M. (1979). More about metaphor. In A. Ortony (Ed.),
knowledge from either empirical laws or formal Metaphor and thought (p. 19-45). Cambridge, England:
quantitative principles. Educators need to dis- Cambridge University Press.
tinguish between activities that are aimed at form- Campbell, D. (1960). Blind variation and selective retention in
ing empirical laws or quantitative principles and creative thought as in other knowledge processes.
those that are aimed at forming explanatory models, Psychological Review, 67(6), 380-400.
Campbell, N. (1920). Physics: The elements. Cambridge:
because the cognitive processes that are involved Cambridge University Press. (Republished in 1957 as The
may be quite different. For example, students are foundations of science. New York: Dover.)
unlikely to learn explanatory models from laborato- Clement, 1. (1981). Analogy generation in scientific problem
solving. Proceedings of the Third Annual Meeting of the Cog- Hesse, M. (1966). Models and analogies in science. South
nitive Science Society, 3, 137-140. Bend, IN: Notre Dame University Press.
Clement, J. (1982). Analogical reasoning patterns in expert Hesse, M. (1967). Models and analogies in science. In P. Ed-
problem solving. Proceedings of the Fourth Annual Meeting wards (Ed.), The encyclopedia of philosophy (pp. 354-359).
of the Cognitive Science Society, 4, 79-81. New York: Free Press.
Clement, J. (1986). Methods for evaluating the validity of hy- Holland, J. H., Holyoak, K. J., Nisbett, R. E., & Thagard, P. R.
pothesized analogies. Proceedings of the Eighth Annual Meet- (1986). Induction: Processes of inference, learning and dis-
ing of the Cognitive Science Society, 8, 223-234. covery. Cambridge, MA: MIT Press.
Clement, J. (1988). Observed methods for generating analogies Karmiloff-Smith, A., & Inhelder, B. (1975). If you want to get
in scientific problem solving. Cognitive Science, 12, 563- ahead, get a theory. Cognition, 3(3), 195-212.
586. Koestler, A. (1964). The act of creation. New York: Macmillan.
Clement, J. (in press a). Nonformal reasoning in experts and in Kuhn, T. (1962). The structure of scientific revolutions (1 st ed.).
science students: The use of analogies, extreme cases, and Chicago: University of Chicago Press.
physical intuition. In J. Voss, D. Perkins, & J. Siegel, Kuhn, T. (1977). Concepts of cause in the development of phys-
Informal reasoning and education. Hillsdale, NJ: Erlbaum. ics. In T. Kuhn (Ed.), The essential tension: Selected studies
Clement, J. (in press b). Use of physical intuition in expert in scientific tradition and change (pp. 21-30). Chicago: Uni-
problem solving. In D. Tirosch & S. Straus (Eds.), Implict versity of Chicago Press.
and explicit knowledge, an educational approach. Norwood, Lakatos, I. (1978). The methodology of scientific research pro-
NJ: Ablex Publishers. grammes. Philosophical papers (Vol. I). Cambridge, En-
Clement, J. (1987, July), with the assistance of Brown, D., gland: Cambridge University Press.
Camp, C., Kudukey, J., Minstrell, J., Palmer, D., Schultz, Langley, P. W. (1979). Rediscovering physics with BACON 3.
K., Shimabukuro, J., Steinberg, M., & Veneman, V. Over- Proceedings of the Sixth International Joint Conference on
coming students' misconceptions in physics: The role of an- Artificial Intelligence, 6, 505-507.
choring intuitions and analogical validity. Proceedings of the Lenat, D. B. (1977). Automated theory formation in mathemat-
Second International Seminar on Misconceptions and Educa- ics. Proceedings of the Fifth International foint Conference
tional Strategies in Science and Mathematics (pp. 84-97). on Artificial Intelligence, 5, 833-842.
Ithaca, NY: Cornell University, Department of Education. Lenat, D. B., (1983). The role of heuristics in learning by dis-
Darden, L. (1982). Artificial intelligence and philosophy of sci- covery. In T. Mitchell (Ed.), Machine learning (pp. 243-
ence: Reasoning by analogy in theory construction. 306). Palo Alto, CA: Tioga Publishing.
Philosophy of Science Association 2, 147-165. Mansfield, R. S., & Busse, R. V. (1981). The psychology of
Darden, L., & Rada, R. (1988). Hypothesis formation via inter- creativity and discovery: Scientists and their work. Chicago:
relations. In A. Prieditis (Ed.), Analogica (pp. 109-127). Los Nelson-Hall.
Altos, CA: Kaufman. Nagel, E. (1961). The structure of science. New York: Harcourt,
Darwin, C. (189211958). In F. Darwin (Ed.), The autobiogra- Brace, & World.
phy of Charles Darwin and selected letters. New York: Nersessian, N. (1984). Faraday to Einstein: Constructing mean-
Dover. ing in scientific theories. Dordrecht, The Netherlands: Mar-
Easley, J. (1978). Symbol manipulation reexamined: An ap- tinas Nijhoff.
proach to bridging a chasm. In B. Presseisen, D. Goldstein, & Peirce, C. S. (1958). In C. Hartshorne, P. Weiss, & A. Burks
M. Appel (Eds.), Topics in cognitive development (Vol. 2, pp. (Eds.), Collected papers Cambridge: Harvard University
99-112). New York: Plenum Press. Press.
Fischbein, E. (1987). Intuition in science and mathematics: An Perkins, D. (1981). The mind's best work. Cambridge: Harvard
educational approach. Boston, MA: D. Reidel. University Press.
Gentner, D. (1982). Are scientific analogies metaphors? In D. Polanyi, M. (1966). The tacit dimension. Garden City, NY:
Miall (Ed.), Metaphor: Problems and perspectives (pp. 106- Doubleday.
132). Brighton, Sussex, England: Harvester Press. Popper, K. (1959). The logic of scientific discovery. London:
Getzels, 1., and Csikszentmihalyi, M. (1976). The creative vi- Hutchinson.
sion: A longitudinal study of problem finding in art. New Rada, R. (1985). Gradualness facilitates knowledge refinement.
York: John Wiley. I.E.E.E. Transactions on Pattern Analysis and Machine Intel-
Gould, S. (1980). Darwin's middle road. In S. Gould, The pan- ligence, 7(5), 523-530.
da's thumb: More reflections in natural history. New York: Rothenberg, A. (1979). The emerging goddess. Chicago: Uni-
Norton. versity of Chicago Press.
Gregory, R. (1981). Mind in science. Cambridge, England. Schweber, S. (1977). The origin of the Origin revisited. fournal
Cambridge University Press. of the History of Biology, 10, 229-316.
Gruber, H. (1974). Darwin on man. New York: E. P. Dutton. Suppe, F. (1974). The search for philosophic understanding of
Hanson, N. R. (1958). Patterns of discovery. Cambridge, En- scientific theories. In F. Suppe (Ed.), The structure of scien-
gland: Cambridge University Press. tific theories. Urbana: University of Illinois Press.
Harre, R. (1961). Theories and things. London, England: New- Toulmin, S. (1972). Human understanding (Vol. I). Oxford,
man History and Philosophy of Science Series. England: Oxford University Press.
Harre, R. (1967). Philosophy of science: History of. In P. Ed- Tweney, R. (1985). Faraday's discovery of induction: A cog-
wards (Ed.), The encyclopedia of philosophy (pp. 289-296). nitive approach. In D. Goodling & F. James (Eds.), Faraday
New York: Free Press. rediscovered: Essays on the life and work of Michael Fara-
Harre, R. (1972). The philosophies of science. New York: Ox- day, 1791-1867 (pp. 189-209). New York: Stockton Press.
ford University Press. Tweney, R. (in press). Fields of enterprise: On Michael Fara-
Harre, R. (1983). An introduction to the logic of the sciences. day's thought. In D. Wallace and E. Gruber (Eds.), Creative
(2nd ed.) New York: St. Martin's Press. people at work. Oxford: Oxford University Press.
CHAPTER 21
Analogical Reasoning and Problem

Solving in Science Textbooks
Shawn M. Glynn, Bruce K. Britton, Margaret Semrud-
Clikeman, and K. Denise Muth
People's frequent use of analogies to explain every- among those concepts. Unfortunately, the concepts
day phenomena underscores their potential value as are often difficult for readers to understand because
instructional tools. Such expressions as "Let me the concepts are either unfamiliar (e.g., entropy) or
give you an" analogy ... ," "It's just like ... ," are defined somewhat differently than they are in
"It's the same as. . . ," "It's no different everyday life (e.g., velocity).
than ... ," "Think of it this way ... ," are Drawing an analogy between concepts in a sci-
commonplace in casual conversation. ence text can serve an explanatory function and a
The meaningful comprehension of science text is creative function. It serves an explanatory function
a kind of problem solving: the reader's problem is to when it puts new concepts and principles into famil-
understand and elaborate the content of the text. iar terms. It serves a creative function when it stim-
Analogies can help readers to solve this comprehen- ulates the solution of existing problems, the identi-
sion problem. Authors can provide analogies for fication of new problems, and the generation of
readers, and readers can be trained to generate their hypotheses. Before we explain exactly how an anal-
own analogies. In this chapter, we will focus on the ogy fulfills these functions, we will explain what an
explanatory and creative roles that analogies can analogy actually is, how it works, and what con-
play in science textbooks. stitutes a good analogy.
Analogical Reasoning and What Is an Analogy?

Comprehending Science Text
In science and technology, an analogy is a corre-
One of the main purposes of a science text is to spondence in some respects between concepts,
explain complex concepts and the relationships principles, or formulas that are otherwise dis-
similar. More precisely, it is a mapping between
Shawn M. Glynn and Margaret Semrud-Clikeman • De- similar features of those concepts, principles, or
partment of Educational Psychology, University of Georgia, formulas. A related term is analog, which is a con-
Athens, GA 30602" K. Denise Muth • Department of
Elementary Education, University of Georgia, Athens, GA cept, principle, or formula that bears an analogy to
30602. Bruce K. Britton • Department of Psychol- something else.
ogy, University of Georgia, Athens, GA 30602. Although the terms analogy and metaphor are
383
frequently substituted for each other, analogy tends ingful learning described by Wittrock. An analo-
to be used more often in scientific and technical gical relation is powerful because it comprises, in
contexts. Metaphor is used more often in literary actuality, an entire set of associative relations be-
contexts (e.g., she's a breath of fresh air). Because tween features of the concepts', principles, or for-
the examples we will present in this chapter are mulas being compared.
from science textbooks, we will use the term analo- In science texts, analogies are used in a relatively
gy rather than metaphor. precise fashion to transfer ideas from a familiar
The terms model and analogy (or analog) are conceptual domain to an unfamiliar one. In our ex-
sometimes used interchangeably. Model is the planation of an analogy, we will call the familiar
more inclusive term, referring to a simplified repre- conceptual domain the analog and the unfamiliar
sentation of the components, operations, and rela- domain the target. The analog, like any concept,
tions of a more complicated object or process. The has characteristics orfeatures; so does the target. If
representation can take several forms, such as an the analog and the target share common or similar
algebraic equation, a diagram, a flow chart, or a features, an analogy can be drawn between them.
physical replica. An analog is one kind of model; We have represented an analogy, with its constitu-
for example, a pump could be used to model the ent parts, in Figure 1. Note in Figure 1 that the
human heart. Models need not be analogs, how- analog and the target are subordinate to a superordi-
ever. Consider the stick-and-ball models of mole- nate concept, principle, or formula.
cules that are used in chemistry texts. These models To illustrate our representation of an analogy, we
are not analogs. will use the target concept of an electric circuit, and
the analog concept will be a hydraulic (water) cir-
cuit. Examine the electric and hydraulic circuit dia-
How Do Analogies Work? grams in Figure 2. An analogy can be drawn be-
tween them because they share common or similar
Wittrock (1985) conceived of meaningful leam- features, such as the following:
ing in science as a
student generative process that entails construction of relations, CIRCUITS

either assimilative or accommodative, among experience, con-
cepts, and higher order principles and frameworks. It is the Hydraulic Circuit Electric Circuit
construction of these relations between and within concepts that Connecting pipes Wires
produces meaningful learning. (pp. 261-262). Narrow pipe section Resistance
Valve Switch
In our view, when an analogy is drawn between Pump Dry cell battery
science concepts, principles, or formulas, a power- Flowing water Electric current
ful relation is constructed which leads to the mean- Water molecules Electrons
SUPERORDINATE CONCEPT, PRINCIPLE, OR FORMULA
ANALOG •••• compared with ...... TARGET
Feature " " Feature
" "
2 " " 2
Figure 1. Our representation of an analogy, with
3 " " 3 its constituent parts. Note that the analog and the
target are subordinate to a superordinate concept,
n " " n principle, or formula.
CHAPTER 21 • ANALOGICAL REASONING AND PROBLEM SOLVING 385
( RESISTANCE
...... SWITCH
DRY CEll-.r
Figure 2. Analogy between a simple hydraulic circuit and an electric circuit. From Conceptual Physics (p. 513) by Paul G. Hewitt,
1987, Menlo Park, CA: Addison-Wesley Publishing Company, Inc. Copyright 1987 by Addison-Wesley. Reprinted by permission.
Note that the electric circuit and the hydraulic cir- (or principle or formula) in the representation of an
cuit are both circuits, that is, they are subordinate to analogy and its constituent parts is an important
the superordinate concept circuit. There are other one. The identification and naming of the superor-
kinds of circuits which, depending upon the number dinate concept can suggest other analogies; it also
of similar features they share with electric circuits, can stimulate readers to generalize what they have
might also be used to draw an analogy. For exam- learned and apply their learning to other contexts.
ple, think of a pinball machine: the pinballs (corre- Therefore, even though it might be difficult to iden-
sponding to electrons or water molecules) travel a tify and name the concept, principle, or formula
circuit and are pushed by a plunger (corresponding that subsumes an analog and its target, it is bene-
to a battery or a pump). The pinballs even meet ficial to do so for purposes of instruction, elabora-
"resistance" in the form of bumpers and funnels. tion, and creativity.
We suspect that other analogies, and more creative
ones, could be drawn by readers who keep our rep-
resentation and its constituent parts in mind when Analogies and Examples
thinking about circuits. Sometimes readers of a science text confuse an
Sometimes, there is not a conventional name for example of a concept with an analogy. An example
the superordinate concept which subsumes an ana- is an instance of a concept, not a comparison be-
log and a target. For example, authors of biology tween similar features of two concepts. Consider
texts frequently draw an analogy between the cam- the relationship between an electric spark and light-
era and the human eye. Here are some of the similar ning. Lightning is not like a big spark, it is a big
features: spark! So, lightning is an example of the concept
electric spark.
Camera Human Eye An analogy can be drawn between two examples
Lens Lens of the same concept (or the same superordinate con-
Inverted image Inverted image cept). Consider the analogy, "A whale is like a
Film Retina dolphin," used in children's science text by an au-
Lens cap Eye lid thor who wants to point out that both are examples
Focus Lens accommodation of sea mammals. Now refer to Figure 3 and consid-
Aperture Pupil dilation er a more elaborate analogy used by the author of a
high school physics text who wants to explain the
What is the name of the superordinate concept that concept of an electric field:
subsumes the analog "camera" and the target
"human eye?" Optical device is one candidate, but Just as the space around the earth and every other mass is filled
it certainly is not as familiar a name as circuit was in with an electric field-a kind of aura that extends through
space .... A gravitational force holds a satellite in orbit about a
the previous example. Some concepts, particularly planet, and an electrical force holds an electron in orbit about a
infrequently used general ones, do not have labels. proton. In both cases there is no contact between the objects, and
Nevertheless, the role of the superordinate concept the forces are "acting at a distance ... Putting this in terms of the
SA1ElLI1E ELEClRON ;!:e _

,-
......
I
/
"\ I
/
'\
f PLANET l' PROlON
, 0;
I
I (f) \ Figure 3. The satellite and the electron both experi-
ence forces; they are both in force fields. From Con-
I ceptual Physics (p. 496) by Paul G. Hewitt, 1987,
/ \
_-
I Menlo Park, CA: Addison-Wesley Publishing Com-
'\ / pany, Inc. Copyright 1987 by Addison-Wesley. Re-
./
....... printed by permission .
field concept, we can say that the orbiting satellite and electron increases as the number of significant, similar fea-
interact with the force fields of the planet and the proton and are tures shared by the analog and target increases.
everywhere in contact with these fields. In other words, the force
that one electric charge exerts on another can be described as the
However, it is possible to draw a "good" analogy
interaction between one charge and the electric field set up by the on the basis of a few (or even one!) similar features,
other. (Hewitt, 1987, p. 496) if those features are directly relevant to the specific
goals of the author. For example, the following
The superordinate concept in this analogy is field, analogy is drawn on the basis of only one similar
the analog is gravitational field (covered by the feature, but it is a very important one in terms of the
textbook author in a previous chapter), and the tar- author's goals: the earth rotates on its axis like an
get is electric field. The analogy is drawn by com- ice skater doing a pirouette.
paring similar features of the two kinds of fields: An analogy is considered "bad" if it is difficult
namely, (a) both. fields embody forces, (b) the to identify and map the important features that are
forces influence objects not in contact, and (c) the shared by the analog and the target. For example,
forces act at a distance. The analogy is also drawn we have seen cartoons in science textbooks in
by comparing a planet-satellite relationship, which which an analogy is drawn between the concept of
is an example of a gravitational field, with a pro- entropy and a teenager's messy room. Usually, the
ton-electron relationship, which is an example of author's point is that systems tend to become disor-
an electric field. These examples, and their compo- dered unless work of some kind is done on them.
nent parts, represent additional similar features on The author applies this principle to the universe.
which to base the analogy. In short, an analogy can Unfortunately, the reader is usually confused about
be drawn between two concepts by comparing ex- the causal agent involved. In the case of the messy
amples of those concepts. room, a teenager is responsible. But who is respon-
sible in the case ofthe universe? (One wag we know
claimed it was still teenagers!) Our point is that a
What Constitutes a Good Analogy? good analogy should not confuse the reader.
How good an analogy is depends upon the pur- If the purpose of the analogy is creative rather
pose for which it is used. In general, analogies can than explanatory, then the criteria by which it is
serve two purposes in science text: explanatory and judged should be based on the products of the analo-
creative. When an analogy is used for an explanato- gy rather than on characteristics of the analogy. In
ry purpose, the goal is for the reader to fully under- other words, when it comes to judging creativity,
stand the target by understanding the role each the "proof is in the pudding." The criteria that
important feature in the target plays. Explanatory should be applied to an analogy used for creative
analogies are instructional; they can be provided by purposes are: (1) the number of its products, (2) the
text authors and generated by the readers them- novelty of its products, and (3) the value of its
selves. products.
If an analogy is serving an explanatory purpose, If an analogy is required to serve an explanatory
then three criteria can be used to judge its appropri- function and a creative function, then both sets of
ateness: (1) the number of features compared, (2) criteria should be applied. An analogy that serves
the similarity of the features compared, and (3) the both functions well would be a very powerful analo-
conceptual significance of the features compared. gy indeed. It is often the case that an analogy can
The explanatory power of an analogy generally only serve one function well. A particular analogy
might be ideal for instructional purposes but might make things easier for the beginner to understand .... Yet say-
lead to no new ideas because no further correspon- ing the electric circuit is "just like" the water circuit does not
prove the electric circuit will have such behavior. "Current
dences can be found between features of the analog flow" is a hindsight description, put in after we have found that
and the target. Or an analogy might suggest new the electric circuit does have experimental properties which re-
ideas on the basis of one important correspondence, semble those of a water circuit. As such, it is good teaching, but
but other correspondences might be lacking, mak- if misused as an attempted proof, would be bad science. (This
ing the analogy a poor choice for authors who want seems a long complaint against a teacher's kindly illustration.
Yet it was the great mistake of medieval science to argue "what
to introduce and explain the target to readers. Thus, must be" from some authoritarian statement; and present-day
some analogies are good for both explanation and popularizers of science make the mistake of building hard-won
creativity, whereas others are good for only one of knowledge into glib analogies from which the science is then
these purposes. apparently produced. Attempts to make physics clearer by analo-
gies may mislead the reader unless he is warned.) (pp. 505-506)
An Analogy Is a Double-Edged Sword A recent recommendation for the use of analogies,

When used for either an explanatory or a creative with caution, can be found in Concepts in Physics,
purpose, an analogy can be a double-edged sword. an introductory physics textbook by Miller, Dillon,
It can lead to misunderstanding as well as under- and Smith (1980):
standing, and poor predictions as well as good ones.
Models and analogies can be of great value in physics if they are
The correspondence of features between two con- used with care and discrimination. It is important, for example,
cepts is never identical, otherwise the correspond- to guard against the danger of believing that a model or analogy is
ing concepts would be identical. When authors or an exact representation of some physical system. One should
readers intentionally or inadvertently compare fea- always regard a model critically and remember that an analogy
means no more than: under certain special conditions, the phys-
tures that do not correspond to one another, misun- ical system being studied behaves as if. ... (p. 253)
derstanding and misdirection can result.
A careful examination of all aspects of an analo- We encourage authors to explain to readers, as
gy is a prerequisite to using it effectively. All too Rogers (1960) and Miller et al. (1980) did, that
easily an analogy can lead readers astray. For exam- analogies are double-edged swords. An analog can
ple, children who believe that electricity in a wire is be used to explain correctly, even predict, some
like water in a hose often conclude, erroneously, aspects of the target; however, eventually every
that if the wire is cut, the electricity will "leak analogy breaks down. At that point, mis-
out. " comprehension and misdirection can begin. Read-
Of course, one way to guard against misunder- ers must understand this.
standing and misdirection simply is not to use anal-
ogies. There always have been a few scientists and
authors who believe that analogies are more trouble Explanatory Function of Analogies
than they are worth and that they should not be used
at all. Most scientists and authors do use analogies, An analogy serves an explanatory function when
despite the potential dangers, citing as their reason it puts new ideas into terms with which readers are
the significant explanatory and creative potential of already familiar. Frequently, authors will use a con-
analogies. cept that they covered in an earlier chapter to intro-
Occasionally, authors warn readers about the po- duce-by way of analogy-a new concept. Con-
tential dangers of analogies. This is a good practice sider the following example in which Paul Hewitt
and we would like to see it adopted by all authors (1987), the author of the text Conceptual Physics,
who use analogies. More than 25 years ago, Eric uses concepts related to the flow of heat (in his
Rogers (1960), in his classic text Physics for the Chapter 21) to introduce concepts related to the
Inquiring Mind, introduced his chapter on electric flow of electric charge (in his Chapter 34):
circuits by praising analogies and then cautioning Recall in your study of heat and temperature that heat flows
his readers about them: through a conductor when a difference in temperature exists
across its ends. Heat flows from the end of higher temperature to
To this day in teaching elementary electricity we liken electric the end of lower temperature. When both ends reach the same
circuits to hydraulic circuits of water pipes full of water all the temperature, the flow of heat ceases.
way around, with pumps, taps, flowmeters, pressure In a similar way, when the ends of an electrical conductor are
gauges ... to correspond to generators, switches, ammeters, at different electric potentials, charge flows from the higher
voltmeters. . .. Like many uses of analogy in teaching, this does potential to the lower potential. Charge flows when there is a
able, they are often misperceived (e.g., trajecto-

