Thought-Based Linguistics

The extent to which language is inseparable from thought has long been
a major subject of debate across linguistics, psychology, philosophy, and
other disciplines. In this study, Wallace Chafe presents a thought-based theory
of language that goes beyond traditional views that semantics, syntax, and
sounds are sufficient to account for language design. Language begins with
thoughts in the mind of a speaker and ends by affecting thoughts in the mind
of a listener. This obvious observation is seldom incorporated in descriptions
of language design for two major reasons. First, the role of thought is usually
usurped by semantics. But semantic structures are imposed on thought by
languages and differ from one language to another. Second, thought does not
lend itself to familiar methods of linguistic analysis. Chafe suggests ways of
describing thoughts, traces the path languages follow from thoughts to
sounds, and explores ways in which thoughts are oriented in time, memory,
imagination, reality, and emotions.

Wallace Chafe is Professor Emeritus and Research Professor at the University

of California at Santa Barbara. His research has focused in part on the Seneca
language in New York and the Caddo language in Oklahoma. He has inves-
tigated differences between speaking and writing; the functions of prosody in
spoken language; the emotion underlying laughter and humor, ways in which
language can be beautiful, and relations between language and thought.
Thought-Based Linguistics
How Languages Turn Thoughts into Sounds

Wallace Chafe
University of California, Santa Barbara
University Printing House, Cambridge CB2 8BS, United Kingdom
One Liberty Plaza, 20th Floor, New York, NY 10006, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre,
New Delhi – 110025, India
79 Anson Road, #06–04/06, Singapore 079906

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9781108421171
DOI: 10.1017/9781108367493
© Wallace Chafe 2018
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2018
Printed in the United Kingdom by Clays, St Ives plc
A catalogue record for this publication is available from the British Library.
ISBN 978-1-108-42117-1 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy of
URLs for external or third-party internet websites referred to in this publication
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.
Chapter 19, “Emotional Involvement in a Conversation,” includes material from the
chapter by Wallace Chafe “Prosody and Emotion in a Sample of Real Speech” in the
book Relations and Functions Within and Around Language edited by Peter Fries,
Michael Cummings, David Lockwood, and William Sprueill, published in 2002 by
Continuum, London.
Chapter 20, “The Feeling of Nonseriousness,” is based on the book The Importance
of Not Being Earnest: The Feeling Behind Laughter and Humor by Wallace Chafe,
2007, with permission from John Benjamins Publishing Company.

The poem “I started early, took my dog” on page 179 is reproduced from The Poems of
Emily Dickinson: Reading Edition, edited by Ralph W. Franklin, Cambridge, MA:
The Belknap Press of Harvard University Press, Copyright © 1998, 1999 by the
President and Fellows of Harvard College. Copyright © 1951, 1955 by the President
and Fellows of Harvard College. Copyright © renewed 1979, 1983 by the President and
Fellows of Harvard College. Copyright © 1914, 1918, 1919, 1924, 1929, 1930, 1932,
1935, 1937, 1942 by Martha Dickinson Bianchi. Copyright © 1952, 1957, 1958, 1963,
1965 by Mary L. Hampson.
Language consists of signs, representing ideas. These signs are
selected by the person who speaks, in accordance with the ideas
prevailing in his own mind, in order to produce the reversed process
in the individual spoken to; they are used for that process – the most
wonderful and important on this earth – of conveying ideas from one
distinct individual to another; for the communion of mind with mind,
through sensuous impressions, made in skillful succession, and in
accordance with general laws.
Francis Lieber 1850. Quoted in Andresen (1990: 115–116)
 Contents

Acknowledgments page ix

Prologue 1

Part I Preliminaries
1 Background 7
2 Ground Rules 12

Part II Thoughts and Their Properties

3 The Priority of Thoughts 23
4 The Path from a Thought to a Sound 27
5 How Thoughts Are Structured 33
6 How Thoughts Are Experienced 42
7 How Thoughts Are Shared 50
8 How Thoughts Flow through Time 57

Part III Verbalization Illustrated

9 From a Thought to a Sound in English 69
10 From a Thought to a Sound in a Polysynthetic Language 84

Part IV Related Issues

11 The Translation Paradox 93
12 Repeated Verbalizations of the Same Thought 100

vii
viii Contents

13 Rethinking Whorf 105

14 Lessons from Literature 113

Part V Common Ways of Orienting Thoughts

15 Small Numbers and Subitizing 123
16 Thoughts and Gender 131
17 Time, Tense, Memory, and Imagination 137
18 Relating Ideas to Reality 151

Part VI The Emotional Component of Thoughts

19 Emotional Involvement in a Conversation 161
20 The Feeling of Nonseriousness 169
21 How Language Can Be Beautiful 175
Epilogue 185

References 187
Index 197
 Acknowledgments

I am grateful to a number of people who contributed to this work, most of them

unknowingly. They include the many students and faculty with whom I was in
contact on the University of California campuses at Berkeley and Santa
Barbara, the many individuals whose work is acknowledged here in references,
and others who contributed in a variety of ways. Among the latter have been
sources of data and analyses of the Caddo and Seneca languages. Indispensable
for Caddo was the late Sadie Bedoka Weller as well as Lynette Melnar, whose
description of Caddo verb morphology contributed fundamentally to
Chapter 18. Numerous collaborators on the Seneca language are listed in
Chafe (2015). Their contributions were essential to Chapters 10, 16, and 18.
Recently additional help with Seneca has come especially from Sandy Dowdy.
Chris Chafe helped with the Mozart example at the end of Chapter 8, and both
Steve Chafe and Subhajit Mitra helped with formatting the many figures.
Advice on Japanese was provided by Akiyo Cantrell. Two anonymous
reviewers encouraged me to complete this work in more or less its present
form. Finally, I can hardly exaggerate the importance of the support I have
continually received from my wife, Marianne Mithun.

ix
 Prologue

The goal of this work is to encourage linguists and other interested parties to
recognize the extent to which language is inseparable from thought. Language
begins with thoughts in the mind of a speaker and ends by affecting thoughts in
the mind of a listener. Although this observation might seem obvious, it is
seldom incorporated in a model of language for at least two major reasons.
First, the role of thought is usually usurped by the semantic component of
language. But semantic structures are imposed on thoughts by languages, they
differ from one language to another, and while they are closely related to
thoughts they are not equivalent to thoughts themselves. Second, thoughts
are not structured in a way that lends itself to familiar techniques of linguistic
analysis. Linguists, if they ever consider the question at all, might well ask
whether thoughts are something their training and experience has prepared
them to deal with.
A way forward cannot help appealing to introspection, which is viewed here
as an essential tool in our observational arsenal, although one that profits from
integration with observations that are more open to public view. Introspection
also demands that we recognize the basic role of consciousness, concerning
which there has been a surprising amount of disagreement and misunderstand-
ing. The flow of thought is seen here as parallel to the flow of consciousness,
a view already espoused by William James (1890).
This work, then, argues for a thought-based linguistics that contrasts with
a bias toward sound (or even worse toward writing) that has upset the
balance of linguistic investigation. The ineffability of thought makes this
a more difficult approach by far, but there is no reason to believe that the
search for truth should be easy. A corollary of this approach is a demotion
of syntax, so often seen as the driving force behind language, and
a recognition that syntax is only one stage along the path from a thought
to a sound. In short, the claim to originality here lies in a stress on the
crucial importance of thought as distinct from semantics, in accepting
introspection and consciousness as essential observational resources, and
in demoting the syntax-centered approach on which so much of linguistics
has been based. Let us see where this leads.

1
2 Prologue

Part I of this book introduces some preliminary considerations. Chapter 1,

Background, tells how this work grew out of my own attempts to understand
the nature of language and the mental processes behind it, as seen against the
background of developments in the field of linguistics. Chapter 2, Ground
Rules, sets forth certain assumptions that underlie the approach taken here,
including what it means to understand something, the importance of introspec-
tion and consciousness, and the value of linguistic diversity.
Part II discusses various properties of thought. Chapter 3, The Priority of
Thoughts, illustrates briefly why priority should be given to thoughts rather
than sounds. Chapter 4, The Path from a Thought to a Sound, provides an
overview of the stages that lead from thoughts to sounds, including the forma-
tion of semantic and syntactic structures and their symbolization by abstract
and overt phonological structures. Chapter 5, How Thoughts Are Structured,
explores what language universals can tell us about types of ideas and the ways
in which ideas are related, as well as “participant roles” like subject and object,
ergative and absolutive, and agent and patient. Chapter 6, How Thoughts Are
Experienced, focuses first on the role of language and then more briefly on the
roles of imagery and emotions. Chapter 7, How Thoughts Are Shared, focuses
on differences between speaking and writing and on the special properties
of reading aloud. Chapter 8, How Thoughts Flow through Time, identifies
coherence at several levels of organization, including intonation units,
sentences, and topics.
Part III traces the verbalization of thoughts in more detail. Chapter 9, From
a Thought to a Sound in English, illustrates the verbalization of thoughts in
English with a brief conversational excerpt. Chapter 10, From a Thought to
a Sound in a Polysynthetic Language, describes the partially different ways in
which thoughts are verbalized in a language of that type.
Part IV discusses a number of related issues. Chapter 11, The Translation
Paradox, asks how it is possible for thoughts that are expressed in one language
to be translated into another language with reasonable success. Chapter 12,
Repeated Verbalizations of the Same Thought, looks at ways in which thoughts
may be verbalized differently when they are remembered at different times.
Chapter 13, Rethinking Whorf, suggests a way of dealing with the controversial
question of whether speakers of different languages think differently.
Chapter 14, Lessons from Literature, illustrates ways in which written literature
can offer insights into the nature of thought, with illustrations of the mimetic
and diegetic options.
Part V discusses a few common ways in which languages orient thoughts.
Chapter 15, Small Numbers and Subitizing, shows how the mind deals with
small numbers of objects and compares findings from linguistics and psychol-
ogy. Chapter 16, Thoughts and Gender, discusses a relation between gramma-
tical gender and a thought pattern that favors masculine singular in Iroquoian
 Prologue 3

languages. Chapter 17, Time, Tense, Memory, and Imagination, shows how
linguistic expressions of time are related to memory and imagination.
Chapter 18, Relating Ideas to Reality, illustrates ways in which different
languages relate thoughts to their speakers’ conceptions of reality.
Part VI is devoted to the emotional component of thoughts. Chapter 19,
Emotional Involvement in a Conversation, illustrates ways in which emotions
are expressed by prosody in a conversational excerpt. The final two chapters
then look at two commonly experienced emotions that are usually overlooked
when emotions are discussed. Chapter 20, The Feeling of Nonseriousness,
deals with the feeling that underlies humor and its expression with laughter.
Chapter 21, How Language Can Be Beautiful, identifies the emotion behind the
experiencing of beauty and explores features that contribute to making lan-
guage beautiful.
The Epilogue mentions some questions that were raised earlier in passing
and that are in particular need of further exploration.
Part I

Preliminaries
1 Background

The human brain has justifiably been called the most complicated object in
the known universe. But the remarkable evolution of the brain has been
inseparable from the evolution of thought and language, which are in turn
inseparable from consciousness, imagery, memory, and imagination. All
these faculties and more combine to form a complex whole that has bestowed
on us, its beneficiaries, the power to radically reshape our environment for
good and evil while surrounding ourselves with transcendent beauty and
stultifying ugliness. The complexity of this feast has overwhelmed our
ability to understand its full nature while providing endless topics for
investigation, frustrating where they seem intractable and exciting where
they open new avenues for discovery.
A number of academic disciplines have taken up the challenge of exploring
selected parts of this complexity, each adding insights supplied by its own
traditions. The present work builds on ways of understanding that have arisen
within certain areas of linguistics, with brief bows in the direction of psychol-
ogy and literary studies. It departs from the mainstream of linguistics by
replacing an approach based on sound and writing with a less explored per-
spective based on thought.
Because this thought-based approach has been a natural outcome of my
own attempts for half a century to understand language and what lies behind
it, the reader deserves to know a little about those attempts and the directions
in which they have led (Chafe 2002a). At first I pictured language as simply
a bridge between meanings and sounds (Chafe 1962). Later I tried to show
how meanings underlie syntax (Chafe 1970a), and later still I explored the
relation between language and consciousness (Chafe 1994). Since then
I have come to appreciate more clearly the difference between thoughts
and the ways in which thoughts are molded by the unique semantic resources
of each language.
When I first became acquainted with linguistics in the mid-1950s it was
a small and relatively homogeneous field. There were heated arguments over
details, but linguists were more or less content to view language as a coherently
designed structure composed of phonemes (basic units of sound), morphemes

7
8 Preliminaries

(sequences of phonemes that functioned as parts of words), and words them-

selves. There was uncertainty about the best way to approach syntax (how
words combined to form phrases and sentences), and there were at least a few
who felt uncomfortable ignoring the fact that words and sentences actually
mean something. A great deal of effort went into describing little-known
languages within this framework, and people took pleasure in discovering
ways in which every language was different from every other language.
Searching for properties that extended across many or all languages was
not seen at that time as a particularly interesting or rewarding way to spend
one’s time.
We who were students of linguistics in those days were sternly warned
to steer clear of a few misguided individuals who tried to relate language
to the mind, consciousness, perception, imagery, or ideas. The prediction
was that “within the next generation [this] terminology of mentalism and
animism [would] be discarded, much as we have discarded Ptolemaic
astronomy, and will be replaced . . . in major part by terms of linguistics”
(Bloomfield 1936: 89). We were required to read a book by Zellig Harris
titled Methods in Structural Linguistics (1951), which promised that we
would sooner or later uncover the underlying structure of languages if we
ignored meaning altogether and studied only the statistical distribution of
sound-based elements. We couldn’t do it just then because it required
huge corpora of linguistic material and powerful computers, but those
resources were glimpsed, realistically enough, on the not-too-distant
horizon. Before long the hidden structure of language would be revealed
in all its glory and we would no longer need to appeal to mentalistic
nonsense. Ironically, advances in technology have now brought us to
a point where Harris’s goal seems realizable, but for some of us at least
it does not offer all the answers he anticipated.
One often hears of a revolution in linguistics that was set in motion by
Syntactic Structures (Chomsky 1957) and solidified by Aspects of the Theory of
Syntax (Chomsky 1965). A revolution must be in the eye of the beholder and
my own experiences suggested something different. The belief that we could
unearth an abstract syntactic structure on which language was built, with
meanings and sounds somehow emanating from it, coupled with the belief
that a major goal of linguistics was to formulate rules that would generate such
a structure – wasn’t that an outgrowth of the distrust of meaning that was
already impeding progress? Two roads now diverged, but for me they were not
an outmoded “structuralist” road and an exciting new “generative” road. Very
different was a road that asked how language was used in the real world while
recognizing the crucial role played by meanings. It was indeed a road less
taken.
 Background 9