potential difference, or difference in potential (voltage), across
the ends of a conductor. The flow of charge will continue until
ries), or confused with everyday uses of the concept
both ends reach a common potential. When there is no potential
difference, no flow of charge will occur. (pp. 509-510)
(e.g., speed and velocity). Because of these charac-
teristics, physics concepts need more "conceptual
support" when introduced in textbooks. Analogies
In the sections that follow, we will present sever- can provide some of this conceptual support by
al more analogies and analogy-based problems de- helping readers to perceive, understand, or rein-
veloped by Paul Hewitt, as we consider his work to terpret physics concepts in terms of concepts with
be exemplary. Analogies drawn by an author be- which they are already familiar. Although we did
tween concepts covered early in a text and concepts find that elaborate analogies were most common in
covered later are particularly effective because the physics textbooks, it is important to keep in mind
author can be reasonably confident that the earlier that textbooks in other sciences, particularly chem-
concepts (which function here as analogs) are part istry, included elaborate analogies.
of most readers' knowledge base. Also, these anal- The 19 high school physics and physical science
ogies are particularly powerful because they prompt textbooks we examined are listed in Table 1. (phys-
readers to connect related concepts and form con- ical science is the study of changes in matter and
ceptual systems. energy; in addition to physics, it includes areas of
To better understand when and how an analogy study such as chemistry and astronomy.) We identi-
should be used to explain a new concept, we exam- fied the elaborate analogies in these texts that in-
ined 43 elementary school, high school, and college volved physics concepts. Some of the texts made
science textbooks for examples of elaborate analo- relatively extensive use of elaborate analogies,
gies. In all the textbooks, we found many examples whereas others made little or no use of them. It
of simple, one-or-two sentence analogies, such as would be difficult to quantify the amount of analogy
"Mitochondria are the powerhouses of the cell;" use in any meaningful way because the texts varied
however, elaborate analogies, those that ran on for considerably in their page length, the number of
a paragraph, a page, or several pages, were rela- concepts covered, the depth of concept coverage,
tively rare. Elaborate analogies compared and con- and the prior knowledge assumed of the reader. In
trasted many features of two major concepts. addition, some textbooks were designed for a termi-
High school physics and physical science text- nal course, whereas others were intended to be the
books appeared to contain the greatest number of first course in a science sequence.
elaborate analogies, so we focused our attention on It is common for the teacher's edition of high
them. Why do high school physics and physical school science textbooks to point out, in the intro-
science textbooks have a greater number of elabo- duction, all the special features that are incorporat-
rate analogies? We have formed some opinions, but ed into the design of the text. These features are
keep in mind they are very speCUlative. We think intended to facilitate readers' comprehension of
that elementary school textbook authors use few text concepts. Here is a sample of these features:
elaborate analogies because the knowledge base of advance organizers; structured overviews; high-
their readers is so limited and because it is not possi- lighted concept names; margin notes; introductory,
ble to cover concepts in the depth needed to support adjunct, and review questions; illustrations; car-
analogies. On the other hand, college textbook au- toons; boxed examples; concept summaries; lists of
thors do use elaborate analogies, but not as often as important terms; conceptual activities; and glossa-
high school textbook authors. College textbook au- ries. These features, along with supplementary ma-
thors may assume that the stronger comprehension terials, such as laboratory manuals, resource books,
capabilities of college students preclude the need videotapes, software, and test item files, are pro-
for as many analogies. College textbook authors moted in the teacher's edition as valuable aids to
may also be under pressure to cover more material, comprehension. Often, the promotion is by an ex-
forcing them to sacrifice some elaborate analogies. pert in reading comprehension, such as the head of a
Why do physics textbooks appear to have more college or university Department of Reading Edu-
elaborate analogies than textbooks in other sci- cation. We believe features such as these can be
ences? Again, we are speculating here. Physics valuable aids to comprehension, under certain con-
concepts are about very basic phenomena that either ditions, as many research studies have shown (e.g.,
do not lend themselves to unaided physical observa- Britton & Glynn, 1987; Britton, Glynn, Meyer, &
tion (e.g., electrons) or, if they are directly observ- Penland, 1982; Glynn, Andre, & Britton, 1986;
Table 1. Physics and Physical Science Glynn & Britton, 1984; Glynn, Britton, & Tillman,
Textbooks 1985; Glynn & Di Vesta, 1977, 1979). We were
surprised to find, however, no mention of analogies
Beiser, A. (1964) The science ofphysics. Palo Alto, CA: in any of the textbook introductions. This was dou-
Addison-Wesley. bly surprising in those textbooks in which authors
Dull, C. E., Metcalfe, H. c., & Williams, J. E. (1963). made excellent use of analogies. Why not promote
Modern physics. New York: Holt, Rinehart & Win- this valuable aid to comprehension in the introduc-
ston. tion to the text?
Genzer, I., & Younger, P. (1969). Physics. Morris- We think that one reason why analogies are not
town, NJ: Silver Burdett. promoted in introductions to textbooks, even in
Haber-Schaim, U., Dodge, 1. H., & Walter, J. A. textbooks that make extensive use of analogies, is
(1986). PSSC Physics. Lexington, MA: D.C. Heath. that the skill of writing good analogies is, at the
Harnwell, G. P., & Legge, G. J. F. (1967). Physics: present time, what psychologists call' 'procedural"
Matter, energy, and the universe. New York: Rein- rather than "declarative." Procedural knowledge
hold. is knowledge of "how to do something" rather than
"how to explain it in words." Because authors and
Hewitt, P. G. (1987). Conceptual physics. Menlo Park,
publishers do not have guidelines or a model for
CA: Addison-Wesley.
what constitutes a good instructional analogy, the
Hirsch, A. J. (1981). Physics: A practical approach. development and evaluation of analogies is fairly
Toronto: Wiley. subjective. In other words, it is more of an art than a
Hirsch, A. 1. (1986). Physics for a modern world. technology. Because the use of analogies is pro-
Toronto: Wiley. cedural and subjective, authors and publishers may
Johnson, G. P. Barr, B. B., & Leyden, M. B. (1988). not feel comfortable about promoting them in their
Physical science. Menlo Park, CA: Addison-Wesley. textbooks as aids to comprehension, at least not
Karplus, R. (1969). Introductory physics. New York: until a model becomes available for designing and
W. A. Benjamin, Inc. evaluating instructional analogies.
Krauskopf, K. B., & Beiser, A. (1986). The physical
universe. New York: McGraw-Hill.
Leyden, M. B., Johnson, G. P., & Barr, B. B. (1988). The Teaching-with-Analogies (TWA)
Introduction to physical science. Menlo Park, CA: Model
Addison-Wesley.
Miller, F., Dillon, T. J., & Smith, M. K. (1980). Con- With the help of members of the faculty in the
cepts in physics. New York: Harcourt Brace Jovano- University of Georgia's Department of Science Ed-
vich. ucation and high school science teachers who were
Murphy, J. T., & Smoot, R. C. (1972). Physics princi- enrolled as graduate students in that department, we
ples and problems. Columbus, OH: Charles E. Mer- reviewed a large number of high school physics,
rill. physical science, chemistry, and biology textbooks
Pasachoff, J. M., Pasachoff, N., & Cooney, T. M. and subjectively identified the most effective analo-
(1983). Physical science. Glenview, IL: Scott, Fores- gies from the standpoint of instructional design.
man. The authors of these analogies performed certain
Physical Science Study Committee. (1965). Physics. key operations that we have incorporated into a
Lexington, MA: D. C. Heath. model that can serve as a guide for authors of sci-
ence textbooks. Our model for using an analogy to
Rutherford, F. J., Holton, G., & Watson, F. G. (1975).
explain a science concept contains the following six
Project physics. New York: Holt, Rinehart & Win-
operations:
ston.
Taffel, A. (1986). Physics: Its methods and meanings. 1. Introduce target concept
Newton, MA: Allyn & Bacon. 2. Recall analog concept
Wong, H. K., & Dolmatz, M. S. (1984). Physical sci- 3. Identify similar features of concepts
ence: The key ideas. Englewood Cliffs, NJ: Prentice- 4. Map similar features
Hall. 5. Draw conclusions about concepts
6. Indicate where analogy breaks down
Figure 4. (Left) Water flows from the reservoir of higher pressure to the reservoir of lower pressure. The flow will cease when the
difference in pressure ceases. (Right) Water continues to flow because a difference in pressure is maintained with the pump. From
Conceptual Physics (p. 510) by Paul G. Hewitt, 1987, Menlo Park, CA: Addison-Wesley Publishing Company, Inc. Copyright 1987
by Addison-Wesley. Reprinted by permission.
The author of a well-designed analogy performs vide a sustained potential difference. Something that provides a
each of these operations for the reader. To illustrate potential difference is known as a voltage source. (p. 511)
these operations, we present excerpts from Paul There is often some confusion between charge flowing
through a circuit and voltage being impressed across a circuit.
Hewitt's (1987) chapters on "Electric Currents" To distinguish between these ideas, consider a long pipe filled
and "Electric Circuits," in which he draws an ex- with water. Water will flow through the pipe if there is a dif-
tended analogy between currents of water (the ana- ference in pressure across or between its ends. Water flows from
log) and currents of electricity (the target): 1 the high-pressure end to the low-pressure end. Only the water
flows, not the pressu·re. Similarly, you say that charges flow
through a circuit because of an applied voltage across the circuit.
Electric Current You don't say that voltage flows through a circuit. Voltage
The last chapter discussed the concept of electric potential, or doesn't go anywhere , for it is the charges that move . Voltage
voltage. This chapter will show that voltage is an "electrical causes current. (pp. 511-512)
pressure" that can produce a flow of charge, or current. within a
conductor. The flow is restrained by the resistance it encounters. Electrical Resistance
(p. 509)
The amount of current that flows in a circuit depends on the
voltage provided by the voltage source. It also depends on the
Flow of Charge resistance that the conductor offers to the flow of charge, or
To attain a sustained flow of charge in a conductor, some ar- the electrical resistance. This is similar to the rate of water flow
rangement must be provided to maintain a difference in potential in a pipe, which depends not only on the pressure behind the
while charge flows from one end to the other. The situation is water but on the resistance offered by the pipe itself. The re-
analogous to the flow of water from a higher reservoir to a lower sistance of a wire depends on the conductivity of the material
one [Figure 4 left]. Water will flow in a pipe that connects the (that is, how well it conducts) and also on the thickness and
reservoirs only as long as a difference in water level exists. (This length of the wire. (p. 512)
is implied in the saying, "Water seeks its own level. ") The flow
of water in the pipe, like the flow of charge in the wire that A Battery and a Bulb
connects the Van de Graaff generator to the ground, will cease It is a bit misleading to say that electrons flow "out of" the
when the pressures at each end are equal. In order that the flow be battery, or "into" the bulb; a better description is to say they
sustained, there must be a suitable pump of some sort to maintain flow through these devices. The flow of charge in a circuit is
a difference in water levels [Figure 4 right] . Then there will be a analogous to a pump, the wires to the pipes, and the bulb to any
continual difference in water pressures and a continual flow of device that operates when the water is flowing . The water flows
water. The same is true of electric current. (pp. 509-510). through both the pump itself and the circuit it connects. It doesn't
"squash up" and concentrate in certain regions, but flows con-
Voltage Sources tinuously. Electric current behaves the same way. (pp. 525-526)
Charges do not flow unless there is a potential difference. A
sustained current requires a suitable "electrical pump" to pro- Electric Circuits
Any path along which electrons can flow is a circuit. For a
IThe excerpts are taken from Conceptual Physics by Paul G. continuous flow of electrons, there must be a complete circuit
Hewitt, 1987, Menlo Park, CA: Addison-Wesley Publishing with no gaps. A gap is usually provided by an electric switch that
Company, Inc. Copyright 1987 by Addison-Wesley. Reprinted can be opened or closed to either cut off or allow electron flow.
by permission. The water analogy is quite useful for gaining a conceptual
understanding of electric circuits, but it does have some limita- likelihood that his readers will overgeneralize from
tions. An important one is that a break in a water pipe results in the analog to the target concept and form some
water spilling from the circuit, whereas a break in an electric
misconceptions.
circuit results in a complete stop in the flow of electricity. An-
other difference has to do with turning current off and on. When Hewitt's use of analogies is excellent. He per-
you close an electrical switch that connects the circuit, you allow formed all of the six operations specified in our
current to flow in much the same way as you allow water to flow model. If an author were to perform only some of
by opening a faucet. Opening a switch stops the flow of elec- the operations, leaving some to the reader, the tech-
tricity. An electric circuit must be closed for electricity to flow.
Opening a water faucet, on the other hand, starts the flow of
nical quality of the analogy would be reduced be-
water. Except for these and some other differences, thinking of cause it is possible that the reader might fail to
electric current in terms of water current is a useful way to study perform an operation or might perform it poorly.
electric circuits. (pp. 526-527) In our survey of textbooks of physics and of
physical science, we have found many instances
In the analogy that Hewitt has drawn between where authors suggested an analog to readers, but
currents of water and currents of electricity, he ef- then left the readers to make sense (or nonsense) out
fectively has performed all the operations specified of it for themselves. Under these circumstances, the
in our model. In the electric current excerpt, he readers could identify irrelevant features of the tar-
introduces the target concepts of voltage, current, get and analog, map them, draw wrong conclusions
and resistance. Next, in the flow of charge excerpt, about the target, and fail to realize where the analo-
he cues the reader to retrieve from memory infor- gy breaks down. In short, the readers' understand-
mation about currents of water. Hewitt realizes that ing of the target concept could be both incomplete
some readers will have only fragmentary knowl- and incorrect.
edge of a water flow system, so he provides a good
deal of information about such a system. In the flow
of charge excerpt, Hewitt identifies and maps sim- The Science Teacher's Use of the Model
ilar features: "The flow of water in the pipe, like Science teachers who keep our model in mind
the flow of charge in the wire. . .. " Plus, his use of can interpret text analogies for their students. If the
water circuit and electric circuit diagrams (refer author fails to perform some of the operations spec-
back to Figure 2) helps readers to visually identify ified in the model, the teacher can perform these
and map similar features. operations for the students. For example, suppose
In addition to explaining new concepts, analo- the author of a physical science text draws an analo-
gies can be used to clear up old misconceptions. In gy between electricity in a wire and water in a hose
the excerpts about "voltage sources" and "a bat- but fails to point out where this analogy breaks
tery and a bulb," Hewitt does exactly this. He uses down: namely, that electricity does not spill out if
his analogy to counter some popular misconcep- the wire is cut. The teacher can remedy this failure
tions that readers either already have or are likely to on the author's part by performing this essential
develop. It is important to keep in mind that analo- operation for the students.
gies can be used to straighten out readers' miscon- We encourage science teachers to familiarize stu-
ceptions, particularly because one of the most pop- dents with this model so that the students willleam
ular arguments against using analogies is that they how to interpret, criticize, and extend an author's
can sometimes cause misconceptions. analogy. When we used our model in science edu-
Hewitt draws conclusions for readers about the cation classes to dissect an author's analogy, the
target concepts. For example, in the electrical re- discussions that ensued between the teacher and
sistance excerpt, he compares electrical resistance students, and the students themselves, had some
to the resistance water encounters in a pipe and wonderful consequences. The discussion of the
draws the following conclusion for the reader: analogy enabled the teacher to identify students'
"The resistance of a wire depends on the misconceptions and knowledge gaps which other-
conductivity of the material . . . and also on the wise would have gone undetected. The discussion
thickness and length of the wire. " also prompted students to tie together concepts that
Finally, in the electrical circuit excerpt, Hewitt they previously viewed as unrelated.
indicates where the analogy breaks down. He ex- We also encourage science teachers and students
plains how a break in an electric circuit differs from to use the model as a guide when generating their
that in a water pipe and how an electrical switch own analogies. They may wish to do this when the
differs from a faucet. By doing so, he reduces the text author has not provided an analogy. Or, if the
author has provided an analogy, teachers and stu- vity::Darwin: _ _ ) investigated by Sternberg
dents may wish to generate another so that they can (1977, 1986) and Rumelhart (Rumelhart & Abra-
examine the concept from more than one perspec- hamson, 1973; Rumelhart & Norman, 1981). In the
tive. For example, the author might use a "flowing four-term problems, there is a built-in prompt: the
water" analogy to explain electricity. The teacher student knows that the A-B relationship is relevant
and students, however, wishing to view electricity and must be mapped to the C-D relationship. With
from a different perspective, might generate a Gick and Holyoak's story problems, in effect the
"moving crowd" analogy, in which electric cur- student is given "C," the problem statement, and
rent is viewed as masses of people, animals, or asked for "D," the problem solution. To solve the
objects moving through passageways. Consider problem, the student must notice that some appar-
this version of a moving crowd analogy generated ently unrelated information ("A-B "), received ear-
by a student: "If you increase resistance in the cir- lier in the experiment, is relevant to the problem and
cuit, the current slows down. Now that's like a must apply it without any prompt or hint. Regarding
highway, cars on a highway where . . . as you the difficulty of problems such as these that require
close down a lane. . . the cars move slower the spontaneous recognition and application of po-
through that narrow point" (Gentner & Gentner, tentially relevant knowledge from a different do-
1983, p. Ill). main, Gick and Holyoak (1983) said:
The advantage of generating alternative analo-
gies and viewing a concept like electricity from It should be noted that in all of our experiments the critical prior
more than one perspective is that each perspective analogs were presented in a context in which their problem-
oriented character was incidental. Subjects were never explicitly
brings particular features of the concept into clearer encouraged to use the stories to learn about a novel kind of
focus. Thus, the teacher and the students who gen- problem. In many situations, such as an instructional context,
erate multiple analogies for a concept will have a more directive guidance in the application of an analogy is often
more comprehensive understanding of that concept given. It is quite likely that more intentionalleaming procedures
and of its relation to other concepts. could improve transfer performance in our paradigm. In particu-
lar, explicit guidance might facilitate transfer from a single ana-
log. In the absence of such guidance, failure to derive a general
schema from a single instance may only reflect appropriate con-
servatism; without either further examples or direct instruction,
Creative Function of Analogies the person may have no principled way to isolate the essential
causal aspects of the situation.
Given the difficulty of schema abstraction from a single ana-
Analogies serve a creative function when they log (at least without the guidance of a teacher), one might ask
help readers to solve problems, find problems, and how anyone could spontaneously notice an analogy between one
generate hypotheses. Authors prompt readers to use initial analog and a semantically remote transfer problem. (p. 32)
analogies creatively by means of questions, hints,
and illustrations. In addition, readers can prompt Gick and Holyoak's findings suggest that text
themselves to use analogies creatively if they are authors should play an active role in helping their
trained to ask certain questions of themselves and readers use analogies to solve problems, find prob-
perform certain analogical operations. lems, and generate hypotheses. We will now exam-
These prompts are essential because spontaneous ine some of the ways that authors help readers to use
analogical problem solving is not common. In a analogies creatively.
series of experiments, Gick and Holyoak (Gick &
Holyoak, 1980, 1983; Holyoak, 1985) demon-
strated that 75% of the college students they tested Analogical Problem Solving
were able to solve a story problem by applying
previously learned information after they received a Authors of science textbooks use questions and
hint to apply it, but only about 30% were able to hints to prompt readers to solve problems by means
solve the problem without a hint. of analogies. Read the following excerpts from
The problems used by Gick and Holyoak were Hewitt's (1985a) Conceptual Physics and then
embedded in stories about military, medical, fire- tackle the problem posed. The temperature and heat
fighting, and other real-life situations. These prob- excerpt was extracted from his Chapter 13, and the
lems are more difficult than the four-term problems electrical potential excerpt and the problem posed
of the form A:B::C:D (e.g., Einstein:Relati- were extracted from his Chapter 20.
Thermal Energy and Temperature readers generate their own analogies and use them
In this and the following two chapters we are going to investigate to solve problems in a science text? Certainly! But
more closely the effects of the chaotic and haphazard motion of readers must know what these problem-solving op-
atoms and molecules that we call thermal motion. We begin by erations are and how to use them. Many, if not
considering that which a body has by virtue of this energetic
most, readers will require training in analogical
motion: thermal energy . ...
The quantity that tells how warm or cold something is with problem solving. To provide a basis for this train-
respect to a standard body is called temperature. We say tem- ing, we will now propose a model of analogical
perature is a measure of the random translational motion of atoms problem solving.
and molecules in a body; more specifically, it is a measure of the
average kinetic energy of atoms and molecules in a body. We
know, for example, that there is twice the thermal energy in 2 The Analogical Problem Solving (APS) Model
liters of boiling water as in I liter of boiling water, because 2
liters of boiling water will melt twice as much ice as I liter. But For the kinds of problems readers often encoun-
the temperatures of both amounts of water are the same because ter in science texts, we propose a model of analo-
the average kinetic energy of molecules in each is the same. So
we see there is a difference between thermal energy, which is
gical problem solving involving three major
measured in joules, and temperature, which we measure in de- operations, each of which has two suboperations.
grees. (pp. 220-221) Here is our APS model:
Electric Potential Energy and Electric Potential 1. State problem
Rather than dealing with the total potential energy of a charged A. Represent problem
body, it is convenient when working with electricity to consider B. Identify important features
the electric potential energy per charge. We simply divide the
amount of energy in any case by the amount of charge .... The
2. Retrieve analog
concept of potential energy per charge is called electric poten- A. Search for analog with similar features
tial; that is, B. Select analog
. . Energy 3. Solve problem
Electnc potential = Charge (p. 333)
A. Map solution
Problem B. Verify solution
Can we say that a body with twice the electric potential energy of
another has twice the electric potential? Why or why not? (Hint: Let us apply this model to the electric potential
Consider the analogy with thermal energy and temperature.) energy and electric potential problem we just
(pp. 337) looked at. In the state-the-problem operation, read-
ers state the problem in familiar terms. To do this
Answer and Explanation
they must perform two suboperations, the first of
We cannot say that a body with twice the electric potential energy which is to form a mental representation of the
of another has twice the electric potential, just as we cannot say
problem. Problems can be represented as diagrams,
that a body with more thermal energy than another has a higher
temperature. For example, a barrel full of warm water may have or sets oflogical statements, or algebraic equations,
considerably more thermal energy than a cup of hot water. But it or geometric figures, or simpler cases, or as exam-
doesn't have a higher temperature, or greater KE/molecule. ples, or in other ways. In fact, the same problem can
Likewise, the body with twice the electric potential energy (PE) often be represented in several ways. In the electric
does not necessarily have the greater PEl charge. It is important
to distinguish between electric potential energy (PE) and electric
potential energy and electric potential problem, it is
potential (pEl charge). The key difference between the two is important to represent electric potential as electric
"per charge. " (Hewitt, 1985b, p. 156) potential energy per charge (pE/charge). It is also
important to recognize that the electric potential
As the preceding excerpts illustrate, readers can energy of one body is to be compared with the
be prompted to use analogies to connect related electric potential of another body. If a reader did not
bodies of know ledge for purposes of solving a prob- attend to this feature of the problem, he or she might
lem. In this illustration, the author simply cued think the problem is asking about what the effect of
readers to retrieve the appropriate analog from doubling the electric potential energy would be on
memory. If the author wished to provide the reader the electric potential of the same body.
with further prompts, he also could have identified In the second suboperation, readers identify the
the important features of the analog and target and important features of the problem. The key features
mapped them for the reader. of this problem are the relationships between the
In the absence of author-provided analogies, can two central concepts, electric potential energy and
electric potential. Both have to do with electric en- pIe for the electric potential energy and electric po-
ergy. More precisely, electric potential energy re- tential problem: a charged 12-volt automobile bat-
fers to the total potential energy of a charged body, tery may have a great deal more electrical potential
whereas electric potential refers to an average, the energy than a balloon rubbed on a hairy head, but it
electric potential energy per charge. This average is does not have more electric potential. In fact, the
important because it permits a definite value for balloon might be charged to several thousand volts!
electric potential energy per charge (also called However, it has only a tiny amount of electric po-
"voltage") to be assigned to a location, say, in an tential energy, because the charge it carries is less
electric circuit, regardless of whether or not a than one millionth of a coulomb.
charge exists at that location. The final suboperation is to verify the solution
Analog retrieval is the readers' next major opera- generated. For the electric potential energy and
tion. It consists of two suboperations: searching for electric potential problem, this can be done by test-
an analog with similar features and selecting an ing the solution by means of actual experiments.
analog for evaluation. In the electric potential ener- For each of the six operations in our analogical
gy and electric potential problem, the author per- problem-solving model, there is associated a poten-
formed both of these sUboperations for the readers tial error that can lead readers to either no solution
by "hinting" that the relationship between thermal or to an incorrect solution. First, readers could mis-
energy and heat is analogous to that between elec- represent the problem. Second, even if readers cor-
tric potential energy and electric potential. Had he rectly represent a problem, it is possible for them to
not done that, readers would have had to search miss an important feature of the problem. Third, the
their memories for analog candidates. They would search for an analog could prove fruitless because
recognize these candidates on the basis of similar readers did not search certain domains. For exam-
features and select the best candidates for evalu- ple, with the electric potential energy and electric
ation. potential problem, if readers had searched only the
It is hoped that readers would have recognized domain of mechanics instead of heat, they might
that some of the relationships between thermal en- not have found a useful analog. Fourth, readers
ergy and temperature are similar to those between might select and use a relatively poor analog simply
electric potential energy and electric potential. because it was the first one they encountered in their
Thermal energy and temperature also have to do search. Fifth, readers might select a good analog,
with energy. More precisely, thermal energy refers but fail to correctly map important features of the
to the total random kinetic energy of atoms and analog to the solution of the problem. Finally, read-
molecules in a body, whereas temperature is a mea- ers might fail to verify that the solution they mapped
sure of the average kinetic energy of atoms and is, in fact, a valid solution.
molecules in a body. If, at any operation in the model, readers detect
The final operation, solving the problem, con- that they have made an error, it is possible for them
sists of two suboperations, the first of which is map- to cycle back to an earlier operation and correct
ping a solution. If readers know that a body with their error. For example, a solution that cannot be
more thermal energy than another does not neces- verified may have been mapped incorrectly, so
sarily have a higher temperature, then they can map readers should reexamine their solution mapping. If
this information onto the problem as a tentative a solution cannot be mapped, this may be the result
solution. However, even if the readers do not have of a poor analog, so readers should select a new
this piece of knowledge immediately available, it is analog. Ifreaders are unable to find a good analog,
still possible for them to induce it and then map this it may be the result of searching the wrong domains,
induction to solve the problem. so readers should examine other domains. If readers
One way to induce a solution is by generating and have difficulty finding the right domains, it may be
mapping an example of the analog to an example of because they have failed to identify important fea-
the target. In this case, for thermal energy and tem- tures of the problem or they have focused on unim-
perature, a reader might generate an example such portant features. And, finally, if readers are not
as this: a huge pot of warm soup may have a great identifying important features, it may be because
deal more thermal energy than a bowl of hot soup, they have misrepresented the problem, and if so,
but it does not have a higher temperature. Then, by they should try some alternative ways of looking
way of analogy, the reader might generate an exam- at it.
The Science Teacher's Use of the Model and hypotheses, looking for relations between con-
cepts where none exist (as far as we know). Critics
We recommend that science teachers train their
will say that the resulting relations will be mean-
students to use this model of analogical problem
ingless, or worse still, confusing to readers.
solving. With it, the students can plan, monitor,
We agree that the relations made between con-
evaluate, and remediate their problem-solving
cepts will be meaningless and confusing at times,
efforts.
but at other times valid relations will be made. We
Science teachers should look for opportunities
believe the process of connecting concepts is
during lectures, demonstrations, and laboratories to
important and must be nourished in readers. To
"model the model. " They should point out to their
reject the process because it sometimes leads to
students when an analogy might be used to gain
meaningless or incorrect relations is like throwing
insight into a problem. The teachers should also
the baby out with the bath water.
demonstrate how to apply the operations in the
Critics of analogies often believe that a concept,
model to the problem at hand.
particularly a science concept, should be learned in
Students should be shown how to criticize the
isolation (from other concepts) so that it is not con-
analogies they generate, as well as those generated
fused with the other concepts. This is a naive view
by the teacher and the text author. Those students
of how human learning occurs. The process of relat-
who learn how to apply the model for themselves
ing concepts by means of analogy is inherent to
will have a powerful intellectual tool they can bring
human cognition; in effect, analogical reasoning is
to bear on complex science problems.
"hard wired" and it is unreasonable to expect read-
ers not to use it. They will use it. Our concern is that
Problem Finding and Hypothesis Generation they use it effectively. One way to ensure they use it
effectively is to train them in the analogical prob-
Analogies can help readers to discover new problem-solving model we presented in the last section,
lems and hypothesize about their solutions. Authors ensuring that readers understand each of the opera-
prompt the problem-finding and hypothesis-gener- tions, the errors that can result from the operations,
ating processes of their readers by posing analogical and the ways to remediate those errors.
questions in and at the end of chapters. Here are Recently, it has been demonstrated experimen-
some examples, again from Conceptual Physics by tally that analogies function as aids to scientific
Hewitt (1987), in which questions are posed which thought. Gentner and Gentner (1983) showed that
prompt problem finding and hypothesis generation college and high school students who used a "flow-
in readers. The questions are based upon an analogy ing fluid" analogy of electricity made different
drawn between gravity and electricity: problem-solving inferences than students who used
a "moving crowd" analogy. These findings sup-
* It is said that a gravitational field, unlike an electric field,
cannot be shielded. But the gravitational field at the center port the view that analogies can serve as important
of the earth cancels to zero. Isn't this evidence that a grav- sources of insight.
itational field can be shielded? Many prominent scientists have emphasized the
* How are a gravitational and an electric field similar? value of analogical reasoning in scientific insight
* How is an electric field different from a gravitational field?
and discovery. For example, Johannes Kepler, the
* The vectors for the gravitational field of the earth point
toward the earth; the vectors for the electric field of a proton eminent seventeenth-century astronomer, wrote:
point away from the proton. Explain. (pp. 501, 507-508) "And I cherish more than anything else the Analo-
gies, my most trustworthy masters. They know all
In addition to responding to the author's ques- the secrets of Nature, and they ought to be least
tions, we believe that readers can and should be neglected in Geometry" (quoted in Polya, 1973, p.
trained to generate such questions for themselves; 12). Kepler attempted to explain astronomical phe-
that is, to discover their own problems rather than nomena, such as the motion of the planets, in terms
rely entirely upon the text author to pose them. of precise physical laws. In a letter he wrote in
Having discovered their own problems, the readers 1605, he drew an analogy between planetary mo-
can then proceed to hypothesize about possible so- tion and clockwork:
lutions. This is what it means to read creatively!
Critics of analogies will point out that readers I am much occupied with the investigation of the physical
will sometimes generate unreasonable problems causes. My aim in this is to show that the celestial machine is to
be likened not to a divine organism but rather to a clock- that analogies, serving as models, can be valuable
work ... insofar as nearly all the manifold movements are car- aids to scientific explanation and discovery.
ried out by means of a single, quite simple magnetic force, as in
the case of a clockwork, all motions are caused by a simple
weight. Moreover, I show how this physical conception is to be
presented through calculation and geometry. (Quoted in Ruther-
ford, Holton, & Watson, 1975, Unit 2, p. 68)
Future Research
We have proposed models for using analogies to

Joseph Priestley (1773-1804), a distinguished
explain concepts and solve problems in science
physical scientist, also found analogies to be useful
texts. These models were derived from an analysis
problem-solving tools. Priestley proposed the "law
of science texts, particularly physics and physical
of electrical force," which was later verified ex-
science texts. It is essential, however, that em-
perimentally by the French physicist Charles Cou-
pirical studies be conducted to validate these mod-
lomb. Here are the events that led to Priestley's
els. We have initiated a program of studies to do just
proposal:
that. We hope to attract the interest of other re-
Priestley verified Franklin's results, and went on to reach a bril- searchers who will study and extend these models.
liant conclusion from them. He remembered from Newton's We have encouraged science teachers to famil-
Principia that gravitational forces behave in a similar way. In- iarize their students with these models. Systematic
side a hollow planet, the net gravitational force on an object (the procedures for training students in the use of these
sum of all the forces exerted by all parts of the planet) would be
exactly zero. This result also follows mathematically from the
models should be developed, and these procedures
law that the gravitational force between any two individual should be validated in classroom settings. Eth-
pieces of matter is inversely proportional to the square of the nographic studies of teacher-student-textbook in-
distance between them. Priestley therefore proposed that forces teractions that involve analogies would be particu-
exerted by charges vary inversely as the square of the distance, larly appropriate.
just as do forces exerted by massive bodies. . . . We call the
force exerted between bodies owing to the fact that they are Because the models we proposed were derived
charged "electric" force, just as we call the force between un- from an analysis of science texts, it is not clear to
charged bodies "gravitational" force. . . . what extent these models generalize across the cur-
Priestley's proposal was based on reasoning by analogy, that riculum to disciplines other than science. We be-
is, by reasoning from a parallel, well demonstrated case. Such
lieve the models might generalize to other disci-
reasoning alone could not prove that electrical forces are inverse-
ly proportional to the square of the distance between charges. But plines; however, textbook surveys and empirical
it strongly encouraged other physicists to test Priestley's hypoth- studies must be conducted to verify this.
esis by experiment. (Rutherford et al., 1975, Unit 4, p. 35) Finally, a great deal more research is needed on
ways to facilitate analogical reasoning and problem
Perhaps the strongest statement in support of solving. More information is needed on how indi-
analogies as aids to insight and discovery in science vidual differences in cognitive development and
is that of the English physicist N. R. Campbell who, content area knowledge influence the generation
in his 1920 book Physics, the Elements, pointed to and use of analogies. More information is also
the billiard ball model of the kinetic theory of gases needed concerning the circumstances under which
and argued: analogies can be counterproductive, resulting in in-
creased confusion instead of comprehension.
Analogies are not "aids" to the establishment of theories; they
are an utterly essential part of theories, without which theories
would be completely valueless and unworthy of the name. It is
often suggested that the analogy leads to the formulation of the Summary and Conclusions
theory, but that once the theory is formulated the analogy has
served its purpose and may be removed or forgotten. Such a
suggestion is absolutely false and perniciously misleading. To promote meaningful comprehension, science
(Quoted in Hesse, 1966, p. 4) texts should incorporate features that help readers to
relate text content to what they already know. One
Testimonials to the creative potential of analo- of the most effective ways for readers to integrate
gies are common in the science literature. These their existing knowledge with text content is to do
testimonials can be found in the personal records of so by means of analogy. Authors can provide analo-
prominent scientists as well as in studies of the gies for readers, and readers can be trained to gener-
history of science (e.g., Darden, 1980; Hesse, ate their own analogies.
1966). Taken together, these testimonials suggest In a science textbook, analogies can serve both
an explanatory and a creative function. They serve analogy is inherent to human cognition; in effect,
an explanatory function when they put new con- analogical reasoning is "hard-wired," and it is un-
cepts and principles into familiar terms. They serve reasonable to expect readers not to use it. They will
a creative function when they stimulate the solution use it. Our concern is that they use it effectively as
of existing problems, the identification of new an aid to understanding and as an insight in science.
problems, and the generation of hypotheses. We have proposed models for using analogies to
In science texts, analogies are used to transfer explain concepts and solve problems in science
ideas from a familiar conceptual domain (i.e., the texts. Our future research will focus on validating
analog) to an unfamiliar one (i.e., the target). In our these models and developing procedures for train-
representation of an analogy, the analog and the ing science students to use them effectively.
target are subordinate to a superordinate concept (or
principle or formula). The identification and nam- ACKNOWLEDGMENTS. We wish to thank Rus-
ing of the superordinate concept can suggest other sell Yeany, Chairman of Science Education at the
analogies; it can also stimulate readers to generalize University of Georgia, for surveying textbooks of
what they have learned and apply their learning to high school physics and of physical science with us.
other contexts. His comments, along with those of Daryl Adams,
When used for either an explanatory or a creative Juliet Allan, Alphonse Buccino, Michael Hale, Pa-
purpose, an analogy can be a double-edged sword. trick Kyllonen, Michael Padilla, Joseph Riley, and
An analog can be used to explain correctly, even Paula Schwanenflugel, on the ideas presented in
predict, some aspects of the target; however, at this chapter, were invaluable.
some point every analogy breaks down. Authors
must therefore make readers aware that, at that
point, miscomprehension and misdirection may
begin. References
Analogies drawn by an author between concepts
that are covered early in a text and concepts covered Britton, B. K., & Glynn, S. M. (1987). Executive control pro-
later are particularly effective because the author cesses in reading. Hillsdale, NJ: Erlbaum.
can be sure that the earlier concepts (which function Britton, B. K., Glynn, S. M., Meyer, B. J. F., & Penland, M. J.
(1982). Effects of text structure on use of cognitive capacity
as analogs) are part of every reader's knowledge during reading. Journal of Educational Psychology, 74, 51-
base. These analogies prompt readers to connect 61.
related text concepts and form conceptual systems. Darden, L. (1980). Theory construction in genetics. In T. Nick-
Textbooks of high school physics and of physical lIes (Ed.), Scientific discovery: Case studies (pp. 151-170).
Dordrecht, The Netherlands: D. Reidel.
science appeared to contain the most elaborate anal- Gentner, D., & Gentner, D. R. (1983). Flowing waters orteem-
ogies, so we focused our attention on them. We ing crowds: Mental models of electricity. In D. Gentner & A.
identified the most effective analogies from the L. Stevens (Eds.), Mental models (pp. 99-129). Hillsdale,
standpoint of instructional design. The authors of NJ: Erlbaum.
these analogies performed certain key operations Gick, M. L., & Holyoak, K. J. (1980). Analogical problem
solving. Cognitive Psychology, 12, 306-355.
that we incorporated into a model. Our model can Gick, M. L., & Holyoak, K. J. (1983). Schema induction and
serve not only as a guide for authors of science analogical transfer. Cognitive Psychology, 15, 1-38.
textbooks, but also as a guide for science teachers Glynn, S. M., & Britton, B. K. (1984). Supporting readers'
and students who wish to interpret, criticize, and comprehension through effective text design. Educational
Technology, 24(10), 40-43.
extend an author's analogy or one of their own. Glynn, S. M., & Di Vesta, F. J. (1977). Outline and hierarchical
Analogies serve a creative function when they organization as aids for study and retrieval. Journal of Educa-
help readers to solve problems, find problems, and tional Psychology, 69, 89-95.
generate hypotheses. Authors prompt readers to use Glynn, S. M., & Di Vesta, F. J. (1979). Control of prose pro-
analogies creatively by means of questions, hints, cessing via instructional cues. Journal of Educational Psy-
chology, 71, 595-603.
and illustrations. In addition, readers can prompt Glynn, S. M., Britton, B. K., & Tillman, M. H. (1985). Ty-
themselves to use analogies creatively if they are pographical cues in text: Management of the reader's atten-
trained to ask certain questions of themselves and tion. In D. H. Jonassen (Ed.), The technology of text (Vol. 2,
perform certain analogical operations. To provide a pp. 192-209). Englewood Cliffs, NJ: Educational Tech-
nology Publications.
basis for this training, we proposed a model of anal- Glynn, S. M., Andre, T., & Britton, B. K. (1986). The design of
ogical problem solving. instructional text: Introduction to the special issue.
The process of relating concepts by means of Educational Psychologist, 21, 245-251.
Hesse, M. B. (1966). Models and analogies in science. Notre Rumelhart, D. E., & Abrahamson, A. A. (1973). A model for
Dame, IN: University of Notre Dame Press. analogical reasoning. Cognitive Psychology, 5, 1-28.
Hewitt, P. G. (1985a). Conceptual physics. Boston: Little, Rumelhart, D. E., & Norman, D. A. (1981). Analogical pro-
Brown. cesses in learning. In J. R. Anderson (Ed.), Cognitive skills
Hewitt, P. G. (1985b).lnstructor's manual to accompany con- and their acquisition (pp. 335-339). Hillsdale, NJ: Erlbaum.
ceptual physics. Boston: Little, Brown. Rutherford, F. J., Holton, G., & Watson, F. G. (1975). Project
Hewitt, P. G. (1987). Conceptual physics. Menlo Park, CA: physics. New York: Holt, Rinehart & Winston.
Addison-Wesley. Sternberg, R. J. (1977). Component processes in analogical rea-
Holyoak, K. J. (1985). The pragmatics of analogical transfer. In soning. Psychological Review, 84, 353-378.
G. H. Bower (Ed.), The psychology of learning and moti- Sternberg, R. J. (1986). Intelligence applied. New York: Har-
vation (Vol. 19, pp. 59-87). Orlando, FL: Academic Press. court Brace Jovanovich.
Miller, F., Dillon, T. J., & Smith, M. K. (1980). Concepts in Wittrock, M. C. (1985). Learning science by generating new
physics. New York: Harcourt Brace Jovanovich. conceptions from old ideas. In L. H. T. West & A. L. Pines
Polya, G. (1973). Mathematics and plausible reasoning (Vol. (Eds.), Cognitive structure and conceptual change (pp. 259-
I.). Princeton, NJ: Princeton University Press. 266). Orlando, FL: Academic Press.
Rogers, E. M. (1960). Physicsfor the inquiring mind. Princeton,
NJ: Princeton University Press.
CHAPTER 22
Toward a Model of Creativity Based