Most people today may be unaware of the extent to which, during the
1960s, investigating languages “in the field” and describing what one
found were scorned as leading to nothing more than “observational ade-
quacy.” At Berkeley, my home at the time, that scorn gradually sapped the
vitality and enthusiasm that had sustained the Berkeley Department of
Linguistics during the 1950s and early 1960s. Fewer talented students
were applying for admission, seeing little to be gained from pursuing an
approach that was only a historic relic. When I became chair of that
department in 1969 I was pressured by faculty in other departments to
abandon our emphasis on “just writing grammars” and to join the march
toward greater enlightenment.
I never lost my enthusiasm for what I thought linguistics could and
should be, but I was unable to transfer that enthusiasm to the perspective
that had so quickly captured the field. I was fascinated by interactions
I thought I saw between language and mental states and processes, con-
vinced that it was futile to separate language from the mind or vice versa.
Whatever linguistic phenomenon I came across raised questions about the
light it shed on how the mind worked. I was not finding answers in the
linguistics of the day.
At first I naively assumed that others shared my concerns. In 1973
I blithely initiated a course titled Language and Cognition in which the
mind–language relation was paramount. It was disappointing to discover
that few of my students were consumed by the same passion for relating
discoveries about languages to the mental insights they offered. Students
who wanted to learn about, teach about, or conduct research in phonology,
morphology, or syntax evidently did not share my curiosity regarding their
mental foundations.
It seemed natural, then, to turn to psychology for more encouragement, but
there too I was disappointed. When I was asked to review a book on psycho-
linguistics (Saporta 1961; cf. Chafe 1964), I was surprised to find that all of its
chapters were reports of experiments. I remarked on this to a psychologist
friend, who said, “Oh, but that’s what we’re supposed to do.” The experiments
seemed unduly narrow, and I was bothered by the way they so frequently relied
on concocted “data” that failed to come close to what I had found language to
really be like. It was especially disconcerting to find the word “empirical”
applied to studies that were intentionally disconnected from reality in order to
achieve control. When I tried to imagine myself a “subject” in one of those
experiments, I found it hard to believe that my responses would show very
much about what real people did under real circumstances. This isolation of
psychology from real people doing real things in the real world is something
I still find disheartening.
10 Preliminaries

The “cognitive revolution” was then being born, and with it I ventured a hope
that psychologists would loosen up a bit. It has been a further disappointment to
see their straitjacket becoming still more confining. I suppose it is not surpris-
ing that calling something cognitive science failed to revolutionize the ways
people actually did their research. We all have our favorite frameworks for
understanding what interests us, and we can’t be expected to shake off those
frameworks just because we’ve acquired a new name. Recent years have given
us impressive new techniques and technologies, which, however, continue to
be laid on top of assumptions about “best practices” that have stubbornly
resisted change.
When friends ask, as they occasionally do, what this book is about, I like to
point out that they just did two things. One was to make sounds. By itself that
was of limited interest. Lots of animals make sounds. It was what those sounds
accomplished that was so magical. As they passed through the air from my
friends to me and made their way into my brain, they let me know a little of
what was happening in their brains: in this case that they were experiencing
a mild curiosity about what I was up to, and that they were inviting me to make
noises of my own that would let them know a little of what I was thinking.
My dog makes sounds that alert me to an intruder in the neighborhood, but the
complexity of what my friends did with their sounds and what they expected
me to do with mine was greater by an order of magnitude. This ability to link
thoughts with sounds defines what language is.
To repeat something said in the Prologue, although many may be willing to
agree that language begins with thoughts in the mind of a speaker and ends by
modifying thoughts in the mind of a listener, that seemingly obvious fact has
seldom been incorporated into an understanding of language structure.
I mentioned two easily understandable reasons for this neglect. One is that
thoughts have been regarded as adequately accounted for in the study of
semantics, and so there is no need to pay attention to thoughts as if they were
something else. The other reason is that thoughts may appear too nebulous to be
incorporated within a satisfying model of language. This work suggests ways
of dealing with both those problems.
Attempting to discover the full nature of language can be compared with
climbing a mountain whose summit is occupied by thoughts but is so
enshrouded in fog that hardly anyone even considers reaching it. Remaining
in the valley where one can devote one’s full attention to sounds can allow
disagreements to be resolved in relatively objective terms. Many linguists,
however, proceed from phonetics to phonology, where they explore what
different languages do with sounds. They may stay at that level, knowing that
climbing further to syntax invites greater risks and disagreements. Many who
do venture on are likely to be satisfied with syntax as their final goal, but there
are some who struggle on to semantics, hoping for a view that will be still more
 Background 11

revealing. The air is thinner and the fog thicker and there are still more
disagreements, but the potential rewards are enticing. Hardly anyone is foolish
enough to venture still further into the realm of thought, where the air is so thin
and the fog so thick that one can easily fall victim to vertigo and hopeless
confusion. This book, nevertheless, undertakes forays in that direction, hoping
for glimpses that are otherwise unobtainable if and when the fog lifts, if only
momentarily. Because even the best-intentioned linguists have usually ended
their explorations with semantics, there is little to guide those forays. I thus beg
the reader’s tolerance for missteps as I trudge onward in a direction I am
convinced must sooner or later be pursued.
2 Ground Rules

This work makes certain assumptions and relies on certain sources of data that
are described in this chapter in the hope of circumventing misunderstandings
that might otherwise arise in the course of the chapters to follow. The aim of
this book is to add to our understanding of language and thought, but many
approaches are possible and none provides a complete or final answer. Before
we go further, it may be useful to step back and consider what it means to
understand anything at all.
I assume that understanding is derived from two essential pursuits. The first
is observing: exploiting our perceptual abilities, often supplemented by tech-
nology, to take in information from the world around us, but also from worlds
that are stored within our brains and are observable through remembering and
imagining. But observations alone can never be enough, because serious
limitations on our capacities to observe provide us with only limited tastes of
what is there. Technology may help, but only up to a point. What we observe
and how we observe it are limited by our genetic endowments, by our cultures,
and even by the circumstances of our individual lives, all of which restrict us to
observations that are not only limited but often distorted. Observations also
suffer from a particularity in time and space that falls short of the general
applicability that understanding requires. Observations alone go no further than
the observations themselves.
We have evolved to believe in the existence of larger, more encompass-
ing systems whose aim is to overcome the limitations of what we observe.
We are endowed with a burning desire to fit our observations into some
larger picture, and to that end we recruit our imaginations, inventing larger
systems within which our observations find a natural place. In science the
observations may be called “data” and the imagined larger visions
“theories,” but attempts at understanding are by no means restricted to
science and pervade all of human life.
Much of what we think and do is governed by folk theories, large and small.
On the large side are religions, political ideologies, folklore, and whatever
other belief systems help us understand broad ranges of experience. On the
small side are ways of categorizing particular experiences that arise constantly

12
 Ground Rules 13

as we think and talk and act. What I see over there is an instance of the dog
category; what I hear is an instance of barking.
Different cultures, subcultures, and even different individuals have internalized
their own folk understandings, and that is what we would expect from products of
imagination. It is typical of folk theories that their adherents believe them to be
fully in accord with reality, thus offering the security of a world within which
people can safely operate. There is often little tolerance for those who believe
differently, with arguments and even wars as a result, and there may be little
motivation to accumulate further observations against which a favorite theory
might be checked. A folk theory exploits a restricted range of observations as raw
material from which some larger reality is imaginatively constructed.
Scientific theories are often regarded as an advance over folk theories. They
are characterized by a high value placed on some mutually agreed-upon method
of data collection coupled with an agreed-upon method of theory validation.
The human sciences – those that aim to increase our understanding of human
experience – all rely on systematically accumulated data of some kind, all try to
understand those data with an imagined theory, and all have an interest in
validating that theory in the light of additional data. Beyond that, ways of
understanding human experience have differed greatly, not only in different
fields of investigation but even within the same field. It is remarkable, for
example, how many theories and subtheories of language blossomed during
the last century, a testament to the complexity and intractability of language
itself as well as of the minds language inhabits.
We might ask whether both data and theories have properties of their own
that favor some over others. So far as data are concerned, I suggest that the
highest value should be attached to observations that occur naturally and are
not themselves products of imagination. In linguistics it has been common to
rely on “grammaticality judgments” applied to snippets of invented language.
Instead of observing language in action, one speculates on whether and why
invented X is “grammatical” and invented Y is not. That procedure has
a usefulness that should not be dismissed, but inevitably it misses something
important, for observing real language as it is produced by real people in the
real world offers insights that are obtainable in no other way. One sometimes
hears that real data of that sort are accidental and thus lack the control required
by experimental methods based on invented data. Sacrificing reality for control,
however, is a devil’s bargain.
A theory can be evaluated with respect to the range of observations it
accounts for and how well it does its job, but also with the more elusive
criteria of coherence and elegance of design. There have been attempts to
establish more precise methods of theory evaluation. One of them can be
traced to the nineteenth-century physicist William Thomson, otherwise
known as Lord Kelvin, who was a prolific source of quotations, including
14 Preliminaries

the statement that “heavier than air flying machines are impossible.”
Frequently quoted has been the following:
I often say that when you can measure what you are speaking about, and express it in
numbers, you know something about it; but when you cannot measure it, when you
cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind; it
may be the beginning of knowledge, but you have scarcely in your thoughts advanced to
the state of Science, whatever the matter may be. (William Thomson 1883)
There is still a common belief that a theory is valuable only to the extent that it
provides numerical measurements, but that requirement applies more appro-
priately in some areas of investigation than in others while it overlooks other
properties that may contribute at least as much to a theory’s worth.
There is also the falsifiability criterion associated with Karl Popper (1959),
which says that a scientific statement is one that can be disproved. Formulated
in that way, this requirement assumes that a theory is a statement, which might
be mathematical or statistical as well as verbal. But if a theory originates in the
realm of thought, as this book suggests, it may or may not be representable in
linguistic form. Some of the most useful theories may in fact have originated
with imagery and neither with language nor with mathematics.
The falsifiability criterion assumes that there are two kinds of theories: those
that deserve to be labeled scientific and those that do not. We can distinguish the
two by trying to imagine an observation that would not be explained by the
theory in question. If we succeed in imagining such an observation, the theory
can be called scientific. Of course the actual existence of an incompatible
observation might lead us to reject a theory, not because it is unscientific, but
because that observation is enough to disqualify it. The theory itself might then
be revised to accommodate the problematic observation, but too many revi-
sions can destroy a theory’s usefulness and coherence. The absence of falsifia-
bility is sometimes used as a cudgel with which to destroy a theory one dislikes,
but its dependence on our ability to imagine a nonconforming observation
means that it is not always easy to apply.
Within linguistics I would argue for the importance of theories that have
a historical dimension. I say this because much of linguistics has confined itself
to synchronic explanations that overlook the history of the data for which
understanding is sought. Chapter 9 includes a complex historical explanation
of the be gonna construction as in we’re gonna go there, illustrating in detail the
value of including history as a dimension of understanding.

Introspection
We cannot go much further without abandoning an orthodoxy that I believe has
retarded the last century of progress in the human sciences. During the
 Ground Rules 15

twentieth century the list of unacceptable practices awarded first place to

introspection, which this book shamelessly retrieves as an essential resource.
The ghost of William James provides us with strong encouragement:
Introspective Observation is what we have to rely on first and foremost and always.
The word introspection need hardly be defined – it means, of course, the looking into
our own minds and reporting what we there discover. Every one agrees that we there
discover states of consciousness . . . [W]e need not anticipate our own future details,
but just state our general conclusion that introspection is difficult and fallible; and
that the difficulty is simply that of all observation of whatever kind. Something is
before us, we do our best to tell what it is, but in spite of our good will we may go
astray, and give a description more applicable to some other sort of thing. The only
safeguard is in the final consensus of our farther knowledge about the thing in
question, later views correcting earlier ones, until at last the harmony of
a consistent system is reached. (James 1890: 185, 191–192, emphasis original)
Was James misleading himself by relying on data we now know to be
hopelessly unreliable? Introspection was vehemently rejected during the
years that followed him, and by the early twentieth century psychologists
were being warned that introspection, consciousness, and in fact any interest
in mental states and processes led nowhere. In his behaviorist manifesto John
Watson proclaimed:
Psychology as the behaviorist views it is a purely objective experimental branch of
natural science. Its theoretical goal is the prediction and control of behavior.
Introspection forms no essential part of its methods, nor is the scientific value of its
data dependent upon the readiness with which they lend themselves to interpretation in
terms of consciousness . . . The time seems to have come when psychology must discard
all reference to consciousness; when it need no longer delude itself into thinking that it is
making mental states the object of observation. (Watson 1913: 158, 163)
Perhaps most psychologists today would distance themselves from so stark
a pronouncement, but skepticism regarding introspection is still ubiquitous in
psychological training and practice.
As for linguistics, during most of the first half of the twentieth century
linguistics in the United States, and to a considerable extent elsewhere, owed
its foundational beliefs to Leonard Bloomfield, who in turn owed a great deal to
his friendship with the behaviorist psychologist Albert Paul Weiss. Partly
because of Weiss, but also because both linguists and psychologists believed
they should emulate the physical sciences, Bloomfield became obsessed with
the need to avoid what he repeatedly castigated as “mentalism.” Statements like
the following are scattered throughout Charles Hockett’s anthology of
Bloomfield’s writings:
Linguistics as actually practised employs only such terms as are translatable into the
language of physical and biological science; in this linguistics differs from nearly all
16 Preliminaries

other discussions of human affairs. Within the next generations mankind will learn that
only such terms are usable in any science. The terminology in which at present we try to
speak of human affairs – the terminology of “consciousness,” “mind,” “perception,”
“ideas,” and so on – in sum, the terminology of mentalism and animism – will be
discarded . . . and will be replaced in minor part by physiological terms and in major part
by terms of linguistics. (Bloomfield 1936 in Hockett 1970: 322)
An insistence that linguistics be recognized as a science was frequently
repeated during the period of Bloomfield’s dominance. It has been interesting
to see psychology following the same path decades later. In 1988 the venerable
American Psychological Association witnessed the defection of the American
Psychological Society, which was soon rechristened the Association for
Psychological Science, the words “psychological science” becoming
a proudly proclaimed self-designation. For a student of linguistic history it is
a development with a familiar ring.
There can be no doubt that introspection is replete with problems that have
led so many to be skeptical of its value. It is by definition internal and hidden
from public view. How, then, can it lead to an understanding that has general
applicability? It suffers, too, from a describability problem. How can we even
discuss an example of introspection unless we are allowed to describe it with
language, often a seemingly impossible task and one that in any case departs
from the introspection itself by imposing on it a linguistic organization? Then,
too, there may be a disturbing lack of agreement across individuals.
My introspection may not be yours. And if measurability is important, how
can introspections be measured? Without wishing to downplay such problems,
this work suggests ways of ameliorating them by supplementing introspection
with observations that are more open to public view.