upon Problem Solving in the Social
Sciences
James F. Voss and Mary L. Means
Although creativity manifests itself in all areas of is affirmative, then one must ask how the person is
inquiry, most of the cited examples of creative able to know and how knowledge of that act is
thought involve the thinking of mathematicians, distinguished from knowledge of other acts. If the
physical and biological scientists, writers, com- answer is negative, then one must ask what criteria
posers, and artists. In typical circumstances, few establish the act as having been creative. Similarly,
examples are cited in reference to social scientists. to what extent does the product of the mental ac-
Yet creative acts may be found in virtually any tivity determine whether the mental processes lead-
domain, as, for example, in sports, when the for- ing to the product constituted a creative act? If the
ward pass was first used as an offensive weapon in product is quite significant in a theoretical and/or
football, or in business, when Henry Ford saw the practical sense, does this mean the act was creative?
implications of an assembly line and also realized Whereas if the product is relatively insignificant,
that mass production would require mass consump- does this mean the act was not creative? Shake-
tion. Similarly, the work of March and Simon speare may be regarded as being more creative than
(1958) on organizational theory and the theoretical Robert Service not because his actual "acts of
developments of Freud embrace creative work. creativity" were fundamentally different from Ser-
Despite the fact that creative acts may be ob- vice's but because his products demonstrated great-
served in many domains, defining what is meant by er insight into the human condition. Is creativity,
creativity and establishing criteria for what con- therefore, a characteristic of only a limited number
stitutes a creative act are certainly nontrivial issues. of people or are all individuals capable of creative
Indeed, the conceptual complexity of creativity acts? As might be expected, the answer to this ques-
may be demonstrated by asking a few questions. tion depends upon one's model of the creative
Does a person who is performing a creative act process.
know that he or she is doing just that? If the answer Numerous writers have, of course, addressed the
issue of creativity (e.g., Bransford & Stein, 1984;
Bruner, 1962; Hayes, 1981; Henle, 1962; Koestler,
James F. Voss and Mary L. Means • Department of Psy-
chology, Learning Research and Development Center, Univer- 1964; Perkins, 1981, 1984), and they have shown
sity of Pittsburgh, Pittsburgh, PA 15260. considerable agreement concerning characteristics
ascribed to products that are taken to be the outcome
399
of a creative process. The product characteristics the information-processing approach by taking into
that are most frequently mentioned are novelty, account the function of value and affect.
usefulness, and harmony or elegance, with most The assumptions of the model also suggest that a
writers also emphasizing that no single criterion is creative act is a function of the various structures
in itself sufficient in defining a creative product. and processes that occur in all higher mental ac-
With respect to the creative process, there is wide- tivity. Thus, creativity is not taken to be
spread agreement among writers that creativity con- qualitatively different from other types of mental
sists of perceiving an important relation where one function, and, similarly, creativity is not regarded
had not previously been known or perhaps even as a special capacity with which only some people
suspected. Poincare (1946), for example, spoke of are endowed. This position does not deny indi-
"an unexpected kinship between facts long known vidual differences in creativity, with such dif-
but wrongly believed to be strangers to one an- ferences being attributed to the operation of the
other" (p. 386). Other similar statements include cognitive and affective mechanisms of mental
"two different lines of thought meet" (Heisenberg, function.
1958, p. 187); "the connecting of diverse experi- The idea of viewing creativity in terms of prob-
ence" (Bruner, 1962, p. 6); "the emergence of a lem solving is, of course, not new (Bransford &
new and fitting organization" (Crutchfield, 1962, Stein, 1984; Hayes, 1981, Henle, 1962; Newell &
p. 123); "new orderings offacts previously unrelat- Simon, 1972; Perkins, 1981). More specifically,
ed" (Henle, 1962, p. 38); "seeing unity in variety" Newell, Shaw, and Simon (1962) described
(Bronowski, 1956, p. 27); and "making the famil- creativity as "a special class of problem solving
iar, unfamiliar and the unfamiliar, familiar" (Mur- activity characterized by novelty, unconven-
ray, 1986, p. 35). tionality, persistence and difficulty in problem for-
Given the general agreement regarding the prod- mulation" (p. 66). Hayes (1981) described
ucts and processes of the creative act, creativity creativity as "a special kind of problem solving,
may be taken to refer to a person's relating two that is the act of solving an ill-defined problem" (p.
previously unrelated concepts and creating a prod- 199), whereas Vaughan (1985) indicated that
uct that is novel as well as elegant and useful. Even creativity typically "applies to situations which
though intuitively appealing, this description of a have no single solution" (p. 40). Furthermore, the
creative act is nevertheless found wanting in at least products sought in the creative process have been
three ways. First, it fails to take into account the regarded as "vaguely and tentatively conceived,
role of context. Creative acts occur in particular groped for, caught at, discovered in the process"
contexts; creativity does not come out of "no- (Perkins, 1981, p. 276).
where" (cf. Perkins, 1981). Second, the descrip- The remainder of this chapter has five sections.
tion, though perhaps missing some components, at We begin in the first section, with an analysis of the
the same time may be too all-encompassing. Using solving of ill-structured problems, followed by an
the above account, a dog of one of the authors has "unpacking" of the model that emphasizes the in-
demonstrated creativity in devising ways to entice teractive role of prior knowledge and mental
the author into playing ball (assuming the process, search. In the third section, we consider how, with-
in its own way, was elegant). Of course, there is in the context of the model, value and affect playa
nothing to argue against the idea that animals can be role in the creative process, then in the fourth sec-
creative, except perhaps man's need to feel superior tion, we address the issue of creativity in the social
to other species. Third, the idea that creativity in- sciences. In the final section, we close with a few
volves observing a relationship that was not pre- remarks regarding how instruction may influence
viously perceived provides little information re- the development of creativity in the social sciences.
garding the processes that underlie the creative act.
In this chapter, we attempt to address some of the
complexities of the creative process by presenting a On the Solving of Ill-Structured
model that views the creative process as a special Problems
case of problem solving, and, in particular, as the
solving of an ill-structured problem. The model has
The Information-Processing Account of
an information-processing orientation, emphasiz-
Problem Solving
ing the use of one's prior knowledge and of search
processes that act upon the prior knowledge (cf. A problem generally is said to exist when an
Newell & Simon, 1972). The model also extends individual wants to reach a particular goal but can-
CHAPTER 22 • PROBLEM SOLVING IN THE SOCIAL SCIENCES 401
not because of some type of barrier. In information- consists of the individual's analyzing the problem
processing terms, the individual who is confronted to determine the givens, the goal, and the con-
by a problem is presumed to be in an initial state, a straints of the problem. Such an analysis may take
state consisting of the individual's knowledge of the various forms, a common one being to try to classi-
givens of the problem, of the problem's constraints, fy the problem as belonging to a particular category
and of the goal of the problem, which is viewed as (a physics mechanics problem) or determining the
the problem's final state. The problem is presumed causes of the problem (a problem of city govern-
to have states intervening between the initial and ment to reduce the crime rate). Then, once the rep-
final states, with the solving of the problem thus resentation is developed, the solver provides a solu-
consisting of moving from one state to the next until tion to the problem. (One can, during the solution
the goal state is achieved. Furthermore, the move- phase, "go back" and "re-represent" the prob-
ment from one state to the next is presumed to take lem, although this does not happen frequently.)
place via use of an operator. If a person is adding a Having considered the general account of the
column of digits, for example, each successive ad- information-processing view of problem solving,
dition of a number is a state that is reached by we now consider the distinction of well-structured
application of the operator "add." Thus, the indi- and ill-structured problems and how solutions are
vidual reaches the goal state when all numbers in obtained for such problems (cf. Reitman, 1965; Si-
the column are added. mon, 1973; Voss & Post, 1988).
The information-processing model also uses the
terms task environment and problem space. The
The Solving of Well-Structured and
task environment refers to the statement of the prob-
Ill-Structured Problems
lem and the environment, external to the individual,
in which the problem statement occurs. The prob- A problem is termed well structured when the
lem space refers to a delineated "mental space" various components of the problem are well spec-
that contains all the states that could occur in the ified and are known to the solver, that is, the goal is
solving of a particular problem. In addition, the well specified, the givens of the problem are either
problem space contains the operators that are avail- in the problem statement or are well known to the
able for use in the problem as well as the constraints solver, and the constraints also are known to the
of the problem. Viewed in this way, the solving of solver. For example, in proving a geometry the-
the problem involves a "walk" through the prob- orem, the givens are stated in the problem state-
lem space, moving along a "path" from one state ment, the "to prove" goal is stated, and the pri-
to the next until the goal is reached (Newell & Si- mary constraint is that the solver may use only
mon, 1972). previously proved theorems in the proof. On the
In order to solve a problem, the solver needs to other hand, ill-structured problems are those in
determine whether the path being followed is in- which the givens, the goal, and/or the constraints
deed leading to the goal. To do this, the solver uses are not well specified or known. Thus, the goal may
some type of evaluation strategy. One commonly be poorly specified, the givens may not be in the
employed strategy is termed means-ends analysis, problem statement and may be unknown to the
in which the solver evaluates whether moving into a solver, and the constraints may be unspecified.
particular state has moved the solver closer to the (The fact that there may be variation in the extent to
goal. This analysis is accomplished by determining which the givens and the constraints are specified
whether the discrepancy of the goal and the present indicates that the ill-structured, well-structured dis-
state is less than it was when the solver was in the tinction is a continuum rather than a dichotomy.)
previous state. The final state is thus reached by Reitman (1965), for example, spoke of a com-
continuing to reduce the discrepancy. Of course, poser's composition of a fugue as an ill-structured
other strategies exist and some are discussed later in problem. In this case, the individual's knowledge
this chapter. of the goal is vague, the givens are the composer's
Given the above description, we tum now to how knowledge of the structure of a fugue and knowl-
the model has been developed in reference to the edge of music theory, with these two factors also
solving of complex problems. The solution of com- constituting constraints. Moreover, as the fugue is
plex problems is typically divided into two phases, composed, additional constraints are set up, such as
that of developing the representation of the problem the key and the fugue passages already written. The
and that of providing the solution to the problem, idea that constraints are generated during the course
given the representation. The representation phase of the problem-solving activity is, moreover, a
characteristic of the solving of ill-structured over, critical conditions may change during this
problems. time.
The well-structured, ill-structured distinction is We now tum to the question of the solving of
especially important when viewed in relation to well-structured and ill-structured problems. Con-
whether consensual agreement is established with sidering the former, Larkin, McDermott, Simon
respect to problem solutions. In particular, well- and Simon (1980) have demonstrated that when
structured problems typically have solutions that physics experts are given a problem from a physics
are agreed upon by individuals who have knowl- textbook, they develop a representation of the prob-
edge of the domain in question. Thus, problems lem by analyzing the elements and relations of the
found in mathematics and physics texts are well problem, sometimes drawing a diagram in the pro-
structured, having solutions that are known and cess. Typically, such examination leads to classify-
agreed upon by mathematicians and physicists. How- ing the problem and, once classified, the problem is
ever, ill-structured problems generally do not have solved by applying the appropriate equations. Fur-
known solutions, and experts in the domain of the thermore, in applying the equations, a means-ends
particular problem frequently not in agreement with analysis is not required because the physicist knows
respect to whether a particular solution is appropri- how to solve the problem in question. Thus, this
ate. Thus, although in economics there may be example indicates that the representation phase
some issues upon which economists would agree, often consists of a pattern-matching procedure in
there are many problems for which agreement which the problem statement is considered in terms
would not be found. Moreover, consensual agree- of its parameters and the relations among the pa-
ment is found even less frequently in such fields as rameters, and these components are matched to a
sociology and political science, as well as in issues pattern in memory. The physicist, having had pre-
of social policy. In fields such as art and music, vious experience with this type of problem, is able
disagreement generally occurs not in deciding to perform the match and, subsequently, provide
whether, for example, a fugue is a fugue, but in the solution. On the other hand, novices in physics
evaluating the quality of the work. often have greater difficulty, for they may not be
Why are there differences in the structure of able to classify the problem and often resort to a
problems across various domains? Probably the process of means-ends analysis in which they try
most important reason is that the data base and the out equations that contain one or more of the param-
theory of such domains as the physical sciences are eters found in the problem statement and eventually
relatively well developed, whereas those of social reach the goal by generating an appropriate se-
sciences are relatively unrefined with respect to the quence of equations (cf. Larkin et aI., 1980).
specificity of concepts, the determination of critical The solving of ill-structured problems is usually
variables, and the development and testing oftheo- more complicated than the solving of well-
ry. Furthermore, there is the question of whether structured problems. Consider a situation in which
social sciences will ever reach such refinement, an a city official is asked to propose a plan for reducing
issue that is discussed later in this chapter. But this the use of drugs within that particular city. The
is not to say that domains such as physics are com- individual will likely develop a representation of
pletely well structured. Ill-structured problems may the problem by determining the causes of the drug
be found at the cutting edge of the research of such problem as well as determining the constraints that
domains. A good example of an ill-structured prob- exist in relation to the problem, constraints that may
lem in physics is found in the research of Tweney include budget, staff, and/or time, as well as other
(1985), who has extensively studied the diaries of factors. The development of the representation may
Michael Faraday. Tweney's analysis shows how thus involve reading reports of the federal govern-
Faraday's discovery of electromagnetic induction ment of the attempts of other cities to deal with the
constituted the solving of an ill-structured problem. problem, and of the historical development of the
Interestingly, Tweney's analysis also points to one problem. In addition, the analysis may involve de-
of the major differences of the physical sciences and termining just how extensive drug use is, in refer-
social sciences; namely, that, in the former, hy- ence to age, socioeconomic class, and other factors.
potheses can be tested by conducting experiments The individual may go to drug rehabilitation cen-
which take a relatively short time to perform where- ters, seeking information and advice. There may be
as, in the latter, experiments, even when they are an attempt to determine how the drug trafficking
feasible, may take years to conclude and, more- takes place and to determine who seems to profit
from the sale of drugs. From the information avail- somewhat of an arbitrary judgment, in the sense
able, the individual may then conclude that there that stopping further revision is not guided by any
are three primary causes of the drug problem in that particular rule except the sense of completion by the
city, or that there is a single cause, for example, composer. The same condition holds for the artist
economic. Given this representation, the individual and the writer. A chemist, however, may have a
may then propose a solution designed to reduce the quite precise goal, and when it is achieved, it is well
impact of the perceived causal factor(s). Moreover, defined.
the solution, if imaginative and workable, may The idea that creativity may be viewed as a spe-
even be called creative. cial case of the solving of an ill-structured problem
In solving an ill-structured problem, both the rep- seems reasonable at a descriptive level of analysis.
resentation and the solution phases involve the pro- However, to provide a more thorough examination
cess of justification. Evidence for the claims of the of this view, it is now necessary to consider the
representation may be provided, or the individual nature of solving ill-structured problems in greater
may develop an argument from which the causes detail, providing an analysis of the components of
may be inferred. Similarly, the solution is typically the process.
justified by indicating why and how the proposed
solution would work as well as by indicating why
the solution may not work and what can be done to Prior Knowledge, Search, and
handle possible shortcomings and objections (cf.
Voss, Greene, Post, & Penner, 1983). In effect, the
Creativity
justification of the solution points to the important
Prior Knowledge
fact that the solution process, taken as a whole, is
basically one of rhetoric. Incidentally, an interest- One of the most critical components of mental
ing aspect of solving ill-structured problems is that, functioning in general and problem solving in par-
in the course of developing the representation and ticular is prior knowledge, and one type of prior
of proposing the solution, the solver does not typ- knowledge is that of the basic concepts, facts, and
ically provide any verbalizations about the nature of principles of a particular subject matter domain.
the strategies being pursued. Instead, solvers tend This knowledge is typically declarative in nature
to deal with problems using the specific termi- (Ryle, 1949) and, though the question of the mental
nology and context of the problem per se. representation of knowledge is an issue currently
The sketch of the solving of an ill-structured receiving considerable theoretical attention, we
problem presented in the preceding paragraph sug- may assume that, for our purposes, such knowledge
gests the lack of consensual agreement that may is stored hierarchically. Furthermore, in addition to
occur in relation to solving ill-structured problems. declarative knowledge, individuals typically have
Another official could argue that the causal factors knowledge of procedures (Ryle, 1949) that are used
that are derived in the analysis are inaccurate and/ or in a particular domain, as, for example, in the use of
inadequate and could then argue that the problem subtraction methods in arithmetic or methods of
should, in fact, be represented in a different way. integration in calculus.
Another person could argue that although the A form of procedural knowledge especially
causes seem appropriate, the solution is not because important to the present analysis is that of knowl-
it will not address the causes. Yet another person edge of strategies. Moreover, the strategies of par-
could argue that though the causes and solution are ticular importance are those described by Newell
satisfactory, the solver did not take into account one (1980) as weak problem-solving methods. Strong
or more constraints that make the solution unwork- problem-solving methods are characterized as
able. being domain specific, whereas weak methods are
Although this account of the "solving" of the not domain specific and include (1) strategies, such
city's drug abuse problem provides an example of as means-ends analysis, analogy, and decomposi-
the solving of an ill-structured problem, other ill- tion, in which a problem is divided into a number of
structured problems may have characteristics that subproblems, (2) conversion, in which one problem
vary somewhat from those found in this case. For is converted into another because the latter is more
example, a composer may have the goal of compos- readily solved, and (3) generate and test, which
ing a particular type of work, but the point at which consists of generating a solution to a problem or
the composer feels that the goal has been reached is subproblem and testing the solution against some
criterion. In addition, examining the history of a because all the germane information is available
problem may also be regarded as a weak method. when one first considers the problem. Instead, it is
What is particularly important about the weak necessary to search, either internally or externally,
methods is that, with the possible exception of eco- for information, with the goal of the search being to
nomics, social sciences have few if any strong obtain information germane to the problem at hand.
methods of problem solving, and the individual Furthermore, when information is obtained, it also
must employ weak methods in the solving process. becomes necessary to evaluate the information with
Moreover, what is of critical importance to the pres- respect to its usefulness to the problem at hand.
ent analysis is that the weak methods are fundamen- Thus, the process of search is regarded as having
tal to search, a point that is developed in the next two functions, the operator function and the
section. evaluator function. The operator function, having
There are two additional points concerning prior essentially the same function as that of the pre-
knowledge that warrant consideration. First, al- viously described operators, refers to the search for
though having prior knowledge is in itself of critical information that enables the individual to enter the
importance, such knowledge also needs to be ac- next problem state. The evaluator function is that of
cessible; that is, in the course of solving a problem, determining the extent to which information that is
the knowledge that may be germane to the solution obtained may be effectively used in the solution
must be accessed, for only if the individual is able to process.
access such knowledge may he or she be able to Considering the operator function in greater de-
relate what is being processed and what is in memo- tail, it is assumed that individuals are able to per-
ry. In other words, prior knowledge needs to be a form this function by using, in part, the previously
"working" knowledge (cf. Voss, 1987), in the described weak methods of problem solving. Find-
sense that the knowledge needs to be accessible ing a useful analogy, for example, would constitute
when the individual may need it. How knowledge a mechanism of search. But other mechanisms are
becomes working knowledge is an important and also available when an individual is reasoning about
unresolved issue, but one thing that quite likely a particular problem. For example, the individual
makes knowledge more "workable" is its use in a may search for examples to support a given conclu-
variety of contexts. When used in this way, it raises sion. Similarly, the individual, in searching exter-
the likelihood of being accessed in a variety of con- nally, may ask questions or go to a library. Also the
texts. The importance of flexibility of access has, of individual may try out ideas on other people and
course, been stressed by writers on creativity, such receive critical feedback. Thus, the individual has a
as Perkins (1984) suggesting that expertise in a do- relatively large number of mechanisms that may be
main allows for "mobility," and Bruner (1962) used to perform the operator function of the search
suggesting that "flexibility" and "combinatorial process.
playfulness" are important. With respect to the evaluator function, the indi-
Another aspect of prior knowledge that warrants vidual needs criteria that permit acceptance or re-
consideration is that though prior knowledge is ex- jection of the information, and an important aspect
tremely important to effective problem solving, it of such criteria is that they typically constitute con-
also may act to hinder problem solving (Henle, straints. Thus, in such fields as music, art, and
1962; Perkins, 1981). For example, prior knowl- literature, criteria may be aesthetic in nature, with
edge may set up categorizations that prohibit the the individual either accepting or rejecting particu-
individual from viewing problems in new ways. Or lar information as "looking good," "sounding ac-
prior knowledge may produce a particular highly ceptable," or "feeling right." Such evaluations,
used problem representation that constrains the in- moreover, may be accompanied by metastate-
dividual with respect to looking at the problem in ments, as with a gut-level reaction, but the indi-
new ways, as in the classical Einstellung effect vidual may not be able to verbalize what such state-
(e.g., Luchins, 1942). ments "mean." In this regard, Perkins (1984)
noted that creative individuals have at their disposal
personal standards pertaining to creative products
Search
that enable them to "strive for products which are
Although it is important that a person have elegant, beautiful, and powerful" (p. 4). In-
knowledge and that the knowledge is working, terestingly, Getzels and Csikszentmihalyi (1976)
solving ill-structured problems does not take place found that, in the context of art, individuals who
upon finishing a painting tended to view their work previously described .operator and evaluator func-
as incomplete were subsequently judged in their tions.
careers as more successful than individuals who One question about the search process that is
viewed their own products as complete. The au- quite difficult to explain is operator selection, that
thors suggested that creative individuals have a crit- is, what determines which weak method is to be
ical capacity that tends to be ahead of their produc- used. Assume that an individual working in the
tive capacity, a notion that places strong emphasis State Department is confronted by the problem of
upon the importance of the evaluative function. helping to develop a policy toward South Africa. Of
course, this is a complex issue involving such issues
as the apartheid policy and United States interests in
On the Interaction of Prior Knowledge and South Africa. The individual may use a number of
Search strategies, such as looking for historical analogies,
either by searching his or her memory or by examin-
In the present section, the interaction of prior ing such sources as books that provide accounts
knowledge and search is discussed by describing germane to the issue. The individual may also focus
four characteristics of the interaction. on the historical roots of the problem as found spe-
1. Search is directed, as opposed to random. The cifically in South Africa or in Africa as a whole.
search process is presumed to be directed, and how The individual may also try to adopt the perspective
the direction is established is of importance. How- of the government officials of South Africa and try
ever, because goals vary in the extent to which they to determine what policies are optimal from that
are defined, the extent to which the search is di- perspective. These various strategies all constitute
rected also varies. In addition, search can be con- efforts to expand the search and, more generally, to
strained by the existence of multiple goals. For ex- develop an appropriate problem representation. As
ample, if an individual is working to develop an previously noted, why one strategy is chosen over
arms agreement with the Soviet Union, he or she another is not clear, but it is likely that operator
may have in mind a number of goals that quite likely selection may be based upon past experience such
are ranked in terms of their importance. The goals that the use of particular strategies may have been
thus act as a hierarchy of constraints. Moreover, the "reinforced." If so, then individuals may have an
individual is also constrained by his or her own operator hierarchy that represents the general order
beliefs and attitudes toward the Soviet Union as in which operators are selected, with the hierarchy
well as by the knowledge of the Soviet Union's being established via prior experience.
track record in past agreements, the policies of the 2. Search processes often lead to ambiguous
current United States administration, the potential and/ or paradoxical propositions. During the
political ramifications of any agreement that is course of the search, it is not uncommon for indi-
reached, and by personal political aspirations. viduals to find themselves holding to ambiguous
Moreover, in an even more complex way, the con- propositions in which, for example, two inconsis-
straints may include a type of cultural framework in tent solutions seem reasonable. The political scien-
which historical beliefs held by individuals within tist analyzing the South African issue may, on the
their own society may serve as constraints. Thus, one hand, conclude that the United States should
all such constraints act to direct the search as well as pressure South Africa to drop the apartheid policy
set criteria for evaluation. immediately but, on the other hand, may also con-
Three points may be made with respect to the role clude that the policy must be gradually eliminated if
of constraints in the search process. First, whereas the country is to avoid a bloody civil war. The
in a well-structured domain the constraints may be literature on creativity contains many references to
viewed more or less as rules of the game, the con- the role of ambiguity in the creative process (e.g.,
straints in an ill-structured situation are typically Perkins, 1984), and two points are emphasized in
more numerous, not as well defined, and they in- this literature. First, creative individuals tend to
volve not only situational factors related to the issue have a high tolerance for ambiguity (Amabile,
but personal factors such as one's own beliefs and 1983; Rothenberg, 1979). Indeed, creative indi-
goals. Second, during the search, an individual is viduals may even seek ambiguities and use the set-
usually not aware of many of the constraints. Third, ting up of ambiguities as an important part of the
the process by which various solutions to goals or creative process. Such a view, of course, is con-
subgoals are generated and evaluated involves the sistent with the idea that thinking is dialectical in
nature (Bassaches, 1984). Second, creative out- however, detachment from a problem also repre-
comes often occur as a result of resolving ambigu- sents a form of perspective-taking, in which an indi-
ities (Bruner, 1962). vidual may respectively be considering a problem
A critical aspect of ambiguity is that the indi- and, at an unconscious level, be considering alter-
vidual is faced with a difficult processing problem native representations of a problem. The present
in trying to resolve an ambiguity. There are two sets view does not suggest, however, that there are any
of constraints that must be dealt with, and the indi- special unconscious mechanisms that facilitate
vidual must therefore generate a representation or a problem solution.
solution that may likely reject some of the existing
constraints while, at the same time, maintaining
other constraints. The complexity of such a pro- A Recapitulation
cess, moreover, may be such that not all individuals
are capable of conducting such mental gymnastics. The interaction of search and prior knowledge, as
3. The search process often involves the violation previously described, is in a sense conceptually
of constraints. As previously noted, acts that are quite simple. What makes the process complex,
regarded as creative quite frequently involve violat- however, is the flexibility of the human mind. The
ing one or more constraints. Moreover, the con- individual has stored in memory an extremely large
straints often violated in such acts are assumptions; amount of information, some of which may be ger-
that is, "breakthroughs" occur when an individual mane to the issue being considered. The question
rejects a widely held theory by attacking its existing then is how to find that information. But, at the
assumptions and formulating an idea based upon same time, finding the information is frequently not
different assumptions. Numerous examples of this enough, for the individual must be able to deter-
phenomenon exist, ranging from Stravinsky's mine how the information fits in a new context; that
"The Rites of Spring" to Paracelsus's insight that is, the information must be related to other concepts
disease may be caused by factors external to the to which it has not been previously related. To ac-
body (Boorstin, 1983). Similarly, President Nix- complish this, the individual has at his or her dis-
on's opening of a dialogue with the People's Re- posal a finite and yet a large number of strategies.
public of China constituted a violation of assump- Furthermore, the individual needs to evaluate the
tions regarding the relations between China and the information and to determine the "fit" of the infor-
United States. Within the present framework, con- mation to the problem at hand.
straint violation thus reflects a process by which the Previously in this chapter, we mentioned that the
individual, in obtaining information from the creative process is an act of the human mind that
search, is forced to re-represent the problem, and involves the use of processes that occur in virtually
this often happens as a resolution to an ambiguous everyone in the course of daily existence. Neverthe-
situation. less, the analysis as thus far presented also indicates
4. Search often involves an alternating immer- that the creative process does not simply involve the
sion and detachment with respect to problem in- use of such mechanisms. To have considerable
volvement. Probably the most widely cited phe- knowledge, to be able to use what one knows, to
nomenon in the creativity literature is that of the define a problem appropriately, to search for infor-
"incubation effect" (Wallas, 1926). This effect re- mation pertaining to the problem, and to evaluate
fers to the observation that individuals may work on information appropriately are indeed highly com-
a problem that defies solution and, upon "setting plex skills requiring considerable integration.
aside" the problem, a solution suddenly occurs to Thus, many people may not be regarded as creative
the solver, often upon the occurrence of an appar- because they lack one or more of these components
ently extraneous stimulus. The problem is thought or because they have not had experience in the use
to be "incubating" during the period of detachment of the mechanisms. One may ask whether we are
and somehow the solution is generated during this essentially saying that creativity is the effective use
period. Attempts to explain this effect have been of the operator and evaluator functions in relation to
discussed by Perkins (1981). one's prior knowledge. The answer to this question
Within the present framework, the incubation ef- is yes, but---. The "but" refers to the fact that
fect is regarded as a continuation of the search pro- although this statement is true, it is not the whole
cess that occurs at a level below the threshold of truth. Thus, we must now consider the roles of
consciousness (cf. Simon, 1966). At the same time, value and of affect.
Value, Affect, and the Creative Process viduals who tend to be creative will likely select
problems that will generate the previously de-
Within the present context, affect is regarded as scribed negative affect. Moreover, the value struc-
the accuracy of positive or negative feelings that an ture of such individuals encourages their selection
individual has about a particular idea, object, or of problems that will be of significance. Thus, the
event. Value is defined as a personalized construct more creative individual seeks out important issues
that presumably includes an individual's life- that are challenging and, most importantly, has the
related goals as well as a sense of the means by ability to discern such issues, a process that once
which goals are to be accomplished. Value is also again involves the use of knowledge and search
taken to refer to a sense of quality regarding the processes. Thus, even though virtually all indi-
goals. When viewed in relation to creativity, one viduals have affect, that is, negative feelings about
senses that individuals who are creative in virtually some things undone an positive feelings about get-
any field have a strong interest in that field, accom- ting things done, and even perhaps about doing
panied by a goal to excel in that field, although what things well, relatively few individuals are attracted
"to excel" means may vary, of course, from per- to issues that have the previously described general
son to person. Furthermore, even though some indi- significance.
viduals have such goals in the arts, in science, or in 2. Value and affect keep the problem at hand in
social science and social policy fields, other indi- the forefront of consciousness. Individuals who are
viduals may have such goals in the office, the facto- working on a problem seem to be continually aware
ry, or in the professions. However, because of the problem (cf. Perkins, 1981), and the effect of
creativity can seldom be found in performing what such awareness is to be searching constantly, inter-
is routine, the observation of creativity is, of neces- nally and externally, for information that is poten-
sity, limited primarily to those individuals whose tially relevant to the problem. The account by
values are such that they strive to excel and who are Gruber (1981) of Darwin's development of the the-
in fields exposed to the critical eye. Hence, the ory of evolution provides an excellent example of
observation of creativity is quite restricted, and it is this process, for whenever Darwin observed ani-
not surprising that sometimes creativity has been mals, he apparently did so with some general ideas
viewed as a special talent. of animal behavior in mind.
We assume that such affect and value relate to The presumed continual awareness of a problem
creative thinking in the following four ways. on the part of the creative individual has some in-
1. Value and affect influence problem selection. teresting implications. For example, the awareness
Some individuals, by their intelligence, curiosity, likely leads to an increased perception of inconsis-
training, and sense of values, do select problems tencies, such inconsistencies being evoked when
that are significant, in the sense that the problems observations conflict with existing beliefs or hy-
have general theoretical significance and/or consid- potheses. Such inconsistencies set up negative af-
erable potential significance to other people (Per- fect which may lead to dialectical analysis and reso-
kins, 1981). Furthermore, such individuals are pre- lution. Also, continual awareness leads to an
sumed to have a relatively strong need to solve the increase in one's knowledge as well as to an in-
selected problem; this need initially setting up nega- crease in the use of search strategies. The latter
tive affect, negative in the sense that individuals takes place because the individual needs to interpret
feel a need to obtain a sense of closure concerning new information, and that process requires compar-
the problem. Progress in the solution process then ing the information to what already exists in memo-
involves positive affect, in the sense that obtaining ry. In a sense, the continual awareness of a problem
closure or moving toward it is rewarding. Thus, in a serves as a heuristic that enables the individual to
very real sense, the solving of the problem con- acquire new information and perceive previously
stitutes a strong affective reward (Amabile, 1983). unknown relationships.
Moreover, although critical recognition of the solu- 3. Value and affect influence persistence.
tion by other individuals may also be important, the Another function of value and affect is that they
more creative individual tends to be motivated more increase persistence in the solution effort. The val-
by such internal criteria than by external criteria ues of the individual and the related negative feel-
(Amabile, 1983; Crutchfield, 1962; Perkins, 1981, ings that exist when a solution is not forthcoming
1984). lead the individual to try new solutions and to
One implication of the role of affect is that indi- search for more information. This point is impor-
tant for it indicates that a close relationship exists sciences have been especially creative were re-
between criteria established by one's values and viewed by Inkeles (1983). He noted that Deutsch,
one's persistence. The more creative individual will Platt, and Senghass (1971) indicated 62 contribu-
not tend to be satisfied by achieving something that tions that were regarded as especially creative and
simply works but will, instead, attain the high stan- that they used two criteria to make this evaluation.
dards of performance that are involved. It seems Specifically, a creative act was taken to produce a
reasonable to assert that most individuals do not new perception or new operations, and these had an
have such strong self-expectations. important influence upon the subsequent develop-
ment of the field in question. As noted by Inkeles
(1983), these criteria largely excluded specific dis-
Value and Affect, Prior Knowledge and Search
coveries or demonstrations of important empirical
When the characteristics of value and affect are phenomena. Furthermore, Inkeles noted that the
considered in relation to prior knowledge and procedure of Deutsch et at. excluded individuals
search, it seems clear that the former produce more who did not begin a movement but who made sub-
search, both internal and external, and serve the stantive contributions within the field. In contrast,
role of providing feelings regarding the ideas that Inkeles pointed out that, in social sciences,
are developed in the course of the solving process; creativity has often consisted of work that has pro-
feelings that can reflect both a sense of accomplish- vided a conceptual framework for analyzing issues,
ment and a sense of despair. The present model thus with such frameworks being neither testable nor
depicts the creative individual as an intellectually falsifiable, for example, socioconscious class. In-
curious person who is at home when working with keles also argued that a number of methodological
uncertainties and paradoxes. This individual has a advancements and discoveries have been quite cre-
substantial knowledge base not only of his or her ative. But, perhaps the most interesting point made
domain of interest but of more general issues and by Inkeles involves the limitations of social sci-
subject matter. This individual is facile in the use of ences, namely, that three limitations act strongly to
search strategies but quite likely is unable to ver- influence the nature of creativity. Moreover, be-
balize when and how they are being used. Finally, cause of these three factors, social sciences would
although virtually all individuals have such charac- be expected to yield fewer creative outcomes than
teristics to some degree and may, to a point, be natural sciences. First, social sciences do not attract
creative, creative acts that are recognized by the individuals who are as creative to the extent that
intellectual community are those in which the indi- such individuals are drawn to the natural sciences.
vidual has typically excelled in the integration of Second, the money provided for social science re-
these characteristics and has generated a product of search is less than that allocated for research in
theoretical significance and/ or practical utility. natural sciences. Third, and the one that Inkeles
regards as the most fundamental, the subject matter
itself contains inherent limitations.
Inkeles "unpacks" his third limitation by citing
Creativity in the Social Sciences three constraints of social science research. The
first constraint is that there is a limit on the size and
The account of creativity that is developed in this range of the subject matter. Thus, although natural
chapter certainly maintains that the nature of the sciences have expanded their inquiry to the universe
creative process does not vary with the subject mat- as well as to particles, social sciences are limited to
ter domain. Differences in content, however, the study of the human population. The second con-
would be expected as a function of domain. The straint is that social science subject matter is condi-
evaluative function performed by the artist, for ex- tional, that is, even though conditions of study in
ample, would be different from that performed by the physical sciences remain relatively stable, the
the chemist. Given that the general process is con- conditions of social science phenomena change
sistent across domains while the specific execution from one situation to the next. Thus, consistency
varies, the execution of the creative act is taken to over social time and social space is rare. The third
be a function of problem context, a matter that is of constraint is that various individuals and public au-
particular concern when considering creativity in thorities refer to social science work in ideological
the social sciences. terms. In this regard, Inkeles notes how, because of
The question of what contributions in the social a presumed ideology, social science development
has been meagre in Socialist countries and also ative thinking. Furthermore, the nature of such in-
notes politically based constraints that exist in struction may be derived from the model presented
Western countries. in this chapter. Specifically, the model leads to at
The Inkeles account raises a number of interest- least five suggestions. (1) Students need to acquire
ing questions regarding creativity in social sci- knowledge of basic facts, concepts, principles, and
ences, and from his comments we are able to gain theory in the domain, and, what is even more
some idea regarding why examples of creative ac- important, such learning needs to concentrate upon
tivity in social sciences are not plentiful. Social explanation and rationale. Textbooks currently em-
sciences seem to evolve not by astounding discov- ployed in elementary and secondary school tend to
eries but by the growth of concept and method. present information in a factual manner with occa-
Social scientists tend to be "concept hunters," in sional causal information, and greater explanation
which the delineation and description of a phe- would be desirable. (2) As acquisition takes place,
nomenon are critical endeavors. However, the de- students need to have experiences in using the
velopment of such concepts sometimes appears to knowledge in the solving of problems. This process
be a statement of the obvious and is sometimes would help to make knowledge accessible as well as
regarded as arbitrary and is met with controversy. providing experience in the use of the search meth-
Thus, relatively few concepts reach a criterion of ods described in this chapter. (3) Students need to
theoretical significance, in the sense of explanation develop skills in evaluating arguments, both with
as opposed to description. In addition, social sci- respect to their own arguments and with respect to
ence developments seldom have met the criterion of the arguments of others. Such experience should
having practical significance. Social scientists have include examining the nature of criteria that are
not cured a disease, invented television or air condi- used in evaluation. (4) Controversial issues need to
tioning, or developed optic fibers. Indeed, one be discussed and the various positions need to be
could question the extent to which social scientists evaluated. There has, of course, been a general
have generally been effective in reducing crime tendency to finesse such issues, and such avoidance
rate, solving international disagreements, and even hardly leads to the development of creative thought.
solving economic problems, although economists (5) Teachers at the elementary and secondary levels
would, of course, argue this point. Furthermore, require much more training in social sciences. Al-
Inkeles's (1983) argument clearly suggests that the though teachers may be well prepared in United
issue is not necessarily one of differences in point of States History and sometimes in World History
development, that is, that the natural sciences are and/or World Cultures, preparation in economics,
more advanced and that social sciences will international relations, and other fields is charac-
eventually "catch up." Instead, the argument is teristically poor (Weiss, 1978). Moreover, such
that social sciences, because of their intrinsic training should lead to a deeper consideration and
nature, may never catch up. Perhaps the best that appreciation of the need for creativity in the social
can be expected is that intelligent individuals will sciences.
realize that the major problems of the world are in
the domain of social sciences, including the social
use of technological developments, and that such Summary
realizations will lead to a deeper analysis of such
issues. At a minimum, social science instruction In this chapter, we presented a model of
should help to increase student awareness of the creativity that is based upon the solving of social
importance of social science questions. The next science problems or, more generally, on the solving
section presents some suggestions regarding such of ill-structured problems. The model emphasizes
instruction. the importance of the individual's knowledge base,
the effective use of a variety of search mechanisms
that provide for obtaining and evaluating informa-
Instruction and Creativity in the Social tion, and the operation of value and affect compo-
Sciences nents that, in a sense, drive the process. We also
noted that the cited examples of creativity are usu-
Although generally not recognized as such, so- ally found in mathematics, natural science, and in
cial science instruction affords a highly challenging the humanities, and are seldom found in social sci-
and exceptional opportunity to help develop cre- ences. This fact may be largely attributed to social
science research not meeting the criteria that are (1980). Models of competence in solving physics problems.
typically interpreted. Finally, we proposed some Cognitive Science, 4, 317-345.
Luchins, A. S. (1942). Mechanization in problem solving.
suggestions with respect to instruction in social sci- Psychological Monographs, 54, (Whole No. 248).
ences that we hope would lead to an increase of March, J. G., & Simon, H. A. (1958). Organizations. New
creativity in students. York: Wiley.
Murray, E. L. (1986). Imaginative thinking and human exis-
tence. Pittsburgh, PA: Duquesne University Press.
ACKNOWLEDGMENTS. Preparation of this chap- Newell, A. (1980). One final word. In D. T. Tuma & F. Reif
ter was supported by the Office of Educational Re- (Eds.), Problem solving and education: Issues in teaching and
search and Improvement via Support of the Center research (pp. 175-189). Hillsdale, NJ: Erlbaum.
Newell, A., & Simon, H. A. (1972). Human problem solving.
for the Study of Learning and of the Learning Re- New Jersey: Prentice-Hall.
search and Development Center of the University of Newell, A., Shaw, J. c., & Simon, H. A. (1962). The processes
Pittsburgh. The opinions expressed in this chapter of creative thinking. In H. E. Gruber (Ed.), Contemporary
do not necessarily reflect those of any of these orga- approaches to creative thinking (pp. 63-119). New York:
nizations. Atherton Press.
Perkins, D. N. (1981). The mind's best work. Cambridge: Har-
vard University Press.
Perkins, D. N. (1984). What else but genius? Six dimensions of
References the creative mind. Paper prepared for the symposium on
creativity in science, Los Alamos National Laboratory, Los
Amabile, T. M. (1983). The social psychology of creativity. Alamos, New Mexico.
Springer-Verlag. Poincare, H. (1946). The foundations of science: Science and
Bassaches, M. (1984). Dialectical thinking and adult develop- hypothesis, the value of science, science and method (G. G.
ment. Norwood, NJ: Ablex. Halsted, Trans.). Lancaster, PA: Science Press.
Boorstin, D. J. (1983). The discoverers. New York: Random Reitman, W. (1965). Cognition and thought. New York: Wiley.
House. Rothenberg, A. (1979). The emerging goddess. Chicago: Uni-
Bransford, J. D., & Stein, B. S. (1984). The ideal problem versity of Chicago Press.
solver. New York: W. H. Freeman. Ryle, G. (1949). The concept of mind. London: Hutchinson.
Bronowski, J. (1956). Science and human values. New York: Simon, H. A. (1966). Scientific discovery and the psychology of
Julian Messner. problem solving. In R. G. Colodny (Ed.), Mind and cosmos
Bruner, J. (1962). The conditions of creativity. In H. E. Gruber (pp. 22-40). Pittsburgh: University of Pittsburgh Press.
(Ed.), Contemporary approaches to creative thinking (pp. 1- Simon, H. A. (1973). The structure of ill-structured problems.
30). New York: Atherton Press. Artificial Intelligence, 4, 181-201.
Crutchfield, R. S. (1962). Conformity and creative thinking. In Tweney, R. D. (1985). Faraday's discovery of induction: A
H. E. Gruber (Ed.), Contemporary approaches to creative cognitive approach. In D. Gooding & F. James (Eds.),
thinking (pp. 120-140). New York: Atherton Press. Faraday rediscovered (pp. 189-209). London: Macmillan.
Deutsch, K. W., Platt, J., & Senghass, D. (1971). Conditions Vaughan, T. D. (1985). The balance of opposites in the creative
favoring major advances in social sciences. Science, 171, process. Gifted Education International, 3, 38-42.
450-458. Voss, J. F. (1987). Learning and transfer in subject matter learn-
Getzels, J. W., & Csikszentmihalyi, M. (1976). The creative ing: A problem-solving model. [Special issue]. International
vision: A longitudinal study of problem finding in art. New Journal of Education Research, 11, 607-622.
York: Wiley. Voss, J. F., & Post, T. A. (1988). On the solving of ill-structured
Gruber, H. E. (1981). Darwin on man: A psychological study of problems. In M. T. H. Chi, R. Glaser, & M. Farr (Eds.), The
scientific creativity. Chicago: University of Chicago Press. nature of expertise (pp. 261-285).
Hayes, J. R. (1981). The complete problem solver. Philadelphia: Voss, J. F., Greene, T. R., Post, T. A., & Penner, B. C. (1983).
The Franklin Institute Press. Problem solving skill in the social sciences. In G. Bower
Heisenberg, W. (1958). Physics and philosophy: The revolution (Ed.), The psychology of learning and instruction: Advances
in modern science. New York: Harper. in research and theory (Vol. 17; pp. 165-213). New York:
Henle, M. (1962). The birth and death of ideas. In H. E. Gruber Academic Press.
(Ed.), Contemporary approaches to creativity. (pp. 31-62). Wallas, G. (1926). The art of thought. New York: Harcourt.
New York: Atherton Press. Weiss, I. R. (1978). Report of the 1977 national survey of sci-
Inkeles, A. (1983). The sociological contribution to advances in ence, mathematics, and social studies education. Report to
the social sciences. Social Science Journal, 20, 27-44. the National Science Foundation and Contract No. C7619848.
Koestler, A. (1964). The act of creation. New York: Dell. Center for Educational Research and Evaluation, Research
Larkin, J. H., McDermott, J., Simon, D. P., & Simon, H. A. Triangle Institute.
CHAPTER 23
The Teaching of Creativity to

Preschool Children
THE BEHAVIOR ANALYSIS APPROACH
Elizabeth M. Goetz
The presentation of the behavior analysis approach behavior principles. Finally, a focused environ-
for the teaching of creativity to preschool children ment for the development of creativity in preschool
requires many subtopics. As a starting point, it will children will be addressed considering the child's
be a fruitful exercise to delineate the obvious dif- home, school, and community.
ferences between the behavior analysis and tradi-
tional approaches to training creativity so the reader
may be aware of the path to be taken. Before trod- Characteristics of Behavior Analysis and
ding down this path, mention will be made of some
other professionals who share with behavior ana- Traditional Approaches
lysts a belief in the early training of creativity,
though they have chosen to take different paths. Is it possible that creativity can be taught to pre-
Crucial to the purpose of this chapter is the defi- school children? Are certain biological prerequi-
nition of creativity, which will be expounded upon sites necessary for such teaching to be effective?
in some depth, followed by the determining bases The behavior analysis approach addresses these
for originality and comments on social and stan- questions in this way. The most important factor in
dardized test validation measures for creativity. teaching any behavior, even creativity, to any per-
With the definition of creativity established, given son, even a preschool child, is the congruence of the
unresolved problems, the major portion of the chap- delivered individualized curriculum with the de-
ter will be devoted to behavior analysis studies for sired behavioral outcome. Although genetic and bi-
the training of creativity to preschool children. ological biochemical variables participate in all be-
Basic experimental designs used in these studies havior, they are considered as givens when teaching
will be covered. All the research will be tied to creativity. The preschool teacher's role is to assess
any child's current degree of creative behavior and
train from that point of reference with a positive
Elizabeth M. Goetz • The Edna A. Hill Child Development expectation that is based on the effects of the use of
Laboratory Preschools, Department of Human Development, behavior principles.
University of Kansas, Lawrence, KS 66045. Traditional ways to training creativity have taken
411
a different approach. Here, the focus is on early identify group norms that tend to mask individual
identification of the creative child followed by performance.
placement of that child in an enhancement or accel- Third, the behavior analyst's individualized
eration program (Horowitz & O'Brien, 1986). This teaching of creativity consists of the precise ap-
approach is nowhere more evident than in gifted plication of techniques that are based on behavior
programs that enroll children based on parent and principles to change operationally defined creative
teacher reports of giftedness (which often includes a behavior. The traditional approach uses more open-
component of creativeness) and standardized test ended teaching to change mental processes, such as
scores. This difference in attitude regarding the re- hypothesis formation and testing.
lationship between established creativity traits and
the effectiveness of the succeeding creativity train-
ing is but one of the differences that set behavior
analysis apart from traditional approaches. Other Cohorts of Early Training
major distinctions between these two views follow.
First, the assessment of creativity in behavior Before proceeding into the details of the behavior
analysis is criterion-referenced for a class of specif- analysis approach, it is proper to acknowledge that
ically defined behaviors closely related to a given behavior analysts are not the only ones to empha-
creative task. Assessment is not based on standard- size the teaching of creativity in the early years
ized test scores interpreted in comparison to group without undue concern for inherited traits, though
norms of creative persons on that test for generic their teaching differs from other methods. For ex-
creativity traits. The assignment of the name ample, Torrance's (1962, 1965) major concern,
creativity to a group of traits, such as fluency, flexi- too, has been the training of creativity in early
bility, originality, and divergent thinking, and then childhood, though the focal point of his training has
the reification of that name (creativity) to an entity been the development of a general skill as measured
or process causing those traits is considered a nomi- by standardized tests (Winston & Baker, 1985).
nal fallacy (cf. Bijou, 1976, p. 85). On a slightly different tangent, other educators
Admittedly, both behavior analysts and tradi- have stressed the early training of basic skills and
tional psychologists recognize that the creative per- certain "habits of the mind" -commitment to and
son may excel in the visual and performing arts, pride in hard work and the joy of learning-as the
dancing and athletics, leadership, creative think- first steps to later creativity. Furthermore, they rec-
ing, or some academic subject with each of these ommend providing early opportunities to become
areas requiring somewhat different skills for "the acquainted with a diversity of pursuits, such as
creative act. " Yet behavior analysts tend to be more dancing, baseball, and camping.
specific in focusing on particular behaviors, either In his research, Bloom (1985) actually found this
observable or nonobservable (internal verbal be- recommendation to be wise advice. His data
havior), whereas the traditional psychologist tends showed that the young and talented had received
to be more generic focusing on internal thinking support (i.e., encouragement and rewards) and in-
processes or constructs. The rationales behind these struction in basic skills and habits from parents and
different foci are that behavior analysts view both teachers before developing their specific creative
observable and nonobservable behavior as behavior skills. Therefore, he exhorts this type of early train-
influenced in the same way by environment, where- ing of the basics and habits along with the freedom
as traditionalists hold a dualistic view separating to explore a variety of experiences that lead
them (Day, 1969; Skinner, 1957). eventually to pairing the child with a highly skilled
Second, behavior analysts stress the individual mentor in the area of promise for that child.
analysis of creative behavior that permits the teach- Susuki, founder of the revolutionary system of
ing to be particularized for a specific person. Even music instruction and the Yoki Gakuen preschool,
when there is group training with group data pre- also dwells on training basic skills (i.e., reading,
sented in behavior analysis research, the group data writing, and math, as well as music, art, gym-
represent a compilation of individual data identified nastics, and dance) and the habits of self-discipline
as a success or failure of the teaching strategy for and extended on-task behavior (Juan, 1985). He
each subject. It follows that group statistical analy- believes that all children are essentially equal in
ses, ever present in traditional research, are not talent and ability at birth, that no child is born
prevalent in behavior analysis studies because they gifted, and that the environment forms the child. He
CHAPTER 23 • TEACHING OF CREATIVITY TO PRESCHOOL CHILDREN 413
espouses repetition at an early age as being the key Random Behavior

element in training any behavior. Montessori The random or accidental act is not a problem in
(1964), also, has long purported this point of view. behavior analysis because intent as a cause is not a
Thus, Torrance, Bloom, Susuki, and Montessori part of the behavior analysis perspective. Behavior
do have something in common with the behavior that occurs is analyzed only in relation to environ-
analysis approach-a belief that early environment ment, not intent.
has a major impact in molding children, possibly
even in the area of creativity. What any child can
learn, almost all children can learn, given the ap- Predictable Behavior
propriate conditions of learning. The behavior anal-
ysis approach, though, has unique characteristics, Because traditionally creative activity is thought
which will be set forth. to be intrinsically motivated and personalized from
the depth of one's unique self (Ghiselin, 1952), it is
an anathema to this notion that predictable, im-
itative, or trained behavior be considered creative.
Definition of Creativity Therefore, even for behavior analysts, a predictable
though novel or original response that is the result
of following a routine formula, as, for example, in
Original Behavior extracting square roots, is not considered an origi-
By definition, creativity has a connotation of nal response.
originality, which may be characterized by novelty,
difference, ingeniousness, unexpectedness, or in-
Imitative Behavior
ventiveness. In behavior analysis, creative behav-
ior may be original verbal or motor behavior, ob- Imitative behavior poses a gray area in judging a
served directly or inferred from a product. For novel response for a preschool child. First, one may
instance, a written poem may be the product from not know whether the response is imitative because
which original verbal behavior or thinking is in- at this elementary level a child making a new per-
ferred; or a blockbuilding construction may be the sonal response often makes a response the teacher
product from which original motor behavior is in- has seen elsewhere. Is the child reinventing the
ferred. In either case, behavior or product, a code of wheel or just copying it? In the case of a profes-
objective definitions with related observation rules sional creative adult, a response seen elsewhere is
is used to measure the originality. Reliability be- considered imitative or derivative, whether it is or is
tween two independent observers, using the same not on the personal level. Second, at this young age,
code simultaneously, is needed for the code (and is not a new response on a personal level, though
the data obtained from using it) to be acceptable. imitative, worthy of commendation as a shaping
The determining bases for originality will be de- procedure toward creativity?
tailed in a later section.
Trained Behavior
Bizarre Behavior
The training of specific diverse responses, which
The meaning of the word creatiVity needs some cumulatively add up to a creativity score, will be
restrictions to disallow the bizarre, random, pre- subjected to the same criticism as imitative re-
dictable, imitative, or trained acts from being la- sponses. Neither the imitative nor trained response
beled creative. To eliminate the bizarre act, the is an original response. But this criticism cannot
creative behavior is required to be relevant to the always be held against the training of various spe-
assigned task. For instance, if the task is block- cific responses, because typically, in behavior anal-
building, the building of different block forms ysis, the training does not focus as much on the
would be relevant, whereas the throwing of a block dimensions of the specific responses as it does on
at an apple to knock it down from a tree, though the differences among the responses. It is the con-
ingenious, would not be relevant to the blockbuild- cept of novelty, doing something different for the
ing task and would be considered bizarre. In other first time in a session or across sessions, that is
words, creative behavior needs to be defined in the trained. For example, when a preschool block-
context of the given task. builder initially builds an elaborated balance and is
praised for it, that child is not merely praised for the Piana, 1980). Formal stimulus control occurs when
dimensions of this specific form but also for making a point-by-point correspondence exists between the
this form for the first time. This descriptive praise discriminative stimulus (before the response) and
not only emphasizes the novelty of the act but also the response-one literal point of reference for
may serve as a prompt for more novelty, of the each response (i.e., an exact translation). Informal
child's choosing, thereafter. stimulus control occurs when the response is under
Whatever the type of creativity training the be- multiple sources of control (the literal point of refer-
havior analyst pursues, for specific forms or for ence and one or more tangentially related points of
doing something different, the training itself is reference for the same response), especially by the-
viewed as a natural process. To be human is to be matic variables or by sources that are unusual in a
shaped by one's environment, planned or un- given verbal community. It should be noted that the
planned. If this shaping is consistent with the pre- words "a given verbal community" bring one back
cise application of behavior principles, it will be to the comparative base used for determining origi-
particularly effective. Because, realistically, life in nality, which was discussed previously; it may be
its entirety cannot be controlled, the interaction of an impossibility to avoid a comparative base in de-
each person's behavioral history with the current fining creativity.
environment is bound to be unique, even for those Winston and Baker (1985) presented an exem-
who receive the same training. Therefore, each plary illustration of these two types of stimulus con-
trained person will bring a personalization to cre- trol. If the child painted a cityscape in bright colors
ative behavior that is in keeping with the traditional immediately after viewing a similar picture using
view. similar colors, the painting is under formal control
of the other picture. But, if the child decided to use
bright colors partly because the city is noisy and
exciting, then it is under informal stimulus control.
Determination of Originality It is doubtful whether one might know the circum-
stances under which the creative act took place; this
hinders one's ability to infer informal or formal
Comparative Base
stimulus control of a response. Certainly, an au-
Not only is originality the hallmark of creativity, dience's statement of the quality of a creative work,
but also the actual determination of originality which is only known outside of the creator's studio,
needs a comparative base-the repertoire of an in- might be a reaction to the product itself, not the
dividual, or the norms of a population, society, or assumed controlling variables.
culture of which that individual is, or is not, a mem- Because the behavior analyst regards thinking as
ber. The comparative base could be worldwide and influenced by environment, as is any other behav-
human-history based, and at the professional adult ior, then the thinking involved in informal stimulus
level, where competition reigns, it is. For preschool control leading to creative behavior should be able
children, which is the thrust of this chapter, com- to be trained by manipulating the environment. In-
petition is not considered a developmentally appro- deed, Della-Piana (1978) has done just that in train-
priate standard (Hendrick, 1975, p. 102). There- ing students to be under informal stimulus control
fore, the creativity of a preschool child is based for revision processes in writing poetry.
typically on that child's own past behavioral history Della-Piana's students were not preschool chil-
for a given task. dren, however, and characteristics of very young
children shed a different light on informal stimulus
control as a training technique or creativity criterion
Informal Stimulus Control for them. Because many three- and four-year-old
children do not respond differentially to appearance
In place of the comparative base for judging the and reality (Flavell, 1986) and think in concrete,
originality of creative behavior, some behavior ana- not abstract or metaphoric, terms (Lavatelli, 1970),
lysts have substituted the stimulus control of a beit is unrealistic to try to train them to be under
havior to determine its originality. Creative behav- informal stimulus control for creative responding.
ior is considered to be under informal rather than It follows that informal stimulus control should not
formal stimulus control (Sloane, Endo, & Della be a creativity criterion for them either.
Validation Measures Standardized Test Validation
Social Validation Some researchers using behavior-analytic meth-