Consciousness
Introspection is often paired with consciousness, as it was in the quotes from
James, Watson, and Bloomfield above. The pairing is easy to understand.
Consciousness is also a private experience hidden from public view and direct
access to it can only come through introspection. Discussions of consciousness
sometimes focus on its phenomenological (i.e. introspective) manifestations
and sometimes on its representation in the brain, both worthy endeavors. This
work views consciousness as an introspectively recognizable mental state
whose properties can help us better understand a variety of phenomena,
linguistic and otherwise.
Although linguists today may not be of one mind on this issue and may
not even have considered it, one notices an undercurrent of discomfort when
consciousness is mentioned. Reactions to my book Discourse, Consciousness,
and Time (Chafe 1994) sometimes treated the word as an embarrassment and
 Ground Rules 17

tried to circumvent it with other terminology. In a book titled Language

Structure, Discourse and the Access to Consciousness (Stamenov 1997: 73)
Ronald Langacker ended his contribution by writing, “it should be evident that
grammar is shaped as much by what we are not consciously aware of as by what
we are.” A more extreme position was taken by George Lakoff and Mark
Johnson:
Conscious thought is the tip of an enormous iceberg. It is the rule of thumb among
cognitive scientists that unconscious thought is 95 percent of all thought – and that may
be a serious underestimate. Moreover, the 95 percent below the surface of conscious
awareness shapes and structures all conscious thought. If the cognitive unconscious
were not there doing this shaping, there could be no conscious thought. (Lakoff and
Johnson 1999: 13)
Percentages carry an aura of authority, and studying phenomena that are
opaque to consciousness suggests an impressive intellectual achievement. But
the philosopher Tim Adamson responded to the above with well-placed
skepticism:
In sum, it is unclear how Lakoff and Johnson justify their claim that the basic structures
of cognition are “utterly and ineliminably inaccessible to direct awareness.” As for an
explicit, clearly-formed argument, Lakoff and Johnson do not provide one. They state
that this claim is the result of much research in cognitive science, but it is not clear which
empirical research has shown that we cannot reflect on and become directly aware of the
structures that have been guiding our experience. (Adamson 2004: 104)
It does not at all follow that the conscious flow of thought is independent of
unconscious forces that affect its content and direction. One thinks of course of
the Freudian tradition that our mental life is dominated by the unconscious.
Without debating the extent to which that might or might not be true, we can at
least admit that the flow of conscious thought may be vulnerable to and
influenced by unconscious forces. But that is very different from saying that
the flow of thought is itself unconscious. We can profit again from William
James, who wrote, “I use the word thinking . . . for every form of consciousness
indiscriminately” (James 1890: 224), thus asserting the parallelism of con-
sciousness and thought, which share at least the following properties (cf. James
1890: 225):
(1) They are experienced by a unique self.
(2) They are continuous.
(3) They are restless.
(4) They have a focus and a periphery.
(5) Their focus may lie either outside or inside the mind of the experiencer.
First, the observation that consciousness and thought are restricted to
a particular self is basic to language, which functions to bridge the divide
18 Preliminaries

between one consciousness and another – one self and another – by pairing
thoughts with sounds in ways explored in this work. Second, consciousness
and thought are experienced continuously throughout all our waking hours.
Consciousness defines what wakefulness is. Third, consciousness and
thought are unavoidably restless, reflecting the constantly changing kaleido-
scope of our lives. In a section titled “Thought is in Constant Change,” James
wrote of the “different great classes of our conscious states. Now we are
seeing, now hearing; now reasoning, now willing; now recollecting, now
expecting, now loving, now hating; and in a hundred other ways we know our
minds to be alternately engaged” (James 1890: 230). This restlessness is
difficult if not impossible to suppress, although meditation may see that as
a goal. Fourth, it may be no accident that consciousness resembles vision in
its possession of a focus and periphery, the focus being verbalized in phrases
that show limitations on how much we can focus on at one time. Fifth and
finally, the focus of consciousness may have its target in the world outside
the thinker, where it is available through perception, or it may lie within the
thinker’s own mind, where it is available through remembering and imagin-
ing. Consciousness is the mind’s way of activating small selections of either
external or internal knowledge. It does that with a moving focus that scans
a real or invented world to create its own facsimile of whatever is of interest
at the moment.

The Value of Linguistic Diversity

At certain points in this work I introduce material from one or the other of two
Native American languages with which I have some acquaintance. One of
them, known in English as Seneca (Chafe 2015), is still spoken by a small
number of people in western New York State. It is a member of the northern
branch of the Iroquoian language family, where its closest relatives are
Mohawk, Oneida, Onondaga, Cayuga, Tuscarora, Wendat (Huron), and
a number of other languages that fell out of active use by the end of the
seventeenth century as a result of the European invasion. It exemplifies
a language type that has been labeled “polysynthetic” (Duponceau 1838;
Fortescue, Mithun, and Evans 2017), or sometimes “holophrastic” (Lieber
1837). Polysynthetic languages package within a single word various elements
that other languages would distribute across several words. How the Seneca
language does that is explored in Chapter 10.
I also introduce some material from another polysynthetic language, which
is known in English as Caddo (Melnar 2004), once spoken in a huge area of
what subsequently became eastern Texas, northern Louisiana, southern
Arkansas, and southeastern Oklahoma. In the history of European colonization
the Caddos had the misfortune to be caught between Spanish, French, and
 Ground Rules 19

English invaders in a three-way vice. A small number of descendants have

lived in western Oklahoma since 1859. The Caddo language is a member of the
eponymous Caddoan language family, which also includes Wichita, Kitsai,
Pawnee, and Arikara. Presently it has few if any active speakers. A conspicuous
legacy of this language is the name “Texas”: first a Spanish and then an English
rendering of the Caddo word for “friend” or “ally,” a word that members of the
Caddo confederacy applied to each other to distinguish themselves from the
hostile tribes that surrounded them.
These two languages are selections from a huge but rapidly fading resource,
some of which is still available to add in important ways to our understanding
of language and thought. Before the end of the current century the number of
the world’s languages that are still actively spoken will have dwindled to
a small fraction of what was once present. It is not being overly dramatic to
say that the results will be tragic from several points of view. Each language
constitutes a vast, complex repository of knowledge and ways of thinking that
have defined the culture of its speakers, leaving a vast emptiness when the
language is gone. Each language provides its own windows into the varied
ways the mind has succeeded in organizing and expressing thoughts, a resource
that is seriously eroded each time a language passes out of active use.
The currently vibrant field of language documentation is preserving some of
that richness, but nothing can substitute for the presence of people who are still
actively using a language.
I also make reference occasionally to the so-called Pear Film, a seven-minute
film showing a boy’s theft of pears and its aftermath (Chafe 1980). Since 1975
it has been shown in many countries where viewers have been asked to tell what
they remember of it as a way of exploring, through these Pear Stories, how
a similar experience is remembered and verbalized in different languages.
In Chapter 8 one of these stories in English is used as evidence for the
hierarchical structure of memory.

Summary
This chapter has established some ground rules for understanding the relation
between thoughts and their expression in sound. Understanding was seen as
derived from an interplay between observations and invented theories.
In evaluating data a preference was expressed for observations that occur
naturally. With regard to theory evaluation, mention was made of the popular
criteria of measurability and falsifiability, and a preference was expressed for
historical explanations. Introspection was accepted as an essential observa-
tional resource that is subject to strengthening by observations that are more
public. Similar recognition was given to consciousness, understood as the
mind’s way of activating selections from external, remembered, or imagined
20 Preliminaries

worlds. The flow of thought and the flow of consciousness were seen as
coextensive. Finally, stress was placed on the crucial importance of linguistic
diversity, now being eroded by an accelerating loss of the world’s languages.
Material from two Native American languages will illustrate that point in this
book.
One last point. Just as members of the population at large vary greatly in
personality type, the same is true of the subset of the population engaged in
scholarly research. Personality differences can significantly affect modes of
research and receptivity to research findings. Commitment, for example, to
qualitative or quantitative investigation, a willingness or unwillingness to
admit introspective evidence, and other biases toward or against particular
avenues of research depend to a considerable extent on personalities, however
much one might want to pretend that they do not. In the end we should accept
that fact and take it into account when passing judgment on what others have
done.
Part II

Thoughts and Their Properties

3 The Priority of Thoughts

In 1976 I joined a friend and fellow linguist, Hanni Woodbury, on a visit to the
Onondaga Indian Reservation south of Syracuse, New York. The Onondaga
language was still being spoken there by a small number of people, and in 1968
I had spent time with several of them, above all with Jessie Pierce and Delia
Carpenter. Later I published a short description of their language in a work
titled “A Semantically Based Sketch of Onondaga” (Chafe 1970b). I was trying
to make a case for an approach to language in which meaning rather than syntax
was the force behind language structure. I explored that approach in a more
detailed way around the same time in Meaning and the Structure of Language
(Chafe 1970a).
Hanni and I paid a visit to Harry and Lotte Webster, who had been her helpers
in assembling materials for her Yale University dissertation on noun incorpora-
tion (Woodbury 1975). One purpose of our visit was to see whether we could
confirm what we thought might be an interesting feature of Onondaga mor-
phological prosody: the distribution of pitch, loudness, and timing within
Onondaga words. It was a feature that distinguished Onondaga from a pattern
hypothesized for Proto-Northern-Iroquoian, the reconstructed ancestor of
Onondaga and related languages. That earlier pattern was characterized by
a prominence assigned to the next-to-last syllable of a word, a prominence that
was realized in higher pitch often accompanied by greater loudness and by
lengthening the vowel of that syllable. What was interesting was that under
certain conditions Onondaga had added a second prominence to the third
syllable from the end of a word, the one preceding the syllable that still retained
the prominence of the proto-language. It can be illustrated with a word pro-
nounced onų́ :dó:da’.1 The development from the reconstructed proto-language
(shown with an asterisk) to the modern Onondaga word can be represented as:
*onųdó:da’ → onų́ :dó:da’

1
The hook under the vowel ų shows that it is nasalized, the colons show lengthening of the
preceding vowels, and the acute accent marks show raised pitch. The final apostrophe represents
a glottal stop.

23
24 Thoughts and Their Properties

nu
o do

da’

Figure 3.1 Fundamental frequency, intensity, and length in an Onondaga

word

On the basis of our unaided “impressionistic” auditory perceptions,

Hanni and I speculated that the third syllable from the end might be
pronounced on a higher pitch than any of the others while the second
syllable from the end might be louder than any of the others. In other
words we suspected that the original prominence on the next-to-last sylla-
ble might still be retained with greater loudness while the preceding
syllable had not only become longer but was also now the highest in
pitch. This separation of pitch prominence from loudness prominence
would be interesting if it were true, and we wanted to test its validity by
listening carefully to the Websters and recording what they said.
When easier and more accurate ways of visualizing sounds became avail-
able we were able to check our guesses against more objective acoustic
displays. We had been partly wrong. Figure 3.1 shows the pitch contour
(above) and the loudness contour (below) in one sample of this word, and it
can be seen that the third syllable from the end was not only higher in pitch
but also louder than the next-to-last syllable. The latter still retained the
prominence it had inherited from the proto-language, but the syllable before
it was even more prominent in both pitch and loudness, at least with respect
to these physical measurements.2
Here we see the value of representing sounds visually and how this ability
may correct initial impressions, but I described our experience for quite another

2
The displays actually show fundamental frequency and intensity, whose relation to the
perception of pitch and loudness may be indirect, but that is not the point at issue here.
 The Priority of Thoughts 25

reason. We wanted to hear this word repeated a number of times so we could

test our hypothesis with careful listening. But each time we asked Harry or
Lotte to repeat it they showed no interest at all in what the word sounded like.
Instead, they kept trying to explain to us what it meant and how it was used.
Harry would say, for example, “It’s a hill or a mountain. Like you’re telling
a story and you start by telling, ‘Over there, there’s a hill.’” And Lotte added,
“For instance, if you’re riding in a car, you look way out in the distance and you
see that onų́ :dó:da’.” While Hanni and I were preoccupied with the word as
a sound, Harry and Lotte were only concerned with what they were thinking
when they used it or heard someone else use it.
This experience was hardly unique. I and others have often had similar
encounters in our work with lesser-known languages. For the people who
speak them, language is a way of organizing and communicating thoughts,
not a way of making sounds. That observation helps to explain, among
other things, why it is so difficult to teach fluent speakers of a previously
unwritten language how to write things they have said all their lives but
have never before associated with letters. Faced with a writing system
whose letters represent sounds, people like the Websters find that divert-
ing their attention from what they were thinking to the sounds they were
making demands an unfamiliar, awkward, and often difficult cognitive
effort.
Years ago the Swedish linguist Per Linell published a book titled
The Written Language Bias in Linguistics (Linell 1982), and it is worth
recalling his insistence on the extent to which linguistics has from the
beginning relied on written data. The very term “grammar” is related to
the Greek for “letter.” From the earliest grammarians to the present,
linguistics has paid unbalanced attention to language as it is written.
Language in written form is, of course, what you see in front of you at
this moment, a physical object that is easy to observe and that lends itself
to analysis. But the heart of language beat for countless millennia before
its pulses were recorded in writing. Speaking, not writing, has always
been the dominant use of language for most people.
But behind this written language bias lies another bias that is equally
entrenched and if anything more insidious. If the function of language is
to associate thoughts with sounds, the bulk of linguistic research has tilted
toward the sounds. Linguistics has been predominantly sound-based and
not thought-based. That is not to say that all linguists are phoneticians, but
rather that linguists pay more attention more of the time to elements of
language that are ways of organizing sounds, if not letters. Postulated
components of language like sentences, words, prefixes, suffixes, and the
rest have tilted toward how they sound or are written, not toward the
thoughts they organize and communicate. Probably few would deny today
26 Thoughts and Their Properties

that a word “has a meaning,” but putting it that way implies that its
meaning is something attached to it – something it has, not something
it is.
In the comic strip Pickles of April 5, 2009 the protagonist, Earl, while
playing checkers with his friend Clyde, keeps repeating the word plinth. He
explains, “It’s my word of the day. Every day I learn a new word. I read an
article that said learning a new word every day will help your mind stay sharp.”
“So,” asks Clyde, “what does plinth mean?”
“Oh, I have no idea. The article didn’t say anything about learning what they
mean.”
If it seems wrong to divorce a word so completely from its meaning, the
priority usually given to sounds lends a pseudo-plausibility to this conceit, and
hence the humor (cf. Chafe 2007). Linguists often seem to have more in
common with Earl than with the Websters. The reasons why a sound-based
perspective has so dominated the study of language are easy to appreciate and
we can return to them. But if thoughts are where language begins and ends, it is
time to reward them with a deeper appreciation and a closer attention to their
nature and role.

Figure 3.2 What Earl didn’t know

4 The Path from a Thought to a Sound

Most of us experience both thoughts and sounds constantly as we live our

daily lives. We spend most of our lives thinking, although of course during
much of that time we are doing other things as well. Often we pair what we
are thinking with sounds. Sometimes we utter those sounds out loud, but
much of the time we experience them silently. Our thoughts themselves are
private, but when the sounds become audible they may allow another person
to gain some understanding of what we are thinking. Pairing thoughts with
sounds has turned out to be a wonderfully effective way of communicating
our thoughts to others, but a second benefit has been the fact that it gives an
organization to our thoughts, whether we try to make them known to others
or keep them to ourselves.
Overt language begins with the thoughts of a speaker and ends (or so the
speaker may hope) by modifying a listener’s thoughts to include at least
some recognition of what the speaker is thinking. If the thoughts are where
language begins and ends, they have an obvious priority over the sounds,
which function in service to those thoughts. But if thoughts enjoy priority
over sounds in that sense, there is an imbalance in the opposite direction
when it comes to the research tractability of these two phenomena. Sounds
are publicly observable, and with modern equipment we can record, edit, and
analyze them in satisfying ways. We know a lot about their physical proper-
ties and even something about how our minds process them. Thoughts are
something else again. Despite their natural priority, despite the fact that we
experience them constantly during all our waking hours and even sometimes
while we are asleep, and despite the fact that they underlie everything we do,
thoughts remain elusive and mysterious. If we spend our lives thinking, what
exactly are we doing?
Those who practice linguistics may ask why they should worry about such
things. Isn’t it enough to devote our research to language itself and leave
thoughts to poets or philosophers or whoever feels more comfortable with
vague speculations? The simple and compelling answer is that language con-
sists of much more than organized sounds, and organizing thoughts and pairing
them with sounds defines what language is.

27
28 Thoughts and Their Properties

thoughts

sounds

Figure 4.1 Language in a nutshell

Figure 4.1 summarizes language at this most basic level. If this is what
language is and does, a major task of linguistics should be to investigate the
arrow in the middle to discover how this thought-to-sound pairing is possible,
for thoughts and sounds themselves have little in common aside from the fact
that both flow through time. One can validly object that this figure ignores
a listener’s task of moving from sounds back to thoughts, but understanding
language depends on having something to understand in the first place.
The direction of the arrow in Figure 4.1 thus comes first and is the major
concern of this work. One can also allow for the replacement of sounds with
visible signs, and sign languages require us to replace various details in
appropriate ways. But that is not our task here, where the goal is to understand
as much as we can about thoughts while investigating how their pairing with
sounds is achieved.
Because the thought–sound association is such a pervasive aspect of daily
experience, it can be surprisingly difficult for anyone who is old enough to
speak to separate the two. From time to time, however, many of us do
experience their separation with disturbing clarity in the “tip-of-the-tongue”
phenomenon (Brown and McNeill 1966; Schwartz 2002). During that mildly
stressful experience we are fully conscious of a thought and everything asso-
ciated with it except for the sound that would otherwise accompany it. For
a certain period of time our access to the sound is inhibited. The thought–sound
association is thus a fragile one.
Quite unexpectedly I recently had occasion to think of a well-known film
director whose name, despite its obvious familiarity, was briefly inhibited.
During that time I was conscious of much about him: that he had directed
many films including Spellbound, Rear Window, Psycho, and others, that he
once had a television show, and that he was British in origin. I even had a visual
image of him that is suggested by the sketch in Figure 4.2, as well as an auditory
image that captured his voice quality. In other words I was conscious of
everything about him except his name. It wasn’t long before the sound Alfred
Hitchcock found its way back into my consciousness, but while it was absent
the thought–sound association lacked its bottom half, as suggested in
Figure 4.3.
Separating a thought from a sound can be experienced in the opposite
direction in rote learning: consciousness of the sound of a poem or song, for
 The Path from a Thought to a Sound 29

Figure 4.2 Visual imagery associated with a thought

thought

sound?

Figure 4.3 The tip-of-the-tongue experience

example, that is disconnected from whatever thought might have been asso-
ciated with it. As a child I learned to sing “America the Beautiful” without
giving thought to a “fruited plain,” whatever that might have been. The sound
was there, but it lacked a thought (Figure 4.4).
I have sometimes had a similar experience while recording samples of
a language with which I was only partially familiar. I found myself recognizing
a word I was sure I knew, at least up to a point, but recognizing it as nothing
more than a sound. This experience can be instructive for someone learning
a new language. The biggest step toward that goal is learning to cross the
thought–sound divide in both directions immediately, spontaneously, and
effortlessly, with no need to fall back on a memorized vocabulary list. Such
an immediate bidirectional connection is what fluent speakers of a language
experience all the time.
30 Thoughts and Their Properties

thought?

sound

Figure 4.4 Rote learning

thought sound

semantic overt
structure phonology

syntactic abstract
structure symbolization phonology

Figure 4.5 The verbalization path

Stages in the Verbalization Path

The word “verbalization” is used throughout this work as a way of referring
to the path that leads from a thought to a sound. Its stages are outlined in
Figure 4.5, the thought-based portion on the left being symbolized by the
sound-based portion on the right. The process is set in motion by a thought.
A language then filters and adjusts that thought by selecting from the
semantic resources the language provides. The resulting semantic structure
preserves a direct relation to the thought, but it may be only one of several
ways in which that thought might have been adjusted to the language.
The semantic structuring of thought is necessary because, as described
further in Chapter 9, the open-ended variety of thoughts could not possibly
be paired with sounds unless there were a way of reducing thoughts to
something manageable, as, for example, when the thought of some object
is interpreted as an instance of a semantic category that applies to any
number of particular thoughts. Syntactic structuring is not “necessary” in
the same sense, but results in part from semantic changes that fail to be
accompanied by parallel changes in sound, leaving behind an absence of
direct thought–sound pairing. Syntactic structures may depart from a direct
relation to thought in any number of ways.
In the bottom tier of Figure 4.5 a syntactic structure crosses the
thought–sound divide to be symbolized by an abstract phonological struc-
ture. Language change is then again at work in separating abstract from
overt phonology. The abstract sequence will not, for example, may lead to
overt won’t, which then provides the input to sound in the top right. In this
 The Path from a Thought to a Sound 31

case will not is sometimes represented more directly by sound in a more

formal style of speaking, but in Chapter 10 we will meet a language where
many abstract phonological structures are not directly represented by
sound at all.
Why should verbalization be represented with this U-shaped path instead of
a straight line? Figure 4.5 captures a symmetry between the two sides. In the
middle tier there is a resemblance between semantic structure and overt
phonology in the sense that both are directly related to thought and to sound
respectively. In the bottom tier syntactic structure and abstract phonology are
more distantly related to thought and sound. In Chapter 9 there is a suggestion
that prosody and gesture may bypass the bottom tier altogether, symbolizing
semantic elements directly without passing through the more abstract stages of
syntax and abstract phonology.
Figure 4.5 calls attention to the several stages of verbalization, but these
stages should not be understood as activated in sequence every time
a speaker speaks. They are not, in other words, temporally ordered stages
of linguistic “performance.” Their ordering, nevertheless, does capture
certain logical and historical priorities. A syntactic structure, for example,
assumes a basis in a semantic structure, while it is itself a necessary basis
for a phonological representation. But later stages may feed back into
earlier ones, and in actual performance all these stages may be activated
simultaneously.

Another View
The view of language design that is represented in Figure 4.5, motivated in
ways illustrated in chapters to come, contrasts with an alternative that is
sketched in Figure 4.6, where syntax is the driving force behind both semantics
and phonology. Some version of this view quickly achieved popularity in much
of linguistics during the second half of the twentieth century. In a confusing
variety of manifestations it has since then been accepted in some quarters as
a foundation for linguistic theorizing.
But language cannot be built realistically on this foundation. Those
who have been pursuing linguistics in other ways have backed away
from what has been characterized as “syntactocentrism” (Jackendoff
2007: 35–38). Alternatives are often brought together under the term
“functionalism” (e.g. Bischoff and Jany 2013), which covers a range of
approaches. The present work suggests ways in which a thought-based
perspective can add to those efforts by showing how adding thought as
the starting point of language design can enrich our understanding of
what language is and does.
32 Thoughts and Their Properties

semantic structure

syntactic structure

phonological structure

Figure 4.6 A syntax-based view of language

Summary
This chapter outlined the essentials of a thought-based approach by identi-
fying stages that intervene between a thought and its association with
a sound. This verbalization path begins by interpreting a thought in accor-
dance with a language’s semantic resources, the resulting semantic structure
then being subject to syntactic patterning before being symbolized, first by
abstract and then by overt phonological structures that lead finally to
a sound. This view of the verbalization path will be assumed throughout
the rest of this work.
5 How Thoughts Are Structured

This chapter identifies aspects of thought structure that are shared by all
languages regardless of whatever structure may be imposed by the semantic
resources of a particular language. We can begin by recognizing a fundamental
dichotomy between the content of a thought and whatever affect (emotion,
mood, attitude) may accompany that content. Presumably content is processed
above all in the neocortex while affect is processed in subcortical areas. Both
aspects of thought are represented constantly in language, content above all in
“segmental” sounds and affect above all in prosody (see Chapter 8). This
chapter is concerned with content, where we can recognize a distinction
between ideas, the basic building blocks of thought, and the way those ideas
are oriented. The distinction between ideas and their orientations will be
fundamental to everything that follows in this book.
What all languages share suggests that ideas can be assigned to three basic
types that are presumably derived from three varieties of human experience.
These three idea types are either “wired into” every human brain or reflect
universal human experiences, but those two sources need not be in conflict.
We all have similar bodies and similar basic needs, we all inhabit the same
planet, and we all share fundamental aspects of our lives like eating and
sleeping. Innate wiring and common experiences must both have been at
work as our brains have adapted genetically and practically to universal
features of the human condition.
Language universals, then, suggest that we experience ideas of events (things
that happen) and ideas of states (the way things are). An event is a dynamic
experience involving some kind of change. It may be an action someone
performs (he ran away) or a change in the state of someone or something (he
got sick, it disappeared). A state is the way something is, a more stable
condition (he’s sick, it’s missing). Ideas of events and ideas of states are almost
always accompanied by a third kind of idea that includes people, objects, and
abstractions that initiate, are affected by, or otherwise participate in them. Ideas
of this third type are sometimes called “referents” or simply “things.” Here they
will be called ideas of “entities.” In the above examples ideas of entities were
expressed by the pronouns he and it.

33
34 Thoughts and Their Properties

Because events and states share certain properties that contrast them with
entities, it would be useful to have a term that combined the two, but
unfortunately the English language does not offer such a term. “State of
affairs” or “situation” are hardly appropriate terms for an event, and perhaps
we can do no better than to settle on simply “event or state” or more briefly
“event/state.” It is awkward also to refer constantly to event ideas, state
ideas, and entity ideas, so “events,” “states,” and “entities” will be used with
the understanding that they refer, not to events, states, and entities them-
selves, but to types of ideas.
Events, states, and entities can be imagined as situated within
a multidimensional matrix of orientations in space, time, epistemology, emo-
tions, social interactions, and the ever-changing context. Every language has its
own ways of verbalizing ideas and their orientations: ideas from an inventory
of verbs, nouns, and perhaps adjectives, orientations with inflections of verbs
and nouns or with separate particles. Language-specific ways of verbalizing
ideas and their orientations constitute the unique semantic resources of each
language.
Distinguishing the verbs that express events from the nouns that express
entities appears to be universal in the languages of the world. Its universality
has occasionally been questioned, for example with respect to languages
spoken in the northwestern part of North America belonging to the Wakashan
and Salishan language families. William Jacobsen (1979) surveyed this issue
for Wakashan, Van Eijk and Hess (1986) for Salishan, and both studies con-
cluded that the languages of the two families do make such a distinction. From
a thought-based perspective it is difficult to imagine that speakers of
a Wakashan or Salishan language would be incapable of distinguishing an
event from an entity.
It is true that in many languages the same sound is sometimes used for both.
Examples in English are words like “cut,” “water,” “hope,” and many others.
English speakers know which they are thinking of, and probably that is true of
the speakers of all languages. There are a few experiences, like lightning,
whose status may be uncertain, but they only highlight the near universality
of the distinction.
Several linguists have noticed that it is impossible to experience the idea of
a particular event or state without including the idea of a particular entity.3
In contrast, it is possible to experience the idea of a particular entity without
assigning it to a particular event or state. As noted by Givón, “a verb-coded
event (‘break’, ‘walk’, ‘talk to’, ‘give’, etc.) cannot be experienced – makes no

3
Exceptions may be “ambient” events like those verbalized in English with phrases like it’s
raining where the word it fills a syntactically imposed requirement in English that a verb needs
a subject.
 How Thoughts Are Structured 35

sense – independently of its noun-coded participants” (Givón 2001: 53).