odology have included scores on standardized tests,
A possibility for identifying the overall creativity such as the Torrance Tests of Creative Thinking and
or aesthetic quality of a product or behavior is called the Unusual Uses Test, in addition to criterion-
social validation. Judges thought to be suitable for a referenced creativity measures (Funderbunk, 1976;
specific evaluation are selected. As an example, Glover & Gary, 1976; Glover & Sautter, 1977;
undergraduate early childhood education majors Maltzman, Bogartz, & Breger, 1958). In this way,
may be chosen to rate the creativity of drawings by they widen their audience to include those who feel
five- and six-year-olds. Social validation has been standardized measurement is necessary for validity,
used by some behaviorists in their studies (Baker & and possibly satisfy their own personal conviction.
Winston, 1985; Ballard & Glynn, 1975; Maloney & Standardized tests do not test overall creativity
Hopkins, 1973; Maloney, Jacobson & Hopkins, per se, however, as shown by their subtests not
1973). being highly correlated (Wallach, 1985; Wallach &
In support of social validation, Wolf (1978) ex- Kogan, 1965). Furthermore, different creative
horts us to be responsible to the fundamental issue tasks require different classes of response sets (e. g. ,
of consumer satisfaction. Bushell and Dorsey blockbuilding response set of arranging hard-
(1985) feel that social validation sensitizes us to the edged, three-dimensional shapes and painting re-
natural contingencies of reinforcement provided by sponse set of drawing fluid one-dimensional lines).
our appreciative and approving society, and that the If there is a match between a response set( s) on a
effects of training in general will be better main- standardized test and a response set(s) trained, then
tained when the training incorporates natural the use of that test is appropriate. But if not, keep in
contingencies. mind that it is impossible for teachers to be found
Winston (1984) has discussed the weakness of accountable if they teach one thing and find them-
social validation in depth, a major difficulty being selves tested on another.
that such ratings generally do not constitute an ex-
perimental analysis of behavior. For instance, more
than one dimension of a child's drawing may be
changed by a training procedure, such as contingent Unresolved Problems
reinforcement for form diversity. If judges rate
drawings produced after this training as more cre- From the above discussion on the nature of
ative than baseline drawings, this change in ratings creativity, it is evident that a definitive answer of
may be due to a change in form diversity or a change what it is does not exist. Additionally, at times
in some other dimension of the drawings that is creativity needs to be viewed in a different light for
correlated with form diversity, such as the resulting the preschool child than for the adult. Although
overall pattern. both are expected to be original, the comparative
Yet another inadequacy of social validation of base for the child is that same child, whereas the
creativity is that it is a matter of taste, and taste is adult is subjected to a broader comparison base of a
fickle, changing from culture to culture and from culture(s). For both the child and adult, bizarre and
time to time within a culture. Even within the same predictable behavior are not considered creative;
culture and the same time period, social validation but imitative behavior for the child may have some
may be idiosyncratic to the particular group of acceptability. The training of creative behavior for
judges. both age groups may be seen as fitting as long as the
Creativity training for preschool children should concept of difference 01 informal stimulus control
not be concerned with the social validation of the takes precedence in the training over explicit exter-
quality of the resulting product. The process is more nal control; though training to be under informal
important than a polished outcome (Hendrick, stimulus control appears developmentally inap-
1980, p.136), which might be an unrealistic stan- propriate for very young children. Social validation
dard for a preschool child's skill level. Punishing or is more appropriate for the adult than the child, but,
putting a preschool child on extinction because of in either case, it is to be remembered that this mea-
poor creative quality based on social validation is sure of taste shifts over time.
obviously undesirable. Other than the instances of disagreement for the
criteria of creativity, depending on whether the term studies, and these changes will be covered
creator is a child or adult, serious problems for further on.
either age bracket revolve around the quantity of
diversity and the definition of informal stimulus
control of the creative behavior or product. The
Individual Analysis Experimental
criterion-referenced measures in the behavior anal-
ysis of creativity sometimes imply, and other times Designs
state outright, that the more diversity in the product,
the more creative it is. This is an oversimplification Because behavior analysis research gives promi-
because minimal art, dance, music, and poetry do nence to individual data, it is appropriate to present
not rest their creative labels on diversity. A very briefly the two common individual analysis experi-
few forms, movements, musical notes or words can mental designs. These designs are used to demon-
merit rave reviews for creativity from the critics. It strate that the independent variable (the training
is possible that criterion-referenced measures for procedure) of the manipulation of the stimulus be-
style rather than amount of diversity are in order. fore or after the response is actually causing the ob-
Winston and Baker (1985) have clearly stated served change in the dependent variable (creativity)
one dispute surrounding the definition of informal of the response. All the behavior analytic creativity
stimulus control: studies with preschool children to be reviewed used
one or a combination of these two basic designs to
the relationship between the Sloan et al. (1980) distinction of verify the experimental control of the observed
formal versus informal control and Skinner's (1957) distinction change.
of formal versus thematic control is not entirely clear. Skinner
(1957), for example, treated alliteration in poetry as a formal
source of strength, whereas Sloane et al. referred to the repetition
Reversal
of sounds as informal control. (pp. 200-201)
A reversal design is used when the behavior
Furthermore, another problem in using informal might easily reverse or revert to its original pattern
stimulus control as a measure of creativity is that, if treatment is withdrawn. In a reversal design, the
by adult standards, several persons under the same normal or baseline rate of behavior is measured
type of informal stimulus control would be called before any attempt is made to change it. For exam-
imitative or derivative, emphasizing the comple, if a teacher wished to increase a child's diver-
parative base of a population, society, or culture for sity in building block forms, occurrences of each
creativity. Another contradiction to the informal different form in several blockbuilding sessions
stimulus control standard for creativity is seen in would be measured before treatment began. Such a
that adult artists in the realistic movement of the baseline measure would be taken until the behavior
1970s received worldwide acclaim for almost the appeared stable.
exact replication of objects and persons, a point-by- During treatment, descriptive verbal praise
point correspondence between each discriminative would be contingent on each different block form
stimulus and each response. Informal stimulus con- per session, and the measurement would continue.
trol, then, cannot always be the criterion for creati- If the number of block forms per session increased
vity. during treatment, the treatment would be discon-
Finally, objectively defining measures for vari- tinued temporarily to find out whether the treatment
ous creative tasks is difficult as implied by the might have been the cause of the increase. When the
sparse applied research in creativity with preschool treatment is withdrawn and if the number of block
children. And, as Winston and Baker (1985) have forms decreases or reverses toward the baseline
decried, this applied research has been concerned rate, this change would demonstrate that the treat-
with short-term specific training rather than the ment was the cause of the behavior change. The
complex chains of behavior that are influenced by more reversals, the stronger the experimental con-
one's home, school or work place, and community trol. The reversal procedure may take several
that would be of tremendous value for our under- forms, such as a return to the baseline condition,
standing of the long-term development of creati- differential reinforcement of any other behavior
vity. In any event, the application of behavior prin- than the creative behavior, or differential reinforce-
ciples does change creative behavior, according to ment of incompatible behavior.
the objective definition of creativity used in short- In the normal course of using this reversal de-
sign, the treatment would be reinstated after the first baseline or control group and another group as the
reversal, and the block forms would once again treatment group.
increase. It has been found that, when behavior is
reversed again and again, it usually will not return
to the original baseline rate and eventually may not Behavior Principles
reverse at all. Baer, Rowbury, and Goetz (1976)
called this phenomenon the behavior trap; that is, Behavior principles are natural laws of behavior
the child finds the changed behavior reinforcing in that are based on objective data. The application of
itself or finds the natural contingencies resulting behavior principles to understand and solve human
from that changed behavior reinforcing and con- problems has become an area of research known as
tinues, therefore, to engage in it in the absence of applied behavior analysis. Creativity is one of the
treatment. human endeavors to which behavior principles have
been applied.
Multiple Baseline The concept of operant behavior helps one under-
stand the derivation of behavior principles. The
The multiple baseline design is used when teach- term operant refers to a class of related behaviors, a
ing a behavior that is not expected to reverse. For response set. In this instance, operant refers to a
instance, if a teacher wished to increase the repre- class of creative behaviors for each different experi-
sentational content in a child's drawing by use of mental task (e.g., blockbuilding and easel paint-
self-instructions, the teacher might feel this pro- ing). In general, an operant is a behavior influenced
cedure may not reverse because the child is, to some by (1) discriminative stimuli that precede the be-
extent, in control of the treatment, using his or her havior and signal the probability of reinforcement
own self-instructions. Again, as in the reversal defor the behavior, and/or (2) reinforcing stimuli that
sign, the normal or baseline rate of the behavior is follow the behavior. All behavior and the associ-
measured before any attempt is made to change it. ated preceding or following stimuli are defined ob-
In the case of the multiple baseline design, base- jectively for the systematic observation of their oc-
lines are taken on the same behavior for more than currence and relationship. Speculation does sur-
one child (or, as alternatives, more than one behav- round thinking which, though not observable, may
ior for one child, or in more than one setting for the be inferred from a related observable behavior or
same behavior for the same child). Then treatment product; it is considered a behavior influenced by
is applied successively to each child. After a defi- the environment in the same way that observable
nite effect has been observed for the first treatment behavior is. But there is no theory regarding hypo-
with the first child, the second treatment with the thetical processes, such as overall creativity, cogni-
second child is introduced, and so on. Thus, each tion, motivation, or personality; behavior is viewed
successive child will have a longer baseline than the and analyzed strictly in terms of functional rela-
previous child. Each succeeding treatment with a tionships with the environment. The operant learn-
different child serves as a replication of the effects ing theory, which is derived from the effects of
of the previous treatment. The more replications, discriminative stimuli and reinforcing stimuli, is
the stronger the experimental control. that behavior that is reinforced increases (see
Figure 1).
Grouped Individual Data Techniques based on behavior principles are far
more numerous than these principles themselves,
When group data are presented in a behavior
analysis, it is typically a collapsing of individual
data in which each individual has had a baseline rate
of behavior taken before being treated. The group
data may be presented, mentioning how many indi-
----
... -- -- - -- ..........
Sign:l:. ~r~~".?~;~ Of
-- .....
viduals did or did not follow the group trend. In the
reversal design, each subject serves as that subject's S - - - _ IR •
D Increases Increases Sn+
n
own experimental control; in the multiple baseline
Discriminative Response Reinforcing Stimulus
design, control subjects are the ones with the longer Stimulus
baselines. By contrast, in traditional research with
matched subjects, one group might serve as the Figure 1. A schematic drawing of the operant learning theory.
~hich embrace reinforcement, punishment, extinc- physical properties or verbal content. In other
tIOn, schedules of reinforcement, shaping, stimulus words, if a particular contingency increased a be-
control, chaining, generalization, and self-control havior, it is a reinforcer (everything else being
(see Goetz, 1982a, for a discussion of these princi- equal), regardless of what it may otherwise be.
ples). For example, some of the techniques based What is reinforcing to a given child in a particular
on reinforcement are attention, general praise, de- situation is determined by that child in that situa-
scriptive praise, food reinforcers, trinket reinfor- tion, not by the teacher.
cers, and token reinforcers.
Goetz and Baer (1971, 1973): Blockbuilding.
The first behavior-analytic creativity study with
Characteristics of Effective Teaching of preschool children was conducted by Goetz and
Creativity to Preschool Children Baer and published in 1971; later, the data were
reanalyzed and published in 1973. Descriptive ver-
Behavior-analytic studies of creativity for pre- bal praise was used that was contingent on each
school children have focused on the principles of different precoded block form built by three chil-
reinforcement, stimulus control, and generaliza- dren in individual training sessions. Descriptive
tion. Additionally, some of these studies have mea- praise consisted of a detailed account of the form, a
sured other variables along with behavior princi- statement that the form was different from any form
ples, such as social validation, material and time built within that session, and praise (e.g., "That is
limits, maintenance, and a flexible creativity code. a very nice arch going up, across, and down, and it
In their comprehensive review of behavior-analytic is the first time you built it today. Good going!").
studies of creativity, Winston and Baker (1985) Form diversity per blockbuilding session for all
mentioned nineteen such studies with eight of them the children increased with this intervention. New
having preschool (pre-kindergarten) children as forms, never before in the child's total prior se-
subjects. In addition to those eight, five more un- quence of blockbuilding sessions, emerged at high-
published studies, many of which were covered by er rates during reinforcement of form diversity than
Goetz (1982b), will be touched upon in this during periods of reversals without praise con-
chapter. tingent on form diversity. The appearance of new
forms was a side effect of reinforcing form diversity
per session. This side effect may be viewed as gen-
Reinforcement eralization along the gradient underlying block-
Reinforcement, a basic behavior principle, inform inventions.
volves the application of an event (a stimulus) after
some action (a response) so that the application will Goetz and Salmonson (1972): Easel Painting.
cause that action to be repeated or increased (see In 1972, Goetz and Salmonson continued the inves-
Figure 2). To be most effective, the stimulus should tigation of the effects of praise on the form-diversity
be applied immediately after the action, although aspect of creativity but used another medium, easel
delayed reinforcement in some cases may be effec- painting, for which forms were precoded. General
tive (Fowler & Baer, 1981; Rogers-Warren & Baer, verbal praise and descriptive verbal praise con-
1976). A stimulus may be considered reinforcing tingent on form diversity per session were com-
only in terms of its effect, and not in its extrinsic pared. General praise acknowledged a form as
being good, but no specific aspect was singled out
for comment. Descriptive praise was the same as
defined in the foregoing Goetz and Baer study. For
\ncreases three children in individual sessions, descriptive
~SR+
praise resulted in higher form diversity per painting
and increased form diversity more quickly than did
general praise, though general praise had a slight
effect. However, there was no systematic increase
of new forms associated with either type of
Response Reinforcing Stimulus
reinforcement.
The stimulus control aspects of the Goetz and
Figure 2. A schematic drawing of the reinforcement paradigm. Salmonson reinforcement procedures deserve com-
ment. Descriptive reinforcement, which included ilar to the Goetz and Baer (1973) side effect of new
describing the form and denoting its difference block forms that increased when form diversity
from other forms currently present and praise for within a single product was reinforced.
that form, may have had the stimulus control func- During the last five sessions of the Romero et al.
tion of signaling that production of a different form study, a reinforcement fading condition was imple-
in the future would be reinforced. Perhaps, in addi- mented in an effort to bring the reinforcement con-
tion to the reinforcement procedure, or instead of it, tingencies of praise to a level that was thought to be
the verbal description of difference and praise for comparable to that of a natural environment. It was
that difference was controlling the probability that found that the complexity of the children's designs
future painting responses would be varied. Thus, was maintained during these last five days and did
descriptive reinforcement seemed to have an in- not differ from the level of complexity under pre-
structional component for future responding. But vious reinforcement.
general reinforcement did not have descriptive
stimuli associated with the experimenter's ver- Goetz (1981): Blockbuilding. In two of the pre-
balization and did not increase form diversity as ceding studies, creativity training consisted of de-
much as descriptive reinforcement. scriptive reinforcement for each different form
(form diversity) in a single block construction or
Romero, Holt, Stella, Baer, and Etzel (1978): painting. A side effect of reinforcing this form di-
Colored Cubes. The comparative effects of gener- versity, for one of the studies, was an increase in
al and descriptive praise were again investigated by new forms that appeared for the first time across all
Romero, Holt, Stella, Baer, and Etzel (1978) with sessions of the study. Still, this variation of
four children. The experimental task was the ma- creativity training was not reinforcing novelty per
nipulation of a single layer of colored cubes within a se, because not only new forms were being rein-
square-shaped container; the surface of various forced across sessions but also some "old" forms
combinations of cubes formed two-dimensional that had been reinforced in previous sessions. It was
shapes (triangle, line, diamond, and arch). The par- thought that reinforcement for only new forms ap-
ticipating children tended to use one solid-colored pearing for the first time in a study would better
side of the cube only, which did not result in making approximate the reinforcement of "true" novelty.
any of the four shapes. Two simple shapes were This procedure, however, would result in a thin
trained: a triangle and a line. Descriptive praise schedule of reinforcement, and it seemed doubtful
(e.g., "That's a nice blue triangle") was more ef- that this minimal descriptive praise initially would
fective in increasing the frequency of the triangle change a behavior.
and line than general praise (e.g., "That's nice"), Considering the above, Goetz (1981) examined
which had a small effect compared to baseline per- the effects of descriptive praise contingent only on
formance. These results supported the Goetz and blockbuilding responses that were considered new
Salmonson (1972) findings that descriptive praise in relation to all forms in previous experimental
was more effective than general praise and again sessions-true novelty. Seven of the nine chil-
suggested that descriptive reinforcement has an ef- dren's new forms increased appreciably during this
fective instructional component. training. One child's data suggested an increase,
The Romero et al. study also had an important and one showed no effects.
side effect. As the children became more skilled in Form diversity per construction was analyzed
making the two trained shapes during the condition also as a possible side effect of reinforcing only new
of descriptive reinforcement, their cube designs forms. Four of the nine children showed a corre-
within the square-shaped container began to display sponding increase in form diversity when new
two complex shapes that were never descriptively forms were reinforced. One child's data suggested a
praised: a diamond and an arch. These two shapes small increase. Four children's form diversity did
not only increased in number but were also used in not change.
different combinations of color and position. The Although the reinforcement schedule was indeed
overall appearance of the designs in the set of a thin one, often with only one reinforcement per
colored cubes then appeared to be more complex, session, this study showed that, overall, minimal
novel, and symmetrical when compared to initial descriptive praise produced marked effects. Most
baseline designs. This side effect-two untrained children's new forms increased over time, and, in
shapes appearing after two trained shapes-is sim- addition, half of the children's form diversity per
construction increased. Romero et al. (1978) had praise described the movement; and (3) creative
shown that minimal reinforcement maintained cre- praise mentioned that the child was engaged in sev-
ative behavior during the last days of training. This eral different movements, and it could only be
study demonstrated that minimal reinforcement given if indeed the child were engaging in different
also could result in initial increases in creative behaviors. The fourth and last condition used cre-
behavior. ative instruction, which told the children to try to
move in different ways. The highest data points for
new movement were under the creative instruction
Stimulus Control
condition. Although this pilot study lacked rigid
As previously noted, an operant response may experimental controls (i.e., no reversals or multiple
not only be effected by the stimulus that follows and baselines, only sequential conditions) and only pre-
reinforces it but also by the discriminative stimulus sented group data, it is included here because it
that precedes the operant and signals the probability suggested that children's dance movements may be
of its reinforcement (see Figure 3). The control of coded and observed and that instructions may in-
the discriminative stimulus over the following re- crease diversity in another type of creative behav-
sponse involves the behavior principle of stimulus ior, dance.
control.
The discriminative stimulus may be visual, au-
ditory, or in the form of a teacher's tactile (manual) Figgs, Dunn, and Herbert (1971): Blockbuild-
guidance or modeling. When these prompts, ining. Four types of verbal prompts were analyzed
structions, or models result in a desired response, by Figgs, Dunn, and Herbert ( 1971) for their effects
that response is then reinforced by the teacher. In on form diversity in blockbuilding with a three-
this way, the prompts, instructions, or models be- year-old subject. The categories of prompts were:
come signals for certain child behaviors that will be (1) suggestions (e.g., "Can you build a farm to-
reinforced. Instruction, verbal prompts, and model- day?"); (2) difference (e.g., "Let me see you build
ing are forms of stimulus control that have been something different today"); (3) repeat (e.g., "Can
used in creativity training. you build another one like that?"); and (4) total
usage (e.g., "Let's use all the blocks today"). The
Reese and LeBlanc (1970): Dance. Reese and most effective prompting technique in increasing
LeBlanc (1970) measured motor behavior when form diversity per construction was the suggestion
they investigated the effects of three types of praise to build a specific structure, such as a farm, garage,
and one type of instruction on the creative spon- or fire station, all of which were familiar to the child
taneous, not trained, dance movements of a group and had a connotation for a wide variety of possible
of four preschoolers. Each child's dance move- forms.
ments were recorded as same, new (several differ- In analyzing their results, Figgs et al. speculated
ent movements), or none during the first few sec- that perhaps their three-year-old subject did not
onds of a 20-sec time sample for individual comprehend the concept of difference, which
children, while they were all dancing in a group to would explain the ineffectiveness of that category
recorded music. Three types of praise were directed of prompt. To compare these results with the effec-
to individual children: (1) general praise did not tiveness of the creative instruction for different
mention the specific movement; (2) descriptive dance movements in the Reese and LeBlanc (1970)
pilot study, one needs to know that three of their
subjects were four-year-olds and one was three-
years-old. The suggestion prompt, then, may be the
starting point for explaining the concept of dif-
ference to a very young child. The suggestion
prompt could be tailored to various types of
creativity tasks.
Discriminative Response Reinforcing Figgs and Herbert (1971): B lock building.

Stimulus stimulus
Figgs and Herbert (1971) conducted another study
Figure 3. A schematic drawing of the stimulus control with a series of conditions that used separately, or in
paradigm. combination, various types of possible discrimi-
native stimuli-modeling, training, and prompts

preceding the desired creative behavior-in addi-
tion to descriptive praise following the creative be- R--....,·SW
havior. For both their subjects, the combined use of Response Reinforcing Discriminative Response
prompts for difference and descriptive praise for stimulus stimulus
difference was the most effective condition for in-
creasing form diversity. These results, however,
might have been due to the sequential effects of the ~
preceding modeling and training conditions. ~-----~.~;:ft+ OR
As stated earlier, the Goetz and Salmonson Fading of Reinforcing Fading of Discriminative
(1972) painting study and the Romero et al. (1978) Stimulus Stimulus
colored-cube study have suggested that descriptive
~
reinforcement has an effective instructional compo-
... R
nent that is a discriminative stimulus for producing
"-.
.',0
future novel behavior. These findings, like those of ... :
Reese and LeBlanc (1970), Figgs et al. (1971), and Fading of Reinforcing Fading of Discriminative
Figgs and Herbert (1971), demonstrated that certain Stimulus Stimulus
discriminative stimuli preceding a creative behav- Figure 4. A schematic drawing of the generalization paradigm.
ior may make it more likely to occur. Although Generalization may occur across behaviors, settings, or persons.
prompts for difference, prompts to use all mate-
rials, modeling, and training did not seem to be very
effective, at least, when used in isolation, sug-
gestions and the in.structional component of de- dren painted independently at the easel for one ex-
scriptive reinforcement did seem to be quite effec- perimenter and then built with blocks for another
tive. The combined use of a prompt for difference experimenter in another room. The training tech-
and reinforcement for difference seemed to be par- nique was descriptive verbal reinforcement con-
ticularly successful. tingent on each different form painted per picture.
There was no training in blocks.
For both children, praise in easel painting in-
Generalization
creased both form diversity per picture and new
Generalization is another behavior principle, and forms that appeared for the first time in the study.
it occurs when a behavior is not tightly controlled Blockbuilding form diversity increased during the
by the stimuli involved in training that behavior (see period in which form diversity in painting was
Figure 4). Generalization may occur across behav- praised-even though there was no praise in block-
iors (tasks), subjects, settings, trainers, and/or building-and decreased when the praise for form
time. When a teacher trains a desirable behavior in a diversity in painting was discontinued. There was,
child, the teacher should plan generalization of the however, no generalization from the increased
behavior to different nontraining conditions for the number of new forms in easel painting to new forms
training to be effective in the child's total life in blockbuilding. So, there was some generaliza-
experience. tion (form diversity, but not new forms) from easel
Does generalization of creative behavior occur? painting to blockbuilding, but the generalization
Specifically, when creative behavior change occurs was incomplete. Because a different experimenter
in one type of task, whether through reinforcement was present for each task, and the session was con-
or stimulus control training, will there be creative ducted in a different room, the experimenter and the
behavior generalization to another task? Several setting themselves were not likely to become a dis-
behavior-analytic studies have looked into this criminative stimulus for making different and/or
question. new forms in the untrained task.
Holman, Goetz, and Baer, Experiment 1 (1977): Holman, Goetz, and Baer, Experiment 2 (1977):
Easel Painting and Blockbuilding. In Experi- Felt-Pen Drawing, Easel Painting, Lego Building,
ment 1, Holman, Goetz, and Baer (1977) investi- and Blockbuilding. In Experiment 2, Holman et
gated the generalization of creative behavior from ai. (1977) implemented a more complex experi-
easel painting to blockbuilding. Each day, two chil- ment on generalization than the above study, which
examined creative responding across topograph- during baseline, and then each child was trained to
ically similar and dissimilar tasks. Each of three improvise by experimenter exemplars within a
children performed four tasks each day, felt-pen given class of tools. All five children were trained
drawing and easel painting (similar tasks), and in hammers, two were trained in two classes of
Lego building and blockbuilding (similar tasks). A tools, and two were trained in the three classes. All
comparison of dissimilar tasks would be block- children showed increased improvisation within the
building and easel painting or Lego building and trained class, but not across classes. Tools fabri-
felt-pen drawing. Descriptive verbal praise was de- cated by item combination were rare in baseline but
livered contingent on felt-pen drawing only. common in training. Follow-up probes immediate-
Generalization from felt-pen drawing to the topo- ly after training showed that the effects were main-
graphically similar task of easel painting was shown tained. The same experimenter participated in all
clearly for two children, and to some extent for the sessions and did not have the effect of being a dis-
third child, both in form diversity and new forms. criminative stimulus for improvisation across tool
Generalization to the topographically dissimilar classes.
task of blockbuilding and Lego building was much
less evident, though for one child it appeared some- Lane, Lane, Friedman, Goetz, and Pinkston
what correlated. One child who later received (1982): Felt-Pen Drawing and Collage Construc-
praise for form diversity in Lego building showed tion. In the Lane, Lane, Friedman, Goetz, and
no generalization to its topographically similar task Pinkston (1982) experiment, generalization was ex-
of blockbuilding. amined across tasks and settings. An experimental
Each child perfOrn'led all four tasks for the same group of nine children received descriptive praise
experimenter in the same setting. Therefore, the and tokens contingent on form diversity in felt-pen
experimenter and the setting could have become the drawing during individual sessions outside the
discriminative stimuli for the making of different classroom on Mondays and Wednesdays. On Tues-
forms in all four tasks. Thus, generalization, depen- days, these children joined the control group of nine
dent on a specific skill that had been trained in one children for felt-pen drawing, without training just
task and could be used in the related task, occurred. practice, in the classroom. On Thursdays, both ex-
Postchecks two months later revealed that the two perimental and control groups engaged in collage
children who had received the greatest number of construction, without training, in the classroom.
reinforcement training sessions, within the multi- The experimental children's form diversity in
ple-baseline experimental design, maintained their felt-pen drawing increased in the training sessions;
high levels of form diversity in the directly manipu- but an increase in new forms that appeared for the
lated task and its topographically similar task. first time throughout all sessions was not a side
effect, which the experimenters speculated might
Parsonson and Baer (1978): Tools. The cre- be related to the ability in receptive verbal language
ative operant analyzed in the Parsonson and Baer of their inner-city black subjects. The experimental
(1978) study was an improvisation using the three group had significantly higher scores for form di-
tool classes: hammer, shoelace, and container. Im- versity during training than the control group.
provisation was defined as finding a substitute to There was some generalization of this increased
replace a specifically designed, but currently un- form diversity to the practice sessions in the class-
available, tool ordinarily used to solve the problem. rooms for felt-pen drawing (generalization across
The procedure was to present a child with a prob- settings) but not for collage construction (gener-
lem, such as hammering a wooden nail into a board, alization across tasks). The experimenters were not
along with a tray of miscellaneous items for possi- present in the classroom, and, therefore, they could
ble improvisation. In the case of hammering a nail, not be discriminative stimuli for increased form di-
the items might include a small rock, a large rock, a versity in the classroom. The two tasks of felt-pen
shoe, a block with a drilled hole, a rod that would fit drawing and collage construction were topograph-
into the drilled hole, and a Styrofoam hammer. ically dissimilar, a characteristic which does not
Some of these items could be used for hammering seem to facilitate generalization as evidenced by the
the nail into the board, some could not, and others partial generalization from easel painting to block-
could be combined to do so. building in Experiment 1 of Holman et al. (1977);
Spontaneous improvisation skills were measured the lack of generalization from felt-pen drawing to
blockbuilding or Lego building in Experiment 2 of and Winston (1978) with judges who were early
Holman et al. (1977); and the lack of generalization childhood education students and parents of pre-
across tool classes in Parsonson and Baer (1978). school children. These experimenters increased
In the four preceding studies, generalization was color diversity of felt-pen drawing for three chil-
monitored, not specifically planned. From the vary- dren and form diversity for two of the same three
ing procedures used, it seems that the same experi- with cartoons contingent on each child's reaching
menter involved in sessions using topographically an accelerated criterion based on previous session
dissimilar tasks does not become a discriminative performance. Both the college students and the par-
stimulus for form diversity across those dissimilar ents gave higher ratings of creativity for drawings
tasks, because specific different skills are used and with increased form diversity but not for drawings
need to be trained for dissimilar tasks. For example, with increased color diversity.
the specific skill of blockbuilding involves the con- Ryan and Winston felt that social validation pro-
struction of three-dimensional pieces for different vided information on what behavior to change, spe-
forms. The specific skill of easel painting involves cifically in the design of the art curriculum. On the
the application of a fluid color on a flat surface for negative side though, they stated that such a narrow
different forms. Generalization across topograph- conception of creativity is inconsistent with art edu-
ically similar tasks, however, was noted in Experi- cation as an occasion for a child to engage in rela-
ment 2 of Holman et at. (1977). Another possibility tively unrestricted expression. Nevertheless, social
is that creative behavior may need to be trained in evaluation may advance our understanding of the
many more different tasks than has been done in interaction between children's creative efforts and
these studies before its generalization will occur. the social environments, which is an aspect, in it-
Generalized side effects within the same trained self, of effective teaching. For, if a teacher under-
experimental task has occurred, however. Goetz stands any possible negative reaction of the child's
and Baer (1973) and Romero et al. (1978) found community or of significant others to that child's
generalization of new forms as a side effect of in- creative efforts, that teacher may prepare the child
creased form diversity per product in blockbuilding to continue the pursuit of individualized creative
and colored cubes. Parsonson and Baer (1978) behavior regardless of the reaction of others.
found that different tools which were fabricated by
item combinations increased within the training of a
single tool class. Material and Time Limits
The findings from these studies emphasize the Several of the previous studies suggest what ef-
need to plan consciously for generalization using fect material and time limits might have on training
the Stokes and Baer (1977) generalization tech- creative behavior. In the Goetz and Baer (1973)
nology: (1) modify sequentially; (2) introduce to blockbuilding study, two of the three children had
natural maintaining contingencies; (3) train suffi- use of an unlimited number of blocks for building.
cient exemplars; (4) train loosely; (5) use indis- The third child was limited to using the same 53
criminate contingencies; (6) program common blocks for each session. Yet the data for the third
stimuli; (7) mediate generalization; and (8) train to child were similar to the data for the other two; form
generalize. Kohler and Greenwood (1986) have diversity and new forms increased for all the chil-
provided an extensive discussion for the identifica- dren during reinforcement for different forms per
tion of natural contingencies of reinforcement that session. In all the other blockbuilding studies, a
seem particularly relevant here for those who are limited number of blocks, approximately 50, was
orchestrating generalization in their training as well used, but that limit did not seem to hamper the
as for those who champion creativity as a natural effectiveness of training.
process. Three studies used time limits for the experimen-
tal creative task. Fallon and Goetz (1975) allowed
their children a 3-min limit for felt-pen drawing;
Social Validation Ryan and Winston (1978), as-min limit forfeit-pen
drawing; and Lane et al. (1982) a lO-min limit for
Ryan and Winston (1978): Felt-Pen Draw- felt-pen drawing. In all these studies, training pro-
ing. Creativity may be evaluated by the commu- duced positive results similar to the Goetz and
nity through social evaluation, as was done by Ryan Salmonson (1972) easel-painting study and the
painting and drawing tasks in the Holman et al. er made on-the-spot decisions as to what constituted
(1977) studies, none of which used time limits. The new forms not in the code and praised them if they
crucial factor in the effectiveness of these studies, were appropriate. After each session, the experi-
then, did not appear to be time but rather the con- menter took a picture of the construction to compare
tingent reinforcement. it with all the preceding pictures to determine if that
form had ever appeared before. If it had, the form
diversity score for a construction would have to be
Maintenance readjusted, but, as it turned out, readjustments did
Accountability of teacher performance requires not have to be made. Using this flexible code, the
one to assess the maintenance of the behavior child's form diversity score increased considerably
trained. Some of these studies shed light on mainte- during reinforcement conditions.
nance at various stages. Romero etal. (1978) found A flexible code seems particularly appropriate
that creative colored-cube designs were maintained for creativity research because it can parallel a
as the contingent reinforcement schedule was child's personal spontaneous efforts, the essence of
thinned during the last five days of training. Parson- creativity. However, the use of individualized
son and Baer (1978) reported that there was mainte- codes for several children does not allow a direct
nance of creative improvisation of tools during comparison among them. But, for the teacher, who
probes immediately following training. Fallon and is not a researcher but rather a person who is com-
Goetz (1975) and Holman et al., Experiment 2 mitted to developing the potential in each child,
(1977) reported maintenance of increased form di- flexibility in assessment may well be used.
versity several months after training for those chil-
dren who had received many training trials (8-12)
in felt-pen drawing, but not for those who had re- Development of a Focused Environment
ceived fewer trials (5-6). It seems that there is gen- for Creativity of Preschool Children
erally short-term maintenance and sometimes long-
term maintenance, at least for two months, if the
training is of considerable length. Whether mainte- Ideally, the child's total environment, including
nance extends beyond several months has not been the home, school, and community, should be holis-
tested. tic for a child's creative behavior to occur to the
At some point, the long-term effects of creativity fullest degree. Although this utopia will probably
training in these studies would no longer be appro- never exist, perhaps one can approximate it and
priate for testing suitable response sets for a child's promote creativity within each of these settings and
interests change. For example, a child may build across settings by keeping the following advice in
with blocks at ages 3, 4, and 5, but not at age 6, mind.
because blockbuilding, at this time, may be consid-
ered childish; therefore, a child may no longer be Prerequisite Behaviors
motivated to build with them.
Children are not miniature adults. Their motor,
language, and cognitive skills are elementary. A
A Flexible Creativity Code creative task should be matched to, or on the outer
edge of, each child's skill level-a developmen-
Elliott and Goetz (1971): Blockbuilding. When tally appropriate task. Therefore, one needs to
behaviors are precoded before a study begins, there know the current level of behavior before selecting
is always the possibility that a child will engage in the next skill in the sequence to be taught. All skills
experimentally relevant behavior that may not be have prerequisite behaviors, though the exact se-
recorded because it was not defined initially in the quence of them may not be invariably fixed.
observation code. Using descriptive praise con- Any child development text will outline the pro-
tingent on blockbuilding form diversity per con- gressive motor, language, and cognitive skill levels
struction, Elliott and Goetz (1971) avoided this prob- to be considered in selecting a developmentally ap-
lem in a single-subject study in which 20 common propriate task, but that is not the intent here. Gener-
block forms were used as a starting point, and addi- ally, one moves from selecting tasks requiring gross
tional forms were added to the code as the child built motor to fine motor skills, simple to complex ex-
them. While the child was building, the experiment- pression, and concrete to abstract thinking. That is,
a child will be given finger painting before easel Disposition