Ronald Langacker also noticed that “an event is conceptually dependent; it
cannot be conceptualized without conceptualizing the participants who interact
to constitute it,” whereas “an object is conceptually autonomous, in the sense
that we can conceptualize it independently of its participation in any event”
(Langacker 2008: 1004). William Croft made the same observation: “For
example, one cannot conceive of an action such as running without the invol-
vement of a runner, or of a property such as height without something that is
tall. On the other hand, one can conceive of a chair or a dog without the
involvement of another concept” (Croft 2001: 87).
As discussed further in Chapter 8, as thoughts flow through consciousness
the focus of consciousness changes rapidly, its transient quality reflected in an
obvious way in language, where the idea of an event or state is typically
expressed for only a second or two before the speaker moves on. We can
observe this movement in the following example, which includes event and
state designations and time measurements.
(1) For one thing I had uh, state 1.0
(2) such a tiny apartment. state 0.9
(3) Um, filler 0.5
(4) when we . . . went back to New York, event 1.9
(5) um .. I always kept a studio there. state 1.6
(6) And .. of course that was small, state 1.6
(7) and then, link 0.6
(8) we got another studio.. in the, event 1.7
(9) in .. our neighborhood that we líked, state 1.6
(10) in the Village, state 0.6
(11) and we moved into that, event 1.0
(12) and we just got so busy, event 1.1
(13) we didn’t get around to- event 1.0
(14) moving into a real apartment, event 1.3
(15) until this past year. state 1.2
Unlike events, entities may persist for a longer time in both thought and
language. In this example the idea of the speaker herself, either alone or
including her husband, was expressed as I or we and was present in eight of
the fifteen phrases. The idea of the first apartment was present in three (a tiny
apartment, a studio, and that), while the idea of New York was present in two
(New York and there), as was the idea of the second apartment (another
studio and that) as well as the idea of Greenwich Village (our neighborhood
and the Village). The event expressed with the words went back in line 4 is
inconceivable without the inclusion of the speaker, her husband, and
New York. But she could certainly think of herself, her husband, or
New York as participants in any number of events. Events, then, differ
36 Thoughts and Their Properties

from entities in two ways: the transience in thought of events compared with
the frequent persistence of entities and the fact that events necessarily
include entities but not vice versa.
States have properties of their own. They share with events the necessary
inclusion of entities, but their time span is not, as it is with events, indepen-
dent of the time span of those entities. The time span of a state typically
mirrors that of the entity with which it is associated. In this example the state
of being small must have characterized the New York apartment whenever it
was thought of or mentioned. The brief time span of an event is independent
of the persistence of an included entity, but that is not true of the time span of
a state.

How Entities Relate to Events and States

Having seen how thoughts include ideas of events and states and the entities
that participate in them, it remains to consider just how entities may participate
in those events and states, or what have been called their “participant roles” (cf.
Mithun and Chafe 1999). We can focus here on “core” roles, as opposed to
“oblique” roles that are less central to an event or state. Core roles are required
by a particular event or state whereas oblique roles are optional. The core roles
in English are those expressed syntactically as subjects and objects, while
oblique roles are typically expressed with prepositional phrases. Also impor-
tant here is the distinction between intransitive clauses with only one core role
and transitive clauses with two.

Subjects and Objects

In languages like English the single core participant of an intransitive clause
plays the role of subject and the two core roles of a transitive clause are
subject and object. These names refer to syntactic roles and we can ask
whether they have a semantic basis. A subject is sometimes said to function
semantically as a “topic” or “what a clause is about.” An early discussion of
such proposals can be found in Jespersen (1924: 146). Lambrecht (1994)
provides a more recent discussion of topicality, aboutness, and subjecthood.
In Chafe (1994: 82–92) the metaphor of a “starting point” captured the
observation that a subject establishes a point of departure for whatever is
included in the rest of a clause.
Subjects exhibit properties that might be expected of starting points.
Most subjects, for example, are given (judged to be already active in
a listener’s consciousness), a smaller proportion are at least accessible (judged
to be semi-active for a listener), and only a few are new (judged to be inactive
for a listener), and then only under limited circumstances. In a sample of
 How Thoughts Are Structured 37

conversational language discussed in Chafe (1994: 85–91), 81 percent of the

subjects were given, 16 percent were accessible, and only three percent were
new. An entity that is already in active consciousness makes the best point of
departure for adding something more. Subjects are also usually identifiable
(or “definite”). When there is a choice, people typically select as subjects
first persons over second or third, humans over non-humans, and animates
over inanimates. People tend to start from their own point of view or from the
point of view of someone with whom they can easily empathize. Agents (see
below) are also more natural starting points than patients. No one of these
properties alone determines what will be a subject, but each contributes to
that role.
The following excerpt from a family conversation illustrates subject use.
The speaker was talking about a windy day of skiing at an area in the Sierra
Nevada mountains of California called Boreal Ridge. The subjects are
italicized.
(1) Doug and I went skìing at Boreal,
(2) It was windy that day,
(3) when we get to the top of the chairlift,
(4) and it was like you were a clapper on a bell you know,
(5) when you go to get off,
(6) a huge gust came by,
(7) just as we were landing,
(8) we were gonna like,
(9) you know,
(10) we were coming ìn like this,
(11) and you know,
(12) so we’d wait to,
(13) All right now!
(14) We jumped off.
(15) It was great.
The starting point throughout most of this excerpt was the idea of the speaker
and his brother Doug, introduced first with the full noun phrase Doug and I and
then continued with the pronoun we. This idea was accessible in line 1 on the
basis of what had been said earlier in the conversation; in the rest of the excerpt
it was given. Other starting points made brief appearances. The word it appears
as the subject of being windy in 2 and of was in 4. Generalized you appeared in
4 and 5, and in 15 there was an it that summarized the whole experience.
Another you appeared as the subject of the formulaic you know in 9 and 11.
The subject of 6 is of special interest because a huge gust expressed new
information that was introduced with the “presentative” predicate came by
(Chafe 1994: 111–113). A new subject like this usually expresses subsidiary
information that fails to be activated again (Chafe 1994: 90–91).
38 Thoughts and Their Properties

Subjects may be expressed with “nominative” case marking on nouns and

pronouns or else with no case marking at all. In languages that have a more or
less fixed word order the orders in which the subject precedes the object (SOV,
SVO, VSO) predominate over orders in which the subject follows the object,
another property that might be expected of a starting point.
The semantic role of object is mentioned below.

Ergatives and Absolutives

A different way of assigning participant roles can be illustrated with examples
from Central Alaskan Yup’ik, a language of the Eskimo-Aleut family that is
spoken in southwestern Alaska (Jacobson 1995). In Yup’ik both the marking of
case on nouns and the pronominal suffixes on verbs show the kind of patterning
that has been termed ergative–absolutive. A survey of views regarding this
pattern is available in Frans Plank (1979). The following examples depart from
colloquial usage but they serve the purpose of illustrating the
ergative–absolutive pattern.
With an intransitive clause the single argument is in the absolutive case:
(1) Arnaq ikayuutuq.
arnaq ikayuu-tu-q
woman (absolutive) help-INTRANSITIVE.INDICATIVE-3rd.SINGULAR
“The woman helped.”
With a transitive clause, what would be a subject in English is in the ergative
case and the other core argument is absolutive:
(2) Angutem arnaq ikayuraa.
angutem arnaq ikayur-a-aa
man (ergative) woman (absolutive) help-TRANSITIVE.INDICATIVE-3rd.
SINGULAR/3rd.SINGULAR
“The man helped the woman.”
What are the semantic roles that underlie absolutive and ergative partici-
pants? An answer suggested by several linguists, including Aleksandr Kibrik,
Edward Keenan, and Marianne Mithun, is illustrated by these Yup’ik examples.
The ergative–absolutive pattern is dominated by the absolutive, the role that is
usually unmarked formally and is obligatorily present in all clauses. Kibrik
(1979: 66) characterized the absolutive as the “closest participant in the
situation . . . an actant [i.e. core participant] who directly takes part in it.”
Keenan (1984: 200–205) characterized it in terms of “bondedness to the
verb.” Mithun (1994: 255) called it the “participant most immediately or
directly involved in the event or state.”
 How Thoughts Are Structured 39

Why should the woman in 1 be assigned the same role as the woman
in 2? What they have in common is suggested by a pair of English
sentences:
(3) The glass broke.
(4) Henry broke the glass.
In both 3 and 4 the glass can be seen as the most involved participant in
the breaking event as opposed to Henry in 4. If this were an
ergative–absolutive language the glass would play the absolutive role
and Henry the ergative. Keenan (1984: 201) noted that the meanings of
verbs often depend on the nature of the absolutive nouns that are asso-
ciated with them. For example, the meaning assigned to the intransitive
verb run varies according to who or what is running (the absolutive
participant) in the following examples:
(1) John is still running.
(2) My watch is still running.
(3) My nose is still running.
(4) The Braque exhibition is still running.

This same dependence on the meaning of the absolutive can be seen with
transitive verbs as well. The meaning of cut varies according to what is cut
(again the absolutive participant) in:
(1) He cut his foot.
(2) He cut the lawn.
(3) He cut the paper.
(4) He cut his whisky with water.

The semantic role of an ergative participant (“he” in these last examples) is

mentioned below.

Agents and Patients

A third way of assigning participant roles has received less attention and is
often misunderstood. The Seneca language, where this arrangement predomi-
nates, provides an illustration. The sentence below, taken from a Seneca
folktale, illustrates both agent and patient roles. In the English translation the
words I and she are both subjects, but in the Seneca the prefix wag- (from wak-)
in sawagadöswe’dë’ “I get hungry again” expresses a first person patient of
becoming hungry, whereas the prefix ag- (from yak-) in wa’a:gë’ “she said”
expresses a feminine agent of the act of saying.4

4
rep=repetitive, fac=factual, 1.pat=first person patient, pun=punctual, fem.agt=feminine agent.
40 Thoughts and Their Properties

Da’áö́ h sawagadöswe’dë’ wë:doh gyö’öh wa’a:gë’.

ta’áö́ h s-a-wak-atöswe’t-ë’ wëtöh kyö’öh wa’-yak-ë-’
impossible REP-FAC-1.PAT-get.hungry-PUN ever hearsay FAC-FEM.AGT-say-PUN
“I can’t ever get hungry again she said.”

Details of this agent–patient pattern can vary with the language, but in every
case there is a characterization of entities according to their direct roles in events
and states. Entities that are treated as agents willfully instigate events over which
they have control. Those treated as patients are not in control but suffer the effects
of events that befall them or of states that apply to them. A discussion of the
features underlying systems of this type and the ways they can vary across
languages is available in Mithun (1991). These roles are determined by the
semantics of the event or state itself, and not by discourse factors as is the case
with starting points and absolutives. It is not surprising, then, that the
agent–patient pattern appears most often as a part of verb morphology, typically
in the pronominal affixes on verbs as in the above example.
Two questions remain: what is the semantic role of the object in the
subject–object arrangement, and what is the semantic role of the ergative in the
ergative–absolutive arrangement? The check marks in Table 5.1 suggest answers,
namely that objects play the role of the most involved participant, and that
ergatives play the role of agent. These parallels capture close resemblances across
the three systems. One might also ask whether these participant roles are universal
aspects of thought. The fact that they vary from language to language suggests
that they are not intrinsic to thought but are assigned differently by the differing
semantics of different languages. The point to be stressed is that subjects, objects,
ergatives, absolutives, agents, and patients all have a semantic basis and are not
merely arbitrary syntactic roles.

Summary
On the basis of linguistic universals we can distinguish ideas (the building
blocks of thought) from orientations of those ideas. Ideas may be ideas of

Table 5.1 The syntax and semantics of participant roles

syntactic roles
subject-object ergative-absolutive agent-patient
starting point √
semantic

most involved √ √
roles

agent √ √
patient √
 How Thoughts Are Structured 41

events, states, or entities. Events and states demand the inclusion of entities
whereas the reverse is not true; an entity may stand alone. In the flow of
thought, events are usually transient whereas entities often enjoy longer time
spans. States may differ from events by adhering to the time span of the entities
with which they are associated. We then looked at ways in which entities are
related to events and states, or participant roles, and contrasted the properties of
subject–object, ergative–absolutive, and agent–patient systems.
6 How Thoughts Are Experienced

At lucky moments we seem on the brink

of really saying what we think we think.
But, even then, an honest eye should wink. W. H. Auden (1965: 12)

In the 1970s a group of us in Berkeley were engaged in a project that was

aimed at discovering semantic prerequisites to machine translation, a hot
topic at the time. During our coffee breaks we chatted informally on other
topics, one of which was the question of what it feels like to be thinking.
There seemed to be a consensus that thinking consists above all and perhaps
entirely of inner speech. We are constantly chattering internally to ourselves,
and maybe that is what thinking is. In this chapter we see reasons why that
answer is too simple and how it overlooks important ways in which thinking
is experienced.
Around the same time I went with a friend to a talk by the psychologist
Ruth Day, who told us of various tests that sorted people into two groups.
The thinking of the people in one group was inextricably tied to language, but
for the people in the other group language was less dominant. As we left the talk
my friend and I discussed Day’s tests and found that we were exact opposites on
every one of them. Day (1977, 1979) called the two kinds of people “language
bound” and “language optional,” but we agreed that “language bound” sounded
as if my friend suffered from a cognitive handicap. Perhaps “language primary”
would have had a more positive ring, in which case I could be called “language
secondary.”
The next time I taught a course in language and cognition I described Day’s
findings to the class, supposing them to be relevant to our subject matter. I was
surprised to be confronted with immediate hostility from some of the students.
I soon discovered, too, that some of my colleagues also found the distinction
distasteful, and I was told that Day’s findings had not been replicated: a kiss of
death. How, I wondered, could I reconcile those negative reactions with my
own sympathy for the distinction? It occurred to me that perhaps language-
secondary people like me were tacitly aware of a difference that lay behind
many of our personal experiences in a world in which we were surrounded by

42
 How Thoughts Are Experienced 43

language-primary people like my friend, whereas for the latter the distinction
failed to resonate and elicited the rejection I had witnessed. I even speculated
that perhaps language-primary people were especially attracted to linguistics,
thus skewing the study of language itself.
As I return to this question these many years later, it seems to me that Day
had discovered something important which was abandoned too quickly, even
though it may have overlooked some of the same complexities that were
overlooked in equating thought with inner speech. It would surely be useful
to revisit the distinction with an open mind, looking for its broader implications
and other ways to investigate it. Its most important lesson is that different
people experience thoughts in different ways, and that while everyone may
share their fundamental properties, people may differ in the weights they assign
to them. This chapter looks at various ways of experiencing thoughts, recog-
nizing that they are open to individual variation that may make them more
complicated than a simple assignment to language-primary and language-
secondary groups would suggest.