painting, a picture book before a storybook, and Dispositions are broadly defined as relatively en-
language-experience story writing before poetry during "habits of mind" or characteristic ways of
writing. responding to categories of experience across t~pes
Selecting creative tasks for children to explore of situations (Katz, 1985). One's past behavlOral
sounds so simple, but it is not quite that easy. Each history plus current behavioral interactions deter-
child is different and surprisingly so. Formulas mine a disposition, given certain biological charac-
using ages or stages cannot be used. Furthermore, a teristics. Creativity is an example of a disposition.
child may show a preference for a task that an adult Therefore, if one wants a child to have a disposi-
feels is out of reach of that child's skill level, and tion toward creativity, that is, to respect, admire,
yet the child's interest in the task may compensate and enjoy creative behavior and be desirous of par-
in the effort that is applied for the lack of skill. ticipating in the creative process, one ne~ds t~ trai.n
Always give the child's preference a try. The cru- this disposition. The child may learn thIS dISPOSI-
cial question is, Is the child primarily participating tion from observation and emulation of models
in this task independently or does the task require when this emulation is appreciated and acknowl-
inordinate amounts of adult help? The child must edged. The adult who wants children to engage in
participate in the task to learn the creative behavior creative behavior will expose them to creative be-
involved whether at home, in school, or in the havior or its products, model the behavior, and rein-
community. force it. Many opportunities for exposure to
creativity exist, such as visits to museums, theaters,
Opportunities to Learn buildings of varying architectures, and gardens. Of
course, it is a must to have a creative ambience in
Research has clearly demonstrated that the op- the school, in the home, and in the community.
portunities to learn are related to the amount of
learning (Berliner, 1985). Opportunities to learn
may be analyzed as (l) how much time i~ allocated Lack of Criticism
to a given task; (2) whether the procedure is so Criticism of creative behavior is undoubtedly a
structured during that time that the child under- punishment that would decrease the behavior.
stands what is to be done; and (3) to what extent the Teachers of young children are not likely to do this,
child is engaged with the task during that time. The but they should be aware that lack of reinforcement
child needs opportunities to learn to be creative. for creative behavior may be construed as criticism,
Creative activities need to be scheduled with especially in an atmosphere in which children are
clear instructions of what may be done, and the accustomed to praise for positive behavior.
teacher needs to be aware of the amount of child In his classic brainstorming study, Osborn
participation during this time. If the time is allo- (1953) found that more original ideas were evoked
cated and the child is not on task, then perhaps the in group ideation in the absence of criticism or judg-
procedures are unclear or the behavior principles of ment of the value of an idea. An implication of
reinforcement, stimulus control, or generalization Osborn's results is that lack of criticism in a group
are not being implemented properly. Perhaps be- of young children seems desirable for an atmo-
havioral techniques that have been used successful- sphere in which creativity flourishes without inhibi-
ly with tasks other than creativity or with slightly tion. Torrance (1980a,b, 1982), in his quality cir-
older children could be tried, such as delayed rein- cles, a group of gifted and talented children
forcement, praise after and not during the task brainstorming and working together to solve their
(Fowler & Baer, 1981; Rogers-Warren & Baer, common problem, similarly downplays criticism to
1976), or self-instruction, that is, the child instructs enhance a proliferation of ideas to be evaluated
himself or herself to be creative (Baker & Winston, later. Children may be taught not to criticize their
1985). peers' efforts toward creative behavior, and teach-
Finally, the family in the home and teachers in ers and parents should know better.
the school or community should capitalize on spon-
taneous opportunities to learn to be creative. As a
case in point, if a child brings a pet gerbil to school Climate
one day, a creative maze and obstacle course for the In learning creative skills, the child benefits mos
gerbil could be built impromptu. from a constant climate that nurtures and sustain
that child's creative behavior. Thus, parents in the at this young age, might be other practical tech-
home and teachers in the school and in the commu- niques for the overworked teacher in that they
nity need to follow the same creativity guidelines of would require only some initial training and possi-
selecting developmentally appropriate tasks, provi- bly intermittent prompts. These self-control tech-
ding opportunities to learn, teaching a disposition, niques, too, are intriguing because they are ger-
and avoiding criticism. mane to the traditional concept of creativity as an
Furthermore, parents and the school and commu- expression from the inner self.
nity teachers need to be aware of what each other is Another avenue for scientific study is the func-
doing. Parents volunteering in school and commu- tion of peers on a preschool child's creativity. For
nity activities is fine, but realistically they often example, can a creative child and a noncreative
cannot and should not be expected to do this. Infor- child be paired at a drawing table with the result that
mal conferences, notes, and telephone calls may modeling by the creative child and subsequent im-
suffice to keep them in touch with each other. In this itation by the noncreative child take place, leading
way, the parents can talk with their children about to the noncreative child's becoming independently
their creative accomplishments in school, and creative? Or can the creative child be asked to help
teachers can talk with the children about such ac- the noncreative child with a positive result? Or can
complishments at home. An open line of commu- the creative child be trained in prompting and rein-
nication across the home, school, and community forcing creativity in the noncreative child with a
settings will keep the child's climate supportive of beneficial effect?
that child's creative efforts, which should then gen- On a different tack, can the types of behavior ex-
eralize from one setting to the other. amined on a one-to-one basis in this chapter be
trained in a group setting? As an exemplar, can a
teacher have a group of three or four children work-
Summary and Discussion ing on individual collages at the same table, with
the teacher circulating around them and praising
Given the foregoing studies in which behavior individual creative acts every so often, and have
principles are applied, it seems reasonable to hold creativity increase? Additionally, will these chil-
that creativity can be taught to preschool children. dren be vicariously reinforced from each other?
Even with unresolved problems in defining Does an overall creative preschool environment,
creativity, it appears that some dimension of in which creative teachers use creative curriculum
creativity, empirically defined, can be taught. activities in a creatively furnished and decorated
Many techniques that are based on behavior prin- classroom, foster creativity indirectly as measured
ciples, however, have yet to be researched in teach- by specific tasks? Or is direct teaching that uses
ing creativity to preschool children. Delayed rein- precise techniques needed for creativity to increase
forcement has not been used and, if successful, in those specific tasks?
would be a practical technique in the typical class- Finally, long-term maintenance of the types of
room where a teacher cannot attend to all children creative behavior presented here has not been as-
immediately. For instance, can a child build with sessed. How long does it maintain? On a kindred
blocks in a comer of the classroom, leave the con- issue, does creativity trained on early childhood
struction there for a teacher to examine later, and at tasks (e.g., blockbuilding) generalize to creativity
the end of the school day be praised for any creative on a related middle childhood task (e.g., carpen-
efforts exhibited in the blockbuilding with the result try)? From the above, it is clear the behavior-analy-
of different forms increasing over time? tic research with preschool children has only begun.
An intermittent schedule of reinforcement for If young children are given many developmen-
creativity deserves experimental examination. tally appropriate opportunities to learn to become
Again, for the realistic reason that a busy teacher creative in supportive environments at home, at
cannot continually monitor a child's creative task, school, and in the community, and if the principles
intermittent reinforcement would be a welcomed of behavior are understood and used, the teaching
technique, if it works. As a case in point, can a child of creativity to all children seems a probability.
work alone at the easel and have a teacher drop in
now and then to praise different forms with the
result of the number of them increasing? References
Self-instruction and self-reinforcement, not yet Baer, D. M., Rowbury, T. G., & Goetz, E. M. (1976). The
used in research as training procedures for novelty preschool as a behavioral trap: A proposal for research (pp. 3-
27). Minnesota Symposium on Learning. Minneapolis, MN: E. Allen & E. M. Goetz (Eds.), Early childhood education:
University of Minnesota Press. Special problems, special solutions (pp. 31-76). Rockville,
Baker, J. E., & Winston, A. S. (1985). Children's creative MD: Aspen Systems.
drawing: Experimental analysis and social validation of a self- Goetz, E. M. (1982b). A review of functional analyses of pre-
instructional procedure. Education and Treatment of Chil- school children's creative behaviors. Education and Treat-
dren, 8, 115-132. ment of Children, 5, 157-177.
Ballard, K. D., & Glynn, T. (1975). Behavioral self-manage- Goetz, E. M., & Baer, D. M. (1971). Social reinforcement of
ment in story-writing with elementary school children. Jour- creative blockbuilding by young children. In G. A. Ramp &
nal of Applied Behavior Analysis, 8, 387-398. B. L. Hopkins (Eds.), A new directionfor education: Behav-
Berliner, D. C. (1985). Effective classroom teaching: The neces- ior analysis (pp. 72-79). Lawrence, KS, University of Kan-
sary but not sufficient condition for developing exemplary sas Press.
schools. In G. R. Austin & H. Garber (Eds.), Research on Goetz, E. M., & Baer, D. M. (1973). Social control of form
exemplary schools (pp. 127-151). Orlando, FL: Academic diversity and the emergence of new forms in children's block-
Press. building. Journal of Applied Behavior Analysis, 6, 209-217.
Bijou, S. W. (1976). Child development: The basic stage of Glover, J. A., & Gary, A. L. (1976). Procedures to increase
early childhood. Englewood Cliffs, NJ: Prentice-Hall. some aspects of creativity. Journal ofApplied Behavior Anal-
Bloom, B. S. (Ed.). (1985). Developing talent in young people. ysis, 9, 79-84.
New York: Ballantine Books. Goetz, E. M., & Salmonson, M. N. (1972). The effect of general
Bushell, D., Jr., & Dorsey, D. (1985). Behavioral models of and descriptive reinforcement on "creativity" in easel paint-
teaching. In T. Husen & T. N. Postlewaite (Eds.), The inter- ing. In G. B. Semb (Ed.), Behavior analysis in education (pp.
national encyclopedia of education (pp. 437-442). New 53-61). Lawrence, KS: University of Kansas Press.
York: Pergamon. Hendrick, J. (1975). The whole child: New trends in early educa-
Day, W. F. (1969). On certain similarities between the philo- tion. St. Louis: C. V. Mosby.
sophical investigations of Ludwig Wittgenstein and the opera- Hendrick, J. (1980). Total learning for the whole child. St.
tionism of B. F. Skinner. Journal of the Experimental Analy- Louis: C. V. Mosby.
sis of Behavior, 12, 489-506. Holman, J., Goetz, E. M., & Baer, D. M. (1977). The training
Della-Piana, G. M. (1978). Research strategies for the study of of creativity as an operant and an examination of its gener-
revision processes in writing poetry. In C. R. Cooper & L. alization characteristics. In B. C. Etzel, J. M. LeBlanc, & D.
Odell (Eds.), Research on composing: Points of departure M. Baer (Eds.), New directions in behavioral research: Theo-
(pp. 105-134). Urbana, IL: National Council of Teachers of ry, methods and applications. In honor of Sidney W. Bijou
English. (pp. 441-472). Hillsdale, NJ: Erlbaum.
Elliott, c., & Goetz, E. M. (1971). Creative blockbuilding as a Horowitz, F. D., & O'Brien, M. (1986). Gifted and talented
function of social reinforcement using an unrestrictive children: State of knowledge and directions for research.
creativity code. Unpublished manuscript, University of American Psychologist, 41, 1147-1152.
Kansas. Juan, S. (1985). The Yoki Gakuen: The Suzuki philosophy in the
Fallon, M. P., & Goetz, E. M. (1975). The creative teacher: The preschool. Childhood Education, 62, 38-39.
effects of descriptive social reinforcement upon the drawing Katz, L. G. (1985). Dispositions in early childhood education.
behavior of three preschool children. School Applications of ERICIEECE Bulletin, 18(2), 1,3.
Learning Theory, 7(2),27-45. Kohler, F. W., & Greenwood, C. R. (1986). Toward a tech-
Figgs, S., & Herbert, E. (1971). The effects of training, primes nology of generalization: The identification of natural con-
and social reinforcement on creativity in blockbuilding. tingencies of reinforcement. Behavior Analysis, 9, 19-26.
Unpublished manuscript, University of Kansas. Lane, T. W., Lane, M. Z., Friedman, B. S., Goetz, E. M., &
Figgs, S., Dunn, J., & Herbert, E. (1971). The effects of primes Pinkston, E. M. (1982). A creativity enhancement program
on creativity in blockbuilding. Unpublished manuscript, Uni- for preschool children in an inner city parent-child center. In
versity of Kansas. A. M. Pinkston, J. L. Levitt, G. R. Green, N. L. Linsk, & T.
Flavell, J. H. (1986). The development of children's knowledge L. Rzepnicki (Eds.), Effective social work practice: Advanced
about the appearance-reality distinction. American Psychol- techniquesfor behavioral intervention with individuals,fami-
ogist, 41, 418-425. lies, and institutional staff(pp. 435-441). San Francisco, CA:
Fowler, S. A., & Baer, D. M. (1981). Do I have to be good Jossey-Bass.
today? The timing of delayed reinforcement as a factor in Lavatelli, C. S. (1970). Piaget's theory applied to an early
generalization. Journal ofApplied Behavior Analysis, 14, 13- childhood curriculum. Boston: Center for Media Develop-
24. ment.
Funderbunk, F. R. (1976). Reinforcement control of classroom Maloney, K. B., & Hopkins, B. L. (1973). The modification of
creativity. In T. Brigham (Ed.), Behavior analysis in educa- sentence structure and its relationship to subjective judge-
tion: Self-control and reading (pp. 197-204). Dubuque, IA: ments of creativity in writings. Journal of Applied Behavioral
Kendall/Hunt. Analysis, 6, 425-433.
Ghiselin, B. (Ed.). (1952). The creative process. New York: Maloney, K. B., Jacobson, C. R., & Hopkins, B. L. (1973). An
Mentor Books. analysis of the effects oflectures, requests, teacher praise, and
Glover, J. A., & Sautter, F. (1977). Procedures for increasing free time on the creative writing behaviors of third grade chil-
four behaviorally defined components of creativity within for- dren. In E. Ramp & G. Semb (Eds.), Behavior analysis: Areas
mal written assignments among high school students. School of research and application (pp. 244-260). Englewood
Applications of Learning Theory, 9, 3-22. Cliffs, NJ: Prentice-Hall.
Goetz, E. M. (1981). The effects of minimal praise on the cre- Maltzman, I., Bogartz, W., & Breger, L. (1958). A procedure
ative blockbuilding of three-year-olds. Child Study Journal, for increasing word association originality and its transfer
11,55-67. effects. Journal of Experimental Psychology, 56, 392-
Goetz, E. M. (1982a). Behavior principles and techniques. In K. 398.
Montessori, M. (1964). The Montessori method. New York: Torrance, E. P. (1965). Rewarding creative behavior: Experi-
Schocken Books. ments in classroom creativity. Englewood Cliffs, NJ:
Osborn, A. F. (1953). Applied imagination. New York: Prentice-Hall.
Scribner's. Torrance, E. P. (1980a). Lessons about giftedness and creativity
Parsonson, B. S., & Baer, D. M. (1978). Training generalized from a nation of 115 million overachievers. Gifted Child
improvisation of tools by preschool children. Journal of Ap- Quarterly, 24, 10-14.
plied Behavior Analysis, 11, 363-380. Torrance, E. P. (1980b). Creativity and futurism in education:
Reese, N. M., & LeBlanc, J. M. (1970). Creative dance: Pilot Retooling. Education, JOO, 298-311.
study. Unpublished research, University of Kansas. Torrance, E. P. (1982). Education for quality circles in Japanese
Rogers-Warren, A., & Baer, D. M. (1976). Correspondence schools. Journal of Research and Development in Education,
between saying and doing: Teaching children to share to 15 (2), l1-l5.
praise. Journal of Applied Behavior Analysis, 9, 335-354. Wallach, M. A. (1985). Creativity testing and giftedness. In F.
Romero, P. M., Holt, W. J., Stella, M. E., Baer, D. M., & D. Horowitz & M. O'Brien (Eds.), The gifted and talented:
Etzel, B. C. (1978, August). Contingency effects of reinforced Developmental perspectives (pp. 99-123). Washington, DC:
and unreinforcedforms: Cube design complexity. Paper pre- American Psychological Association.
sented at the American Psychological Association, Toronto, Wallach, M. A., & Kogan, N. (1965). Modes of thinking in
Canada. young children. New York: Holt, Rinehart & Winston.
Ryan, B. A., & Winston, A. S. (1978). Dimensions of creativity Winston, A. S. (1984, August). Beyond reinforcement of novel-
in children's drawings: A social validation study. Journal of ty. In A. S. Winston & E. M. Goetz (Chairs), Current issues
Educational Psychology, 70,651-656. in the study of creativity as operant behavior. Symposium
Skinner, B. F. (1957). Verbal behavior. New York: Appleton- conducted at the meeting of the American Psychological As-
Century-Crofts. sociation, Toronto, Canada.
Sloane, H. N., Endo, G. T., & Della-Piana, G. (1980). Creative Winston, A. S., & Baker, J. E. (1985). Behavior analytic studies
behavior. Behavior Analyst, 3, 11-22. of creativity: A critical review. Behavior Analyst, 8, 191-
Stokes, T. F., & Baer, D. M. (1977). An implicit technology of 205.
generalization. Journal of Applied Behavior Analysis, JO, Wolf, M. M. (1978). Social validity: The case for subjective
349-368. measurement or how applied behavior analysis is finding
Torrance, E. P. (1962). Guiding creative talent. Englewood heart. Journal of Applied Behavior Analysis, 11, 203-214.
Cliffs, NJ: Prentice-Hall.
CHAPTER 24
Mental Management and Creativity

A COGNITIVE MODEL OF TIME MANAGEMENT FOR
INTELLECTUAL PRODUCTIVITY
Bruce K. Britton and Shawn M. Glynn
Introduction Why Intellectually Productive People Need to

Manage Their Time
Defining Creativity
A wise Man should order his Designs, and set all his interests
"Creative" is a tenn of praise much affected .... It is presum- in their proper Places: This Order is often confounded by a
ably intended to mean original, or something like that, but is foolish Greediness, which, while it puts us upon pursuing so
preferred because it is more vague and less usual (cf. "semi- many several Things at once, that in Eagerness for Matters of less
nal"). It has been aptly called a "luscious, round, meaningless consideration, we grasp at Trifles, and let go Things of greater
word, " and said to be "so much in honor that it is the clinching value. (La Rochefoucauld, 1706, p. 43)
tenn of approval from the schoolroom to the advertiser's stu-
dio." (Fowler, 1965, p. 114) Intellectually productive people usually have
more things that they would like to do, or need to
Creativity is a debased term. Intellectual produc- do, than they have time (Roe, 1952); that is, they do
tivity is the best term for what we have in mind in not have enough time to do the things they have to
this chapter. Intellectually productive people work do. One part of the solution to this problem is men-
to produce useful new ideas. Some examples of tal management. To produce large quantities of
intellectually productive people are scientists doing high-quality mental products, people must effec-
original research, authors writing imaginative liter- tively manage the limited resource that produces
ature, political workers devising new solutions to those products: their minds. Managing the mind has
social problems, and mediators proposing ways to several aspects. One aspect, which is considered
resolve conflicts between parties. elsewhere, is memory management (Britton, Smith,
& Glynn, 1985). The aspect that this chapter con-
siders is the management of mental time.
The mind can be regarded as a single processor
that can do only one thing at a time. The objective of
Bruce K. Britton • Department of Psychology, University
of Georgia, Athens, GA 30602. Shawn M. Glynn • the mental time-management system described in
Department of Educational Psychology, University of Georgia, this chapter is to maximize the use of the resource of
Athens, GA 30602. mental time. The use of mental time is maximized
429
when the largest number of the highest quality intel- External forces include deadlines, the goals desired
lectual products are produced in the shortest time. by other people, and the order in which tasks arrive
To maximize mental time, it is necessary to make at their desk; internal forces include inertia and mo-
use of a metacognitive system that oversees and mentum. The output of persons who are not in con-
supervises the operation of the rest of the cognitive trol of their processing resources is likely to be
system. Some people have a mental time-manage- determined fortuitously by the interplay of external
ment system, and others do not. It is the faith of this and internal forces. In contrast, the output of people
chapter that those who have such a system are better who manage their mental time is characterized by
off, in the sense that their intellectual productivity is the formulation of ordered precedence relations that
greater. This is called a "faith" because it is based are relevant to the achievement of specific ex-
on very little empirical research. The empirical lit- plicitly formulated goals, and by discrete steps that
erature on the effects of time management is very are taken toward the achievement of those goals.
small: the only experimental empirical paper we In the next sections of this chapter, we describe a
have found is by Hall and Hursch (1982). Literature time-management system that is based on a syn-
on the relationship between creativity and time thesis of the computer science literature with the
management does not appear to exist (Paul Tor- popular literature. We then describe the specific
rance, personal communication, October 5, 1987). time-management problems that intellectually pro-
Although there is a large popular literature in this ductive people are likely to have. In the concluding
area, it is almost entirely exhortatory in tone, with sections, we point out that time management is it-
only anecdotes as evidence. The best example of self an intellectually demanding task for which suf-
this literature is Lakein (1973), and though it is very ficient time must be set aside. After providing a
good indeed, the claims it makes are not supported research agenda, we summarize briefly. An Appen-
by the kind of evidence usually adduced in behav- dix provides a time-management instrument.
ioral science investigations.
There is, however, a very substantial empirical
literature in computer science that demonstrates
The Time Manager
that the productivity of computers can be substan-
tially increased by applying some of the principles
Top-Level Components
of time management described here. Most present-
day computers, like human minds under most cir- Mental time management is performed by a met-
cumstances, have a single processor that can do acognitive system that has three parts, as shown
only one thing at a time, and computer scientists schematically in Figure 1: the Goal Manager, the
have been much concerned with maximizing the Task Planner, and the Scheduler. The Goal Manag-
use of that resource. This literature was thus the er takes as input the person's desires and produces
original source of many of the ideas that are pre- as output a list of goals and subgoals with priorities
sented in this chapter (Greenstein & Rouse, 1982; attached.
Pattipati, Kleinman, & Ephrath, 1983; Tulga, The Task Planner takes as input the output of the
1979; Tulga & Sheridan, 1980). Goal Manager-the prioritized list of goals and
The alternative to having a mental time-manage- subgoals. The planner operates upon those goals
ment system is not having one. People who do not and subgoals and produces as output a list of tasks
have such a system are not supervising their own and subtasks with priorities attached. Goals are dis-
cognitive system; instead, they are driven by exter- tinguished from tasks in that goals are objectives,
nal and internal forces that they do not control. whereas tasks are activities. Objectives are state-
Prioritized Prioritized
Desires_ _ Goals and _ _ Tasks and _ Scheduler
Subgoals Subtasks
Figure 1. Top-level components of a mental time management system.

CHAPTER 24 • TIME MANAGEMENT 431
ments of a desired end product or a desired state of his motions accordingly. Our judgments go astray if they have no
affairs; objectives are not something that can be direction and no aim. No wind is fair for the sailor who has no
purposed port. (Montaigne, 1925, pp. 48-49)
"done" with the limbs of the body or the parts of
the mind. In contrast, tasks are activities that can be
done with the limbs of the body or the parts of the . By far the most important thing about any system
mind. Tasks are instrumental to goals in that they IS the goals to which it is directed. In the context of
are activities that can bring about a desired state of time management, this means that the quality of a
person's goals is much more important than the
affairs. For example, a goal could be to achieve the
ability to be fluent in French; a task relevant to that efficiency with which they are pursued. Even an
goal would be to enroll in a French class. optimally efficient time-management system, if it
SUbgoals and subtasks are subordinate steps that operates to achieve low-quality goals, cannot lead
must be taken to achieve their superordinate goal or to high-quality results. And even a very poor time-
task. For example, subgoals of becoming fluent in management system can produce important work if
French are (a) learning French vocabulary and (b) it is directed toward important goals. The goals of
learning French syntax. Subtasks for the task of one's cognitive system are an order of magnitude
~ore important than one's time-management prac-
enrolling in a French class might be (a) looking in
the telephone book for educational institutions that tIces. The most efficient possible way of using
might offer French classes, (b) recording their num- ?ne's time on a valueless goal is utterly valueless; it
IS a great deal worse than the least efficient possible
bers, and (c) dialing the numbers. It is important to
note in these examples that these subgoals and sub- way of using one's time on a valuable goal. So the
tasks are arranged in precedence relations; that is, goal-management system is more important than
certain subgoals must be achieved before others can the time-management system; or to put it another
~ay, the time-management system's only purpose
be achieved. Learning some French vocabulary, for
IS to s~rve the goals. The importance of high-quality
example, is a prerequisite for learning French syn-
tax, and certain subtasks, such as looking up the goals IS the thing we will be left pointing at after we
phone number, must be completed before later have said everything we have to say about time
ones, such as dialing the number, can be begun. management.
The third component of the time manager is the People often resist specifying their goals. Fre-
Scheduler, which takes as input the output of the quently stated objections are that they find the task
Task Planner-the list of tasks and subtasks. The overwhelming or embarrassing, that they feel the
Scheduler produces a "To-Do List," which is used task limits their options, or that it takes away some
to make the decision about what to do next. This of the creativity, spontaneity, and fun from life
decision is then input to the mind, which then car- (Lakein, 1973). Against these objections must be
ries out the activity. When each activity is com- placed the fact that goals that have not been spec-
pleted, the Scheduler is consulted about what to do ified are unlikely to be achieved. The number of
next. paths a life can take is infinite; so the number of end
points that can be reached is infinite in number. But
Description of Each Time Manager Component the value of those end points differs greatly. So the
expected values of the paths differ greatly as well.
Goal Manager The Goal Manager is composed of several parts,
He who has not directed his life in general to a certain end, for as shown schematically in Figure 2.
him it is impossible to adjust the separate acts; for him it is
impossible to arrange the pieces, who has not a figure of the
whole in his head . . . the archer must first know at what he Goal Generator. When persons want to find
aims, and then adapt his hand, the bow, the string, the arrow, and out what their goals should be, the first thing they
1
~oal Generator
. '
1_ Goal Prioritizer _ Subgoal _
Generator
Subgoal
Prioritizer
Figure 2. Components of the Goal Manager.

should do is to produce a long list of possible goals, on feasibility and other constraints. The final output
which can then be narrowed down for the next is a prioritized list of subgoals.
stage. The input to the Goal Generator is the per-
sons' desires, and everything they know about
themselves and the world. The output is a list of Task Planner
candidate goals, which might include, for example, The Task Planner has several components, as
professional goals, such as becoming a prominent shown in Figure 3. When beginning work, the plan-
psychologist; personal goals, such as learning ner has a list of prioritized goals and of the subgoals
French; and social, political, intellectual, spiritual, for each goal. For simplicity, we can consider the
and other types of goals. list as being ordered from top to bottom, with the
first goal on the list being the highest priority goal.
Goal Prioritizer. It is obvious that all goals are Underneath each goal is a list of sub goals that must
not of equal importance. Also, because usually the be achieved before the main goal can be achieved.
time available is not sufficient to achieve them all, These subgoals are also in priority order. The plan-
they will have to be placed in order of importance. ner starts with the highest priority goal and looks for
The Goal Prioritizer orders the goals. The input to its highest priority subgoal. The planner's job is to
the Goal Prioritizer is the list of goals with priorities specify what tasks must be done to achieve that
attached, which can be visualized as a list of goals subgoal. It will be useful to recall here that, al-
with the most important ones at the top. The pri- though goals and subgoals are statements of desired
oritizing is done by heuristics. These heuristics are end products, tasks and subtasks are activities that
very important determiners of the quality of the can be performed with the limbs of the body or the
output of all the components described below. A parts of the mind, often with the help of physical
description of these heuristics for the general case is objects or other external resources.
beyond the scope of this chapter, but some heuris-
tics for intellectually productive people are in- Task Generator. The Task Generator consid-
cluded in later sections. ers the subgoal and thinks of different possible ac-
tivities that could be performed to accomplish it. In
Subgoal Generator. For each goal, there are the time available for this phase of planning, the
usually several prerequisite subordinate goals that generator then thinks of as many of these activities
have to be achieved in order to achieve the superor- as is possible and ends up with a list of possible
dinate goal, and often there are several possible activities that could be done to achieve the goal. In
routes to the same goal. For example, one might general, these activities are too molar to be per-
become fluent in French by taking French courses formed with specific parts of the body or mind; the
locally or by going to France. The Subgoal Gener- Subtask Generator has the responsibility of break-
ator takes as input each goal separately, and, using ing them down into molecular subcomponents.
heuristics and knowledge about the world, gener-
ates various possible sets of subgoals that could Task Prioritizer. On the generated list of pos-
serve as prerequisites to achieve that goal. The out- sible activities, there will be some that are clearly
put is the various sets of possible subgoals. infeasible or that are very costly in terms of suffer-
ing; some that use a scarce resource; and some that
Subgoal Prioritizer. Taking the list of possible are efficient, easy, or pleasant to do. The Task Pri-
subgoals, the Subgoal Prioritizer rates them based oritizer selects which ones are the first choices.
Figure 3. Components of the Task Planner.