Thoughts That Resist Verbalization

Equating thinking with language overlooks the fact that people may have
a problem finding language to express what they are thinking. They might
recognize this difficulty explicitly by saying, “I don’t know quite how to say it,”
or, “that’s not exactly what I meant.” One well-known scholar described the
experience as follows:
Now what seems to me obvious by introspection is that I can think without language.
In fact, very often, I seem to be thinking and finding it hard to articulate what I am
thinking. It is a very common experience at least for me and I suppose for everybody
to try to express something, to say it and to realize that is not what I meant and then to
try to say it some other way and maybe come closer to what you meant; then
somebody helps you out and you say it in yet another way. That is a fairly common
experience and it is pretty hard to make sense of that experience without assuming
that you think without language. You think and then you try to find a way to articulate
what you think and sometimes you can’t do it at all; you just can’t explain to
somebody what you think. (Chomsky 2000: 76)

This disconnect between thinking and language can show itself in writing
too. Despite the fact that writing provides extra time in which to find language
that is appropriate, a writer may still be dissatisfied. One writer described the
experience this way:
At the end of a day of writing-work, I emerge with pages of what I am accustomed
to call what I wanted to say. But in more cautious spirit I now ask myself: Are these
words, printed out on paper truly what I wanted to say? Is it ever good enough, as
a phenomenological account, to say that somewhere deep inside I knew what
44 Thoughts and Their Properties

I wanted to say, after which I searched out the appropriate verbal tokens and moved
them around until I had succeeded in saying what I wanted to say? Would it not be
more accurate to say that I fiddle with a sentence until the words on the page
“sound” or “are” right, and then stop fiddling and say to myself, “That must be what
you wanted to say”? (Coetzee 2007: 196)

Disfluencies
Hesitations, false starts, and rewordings are scattered throughout much of
spoken language. The English language, and probably every language, pro-
vides words and phrases that recognize this problem with expressions like you
know, I mean, and the currently ubiquitous like (it’s, like, too hot to handle).
In the following example the language that was begun in 1 was truncated and
then reformulated in 2, the phrase you know was inserted in 3, the attempted
continuation in 4 was reformulated in 5, and then came another truncated
continuation with I mean in 6, and finally another reformulation in 7. Clearly
this person was having more than a little trouble finding language that would
adequately express what she was thinking.
(1) Because and you know she hasn’t been
(2) . . . and she’s been eating,
(3) you know,
(4) probably m-
(5) .. same if not more,
(6) but I mean hasn’t,
(7) .. it’s just her body’s just.. not.. doing well.

Thus, both introspection and disfluencies provide evidence that thoughts

can be disconnected from whatever language might be chosen to express
them. Evidently people routinely compare what they are thinking with the
thoughts that would be expressed by whatever language they might be
considering. As thoughts flow through their consciousness they monitor
the relation between those thoughts and what might be expressed by
language. Aware of both what they are thinking and what would be
expressed by any possible language, they may surrender in frustration or
may modify what they are thinking to bring it in line with linguistic
possibilities.
In the seventh proposition of his Tractatus Logico-Philosophicus Ludwig
Wittgenstein famously proclaimed, “Whereof one cannot speak, thereof one
must be silent” (Wittgenstein 1955 [1922]: 108). Because “whereof one cannot
speak” might be understood to describe a thought that cannot be verbalized,
further discussion of such a thought would seem to be ruled out. But perhaps we
need not take Wittgenstein at his word. Is there, after all, something that can be
said about a thought that resists verbalization?
 How Thoughts Are Experienced 45

Why should such a situation arise? A relatively common situation is one in

which a language’s semantic resources simply fail to provide what is needed.
Perhaps, for example, the language offers no satisfactory way of categorizing
an event or an entity that is included in the thought. When such a situation
arises, one might settle on a categorization whose coverage comes close to what
is needed while still leaving a feeling of dissatisfaction. In the following
exchange the idea of a certain part of a penguin’s anatomy led to uncertainty
as to how the idea of that part could be verbalized, and each of the four speakers
had a different suggestion:
a: Each penguin is numbered,
And on . . . the arm,
b: Wing.
a: Okay,
the joint here,
well,
what do you call this?
b: The wing.
c: Humerus.
d: It’s a flipper.

Or one might choose a sequence of linguistic expressions that together zero

in on the thought. The speaker of the following example was searching for
a satisfactory way to verbalize her status as a beginning teacher:
(1) I’m the only teacher who’s not experienced,
(2) who’s not certified,
(3) who just started teaching.
This example is reminiscent of the “couplets” (or in this case “triplets”) that are
a frequent feature of the discourse of many languages (Lengyel 1988), a fact
that suggests the ubiquity of this situation.
It may be the case that a thought remains at a level where language is
simply not an option. Thought at this level is the immediate, minimally
processed result of a sensory experience, an action, or an emotion. One
might, for example, be aware of something that is seen, or of moving
one’s body in a certain way, or of experiencing an emotion without
reflecting on that awareness any further. A great deal of what we experi-
ence may in fact remain in this raw, unreflective state, a state that has
been recognized with the label “anoetic.” The following description is
helpful:
Anoetic consciousness is defined as the rudimentary state of affective, homeostatic,
and sensory-perceptual mental experiences. It can be considered as the autonomic
flow of primary-process phenomenal experiences that reflects a fundamental form of
first-person “self-experience,” a vastly underestimated primary form of phenomenal
46 Thoughts and Their Properties

consciousness. We argue that this anoetic form of evolutionarily refined consciousness

constitutes a critical antecedent that is foundational for all forms of knowledge
acquisition and via learning and memory, giving rise to a knowledge-based, or noetic,
consciousness as well as higher forms of “awareness” or “knowing consciousness” that
permits “time-travel” in the brain-mind. (Vandekerckhove, Bulnes, and Panksepp
2014: 1)
Experiences at this level fail to reach a stage at which they might be
expressed with language, but “even if one does not talk about it, anoetic
consciousness still may be expressed in body posture, gestures, and tone of
speech” (p. 3).

Imagery
Much of thought, however, involves “higher forms of awareness” of several
kinds, not all of which are language-related. For many people thought is
experienced as imagery: the attenuated replaying of experiences that were
perceived directly at an earlier time or that may simply be products of
imagination. Associating thought with imagery may have a special appeal
for those whose mental life overflows with visual images (as experienced
strongly by visual artists), or with auditory images (as experienced strongly
by musicians), or with tactile images (as experienced strongly by the blind),
or with olfactory images (as experienced strongly by dogs). Individual
differences in the experiencing of imagery (Poltrock and Brown 1984)
help to explain the variation that is found in the role it plays in thought.
William James, having asked people about their image of their breakfast
table, reported:
an exceptionally intelligent friend informs me that he can frame no image
whatever of the appearance of his breakfast-table. When asked how he then
remembers it at all, he says he simply knows that it seated four people, and
was covered with a white cloth on which were a butter-dish, a coffee-pot,
radishes, and so forth. The mind-stuff of which this “knowing” is made seems
to be verbal images exclusively. (James 1890: 265)
There may be two ways to understand what James meant by “verbal images.”
Perhaps he concluded that his friend experienced the auditory component of
silent language while lacking any visual accompaniment. More interestingly,
perhaps his friend was conscious of a thought that had neither a visual nor an
auditory accompaniment. That is the more interesting possibility because it
implies an imageless consciousness that was sufficient to let James’s friend
know the circumstances of his breakfast table. Imageless thought was a major
concern of the Würzburg school in Germany a century ago (Humphrey 1951)
and it could also profit from revisiting. However that may be, it is important to
 How Thoughts Are Experienced 47

recognize the variability of the imagery experience among people who are
clearly not lacking the ability to think.

Affect
Another ingredient of thought is certainly the experiencing of an emotion,
mood, or attitude. Antonio Damasio (1994) called special attention to the
crucial importance of emotion as a contributor to the totality of thought.
Let us assume here the obvious presence of affect as another ingredient of
thought while postponing further discussion to the final three chapters of
this work.

Silent Language
As we return now to language we need to distinguish audible spoken
language from language that is not spoken out loud but remains silent within
the consciousness of the thinker – in effect, mental imagery of language.
It has been studied by psychologists as part of the study of imagery (see the
many references in Tian and Poeppel 2012). Those studies could benefit
from looking beyond imagery of sounds to the thoughts behind the sounds.
Silent language is no less thought-based than audible language, and it seems
impossible to divorce linguistic sounds from the ideas that are paired with
them.
There are important ways in which silent language differs from audible
language. Because it is private its properties have not been well studied, but
it is a totally natural form of thought and a common experience for everyone.
Much of silent language stops short of being fully formed but may be limited to
experiencing ideas of events, states, and entities, as discussed in Chapter 6. But
it is also possible to experience fully formed language that is immediately
available for expression in sound. Because silent language is not created in
interaction with others it is freer to follow its own path, which may be less
constrained by a coherent or consistent trajectory.

How Thoughts Take Precedence over Language

We have seen that a thought may be experienced as anoetic consciousness, as
imagery, as affect, or as silent language. We can turn now to cases where
thoughts are expressed with overt language but where the thoughts themselves
carry more weight than the language. In the 1920s a German psychologist
named Hanns Georg Hartgenbusch read sentences aloud to people who were
asked to write down immediately what they had heard, preserving the exact
wording (Hartgenbusch 1933). These first responses were then read aloud
48 Thoughts and Their Properties

to second persons, their responses to third persons, and so on to a total of

twenty-five responses, constituting a series. One of the sentences Hartgenbusch
used at the beginning of a series was:
Der Bäcker backt uns in der Nacht das Brot und ruht bei Tage aus.
The baker bakes bread for us at night and rests during the day.
After this sentence had passed through twenty-five readings and responses, the
following are typical final results:
Der Bäcker backt am Tage das Brot und ruht in der Nacht aus.
The baker bakes bread during the day and rests at night.
Der Bäcker backt uns das Brot bei Tag and schläft bei Nacht.
The baker bakes bread for us during the day and sleeps at night.
Der Bäcker backt Brot, in der Nacht ruht er.
The baker bakes bread, at night he rests.
Noteworthy is the fact that the baking was routinely shifted from night to day
and the resting from day to night, presumably because most people, unlike
bakers, think of working during the day and sleeping at night. The instruction to
write verbatim what was heard had less influence than the motivation to adhere
to normal ways of thinking.
In the 1960s Jacqueline Sachs investigated how well people remembered the
form of a sentence as compared with how well they remembered the thought
behind it. Ninety-six undergraduates heard twenty-eight prerecorded passages
of various kinds – folktales, history, biography, etc. One passage was the
following (Sachs 1967: 438–439):
There is an interesting story about the telescope. In Holland, a man named
Lippershey was an eye-glass maker. One day his children were playing with some
lenses. They discovered that things seemed very close if two lenses were held about
a foot apart. Lippershey began experimenting and his “spyglass” attracted much
attention. He sent a letter about it to Galileo, the great Italian scientist [italics
added]. Galileo at once realized the importance of the discovery and set about to
build an instrument of his own. He used an old organ pipe with one lens curved out
and the other in. On the first clear night he pointed the glass toward the sky. He was
amazed to find the empty dark space filled with brightly gleaming stars! [80
syllables after the italicized sentence] Night after night Galileo climbed to a high
tower, sweeping the sky with his telescope. One night he saw Jupiter, and to his
great surprise discovered near it three bright stars, two to the east and one to the
west. On the next night, however, all were to the west. A few nights later there were
four little stars. [160 syllables after the italicized sentence]
This passage was interrupted either immediately after the italicized sentence
or eighty syllables later, or was continued to the very end of the selection.
Following the interruption a test sentence like one of the following was read.
 How Thoughts Are Experienced 49

The test sentence was either a verbatim repetition of the target sentence,
a modified version that exhibited a different word order or a shift from active
to passive voice or vice versa, or a new version in which the meaning had been
changed.
Verbatim: He sent a letter about it to Galileo, the great Italian scientist.
Change in word order: He sent Galileo, the great Italian scientist, a letter about it.
Change to passive: A letter about it was sent to Galileo, the great Italian scientist.
Change in meaning: Galileo, the great Italian scientist, sent him a letter about it.
The students were asked to mark whether the test sentence was identical to
what they had heard or whether it was changed and, if they thought it was
changed, whether there was a change in meaning or a change in grammar.
The findings indicate that recognition memory for the form of a sentence declines much
more rapidly than recognition memory for the meaning. If the test sentence followed the
original immediately, recognition was high for all sentences. After only 80 syllables of
connected discourse had been interpolated between the original and test sentences, Ss’
recognition of syntactic changes was close to chance. Changes in the meaning of
a sentence, on the other hand, were recognized quite well, even though these changes
were subtle . . . Thus, very slight changes in the words of a sentence had vastly different
effects on the experimental task, depending on whether or not the change affected the
meaning. (Sachs 1967: 442)

Summary
Thoughts may be experienced (1) as unreflective anoetic consciousness, (2) as
imagery, (3) as affect, (4) as silent language, or (5) as audible language. There
may be a disconnect between silent thought and overt language, as shown
by thoughts that resist verbalization and by the frequency of linguistic
disfluencies. Psychological studies have shown that memory for thoughts
carries more weight than memory for overt language.
7 How Thoughts Are Shared

A question one might validly ask at this point is just why thoughts should be
turned into sounds in the first place. Why isn’t it enough for people just to think,
without following the pathway to sound that was described in Chapter 4?
The answer may seem obvious, but it depends on a recognition that people
depend on other people, and that the ability to share thoughts is a fundamental
human need. Because thoughts themselves are not open to sharing in the private
form in which they are experienced, they need to be converted into a medium
that allows them to bridge the gap between one mind and another.
Why, then, should that medium be sound? Sign languages show that it does
not have to be, but the ubiquity of sound as a primary medium for sharing
thoughts suggests that sound must serve that purpose especially well. Humans
have evolved to use their vocal organs to produce fine distinctions in sound, and
their ears and brains to process those distinctions accurately. It is also an
advantage that sounds fade rapidly, to be quickly replaced by other sounds,
a property that allows those sounds to follow closely the pace of constantly
changing thoughts. Furthermore, the conversion of thoughts into sounds
requires that thoughts be organized in ways that might not otherwise be
necessary. Sound thus serves both to communicate and to organize our
thoughts.
Producing and listening to sounds was what led language to evolve as it did.
Being able to share thoughts in this way has given humans a huge advantage
over other animals, and it has long been recognized as a fundamental property
of humanness. Speaking and listening are abilities that most people acquire
naturally without special training. Not all thoughts are shared, of course, and
silent thinking is another fundamental aspect of human experience – so funda-
mental, in fact, that it has been tempting to equate silent language with thinking
itself, as noted at the beginning of Chapter 6.
A very different way of sharing thoughts arose only during the past few
thousand years and did not come into widespread use until relatively recently.
Instead of using our vocal organs to make sounds and our ears to listen to them,
we use our hands to make visible marks on paper or some other medium and our
eyes to look at those marks; in other words, we write and read. Writing is

50
 How Thoughts Are Shared 51

facilitated by our ability to make fine movements with our hands, and reading
has benefited from a highly developed sense of vision. But humans did not
evolve to write or read and these abilities must be taught and learned. Writing
has conferred important benefits that supplement what is possible with speak-
ing and listening. It is now possible for thoughts that were experienced in one
place and time to be shared at a totally different place and time, and for thoughts
to be preserved more or less indefinitely. Writing even lets us connect with
thoughts that were first experienced centuries ago. Chafe and Tannen (1987)
reviewed earlier work on differences between speaking and writing, and for
a comparison of these two ways of using language see Chafe and Danielewicz
(1987) and Chafe (1994: 41–50).