Subtask Generator. Starting with the first- and amount of mental resources that are required to
choice task, the Subtask Generator breaks it down achieve a goal or task.
further into components that can actually be done These additional properties are mentioned here
with parts of the body or the mind. There are many because they are needed as part of the input to the
different ways to accomplish the task, and, at this next component of the time-management system,
stage, as many possibilities will be generated as which is the Scheduler. At this point, it will be easy
time is available. to think back to see how they could have been incor-
porated into the processing from the very earliest
Subtask Prioritizer. The task possibilities are stages. For example, the deadline property could
then narrowed down by the Subtask Prioritizer to have been attached to certain goals as soon as the
produce a list of practical, performable activities. Goal Generator produced them, or as soon as some
outside force imposed them on the system. The
deadline property would then have influenced the
Task Properties work of the Goal Prioritizer and would have been
inherited by the goals that are produced by the Sub-
Up to this point, we have emphasized one proper- goal Generator and that are chosen by the Sub goal
ty of the goals and tasks: the property we have Prioritizer. In tum, the deadline property would be
named priority. We need now to consider three inherited through the task and subtask planning pro-
other important properties of the tasks. One impor- cess. In similar ways, the other properties could
tant property is whether the task has a deadline or have been incorporated into the processing from the
not. Whether a particular task has a deadline often earliest stages.
depends on whether its superordinate subgoal and
goal have a deadline, that is, the deadline property
Scheduler
may be "inherited" by any progeny of a goal. For
example, one common deadline maintenance goal The Scheduler is in charge of choosing a person's
is to pay one's income taxes by April 15. The dead- next activity. Typically, there are a large number of
line property is inherited by its subgoals, like pay- different things that can be done, but because the
ing both Federal income tax and State income person can usually do only one thing at a time, the
tax,and by their tasks, like gathering one's financial Scheduler has to decide which it is to be. In making
records and procuring the appropriate extra forms the decision, the Scheduler has access to all the
from the government, and also by their subtasks, information that is retained from the previous
like getting into the car and driving down to the stages, including priorities, deadlines, time re-
local Internal Revenue Service office to get the quired, interruptibility, and mental and other re-
forms. sources. The Scheduler has two components, as
Another important property of tasks is the time shown schematically in Figure 4.
required to complete them. Very long tasks need to
be treated separately from very short ones. Some Job Selector. The Job Selector produces a To-
tasks should not be begun unless there is sufficient Do List for each day. This is a run down of the
time to complete them on a particular day: waxing different things the person has to do that day. Each
the kitchen floor is a maintenance goal of this type. of the things on the list will be called a job. (This
Others can be interrupted and resumed later without new term is used to avoid having to distinguish
penalty. This property is called interruptibility. between subtasks and tasks.) The input to the Job
Another important property is the type of mental Selector is the list of tasks and subtasks. Each task
resources that a task requires. For some tasks it is and subtask has attached to it its entire list of proper-
necessary to have a very alert, awake, and active
mind-high-Ievel intellectually productive work
often requires this property. Such tasks can only be
done at times of the day when the needed resource is
available, for example, in early morning. Other
subtasks, such as dealing with one's mail, can often IJob Selector I --+ Run Manager
be done with less mental resources, in late after-
noon, for example. This property specifies the type Figure 4. Components of the Scheduler.
ties, including priority, deadlines, time required, intellectual productivity that we know of occurs in
interruptibility, and resource requirements. The Job response to job goals. This is particularly likely if
Selector applies to these job properties a heuristic the job has intellectual productivity as one of its
program to select which jobs will be placed on the explicit or implicit requirements. However, some
list for that day. jobs have no scope for intellectual productivity or
positively discourage it; persons in such jobs are
Run Manager. The To-Do List is then per- unlikely to be intellectually productive.
formed by the Run Manager, who decides which
job to do first. If that job is not interrupted, it will be
completed, and then the Run Manager must decide Content Goals versus System Goals
which job to do next. If there is an interruption, the Content goals state the content of desired states
Run Manager must decide what to do next. of affairs, such as knowing French, becoming a
prominent psychologist, or owning a sailboat. In
contrast, system goals state the operating criteria by
Special Properties of Time Management which the performance of the cognitive system is to
for the Intellectually Productive be judged. For example, one common system goal
of some people's cognitive system is to maximize
Distinctions between Types of Goals the accuracy of the system. Each system goal has
associated with it a particular measure of effective-
There are many different types of goals. Intellec- ness; for maximizing the accuracy of the system,
tual productivity is likely to be associated more with the measure of effectiveness would be error rate.
some types of goals than with others. Another system goal might be to work as fast as
possible, for which one measure of effectiveness
Intellectual Productivity versus Noncreative would be the number of jobs completed per unit
Achievement and Maintenance Goals time. Each system goal might be appropriate for a
particular profession or job mix. Also, it is possible
If people are to achieve the goal of being intellec- to combine such goals; for example, one common
tually productive (creative), then they must set goal might be to combine speed with accuracy, that
themselves appropriate goals and allow consider- is, to be as fast and accurate as one can.
able periods of time to work toward achieving One reason that system goals are mentioned here
them. But most people also have to devote consid- is to note that the Goal Manager is not concerned
erable periods of time to tasks that do not lead to with them. The Goal Manager is concerned with
intellectual productivity. It is very easy to spend all content goals. However, system goals can be
one's time attending to noncreative or essential important to the scheduling system, because the
maintenance activities, routine tasks, previous order in which jobs are done can be partly deter-
commitments, interruptions, and crises (Lakein, mined by system goals.
1973). In fact, most people devote all their time to The other reason system goals are distinguished
tasks that are not creative. Although these tasks from content goals is to point out that intellectually
certainly lead to the maintenance of one's body, productive people must be concerned primarily
home, possessions, and relationships, and may also with content goals and, moreover, with content
lead to considerable job achievement; they are un- goals that lead to intellectually productive results.
likely to lead to intellectual productivity, except People who do not intend to be intellectually pro-
fortuitously. ductive may, however, be concerned primarily with
system goals, such as doing a particular job more
and more quickly, efficiently, and accurately. But
Personal Goals versus Other Goals and Job Goals
people concerned primarily with system goals will
Some goals originate within the self, and other have less time to devote to content goals, and, for
goals are imposed from outside, either by real- that reason, are less likely to be intellectually pro-
world events, such as natural disasters and wars, or ductive. By the same token, intellectually produc-
by other people who have some influence over us, tive people, because they must concentrate pri-
such as children, spouses, bosses, and political or marily on content goals, will generally have less
religious leaders. Conceivably, intellectual produc- time to focus on system goals. The result is that
tivity can result from any of these goals, but most their intellectually productive work is sometimes
performed slowly and inefficiently, or otherwise tions, so it is often not possible to calculate any of
does not meet certain desirable system goals. these properties. Often, the assignment of high pri-
ority to creative goals represents a leap of faith.
Intellectual Productivity Effects on Components
of the Goal Manager Subgoal Generator and Prioritizer
The role of the subgoal component is to break
Goal Generator
down the highest priority goals into subparts to the
People who wish to be intellectually productive point where they can be restated as tasks, that is, as
need to generate goals that are appropriate to that activities that can be performed by the limbs or the
objective. But because creative ideas are by defini- mind. The breaking down and prioritizing pro-
tion new ones, it is usually impossible to specify in cesses are essentially the same whether the goals are
advance what the end point idea to be produced will creative goals, noncreative achievement goals, or
be, for it has not yet been invented. So intellectually maintenance goals. Of course, some decomposi-
productive people often need to specify problems in tions into subgoals are better than others, in the
place of goals. In such problem-goal statements, sense that they will lead to the goal more effectively
the goal idea appears as an unknown entity that is or more cheaply. The job of figuring out the best
the solution to a particular problem. The generation decomposition is a problem-solving task that uses
of such problem-goal statements is often unlikely essentially the same heuristics whatever the type of
to occur because the existence of even very imper- goal.
fect solutions to a problem can cause reasonable
people to consider that the problem is already
solved. Then the problem becomes part of the back- Task Planner
ground, and it is difficult even to see it as a problem. Nothing more will be said about the Task Planner
Intellectually productive people must incorporate here, except that the planner's operation is the same
heuristics in their goal generators that deal with for creative goals as for any other type of goal.
this.
Scheduling for Intellectual Productivity
Goal Prioritizer
Job Selector. On each day in which the To-Do
If creative products are to result, priority must be List includes jobs that are priorities for creative
given to creative goals. Maintenance goals and non- work, there is a likelihood that some creative work
creative achievement goals must be given lower will be accomplished. On other days, creative prod-
priority. This particular conflict between goal types ucts can be achieved only fortuitously. The Job
is an instance of goal conflict. When people have Selector, who is in charge of the To-Do List for
more things to do than they can possibly do in the each day, must select creative jobs if any progress is
time that is available, goal conflict is inevitable. In to be made toward creative goals.
general, goal conflict can only be resolved by set-
ting priorities. Run Manager. The Run Manager is in charge
When creative goals are involved, unique prob- of when the selected jobs are to be done during the
lems arise in setting priorities and in dealing with day. Intellectual productivity typically requires (a)
the other goal and task properties. These can be the absorption of large amounts of information, (b)
seen most clearly by considering how priorities are its integration by thinking, and (c) the production of
set for maintenance and noncreative achievement the desired end product. All these activities are per-
goals. The criteria that are used there include formed most effectively by the processor when two
threshold considerations of feasibility, practicality, conditions are met: the optimum time of day, and
and probability of successful achievement, and also large enough blocks of time that are not interrupted
incremental calculations of expected value, re- by outside forces or by lack of resources needed for
source requirements, and time to achieve. Because the job.
maintenance and noncreative achievement goals
can be formulated as end-product content state- Time of Day. Most people have a certain time
ments, it is usually possible to estimate these prop- of day when they can do intellectually productive
erties. In contrast, creative goals are unknown solu- work most effectively. For most intellectually pro-
ductive people with whom we are acquainted, the from the progress of the task at hand. Often intellec-
morning is the best time. Whatever the time, pro- tually productive people will reach a particular
ductive work seems usually to be limited in length point and then find that they need something that is
to no more than a few hours. This time will be called not immediately available: a document, a word, an
high time. During high time, high-quality intellec- idea, or some other thing. Usually these needs
tual work can be performed, whereas at other times spawn a new subtask. Sometimes these subtasks are
it either cannot be performed at all or only with the not in themselves intellectually demanding and so
greatest difficulty and waste of time. People who can be put off to lower quality time periods. At
can schedule their intellectually demanding tasks other times, they are fully as intellectually demand-
during high time will achieve more of the tasks than ing as the main task in progress and so must be
people who cannot. pursued during high time, but they can be put off to
later blocks of high time.
Large Blocks o/Time. Typically, intellectually
productive tasks require the absorption of amounts Scheduling Creative Subtasks Not Requiring
of information that are very large, the following of High Time. Often creative production requires
chains of thoughts that are very long, and end prod- large amounts of tedious work. For example, re-
ucts whose production is very time consuming. The search psychologists may have to collect, score,
amounts of time required for these activities are and analyze data from subjects. If such work can be
very large. In most cases, the completion of anyone delegated to someone who can be trusted to do it
unit of an intellectually productive task will require right, then time can be saved. But all too often,
at least all of one day's high time. People who can intellectually productive people have to do tedious
devote a large amount of their high time to intellec- tasks themselves because the tasks have never been
tually productive tasks will complete more of them. done before and only they have the understanding to
do them right, or because the progress of the tasks is
Uninterrupted Blocks o/Time. When intellec- inherently uncertain, and novel decisions must be
tual activity is interrupted, the task that is in pro- made in real time. If possible, such tasks should be
gress must be stored away, along with the results scheduled at other than high time.
that are reached thus far, and all the information that
is required to reactivate the task later on when the
interruption has been dealt with. When the task is Scheduling Management Time
taken up again, the material needed for its resump- The most important goal of our cognitive system
tion must be reactivated. If the task is not dealt with
is to pursue goals that are important to us. The
in this way, it will have to be started from the begin-
purpose of the entire time-management system de-
ning, or not taken up at all.
scribed in this chapter is to achieve that goal. Be-
Interruptions are particularly disruptive to intel- cause management of our mental time is itself an
lectually productive work because such work often
intellectually demanding task, high time should be
involves lengthy, novel chains of thought that are allocated to it. Allocation of a brief period of plan-
fragile in the sense that their disruption, even tem- ning time at the beginning of each high-time period
porarily, can cause a complete loss of critical ele- will ensure that the time-management system will
ments or ordered links between them. Similarly, in
have an opportunity to work.
intellectually productive work, very large bodies of
diverse prior knowledge (driving the production of
new ideas) often are configured in novel and there- Research Agenda
fore very unstable ways, and, when interruptions
disrupt these configurations, they may not be re- Research on time management and on intellec-
coverable. Thus, interruptions may have a substan- tual productivity requires measurement and training
tial cost if they are permitted to occur during times of time-management practices, skills, and abilities
when intellectually productive activities are in on the one hand, and measurement of intellectual
progress. productivity on the other. Britton and Tesser have
Two types of interruptions can be distinguished: developed a time-management questionnaire for
external interruptions, such as telephone calls or student populations, which is included in the Ap-
unexpected visits by people, are generally not rele- pendix. A factor analysis (N = 90) yielded three
vant to the task at hand and are sometimes uncon- factors, as shown in Table l. A project is now un-
trollable; whereas job-internal interruptions arise derway to relate these scores, taken on freshmen, to
Table 1. Items Loading on Three Factors of intellectual productivity in connection with these
Time Management Questionnaire instruments.
The exemplary work on training time-manage-
Planning factor ment skills and measuring intellectual productivity
Do you make a list of things you have to do each day? can be found in Hall and Hursch (1982). Numerous
(.80)a workshops on training time-management skills are
Do you plan your day before you start it? (.72) offered at universities (e.g., through the Counsel-
Do you make a schedule of activities you have to do on ing and Testing Center at the University of Georgia)
work days? (.71) and by private organizations, but we know of none
Do you write a set of goals for yourself each day? (.67) that has attempted to measure the effectiveness of
Do you spend time each day planning? (.66) the programs. Such measurements would be easy to
Do you have a clear idea of what you want to accomplish take with the instruments in the Appendix and
during the next week? (.54) would provide important information not otherwise
Do you set and honor priorities? (.50) available.
A final research tool in this area, which may be
useful for both measurement and training, is a com-
Grind factor puter gamelike program that was devised by Tulga
(1979; Tulga & Sheridan, 1980), in which a com-
Do you rarely do things which interfere with your puter screen display shows several boxes moving
schoolwork? (.63) right toward a "deadline." Each box represents a
Do you feel you are in charge of your own time, by and task. The length ofthe box represents the amount of
large? (.60) work remaining to do on it; the height of the box
On an average day do you spend more time with school- represents the number of points gained per unit of
work than with personal grooming? (.57) work; and the symbols in the box represent various
Do you believe there is little room for improvement in other properties of the task, such as interruptability.
the way you manage your time? (.55) The subject has control of a cursor that can be
Do you make constructive use of your time? (.55) moved to a box to "work" on it. The subject gets
Do you stop unprofitable routines or activities? (.51) points depending on which task he chooses to work
on. For each configuration of boxes (tasks), it is
possible to calculate (by enumeration) which
Orderly factor task(s) to work on to maximize some criterion. Cri-
Do you usually keep your desk clear of everything other teria might be maximizing the number of jobs com-
than what you are currently working with? (.66) pleted, or the number of points gained, or minimiz-
Do you have a set of goals for the entire quarter? (.49) ing the number of jobs that are uncompleted at
The night before a major assignment is due, have you deadline. The measure of performance is how
usually completed it? (.47) closely the subject comes to maximizing the chosen
When you have several things to do, do you think it is criterion.
best to do a little bit of work on each one? (.47) Such a task is a gamelike simulation of real time
Do you keep your important dates (e. g., exam dates, management tasks, and could be useful not only for
research paper due dates, etc.) on a single calendar? measuring ability but also for training time-man-
(.43) agement skills.
Do you regularly review your class notes, even when a
test is not imminent? (.42)
Do you skim reading materials first to see if they are
worth reading in depth? (.40) Summary
aFactor loadings in parentheses.
A general model for time management is de-

scribed in detail. Special properties of time man-
cumulative grade point average over the college agement that are particularly relevant to
years as an index of intellectual productivity. intellectually productive people are pointed out,
Instruments that are appropriate for business peo- and some research tools are described. An App~n
ple are contained in Greene (1969) and Bliss dix provides an instrument for measuring time-
(1976). We do not know of any attempts to measure management practices.
438 PART IV . APPLICATIONS
Appendix
Britton and Tesser Time Management Instrument
This questionnaire includes 35 items that might be descriptive of you. Please read each question and
then place a check mark in one of the parentheses next to the question, corresponding to the category that
best describes how the question applies to you. For example, if the first question always applies to you,
put a check mark in the parentheses beneath "always." Remember, we are interested in how you think
you actually are, not how you would like to be. Be sure to answer all 35 questions.
Always Frequently Sometimes Infrequently Never

I. Do you believe that there is
room for improvement in the
way you manage your time? ( ) ( ) ( ) ( ) ( )
2. Do you write a set of goals for
yourself for each day? ( ) ( ) ( ) ( ) ( )
3. On an average class day do you
spend more time with personal
grooming than doing
schoolwork? ( ) ( ) ( ) ( ) ( )
4. Do you have a set of goals for
the entire quarter? ( ) ( ) ( ) ( ) ( )
5. Do you set and honor priorities? ( ) ( ) ( ) ( ) ( )
6. Do you usually have a radio,
television, or stereo playing
while you study? ( ) ( ) ( ) ( ) ( )
7. Do you usually keep your desk
clear of everything other than
what you are currently working
with? ( ) ( ) ( ) ( ) ( )
8. Are you able to put schoolwork
out of your mind when you are
socializing? ( ) ( ) ( ) ( ) ( )
9. The night before a major
assignment is due, are you
usually still working on it? ( ) ( ) ( ) ( ) ( )
10. Are you able to make minor
decisions quickly? ( ) ( ) ( ) ( ) ( )
II. Do you clip or xerox articles
which, although not presently
important to you, may be in the
future? ( ) ( ) ( ) ( ) ( )
12. Do you smoke an average of at
least one pack of cigarettes per
day? ( ) ( ) ( ) ( ) ( )
13. Do you regularly review your
class notes, even when a test is
not imminent? ( ) ( ) ( ) ( ) ( )
14. Do you keep your important
dates (e.g., exam dates, research
paper due dates, etc.) on a
single calendar? ( ) ( ) ( ) ( ) ( )
15. Do you often find yourself
doing things which interfere
with your schoolwork simply
because you hate to say "No" to
people? ( ) ( ) ( ) ( ) ( )
CHAPTER 24 . TIME MANAGEMENT 439
Always Frequently Sometimes Infrequently Never

16. Do you plan your day before
you start it? ( ) ( ) () ( ) ( )
17. Are you concerned about how
well or poorly you use your
time? ( ) ( ) ( ) ( ) ( )
18. Do you try to schedule your best
hours for your most demanding
work? ( ) ( ) ( ) ( ) ( )
19. Do you make a schedule of the
activities you have to do on
work days? ( ) () ( ) ( ) ( )
20. Do you find yourself waiting a
lot without anything to do? ( ) ( ) ( ) ( ) ()
21. Do you skim reading materials
first to see if they are worth
reading in depth? () ( ) ( ) ( ) ( )
22. Do you have a clear idea of
what you want to accomplish
during the next week? ( ) ( ) ( ) ( ) ( )
23. Do you make constructive use
of your time? ( ) ( ) ( ) ( ) ( )
24. Do you set deadlines for
yourself for completing work? ( ) ( ) ( ) ( ) ( )
25. Do you spend time each day
planning? ( ) ( ) ( ) ( ) ( )
26. Do you continue unprofitable
routines or activities? ( ) ( ) ( ) ( ) ( )
27. Do you keep things with you
that you can work on whenever
you get spare moments? ( ) ( ) ( ) ( ) ( )
28. Do you feel you are in charge of
your own time, by and large? ( ) ( ) ( ) ( ) ( )
29. Do you have a set of goals for
each week ready at the
beginning of the week? ( ) ( ) ( ) ( ) ( )
30. When you have several things to
do, how often do you try to
figure out which are most
important? ( ) ( ) ( ) ( ) ( )
31. When you have several things to
do, do you think it is best to do
a little bit of work on each one? ( ) ( ) ( ) ( ) ( )
32. Each week do you do things as
they naturally occur to you,
without an effort to make a plan
in advance and compulsively
follow it? ( ) ( ) ( ) ( ) ( )
33. Generally, do you think you can
usually accomplish all your
goals for a given week? ( ) ( ) ( ) ( ) ( )
34. Do you make a list of the things
you have to do each day? ( ) ( ) ( ) ( ) ( )
35. When you have an idea, do you
usually try to remember it
mentally until you need it,
rather than write it down
somewhere? ( ) ( ) ( ) ( ) ( )
References Lakein, A. (1973). How to get control of your time and your life.
New York: Signet.
Bliss, E, C, (1976). Getting things done: The ABCs of Time La Rochefoucauld, F., Duc de (1706). Moral maxims and reflec-
Management. New York: Scribner's. tions. London: Sare.
Britton, B. K., Smith, J. W., & Glynn, S. M. (1985). A cog- Montaigne, M. (1925). Essays (George B. Ives, Trans.).
nitive workbench model for understanding expository text. In Cambridge: Harvard University Press.
B. K. Britton & J. B. Black (Eds.), Understanding expository Pattipati, K. R., Kleinman, D. L., & Ephrath, A. R. (1983). A
text (pp. 277-248). Hillsdale, NJ: Erlbaum. dynamic decision model of human task selection perfor-
Fowler, H. W. (1965). A dictionary of modern English usage mance. IEEE Transactions on Systems, Man, and Cyber-
(2nd ed.). Oxford: Oxford University Press. netics, 13, 145-166.
Greene, R. M. (1969). The management game: How to win with Roe, A. (1952). The making of a scientist. New York: Dodd,
people. Homewood, IL: Dow Jones-Irwin. Mead.
Greenstein, J. S., & Rouse, W. B. (1982). A model of human Tulga, M. K. (1979). Dynamic decision making in multi-task
decision making in multiple process monitor situations. IEEE supervisory control: Comparison of an optimal algorithm to
Transactions on Systems, Man, and Cybernetics, 12, 182- human behavior. Unpublished doctoral dissertation, Mas-
193. sachusetts Institute of Technology, Cambridge, MA.
Hall, B. L., & Hursch, D. E. (1982). An evaluation of the effects Tulga, M. K., & Sheridan, T. B. (1980). Dynamic decisions and
of a time management training program on work efficiency. work load in multitask supervisory control. IEEE Transac-
Journal of Organizational Behavior Management, 3, 73-96. tions on Systems, Man, and Cybernetics, 10, 217-232.
Index
Abstract ability, 220 Analogies (cont.) Association, 5-6, 7, 212, 260; see also
Abstraction, reflective, 193, 196, 197 story problems, 392 Remote Associates Test
Adjective Check List, 55, 62, 255 survcy of textbook use of, 388-389 Associative listing, 315
Adolescents, 54, 81 transformations and, 371-372 Associative theories, 19-20,20-21,
Aesthetic sense, 223 Analogous problem-solving tasks 114-115, 120, 217, 257
Affective disorders: see Depression transfer of information to, 166, 168- Attention
Age, 26,100-101,199,221 169,170-171,173 defocused, 216-217, 217
Age-within-specialty, 221-222 Anomalies, 370, 371, 379; see also deployment of, 21, 259, 277
Ages of study subjects, 235 Contradiction and paradox focused, 154,218,273-274
Aiken, Conrad, 180 Appropriateness of product, 11-12,94, sustained, 223, 228
Alcohol abuse, 226 413 Attitude and interest inventories, 54-55,
Altered states of consciousness, 152- Aptitudes Research Project, 37 65,67,69
153 Architects, 98, 99 Automated Mathematician (AM), 138
Alternate Uses Test, 47, 48,54,260 MacKinnon's research, 3, 9, 10, 12, Automatizing of writing, 312-313, 319
reducing fluency as confounder, 48 20,26,57,63,78,88,94,98,99, Autonomy and independence, 137,
Amabile's interactive model, 26-28 104, 105, 108, 136, 137,254 237-238, 253
Ambiguity or disorder Architecture, 60, 78, 94
toleration of, 223, 256, 259, 405-406 Aristotle, 157,244,245,316
Analog, 383, 384 Artistic creativity, 47, 61-62, 207 Bach family, 103
Analogical Problem Solving (APS) Artists BACON I, 138
Model, 393-395 brain-hemisphere influences on, 100 Barrie, Sir James, 248
Analogies cognitive processes and traits of, 96, Barron-Welsh Art Scale, 59, 62, 98
in case study of solution, 351-359 257 Baudelaire, Charles, 222, 224, 246
creative aspect of, 212, 350, 383, IQs of eminent, 98 Beethoven, Ludwig, 247
386-387,392-396 judging of products of, 49, 60, 61 Berkeley, Bishop George, 157
Darwin's use of, 374 personality of, 253, 254 Binet scale, 121
defined, 383-384 preparation needed, 139 Biographical studies, 55, 56, 63, 64,
as different from examples, 385-386 psychiatric disturbances of, 246-247, 104,313
evaluating proposed, 376 249, 250 Birth order, 26, 253
explanatory, 383, 386, 386-387, re-representing problems by, 285 Bisociation, 203, 206, 272
387-389 sense of completion in, 404-405 Blake, William, 215, 225
and explanatory models, 346, 348, study of, 22, 23-24 Blind variation and selective retention,
359-361,379 women, 101-102 24, 25-26, 372-373, 378
importance of naming superordinate Assessment methods in creativity Boulton, Matthew, 251
concept, 385, 397 research Brain hemisphericity, 100, 112, 121,
invented rather than retrieved, 372 gaps in existing knowledge, 87-88 260-261, 265
mechanism of, 384-385 lack of external criterion of, 255-256 cerebral dysfunction and depression,
models of, to explain scientific con- need for common, 239 250
cept, 389-392 non parametric analysis of, 95 Brainstorming, 105, 141,239,315,
pitfalls of, 387 nontest, 54-58 425
quality of, 386-387 problems with multiplicity of, 238 Bransford and Stein's IDEAL learning
in solving scientific problems, 348, taxonomy of, 53-59 framework, 172, 174
349 See also Tests and testing of Bridge statements and terms, 7, 29
spontaneous, 349, 351,378-379 creativity Bridging between analogous cases, 358n
441
442 INDEX
Britton and Tesser Time Management Constructivist theory of rationality, Creativity (cont.)
Instrument, 438-439 193-195 107, 114, 126,211-212,238,239,
Browning, Elizabeth Barrett, 248 and creativity, 195-197 413-416
Constructivist use of writing, 323-324, as debased term, 429
Campbell, N. R., 346, 348, 396 327,329-330,335 as distinct construct, 29, 68
Category width, 220, 265 Content validity, 42 distribution of, 12-13, 13, 95, 196
Chance, 22, 24-25, 95, 135-136, 140 Contextual cues and elements, 149, 154 as goal, 289
Cheever, John, 180 Contradiction and paradox, 203, 204- kinds of theory of, 85
Childhood Attitude Inventory for Prob- 205, 205-206, 239; see also knowing-levels approach to, 195
lem Solving, 55 Ambiguity or disorder as intervening variable, 4
Children Convergent thinking, 17-18, 18-19, as layperson's concept, 77
adaptation of tests for, 54 22, 84, 236, 240, 256, 257 need for theoretical viewpoint in stud-
awareness of own cognitive systems, defined, 15 ies of, 238, 239
207 judging tests requiring, 45, 46-47 practical, 207-208, 283, 302
factors affecting creative thinking in, Correlates, Spearman's principle of, 17 as quasi-random combination of men-
86,126 Cortical arousal, 218- 219, 225, 227, tal elements, 218, 228
as most common test subjects, 88 258,260 question of whether unitary charac-
peer nominations of, 56 Counseling process teristic, 62, 80
performance on tests of, 49 creativity in, 272-279 renewal of interest in, 50
problem-solving styles of, 97 defining, 272 social or spiritual, 94
reasoning in, 185-187, 189, 190, 194 Cowper, William, 248 Criterion Committee Reports of the
reliability of testing of, 45 Cox, C. M., 58, 82, 97-98, 101, 104, Utah Conferences on Creativity, 11
slow and complex rate of personal 112 Criterion-related validity, 23, 24, 42-
development of, 237 Creative Ability in Mathematics Test, 44,62,64
teaching of creativity to preschool, 85 Cultural factors, 44, 45-46, 96
411-426 Creative activity and behaviors, 233, Curie, Marie, 99, 136
writing by, 308, 313, 315 238 Curie, Pierre, 136
Chopin, Frederic, 248 need for taxonomy of, 265 Curiosity
Churchill, Winston, 255 predictors of, 22, 24, 44, 63, 88, "why" questions, 375, 376
Clemons, Alonzo, 112 104, 107 See also Problem finding
Clicks, 179-180 "ultimate" and "concurrent" criteria
Closure of, 255, 256 Dali, Salvador, 224
deferral of, 223 Creative Attitude Survey, 55 Darwin, Charles, 206, 250, 251, 342,
premature, 24, 239-240, 315 Creative Behavior Disposition Scale, 343, 373-374, 407
Cognition 54-55 Darwin, Erasmus, 251
defined, 15 Creative Perception Inventory, 54 da Vinci, Leonardo, 99, 246
primary process, 96-97, 216, 217, Creative persons, animosity toward, 241 Deductive reasoning, development of
218,225,226,228 Creative process in children, 183-187, 194, 195
secondary process, 216, 218, 228 agreement re, 400 concept of inference validity, 183-
Cognitive processes, 96-97, 108 analyzabilityof, 138,380 185
in writing, 308-309, 328-330 perceptual components of, 159-160 empiricist view of, 189
Cognitive style, 12,26, 27,78-79,83- as problem solving, 84, 97, 118, linguistic transmission and, 189
84,85,97,219-220,244,315- 138-139, 203-204, 399-410 via reflective abstraction, 188
316,408 product vs., 11-13,399-400 Depression
personality and motivation more as same in all domains, 212 base rates in general population, 250
important than, 213 as "special," 138 bipolar, 226
Coleridge, Samuel Taylor, 223, 246 stages of, 214-216 among eminent people, 246-247
Composers: see Musicians Wallas's model of, 178-181,214 and inspiration, 249
Composite Creative Personality Scale, Creative Process Check List, 9 most likely psychiatric disturbance of
225 Creative product writers, 249-250
Computerized axial tomography scans, conditions for defined, 123 Dewey's problem-solving model, 115
40-41 core characteristics of, 11-12, 135, Dialectic process of model construction,
Computers 203, 399-400 370
designing for creativity, 152 as criterion of creativity, 64 Dialectic tension, 370
time management and productivity of, exemplary, 408 Dialectic thinking, 200-208, 352,405-
430 judging, 28, 45-46, 47, 49, 57, 60, 406
use of search heuristics in, 138, 141 61; see also Appropriateness of Diligence, 137,213; see also
Concept Mastery Test, 254 product; Evaluation; Evaluators and Persistence
Condensation, 12, 128 judges Dimensional discrimination, 60-61, 68
Consequences Test, 47, 48, 54 vs. person, 129 Discriminant validity, 34-35, 38, 39,
Construct validity, 18,34-41,87-88, vs. process .issue, 11-13,399 40,47-48,48-49,60-61,62-63,
125-126,255; see also Factor Creativity 68, 69
analysis; Nomological validity criteria for and definitions of, 63, 94, Disinhibition, 225, 226, 228, 258, 260
INDEX 443
Divergent thinking Eminent persons (cont.) Flexibility (cont.)