The Timing of Thoughts

Speaking and listening take place at a relatively rapid pace that may come
close to the pace of silent thinking. The fact that spoken intonation units
typically occupy about a second and a half may suggest approximately
what that pace might be, although the time necessary for the articulation of
sounds may hold speaking to a maximum rate, with silent thinking pro-
ceeding still more rapidly. In any case it is clear that the act of writing
requires significantly more time. Writers are under no obligation to main-
tain continuity and they may easily revise their thoughts as they write.
Readers, unlike writers, are free to process what they are reading at a pace
that may even exceed the pace of speaking.
In writing the flow of thought is more open to interruption and backtracking,
and because writers are usually isolated from direct contact with an audience
they may not feel the same pressure to hold an audience’s attention. Readers,
for their part, are free to follow wherever their own thoughts may lead, whereas
with listeners a lack of attention to what is being said may deviate from
expected norms of cooperation. It is a remarkable fact that reading can be,
and often is, interrupted for hours, days, or even longer and then resumed with
little interruption in topic continuity.

The Covert Prosody of Written Language

Prosody plays an important role in speaking, signaling boundaries in the
flow of thought and ways in which thoughts are evaluated. It is only
minimally and inconsistently represented in writing, but many people
when they read assign prosodic boundaries, accents, and intonation con-
tours to what they are reading. The following statements to that effect are
from a textbook on linguistics by Dwight Bolinger and an autobiographical
book by Eudora Welty:
52 Thoughts and Their Properties

We monitor our writing sub-vocally, reading in an intonation, and the fact that the
intonation is not actually shown and our reader is going to have to guess at it is as likely
as not to escape our attention. (Bolinger 1975: 602)
Ever since I was first read to, then started reading to myself, there has never been a line
read that I didn’t hear. As my eyes followed the sentence, a voice was saying it silently to
me . . . My own words, when I am at work on a story, I hear too as they go, in the same
voice that I hear when I read in books. (Welty 1983: 12–13)

Reading Aloud
Reading aloud is a special method of thought sharing that has properties of its own
(see Esser and Polomski 1987, 1988 and Chafe 2006). Reading aloud takes
language that was initially produced as writing but at some later time delivers it
as speaking. Its audience does not read it but listens to it. Why do people read
aloud? There are situations in which listeners are incapable of reading to them-
selves, either because they have not yet learned to read, as with young children, or
because they are blind. But in other cases there may be a wish to share a piece of
writing with another person as a way of capturing the intimacy of conversation, as
when a husband and wife read to each other. An orator or politician may deliver
a written text by reading from a teleprompter. Some authors read aloud their own
works professionally, a common practice in the nineteenth century but one that
continues to this day. Different in motivation and effect is the presentation of so-
called papers at academic conferences. Practices differ with different disciplines.
In general, scholars in the humanities most often read their papers aloud with
minimal spoken prosody while those in the sciences and engineering most often
speak from notes or with no prompting at all, but there are always exceptions.
Read-aloud papers may be the most difficult kind of language for an audience
to process, although that depends on the ability of the reader to mimic sponta-
neous speech. In spontaneous speaking a speaker’s consciousness is focused on
the flow of thoughts and not on the sounds, and the prosody flows naturally and
unconsciously out of the thoughts. When a speaker, reading aloud, focuses more
on producing the sounds and less on the thoughts behind them, prosody no longer
emerges naturally, often sounding artificial or inappropriate.
Some years ago I conducted an informal survey of approximately 100
members of the Linguistic Society of America, asking whether, when they
presented papers at conferences, they spoke their presentations or read them.
According to their answers 56 percent were readers and 37 percent were
speakers, while 7 percent said they did both equally. When asked why they
did what they did, the readers mentioned adhering to time constraints, feeling
more secure, and having a concern for the elegance of their language.
The speakers referred overwhelmingly to a belief that they were easier to
listen to.
 How Thoughts Are Shared 53

Speaking Writing Reading

and Listening and Reading Aloud

Production Making sounds Making marks

fast slow
interactive isolated
natural unnatural

Product Lasting marks

special lexicon
special syntax
covert prosody

Delivery Making sounds

unnatural prosody

Reception Listening Reading Listening

effortless more difficult most difficult

Figure 7.1 Properties of speaking, writing, and reading aloud

Listenability is difficult to study, but its companion process, readability, has

been studied a great deal because of its practical applications. A long tradition
of readability research dating from the 1920s was surveyed in Klare (1974).
The goal of much of that research was to discover a readability formula that
could be applied to written works in order to determine their suitability for
school children at different grade levels. These mechanical formulas fell short
of capturing everything that makes something readable and other, more recent
attempts have tried to identify cognitive factors that affect the ease or difficulty
of written selections (e.g. Hirsch 1977; Holland 1981; Chafe 1991).
Figure 7.1 summarizes differences between speaking and listening, writing
and reading, and reading aloud. In the left-hand column, speaking, there is
a two-way distinction between the production of the language and its reception.
It is typical of speaking that it is relatively fast, that it is interactive (the
producer at one moment becoming the receiver at the next), and that it is
natural (learned without special training). As noted at the bottom of this
column, listening to speaking is something humans are able to process without
special effort.
In the middle column, writing, the production is separated from the product.
Producing written language involves making marks of some kind, a slower
process than speaking. Writing is typically performed in isolation, lacking the
possibilities for interaction that are typical of speaking. And writing is unnatural in
54 Thoughts and Their Properties

the sense that it must be taught and learned. The product of writing consists
of marks whose relative permanence contrasts with the evanescence of
speaking. This permanence has made it possible to create language in one
place and time and have it received at a totally different place and time, an
ability that has had an enormous impact on recent human history.
The written product is likely to exhibit its own special lexicon and syntax,
which may differ from those of ordinary speaking. Writing is relatively
impoverished with respect to the prosodic qualities that contribute impor-
tantly to speaking: the variations in pitch, loudness, tempo, and voice quality
that are only covertly present. At the bottom of this second column is
a suggestion that the reception of written language, in other words reading,
is more difficult than listening to ordinary speech.
The right-hand column shows properties of reading aloud. The production
and product are those of written language, but reading aloud is special in that
the language is ultimately delivered by making sounds, as with ordinary
speaking. The result, however, is not equivalent to ordinary speaking, in part
because of a special lexicon and syntax, in part because reading-aloud
prosody is likely to have its own properties that diverge from those of
ordinary speaking. At the bottom of this third column is a suggestion that
listening to language read aloud may be the most difficult of all three
conditions. Because of its special lexicon, syntax, and prosody, listening to
such language may require a mental effort exceeding that which is necessary
for silent reading, and far exceeding that which is necessary for listening to
conversation.

An Example
It can be interesting to compare language that is read aloud with spontaneous
speaking by the same person, as may happen during a press conference. In this
example the former American president George W. Bush opened a press con-
ference with a read-aloud prepared statement that exhibited a flat prosody
conveying minimal involvement with his subject matter. In this case an entire
syntactic sentence was uttered as a single prosodic phrase:
The House and the Senate are now considering my supplemental request for operations
in Iraq and Afghanistan.

Figure 7.2 shows the fundamental frequency of this long phrase, with a final
falling contour. This same pattern was repeated throughout the President’s
prepared remarks.
In contrast, at one point during the question period that followed, President
Bush spoke as follows:
 How Thoughts Are Shared 55

Figure 7.2 Fundamental frequency of President Bush reading aloud

Figure 7.3 Fundamental frequency of President Bush speaking

spontaneously

(a) Yeah.
(b) I think it’s-
(c) I think it’s a very interesting point you make in your question.
(d) They’re trying to send a warning.
(e) . . Basically what they’re trying to do is-
(f) . . . . is uh,
(g) . . . . cause people to run.
(h) . . You know?

The difference in style is obvious. The initial response in (a) was followed by
a fragmentary intonation unit in (b) that was repeated at the beginning of the
successful but less spontaneous answers in (c) and (d), whose wordings were
employed by President Bush on other occasions. The more spontaneous explana-
tion in (e) through (g) exhibited disfluencies that included a repetition and two
unusually long pauses. The sequence ended with a softly spoken you know.
The considerable pitch variation that is visible in Figure 7.3 is in marked contrast
to that shown in Figure 7.2.
56 Thoughts and Their Properties

Summary
This chapter began with natural speaking and listening, features of humanness
that are fundamental to the way thoughts are shared. Speaking and listening
stand in sharp contrast to writing and reading, a recent development in human
history that has revolutionized the ways thoughts can be shared, allowing them
to be transmitted across time and space and preserved for considerable lengths
of time. Reading aloud, which takes language that was first produced as writing
and delivers it subsequently as speaking, may be the most difficult kind of
language for a listener to process.
8 How Thoughts Flow through Time

As we take, in fact, a general view of the wonderful stream of our conscious-

ness, what strikes us first is this different pace of its parts. Like a bird’s life, it
seems to be made of an alternation of flights and perchings. The rhythm of
language expresses this, where every thought is expressed in a sentence, and
every sentence closed by a period. William James (1890: 243)

James’s comparison of thoughts to a perching bird combines several insights

regarding the flow of consciousness. First, consciousness is in constant
change. Second, those changes are reflected in “the rhythm of language.”
Third, this rhythm is manifested in two linguistic areas, one of which is syntax:
“every thought is expressed in a sentence.” The other area is prosody: “every
sentence [is] closed by a period,” or by some other phrase-final pitch contour.
These were prescient observations, but we know more about language today
and it is possible to elaborate on what James wrote.
A different way of putting it would be to say that thinking occurs in spurts,
which find expression in both the syntax and the prosody of language. “Prosody”
covers four ways in which speech sounds may vary: in pitch, loudness, timing,
and voice quality. Pitch is the way we perceive variations in the fundamental
frequency of periodic sounds. Loudness, or volume, is the way we perceive
acoustic intensity. Timing covers two phenomena: the duration of units like
vowels, consonants, syllables, words, and phrases, but also the rapidity with
which sequences of those units are produced. Voice quality includes properties
such as whispering, laughing, “creaky voice,” and harshness (Laver 2009).
These prosodic qualities occur simultaneously with qualities that have been
called “segmental” because they divide sound into segments like vowels,
consonants, syllables, words, and phrases. Prosodic and segmental qualities
combine to create the totality of speech sounds, just as color and form con-
tribute simultaneously to the totality of a painting. Prosodic qualities have
sometimes been imagined as riding “on top of” segmental sounds and for that
reason they have sometimes been called “suprasegmental.” Most segmental
qualities are produced by physical events in the vocal tract above the larynx:
fine movements of the lips and especially the tongue. Prosodic qualities are
produced largely by events in the lungs and larynx.

57
58 Thoughts and Their Properties

Table 8.1 Aspects of thought and sound and their areas of control

Organs of sound
Aspects of thought Brain location Aspects of sound production
ideas cortical segmental tongue and lips
emotions subcortical prosodic lungs and larynx

These two qualities of speech sounds – segmental and prosodic – are related
to ideas and emotions respectively. Presumably ideas are processed above all in
the neocortex and emotions in the older brain or “limbic system.” Aspects of
thought, the principal locations in the brain where they are processed, the
aspects of sound by which they are expressed, and the principal organs
involved in their expression are summarized in Table 8.1. These observations
suggest the existence of devoted neural connections between the widely dis-
tributed areas of the neocortex that process ideas and the smaller area that
controls tongue and lip movements, as well as special connections between
centers of emotion and the brain areas controlling the lungs and larynx. It is
worth noting, too, that the items in the bottom row of Table 8.1 are highly
developed in many sentient creatures in contrast to those in the row above,
whose elaboration is confined largely to humans.
But something important is missing from Table 8.1. While prosodic sounds do
express emotions, they also organize and track the flights and perchings of
consciousness described by James. The top of Figure 8.1 shows pitch variations
within the utterance we’re gonna go hang out there as discussed in detail in
Chapter 9. At the bottom of Figure 8.1 is a spectrogram, where the horizontal axis
shows time and the vertical axis shows fundamental frequency, with the degree of
blackness showing the amount of energy at that time and frequency. The relative
duration of the words is shown in the middle, where it can be seen that the
sequence we’re gonna go was spoken very rapidly. Beginning with the word
hang there was increasing deceleration as her utterance approached its end.
The relation of one thought to another within the flow of speech is signaled in
part by a terminal pitch contour, observable here in the falling pitch on there.
This contour often expresses closure of a thought or thought sequence, as
opposed to the anticipation of more to come that is signaled by a level or rising
pitch. It appears to be true of all languages that the flow of spontaneous speech
can be segmented into intonation units of this kind (Chafe 1987, 1994; Du Bois
et al. 1993). Usually these intonation units occupy no more than a second or
two, suggesting that foci of consciousness are experienced at that rate. Each
exhibits a coherent pitch contour, with each language offering a variety of
specific contours. Many intonation units are preceded by a pause – anything
 How Thoughts Flow through Time 59

we’re gonna go out

hang

there

we’re gonna go hang out there

Figure 8.1 Pitch and timing in Jane’s statement

from a brief break in timing to a conspicuous period of silence – and they may
exhibit other prosodic properties such as an initial resetting of the pitch level
and an initial acceleration followed by a deceleration and/or decrescendo,
sometimes accompanied by creaky voice or “vocal fry.”
Because speakers must breathe in order to stay alive, there might be
a temptation to suppose that each spurt of language is nothing more than
an expulsion of oxygen-depleted air followed by a pause for inhalation
before the next exhalation. But intonation units occur so rapidly and are
so intimately related to the flow of thought that explaining them by the
need to breathe is clearly inadequate. They can be observed in casual
listening to ordinary talk (listen to the person behind you on the bus), but
their properties and functions could not be investigated more system-
atically before the advent of electronic sound processing. There are still
disagreements about their boundaries, with different investigators relying
on partially different criteria (Stelma and Cameron 2007), but their ubi-
quity in speech, regardless of the language being spoken, seems well
established.
If we shift our attention from their acoustic properties to their content,
intonation units give us a window on minimal units of thought, or foci of
consciousness, allowing us to view the sequencing of such foci through time.
Their further nature and the way they group into larger units is most easily
discussed with reference to an illustration. Quoted below is the beginning of
one of the Pear Stories discussed in Chafe (1980). The person speaking will be
called Ellen.
Each line transcribes an intonation unit. Accent marks show prominent
syllables. The sequences of dots indicate pauses of varying lengths
60 Thoughts and Their Properties