cognitive traits comprising, 256 major studies of, 57-58, 82, 83, 86, and achievement, 24, 137-138
and convergent thinking, 17-18, 18- 104-105, 112,253-254 confounding in studies of, 47, 48
19, 257 mental disorders in, 246-247, 249- defined, 16, 35, 115,256
discriminant validity of judgments of, 250, 252 Fluency, 84,115,115-116
60-61 use of primary process cognition by, and artistic achievement, 24
in Guilford's SI, 15, 16, 35-36, 83 216 as confounding factor, 35, 47-49,
heuristic searching vs., 141 Emotional factors 119
in hypothesis generation, 372, 375, in cognition, 332-333 defined, 16,256
379-380 in creativity, 97, 100,206,226,407- figural, 68
and Kuhn's revolutionary science, 13 408 ideational, 6, 7, 14,20,21-22,47-
relationship with creativity, 62, 98- in reader's response, 332, 337 49,78, 84,88,260,328-329;see
99,99, 136-137,236-237,239, Energy level, inertial quality of, 222- also Idea generation and
259-260,412 223 development
Divergent-thinking tests, 18-23,44-45, Engineers, 55, 56, 57, 125, 138 verbal, 6, 35, 67, 68, 68-69, 78
46, 47-48, 54 Environment: see Situational factors Foreign language learning, 221
predictive validity of, 88 Eureka event: see Insight Freud, Sigmund, 96, 251
suggestions for revision of, 54 Evaluation Frost, Robert, 180,250
See also Guilford tests; Torrance in art students, 24
Tests of Creative Thinking; Wal- as a creative process, 24 Galton, Francis, 8, 82, 94, 103
lach-Kogan creativity tests defined, 15 Garcia Marquez, Gabriel, 180
Diversity in product, 416 and problem finding, 140 Generalization, 421-423
Dobzhansky, Theodosius, 219 testing, 14 Generative processing, 123
Dogmatism, 78 See also Closure; Hypothesis evalua- Genetic factors, 98, 99-100, 107,226,
Domains tion; Revision 226-227, 243, 245, 263-265
of creativity, 94 Evaluators and judges and some artistic differences, 96
maintaining creativity by shifting, consensual techniques of, 12, 13 epigenetic development of reasoning,
222 reliability of, 45-47, 59-61,63-64, 191-192
skills related to: see Knowledge 67-68,69, 82 heritability of divergent thinking, 257
Donizetti, Gaetano, 248 social validation by, 415, 423 Genetic theory, 93-94
Dramatist method, 316-317 Evolving systems approach, 206-207, Genius, 94, 96, 243, 244-252
Dream recall, 153, 160, 216 207 Germain, Sophie, 99
Dream states, transitional, 152 Exvia,99 Gifted, the, 94, 235, 239
Drug usage, 160 Eye movement laterality, 265 Glass, William, 142
Eye tracking, 264-265 Goal Generator, 431-432, 435
Eccentricity, 255, 259, 262 Eysenck Personality Inventory, 224, Goal Manager, 430-431, 431
Edel, Leon, 249 258 Goal Prioritizer, 432, 435
Edison, Thomas Alva, 213 Goals, 28,167,178,308,309,405,
Education and scholastic success Factor analysis, 14-15,36-38 434-435
differences between high and low confirmatory, 35, 36-37, 64-68, 69 Goals and gists strategy, 30 I
achievers, 24, 253 of criterion measures, 43 Goal-setting, 140, 143, 172, 177,206,
formality of schools as factor, 257 exploratory, 35, 36, 68 274, 285; see also Problem finding;
as independent variable, 129 limitations of, 37-38, 121 Problem selection
relationship with creativity, 112, 119, rotation of factor axes, 36 Goethe, Johann, 219, 246
222, 279 targeted rotations, 18 Gough's California Psychological In-
See also Learning Families ventory, 55
Einstein, Albert, 212, 247 creativity in, 103 Gough's Creative Personality Scale, 55,
Einstellung effect, 370-371, 379,404 resemblances in, 107 62
Elaboration size of, 83 Graves, Robert, 225
defined, 16, 35, 256 Famous persons: see Eminent persons; Group Inventory for Finding Interests
interests of students surpassing in, 85 name of individuals; specific oc- (GIFFI), 54, 65, 67, 69
in learning process, 167, 171-172 cupations Guilford, J. P.
in transfer of information, 167-168, Fantasy, 216 Aptitudes Research Project, 37
172 Faraday, Michael, 247, 250, 368, 402 views on creativity, 13-19,256-
Eliot, T. S., 225, 250 Father, absent, 221 257
Ellis, Havelock, 58, 252 Femininity, psychological, 78, 254 Guilford's Structure of Intellect, 14, 15,
Emerson, Ralph Waldo, 249 Festinger, Leon, 222 35,83, 115-116, 122,256-257
Eminence, Galton's, 94 Field independence, 78, 84 evaluated, 18-19,37,84-85
Eminent persons Fixity-mobility, 84 See also Divergent thinking
creative process in, 23, 64, 214 Flaubert, Gustave, 222 Guilford tests, 18-19,38,39,56,95,
IQs of, 97-98, 125 Fleming, Alexander, 25 125, 125-127,255
familial resemblance in, 103 Flexibility, 14, 115-116, 142-143, Guilford-Zimmerman Interest Inventory,
grades of and test scores of, 98 375, 376 55
444 INDEX
Handedness, 82 Inductive reasoning Intelligence tests and testing (cont.)

Hanson, N. R., 343, 344, 347, 348 development of, 187-188, 189-190, development of, 8
Helmholtz, H. von, 96, 214 194-195 earlier inclusion of creativity mea-
Henmon-Nelson Tests of Mental Abil- in forming hypotheses, 342-343, sures, 5
ity,48 345, 347 inadequacies of, 118, 119
Hewitt, Paul, 387, 388, 390-391, 392 Information: see Knowledge no allowance for divergent response,
Homosexuality, 221 Information processing 122
Homospatial thinking, 261 capacity limitations, 312-313, 318- Interaction with others, 330
Hugo, Victor, 222 319 Interactionism, dynamic
Humor, 240, 335 defined, 15 and development of reasoning, 192-
Huxley family, 103,246 empathic, 276-277, 279 193
Hypnotic ability, 216, 237 framelike structural features in, 328 Interactionist models of creative behav-
Hypothesis evaluation, 364-365, 378, models of, 306-308, 401 ior, 26-28, 79-81, 83, 85, 87, 88,
379-380 "optimum deviation" in continuum, 89
explanatory models and, 346-347, 244, 261-263 Interactive nature of counseling process,
347-348, 362, 375 in perceptually handicapped, 151 272, 274, 275, 277
falsification views, 187, 188, 189- transformation ability in, 157, 159 Interests
190, 190, 343-344 Innovation cross-sexual, 220-221
Hypothesis formation, 348 as not creative, 233, 238 intensity of, 222
by abduction, 347, 377 Innovativeness inventories of, 54-55, 65, 67, 69
analogies and, 346-348 measures of, 59 patterns of, 85
by creative intuition, 345 Insight range of, 219-220
by induction, 345, 347 Eureka type, 342-343, 348 Interpersonal influence research para-
model vs. anomaly and, 371 sudden vs. incremental, 365-369, 379 digm, 278
by non inductive reasoning, 344, 345, transformation abilities and, 157, 160 Invention techniques, 315-316
351-359, 361-362, 375, 377 See also Illumination and inspiration Isadora, Rachel, 326-327
Popper's conjectural viewpoint, 343- Insight behavior, 358, 368-369 Iser, W., 325
344 Institute for Behavioral Research in
prior theoretical framework and, 348 Creativity, 55 James, William, 192, 197
rationalist-Eurekaism VS. inductivist- Institute of Personality Assessment and Janusian thinking, 6, 203, 206, 261
accretionism, 342-343, 365-369 Research (IPAR), 20, 57, 254, Jenkins's learning framework, 165, 174
successive refinement VS. blind varia- 254-255 Journals (research publications), 147,
tion, 24, 25-26, 372-373, 378 Intellect, Structure of: see Guilford's 238, 239
See also Idea generation and Structure of Intellect
development Intellectance, 26 Kafka, Franz, 249
Intelligence Keats, John, 276
Idea "certification theory" of, 136 Kepler, Johannes, 395-396
question of whether product, 57, 211, correlations with ratings of creativity, Kirton Adaptation Inventory, 59
233 60,61 Knowledge, 206
Idea generation and development developmental theory of, 122-123, acquiring, 118, 139-140, 143, 178-
as case of "ill-structured problem," 123 179
307-308 distinction between creativity and, 41 and chance, 136
effect of audience on, 306-307 "excessive," 112 construction of, 377, 380
methods of, 314-317, 319 manifestations of, 115, 122 in counseling, 275-276
models of, 308-311, 318 as multifaceted, 19 dialectical epistemology, 203, 205-
purpose and, 308 psychometric theories of, 121-122 206, 207
toward a theory of, 305 threshold hypothesis, 87, 98, III, domain-relevant, 26, 27, 28, 140,
See also Fluency, ideational 125, 126-127, 129, 136,212-213 173, 213,403
IDEAL learning framework, 172, 174 Intelligence-creativity relationship, 97- flexible representation of, 178, 179,
Idiots savant, 100 98,107,111-114,123-124,127 181
Illumination and inspiration synergism in, 127-129 as impediment, 221-222, 278, 404
cognition and, 214-216 Thurstone's factor-analytic model of, informed transfer of, and set acquisi-
depression and, 249 6 tion size, 169-171
evolving systems VS., 206 See also Guilford's Structure of levels of knowing, 193-194, 195
low cortical arousal during, 218 Intellect needed for creative solutions, 140,
as noticing an analogy, 218 Intelligence quotients 164-165, 173
during "relaxed tension," 277 and creativity scores, 124, 129 new vs. old in explanatory-model
Wallas's stage, 115, 179-180,214 and divergent-production scores, 20- construction, 364, 372
See also Idea generation and develop- 21,21,116 readers', 325, 330
ment; Insight of eminent persons, 97-98, 136 in search process, 405-406
Incubation, 5, 6, 9, 179,214, 218 See also Threshold hypothesis under self-understanding and, 193, 196
in therapeutic counseling, 274-275 Intelligence spontaneous transfer to problem-solv-
Wallas's creative-process stage, 5, 6, Intelligence tests and testing ing tasks, 166-167, 170-171,
96, 141-142,406 correlation with creativity tests, 13, 172-175
and writing revision, 314 43, 48 and use of heuristics, 26
INDEX 445
Knowledge (cont.) Memory (cont.) Newton, Sir Isaac, 222, 250, 250-251
writers', 286, 293, 294, 313, 318, negative effects on creativity, 164, Nietzsche, Freidrich, 215
319 164-165 Nightingale, Florence, 251-252
See also Memory and perceptual processes, 153 Nobel prizewinners, 105, 137,227,261
Koestler, Arthur, 6, 180, 224, 272, remembering as creative activity, 163 Nomological validity, 61-63, 68-69
345, 346 search process, 404-406 Nonconformity, 24, 225
Krippner's systems approach, 112-113 short-term, 218 Novelty, 207
Kuhn, Theodore, 7, 13,215,263,342, structure of, 332 as criterion of creativity, 11, 14,94,
343, 344, 348, 370, 371 working, 306, 318,404,406 203
in writing process, 328, 329, 335 liking for, 224
Lateral thinking Mendel, Gregor, 250 of point of view, 206, 207
de Bono's "new think," 105, III, Mentors and role models, 22, 227, 253 in preschoolers' responses, 413-414
112,119 Meta-analysis, 88
Lateralization: see Brain hemisphericity Metacognition, 118, 177-181,287,
Object-sorting tasks, 224-225, 249
Lawrence, Ruth, 99 403-405,430,430-433; see also
Odd or bizarre responses, 21, 224, 413
Leadership Self-evaluation
Openness to experience, ideas, or infor-
creativity as variety of, 26 Michael and Colson's biographical
mation, 127,224,276
Learning questionnaire, 55
serendipity, 25, 240
by construction conceptual rela- Michael Test, 38, 39
Open Processing Scale, 59
tionships, 384 Michelangelo, 218, 245, 252
Operant conditioning, 80,417
and development of reasoning, 189 Minnesota Multiphasic Personality In-
Operational thinking
via model construction, 376-377, ventory, 224, 254
formal, 199-200,205
380 Model
postformal, 200, 200-202
reading, summarizing, and, 287 defined, 384
Opinion, Attitude, and Interest Survey,
Teaching-with-Analogies Model, expedient, 359
55
389-392 explanatory, 350, 359, 376, 377,
Order
theories of, 120, 123,417-418 378, 380
need for while tolerating disorder,
Leisure activities, 241 formation from analogies, 359-361
223
Lifetime Creativity Scale, 59, 64 generating hypotheses using, 346,
Origence, 26
Light shows, 158 346-347,361-365
Originality
Lithium, 250 persistence of original, 370-371, 379
in behavior analysis, 413, 414
Locus of control, 78, 85, 235-236, 238 scientific, 349-350, 378
confounded by fluency, 47-48, 48
Logical necessity, 189-190, 192 for simplification, 360-361
correlation with RAT scores, 20
Logical positivists, 343 structural problem-solving, 11, 15,
data on, 84, 85
London, Jack, 250 16-18
defined, 16,35, 256
Longitudinal studies, 40, 108 Montessori, Maria, 413
determining a preschool child's, 414
Lowell, Amy, 179 Moore, George, 224
drive for, 137
Morel, Benedict Augustin, 245
and psychoticism, 258-259
McCormick, Cyrus, 212 Motivation by extrinsic factors, 27, 28,
as simulated product, 119
MacLeish, Archibald, 142 86; see also Reinforcement
validity question, 18,60
Maladjustment and creativity, 219, 235, Motivation by task, 26, 27, 28, 277
237-238, 243-244, 253, 253-254 Amabile's studies of, 223-224
historical overview of, 244-252 anomaly as source of, 371 Paradox and contradiction, 203, 204-
Manet, Edouard, 140 and creativity, 237, 239 205, 205-206, 239; see also
Marijuana use, 226 in leisure behavior, 240 Ambiguity or disorder
Masters, Edgar Lee, 250 Motivational and emotional factors, 13, Pascal, Blaise, 247, 249
Mathematicians 108, 126, 137, 143, 143-144,213, Passive avoidance tasks, 228
cognitive traits of, 257 222-224, 228, 261 Pauling, Linus, 222
"flexibility" of creative, 137-138 Mozart, Wolfgang Amadeus, 99, 248 Pavlov, I. P., 140, 260
Stanley'S students, 99, 106 Multitrait-multimethod analysis, 38-40, Peer nominations, 56, 60-61
women, 102,221 47-48, 63, 88 Peer role in fostering creativity, 426
Mathematics, 78, 85 shortcomings of, 39-40, 47 Perception and creativity, 148, 150-
de Maupassant, Guy, 248-249 Murray, Donald M., 142 151, 156-157
May, Rollo, 113 Murray, E. L., 330 research on, 147-148, 152-153
Meditation, 240 Musicians, 99, 103, 139,246-247,257 Perceptual constancy, 148
Memory Perceptual organization, 149-150, 160
accessibility of information in, 167, Nadia, 99-100 assessing individual variation in,
169,172,175,393 Narcissism, 78 154-156
episodic, 324, 331-332, 333, 335, Nativism, 191-192 and creativity, 159, 161
337 Necker cube, 149 decoupled from sensory input, 152-
forgetting, 228 Nelson, George, 179 154
free recall from, 169, 175 Neurological factors, 40-41, 100-101, "loose," 150, 153, 159, 159-160
in Guilford's Structure of Intellect, 15 120-121,123,217-218,227; see Persistence, 87, 375, 407-408
long-term, 306, 309, 318; see also also Brain hemisphericity; Cortical Person
Knowledge arousal in interactionist model, 81
446 INDEX
Personality Problem selection, 220, 277, 376n, 407 School environment, 106, 331, 424-
double, 225 Prodigies, 94, 99 426,426
theory of, 78, 85 Proust, Marcel, 227, 248 Schumann, Robert, 247-248
traits, 4, 13 Psychologists Science education, 376-377, 380, 383,
Personality factors, 12,26,26-27,77- surveyed, 105 388-389,390-391,392-393
78,85-86, 104, 129,219-226, Psychosis, 224, 261, 263; see also Sciences
252-256 Maladjustment creativity in, 5, 94, 107, 138, 375-
confounded by cognitive style, 261 Psychotherapeutic counselors 376
factor-analytic studies of, 224 training of, 278-279 ill-structured problems in, 402
and interactions with intelligence, 118 Psychotic thinking, 257-258, 259, 261 situational factors affecting creativity
See also Cognitive style; Motivational in, 95, 107
and emotional factors; specific Racial factors, 82, 85, 96, 236 Scientists
traits Ratings: see Evaluators and judges biographical inventories for, 55
Personality inventories and studies, 9, Readers characteristics of, 98, 99, 104, 105,
10, 54, 55, 56, 105, 225 adapting to needs of, 286, 289, 290 137,253-254
Persons, famous: see Eminent persons; responding to literature, 323-333 creative process in, 96, 215
names of individuals; specific oc- as rhetorical goal for writers, 291 IQs of, 98
cupations teaching, 334-335 mental illness among creative, 247,
Petrosky, A., 325, 330 Reason, and emotion, 333 250
Phaedrus, 245 Reasoning, development of: see peer nominations of, 56
Physics, 78 Deductive reasoning; Inductive supervisors' ratings of, 57
Piaget, Jean, 122, 193, 196 reasoning women, 101, 102
Plato, 225 Redefinition, defined, 256 Selective retention, 24, 25-26, 372-
Plot Titles test, 47, 48, 54 Reinforcement, 417-420, 425 373
Poe, Edgar Allan, 158,214,219,246 delayed or intermittent, 426 Self, 233-234, 239
Poets, 139,214-215, 221, 223, 249 Religious background, 83, 105 Self-actualization, 235
Poincare, Henri, 5, 96, 212, 215, 218, Remote Associates Test (RAT), 20, 40, Self-confidence, 213, 222, 276
400 56,59,62,99, 115, 124-125,265 Self-esteem, 78, 234-235, 236, 239
Popper, Karl, 7, 138, 189-190,343- problems with, 118-119 Self-evaluation, 276, 375
344, 345 Revised Art Scale, 260 Self-identity
Pound, Ezra, 250, 326 Revision contrasting and dissociated subselves,
Preparation stage, 214; see also goals and gists strategy, 30 I 225
Knowledge standards of, 142-143, 180 Self-reflection, 185, 238
Priestley, Joseph, 396 "verification," 180-181,214 in development of reasoning, 188,
Problem finding, 14, 23-24, 24, 140, of writing, 141, 286, 287,295-301, 191
220, 395 313-314,319,330-331,414 Self-reports, 9, 33-34, 45, 58, 104,
Dewey's problem recognition, 115 Rewards: see Motivation by extrinsic 214-215
Problem solving factors Sensory impairment, 150-151
analogical, 392-395, 395-396 Rimbaud, Arthur, 222, 246 Sensory stimuli
analysis of protocols, 196-197 Risk-taking, 225 control of, 414, 416, 420-421
creativity as, 5, 84, 97, 118, 138- increasing failure rate, 240 heightened sensitivity to, 227
139, 399-410 Role models and mentors, 22, 227, 253 information processing of, 262-263
functional fixedness in, 164 Rossetti, Dante Gabriel, 213, 246 intentional blocking of, 227
Guilford's SIPS model, 15, 16-18, Rush, Benjamin, 249 interfering with creative thought pro-
19 cesses, 151
measures of, 59 Schaefer's Biographical Inventory, 55, Serendipity, 25, 240
models of, 6, II, 15, 16-18,26-28 56 Sex as factor, 45, 55, 81-82, 85, 100,
previous experience with same tasks, Schemata analysis, Martin's, 278 101-103, 107,234-235,236,237
164, 173 Schema theories, 122-123, 178, 179, cross-sexual interests and identities,
the psychotherapeutic process as, 271 181 220-221
role in learning, 172, 174 Schizophrenia Shelley, Percy Bysshe, 246, 248
skills, 106, 141, 377 common cognitive configuration, Simon, Herbert, 137
strategies, 286-288, 295-301, 403, 153, 225-226 Singer, Isaac Bashevis, 179
405; see also Revision and creativity, 150, 161,244,254 Situational factors, 14,22,28,214,
styles of children, 97 and eye-tracking dysfunction, 264- 218,227,405,430; see also Work
transferring knowledge to tasks, 165- 265 environment
172, 173-174 survival genetically, 263-264 Sixteen Personality Factor Question-
in writing processes, 305, 307-308, Schizophrenics naire, 55, 61,253,254
314-316 categorization by, 155-156, 220, Skills, 26-27, 28, 177
Problems 249, 259 Social climate (Zeitgeist), 95, 105, 236,
ill defined, 284, 284-286, 287, 302, creativity of, 260 251,263
307, 400-403 eminent, 246-247, 250-251 Social and environmental factors, 79-
solution as ill structured also, 403 similarities with creatives, 259 80,81-83,86-87, 103, 104, 107,
well structured, 40 I, 402 Scholastic Aptitude Tests (SATs), 124- 344n
See also Task representation 125 and neurological growth, 100
INDEX 447
Social and environmental factors (cont.) Tests and testing of creativity (cont.) Transformations (cont.)
See also School environment; Situa- deficiencies of, 9-10,118-119,125 generating analogies, 371-372
tional factors; Work environment development of, 8-9 intramodal, 156-157, 160
Social sciences list of tests, 59 Twain, Mark, 250
creativity in, 408-409 performance norms, 49, 50 Twin studies, 103, 108, 258, 264
structure of problems in, 402 problems with laboratory, 165, 168,
Social scientists, 104, 253 169,172,174-175 Unconscious or subconscious processes,
Socioeconomic status, 82 reliability of, 13, 18,43,44-46,59, 5,96-97, 107; see alw Incubation
Socrates, 244 67
Spearman's g, 4, 29, 94, 121 scoring or judging of, 46-47, 61 Valery, Paul, 222
Spring problem, introduced, 350-351 standardized, 415 Value or usefulness of product, 127-
Standards for Educational and Psycho- validity of, 18, 23, 24, 34-41,42- 128, 135, 203
logical Testing, 34, 35 44, 47-49, 60-64, 68-69, 87-88, Verhaeren, Emile, 227
Stevenson, Robert Louis, 225 125-126, 255 Verification: see Revision
Structure of Intellect: see Guilford's See also Divergent-thinking tests; Vernon's model, 122
Structure of Intellect Guilford tests Visual relationships and abilities, 152-
Structure of Intellect Learning Abilities Thackeray, William Makepeace, 215 153, 157, 160
Test, 40 Thomas, Dylan, 142
Study of Values Scale, 55, 61, 223 Thought processes Wallach-Kogan creativity tests, 20-21,
Suicides, 246-247, 249, 250 allusive, 257-258 54,64, 65, 67
Supervisor ratings, 56-57 complex, 14, 127,237,376 Wagner, Richard, 248
Susuki,412-413 continuum of, 261-263 Watt, James (Scottish engineer), 251
Swift, Jonathan, 248 extraordinary, 345-346, 367-368 Wechsler scale, 121
Synectics, 141,315-316 lateral, 105, III, 112, 119 Wedgwood, Josiah, 251
Synesthesia, 157-159, 159, 160 operational, 199-202,205 Westinghouse Science Talent Search,
Onomatopoeia and Images test, 59 See also Cognition; Cognitive 137
Synthetic abilities, Sternberg's, 118, processes Wheel problem, posed, 348-350
122 Thurstone's multiple-factor theory, 121 Williams, Tennessee, 179
Time of day, 435-436 Woolf, Virginia, 249
Tagmemics, 317 Time management, 430-436 Work environment, 79, 86-87, 112
Talents, 26, 94 questionnaire, 438-439 Writer's block, 286-287, 309-310
Task management, 430-436 Torrance Tests of Creative Thinking, 4, Writers
Task representation, 140-141, 143, 38,39,47,54,56,65,67,82,95, alcoholism in, 226
173-174,401,402,403 117,124,236 for children, 306
re-representation, 285, 287,406 TOTE (test-operate-test-exit) model, 16 cognitive traits of, 98, 249, 257
by writers, 284-286, 288, 301, 302 Trainability of creative skills, 105-106, nature of inspiration, 214-215
Tasks 107-108,181,226,236,237, personality profiles of, 254
need for a theory of, 19 239-241,278-279,279,302,409, psychiatric disturbance among, 246,
properties of, 433 412-413 249-250
Tchaikovsky, Peter Ilich, 142,246 assessing as independent variable, women, 102
Teachers' evaluations, 24, 55-56, 60, 416-417 Writing, 60, 65, 66, 67, 68, 68-69
239 behavioral approaches, 120,411- Amabile studies of, 237
Teaching-with-Analogies (TWA) 412, 413-414, 426 creati ve defined, 305
Model, 389-392 Transfer recall, 17, 125 models of, 308-311, 318
Terman Concept Mastery, 98 Transformation of constraint, 12 as response to reading, 323-337
Terman Genetic Studies of Genius, 250 Transformations, 83-84, 128, 159 See also Revision
Tests and testing of creativity convergent thinking and, 35, 46 Writing process, 180, 181,214-215,
administration of, 49, 50, 61 in counseling process, 272-273 312
adult subjects needed, 88, 108 cross-modal, 157-159, 160
completion-type needed, 13 defined,35 Zola, Emile, 249

(Perspectives On Individual Differences) Robert T. Brown (Auth.), John A. Glover, Royce R. Ronning, Cecil R. Reynolds (Eds.) - Handbook of Creativity-Springer US (1989) PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Perspectives On Individual Differences) Robert T. Brown (Auth.), John A. Glover, Royce R. Ronning, Cecil R. Reynolds (Eds.) - Handbook of Creativity-Springer US (1989) PDF

Uploaded by

Copyright:

Available Formats

Handbook of

CECIL R. REYNOLDS, Texas A&M University, College Station

Recent volumes in the series

THE CAUSES AND CURES OF CRIMINALITY

HANDBOOK OF MUL TIVARIATE EXPERIMENTAL

HISTORICAL FOUNDATIONS OF EDUCATIONAL

THE INDIVIDUAL SUBJECT AND SCIENTIFIC PSYCHOLOGY

LEARNING STRATEGIES AND LEARNING STYLES

METHODOLOGICAL AND STATISTICAL ADVANCES IN THE

THE NEUROPSYCHOLOGY OF INDIVIDUAL DIFFERENCES

PERSONALITY DIMENSIONS AND AROUSAL

THEORETICAL FOUNDATIONS OF BEHAVIOR THERAPY

Springer Science+Business Media, LLC

© 1989 Springer Science+Business Media New York

All rights reserved

E. Paul and Pansy Torrance

Bonnie B. Armbruster Roger Bruning

Patricia Bachelor Linda J. Carey

Carolyn A. Ball John Clement

Suzanne Benack E. Thomas Dowd

Bruce K. Britton Karen Fitzgerald

Robert T. Brown Linda Flower

John H. Flowers Colin Martindale

Calvin P. Garbin Linda F. Mattocks

Shawn M. Glynn Mary L. Means

Leslie E. Lukin Lyle F. Schoenfeldt

Margaret Semrud-Clikeman James F. Voss

PART I THE NATURE OF THE BEAST

Chapter 1 Creativity: What are We to Measure? 3

What is Creativity? .............................................................. 3

Chapter 2 Psychometric Issues in the Assessment of Creativity 33

Chapter 3 A Taxonomy and Critique of Measurements Used in the Study of Creativity. . . . . . 53

A Taxonomy of Creativity Measurements ............................................ 53

Chapter 4 Individual Differences in Creativity: An Interactionist Perspective 77

Theoretical Explanations for Differences in Creative Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

Chapter 5 The Nature-Nurture Problem in Creativity 93

Chapter 6 Creativity and Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

PART II COGNITIVE MODELS OF CREATIVITY

Chapter 7 Cognitive Processes in Creativity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 5

Characteristics of Creative People . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

Chapter 8 Creativity and Perception. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

Chapter 9 Memory and Creativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

The Effects of Memory on Creativity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 164

Chapter 10 Metacognition in Creativity 177

What is Metacognition? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 177

Chapter 11 The Creative Construction of Rationality: A Paradox? 183

Lori: The Creative Construction of Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

Empiricist and Nativist Accounts of the Development of Reasoning. . . . . . . . . . . . . . . . . . . . . . .. 188

Chapter 12 Dialectical Thinking and Adult Creativity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199

PART III PERSONALOGICAL VARIABLES AND CREATIVITY

Chapter 13 Personality, Situation, and Creativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 211

Introduction .................................................................... 211

Definition of Creativity ........................................................... 233

Chapter 15 Creativity and Psychopathology: Gamboling at the Seat of Madness . . . . . . . . . . .. 243

An Historical Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ., 244

Chapter 16 Examining Counselors' Creative Processes in Counseling . . . . . . . . . . . . . . . . . . . .. 27 1

The Counseling Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 272

Chapter 17 Foundations for Creativity in the Writing Process: Rhetorical Representations

Introduction: An Approach to Studying Creativity in Writing . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 283

Human Information-Processing Model ............................................... 306

Creative Dimensions of Reader Response. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 327

Introduction .................................................................... 341

Analogical Reasoning and Comprehending Science Text ................................ 383

Originality r#.__~:~===r-----·-ii61 (.94)