corresponding roughly to the number of dots. A comma at the end of a line

shows a nonfinal terminal pitch contour. Various nonfinal contours exist and
their use varies to some extent from one language to another. A period at the
end of a line shows a falling pitch that is perceived as a sign of closure, as in
Figure 8.1. The two words in italics at the end of line 23 were whispered.
Most of these intonation units were substantive in content, expressing
a single focus of consciousness, but lines 1, 4, and 10 were regulatory, func-
tioning as guideposts to the flow of thought, while lines 3 and 5 were fragmen-
tary – begun but not completed (Chafe 1993). On the right are the lengths of
substantive intonation units in seconds, excluding the time occupied by initial
pauses. Lines 12 and 14 have internal pauses, with the duration of the entire unit
given in parentheses, followed by the length of the coherent portion after the
last pause (Latin-looking and climbs up the ladder). If we count only the
portions just described, the mean length of the substantive units was 1.24
seconds, ranging from 0.08 seconds for line 2, it starts out, to 2.03 seconds
for line 22, and he’s already got two baskets full.
(1) . . . . . . . . . Uh, regulatory
(2) it starts óut, substantive 0.08
(3) there’s a fragmentary
(4) . . . . . . . . . .uh, regulatory
(5) . . . . . the lándscape is like a fragmentary
(6) . . . . . . . it’s a péasant landscape, substantive 1.27
(7) but it isn’t really fármland, substantive 1.51
(8) it’s like an órchard. substantive 0.94
(9) . . . It’s a smáll orchard, substantive 1.08
(10) . . . and, regulatory
(11) . . . it’s gréen. substantive 0.54
(12) And there’s this sort of . . . . . . Látin-looking, substantive (4.36) 1.42
(13) middle-aged mán, substantive 1.16
(14) who’s . . . . um . . . clímbs up the ládder, substantive (3.55) 1.25
(15) . . . . . uh that’s léaning against a trée, substantive 1.49
(16) and picks péars, substantive 1.22
(17) . . . puts them in.. his ápron. substantive 1.47
(18) . . . Clímbs down the ládder, substantive 1.19
(19) . . . and émpties the pears ínto, substantive 1.96
(20) . . . a bíg básket. substantive 0.95
(21) . . . . He’s got three baskets sítting there, substantive 1.45
(22) and he’s already got two baskets fúll. substantive 2.03
(23) . . . He dóes this a couple times. substantive 1.28
 How Thoughts Flow through Time 61

Ideas and Their Types

Substantive intonation units expressing foci of consciousness are built on the
minimal thought units that were identified in Chapter 5 as “ideas.” We saw there
that ideas may be of events, states, or entities. Many events involve an action,
something someone does, as is true of most of the events in lines 14–23:
(14) who’s . . . . um . . . clímbs up the ládder,
(16) and picks péars,
(17) . . . puts them in.. his ápron.
(18) . . . Clímbs down the ládder,
(19) . . . and émpties the pears ínto.
(20) . . . a bíg básket.
(23) . . . He dóes this a couple times.
Other events involve a change of state. The only example above is:
(2) it starts óut,
But at a later point in the narrative all the intonation units except line 25 in the
following sequence express changes of state:
(24) And his hat blows off,
(25) when they cross,
(26) and his bike hits into a rock,
(27) and he falls down,
(28) and all the pears scatter.
States perform various functions. They may, for example, assign an entity to
a category or property:
(6) . . . . . . . it’s a péasant landscape,
(7) but it isn’t really fármland,
(8) it’s like an órchard.
(9) . . . It’s a smáll orchard,
(11) . . . it’s gréen.
Or they may describe the temporary “disposition” of an entity rather than an
integral property. Leaning against a tree was a disposition of the ladder in line
15, and sitting there was a disposition of the baskets in line 21:
(15) . . . . . uh that’s léaning against a trée,
(21) . . . He’s got three baskets sítting there,

Split Thoughts
The following sequence shows that sometimes a unitary thought may be split
across two intonation units:
62 Thoughts and Their Properties

(12) And there’s this sort of . . . . . . Látin-looking,

(13) middle-aged mán,
These two intonation units together introduced the idea of a person who went
on to perform a series of actions. Introducing him was a unified thought but it
straddled two foci of consciousness. A focus of consciousness is normally
limited to one new idea: an idea the speaker assumes is being introduced into
the listener’s consciousness for the first time within the current conversation
(Chafe 1994: 108–119). In line 12 the idea of being Latin-looking was new, as
was the idea of being a middle-aged man in line 13. The “one new idea at
a time” constraint distributed these two ideas across two foci of consciousness
and thus two intonation units. The speaker’s thought of introducing this man
evidently included too much for one focus of consciousness. The same con-
straint led shortly after that to the splitting of another thought, in this case the
thought of an action:
(19) . . . and émpties the pears ínto,
(20) . . . a bíg básket.
The idea of emptying the pears was one new idea, the idea of the big basket
another, and hence the separation. Although a minimal thought is usually
coextensive with a single intonation unit, examples like these show that this
restriction does not always apply.

Expanded Thoughts and Their Verbalization in Sentences

Moving from minimal thoughts, usually limited to one new idea, to larger
segments of speech, we can learn something about expanded thoughts by
examining how they are verbalized. There is no entirely satisfactory name for
such an expanded thought, but in Chafe (1994) it was called a “center of
interest,” borrowing a term from Buswell (1935) where it referred to the way
people look at pictures.
It is tempting to see this situation in evolutionary terms. The presence
of intonation units in all languages suggests a prehuman stage when the
processing of experience was limited to one new idea at a time. As the
thinking of our remote ancestors evolved to accommodate larger segments
of experience, the limitations imposed by foci of consciousness became
too constraining and there arose a need to accommodate thoughts that
embraced more. The result was a verbalization pattern in which a series
of minimal thoughts were combined into a more inclusive center of
interest that was expressed in what we know as a sentence. It is no
accident that a sentence has been characterized in popular terms as the
expression of “a complete thought.” A complete thought can thus be seen
 How Thoughts Flow through Time 63

as a sequence of foci of consciousness that together capture a coherent

center of interest.
It is not unusual to find uncertainty with regard to the point at which
a sequence of foci of consciousness forms a coherent whole. As speakers
make their way through such a sequence, they decide at some point that
a center of interest has been adequately covered and signal their decision
with a falling pitch. In the following sequence Ellen’s goal was to verbalize
what she remembered of the opening scene of the film:
(1) . . . . . . . . .Uh,
(2) it starts óut,
(3) there’s a
(4) . . . . . . . . . .uh,
(5) . . . . . the lándscape is like a
(6) . . . . . . . it’s a péasant landscape,
(7) but it isn’t really fármland,
(8) it’s like an órchard.

By the end of line 8 Ellen was momentarily satisfied with the words it’s like
an orchard and so she closed this sequence with a falling pitch. But the word
like in line 8 already suggested that, after her attempts at describing the scene
with peasant landscape and farmland, even orchard left more to be desired, so
she added an “afterthought” with two more new ideas:
(9) . . . It’s a smáll orchard,
(10) . . . and,
(11) . . . it’s gréen.
Lines 9–11 provided a fuller and more satisfying closure to this center of
interest. Sentence boundaries like the one at the end of line 8, and then more
satisfactorily at the end of line 11, signal opportunistic on-line judgments that
a center of interest has been adequately covered. Such a judgment has both
prosodic and syntactic consequences that may, but need not coincide (Chafe
2000a).
Lines 12–23 illustrate repeated attempts to provide closure to another center
of interest, in this case a description of the pear picker’s activities:
(12) And there’s this sort of . . . . . . Látin-looking,
(13) middle-aged mán,
(14) who’s . . . . um . . . clímbs up the ládder,
(15) . . . . . uh that’s léaning against a trée,
(16) and picks péars,
(17) . . . puts them in.. his ápron.
(18) . . . Clímbs down the ládder,
(19) . . . and émpties the pears ínto,
(20) . . . a bíg básket.
64 Thoughts and Their Properties

After temporary closure was achieved in line 20, Ellen proceeded to add two
afterthoughts. First:
(21) . . . . He’s got three baskets sítting there,
(22) and he’s already got two baskets fúll.
and then finally:
(23) . . . He dóes this a couple times.

Topics
The whispering at the end of line 23 signaled the completion of a topic that
covered the pear-picking activity in its entirety. William James is helpful here
again:
In all our voluntary thinking there is some topic or subject about which all the members
of the thought revolve. Half the time this topic is a problem, a gap we cannot yet fill with
a definite picture, word, or phrase, but which . . . influences us in an intensely active and
determinate psychic way. Whatever may be the images and phrases that pass before us,
we feel their relation to this aching gap. To fill it up is our thought’s destiny. Some bring
us nearer to that consummation. Some the gap negates as quite irrelevant. Each swims in
a felt fringe of relations of which the aforesaid gap is the term. (James 1890: 259; cf.
Chafe 2000b)

In linguistics the word “topic” has been used in several quite different ways,
but its use here, sometimes qualified as “discourse topic” (Brown and Yule
1983: 71–106), refers to an aggregation of thoughts that (1) remains relatively
stable through repeated verbalizations of the same or similar thoughts, and
(2) lends itself to summarizing and labeling. Topics of this kind vary greatly in
the amount of material they cover. The example discussed above illustrates
a topic that can be labeled “pear-picking.” It is, however, a subtopic within
a larger “basic-level” topic that was coextensive with the entire narrative.
Topics thus form hierarchies, with a basic-level topic often containing several
subtopics and perhaps constituting one element within a supertopic. Often there
is significant pausing before a new topic at any one of these levels is introduced.
A topic may open with heightened volume, higher pitch, and faster tempo, and
it may peter out when the speaker decides that James’s aching gap has been
adequately filled.
Once a topic has been introduced, a speaker’s focus of consciousness
navigates through it, activating first one included thought and then another
until closure is judged appropriate. The navigation process may be guided by
a “schema” – a familiar pattern that provides a path through the topic (Bartlett
1932; Chafe 1986) – or it may be driven by less predictable interactions
between conversational participants (Chafe 1994: 120–136). In the above
 How Thoughts Flow through Time 65

Table 8.2 Discourse segments and their properties

Place in the flow of Focus of consciousness Center of interest Topic

thought
Prosodic manifestation intonation unit prosodic sentence topic boundary
phenomena
Remarks severely limited in opportunistic boundaries basic level with
duration and content and afterthoughts subtopics and
supertopics
Examples above 1, 2, 3 . . . 23 1–11, 12–23 1–23

Figure 8.2 Mozart Köchel 331

example the trajectory of the topic was largely determined by the sequence of
events within the film.

Summary
Table 8.2 summarizes the discourse segments that were described in this
chapter, showing their place in the flow of thought and their prosodic manifes-
tations, with remarks on their limits and organization.
Chafe (1994: 186–191) suggested that the same pattern is found in European
music of the classical style, but also in Seneca songs from a very different
tradition. The implication was that a hierarchy that consists of foci of
consciousness combined into centers of interest and then into larger topics,
discoverable not only in language but also in music, illustrates a common basis
for thought organization through time. In language this organization underlies
the flow of ideas, in music the flow of the emotions that music expresses.
As an illustration we can look at another Mozart piano sonata, Köchel 331,
shown in Figure 8.2 with only the treble voice. The slurs indicate phrases that
66 Thoughts and Their Properties

correspond to spoken intonation units, expressing brief and continually shifting

foci of active consciousness. Each of the four lines captures a sequence of those
phrases corresponding to a spoken sentence. The entire excerpt corresponds to
a spoken topic, with lines one and two and then lines three and four constituting
subtopics. The first subtopic ended on the tonic A, and the entire topic ended
more definitively with the sixteenth notes preceding the final A. The majority of
phrases occupied about a second and a half, corresponding to the length
of spoken intonation units, but each line began with two longer figures.
This resemblance between the hierarchical structures of speech and music is
deserving of further study.
Part III

Verbalization Illustrated
9 From a Thought to a Sound in English

Jane and Molly were college students in their early twenties. As Jane was
talking on the telephone with her boyfriend Fred, he heard Molly say in the
background:
molly: I’m gonna go over there.

which led to the following exchange:

fred: Where’s Molly going?
jane: She’s going next door, to Steve’s?
He’s having some people over.
fred: Oh.
At which point Jane said:
jane: Yeah we’re gonna go hang out there.
This chapter traces Jane’s last sentence from its beginning in her thought to the
sound she uttered. We can look in turn at its semantic, syntactic, and phonolo-
gical manifestations, at the processes responsible for each of them, and at the
reasons behind those processes.
A book is confined to visual representations, but neither Jane’s thought nor
her sound were anything you could see. At best we can only attempt to describe
certain aspects of her thought, while representing her sound in one or more of
the various ways that are available for that purpose, including the orthographic
representation above and any of various acoustic representations such as the
display of fundamental frequency at the top of Figure 9.1.
Jane’s first word, yeah, was a “discourse marker” (Schiffrin 1987), which in
this case did not occupy a separate focus of consciousness but was included
with the rest of Jane’s utterance within a single intonation unit. Yeah typically
functions as a positive response to a yes–no question. In this case the question
had not been explicitly asked, but Jane evidently assumed that Fred was
wondering whether she and Molly were both going to Steve’s and not just
Molly. The next word, we, expressed an idea that was given, but it was spoken
with a high pitch because it contrasted Jane and Molly as a pair with Molly

69
70 Verbalization Illustrated

yeah we’re
gonna go out
hang
there

Figure 9.1 Tracking the fundamental frequency of Jane’s sound

{event} {location} {participants}

Figure 9.2 The ideas included in Jane’s thought

alone. Contrastive ideas are expressed with prominent prosody even when they
are given.
There is no accepted way to picture a thought, but Figure 9.2 shows the three
ideas that were active in Jane’s thought at this moment. The event was the idea
of hanging out at Steve’s, its location was Steve’s apartment, and the partici-
pants in it were Jane and Molly. The curly brackets are a way of locating these
three ideas in the realm of thought. The smaller size of the location and
participant ideas shows their status as given, as opposed to the new status of
the event. These three ideas were of course not independent of each other but
combined to form the totality of Jane’s thought. The event of hanging out was
the nucleus of that thought, but it called for the accompaniment of a location
and participants.
A successful conversation requires not only that speakers verbalize their own
thoughts but also that they imagine the thoughts of their interlocutors, integrat-
ing what they themselves are thinking with what they imagine those others to
be thinking. Conversationalists vary in the extent to which they fulfill this
requirement, occupying a position on a scale from maximum empathy to
maximum solipsism. The difference can appear in the choice of topics and
the ways topics are elaborated, but also in the use of pronouns, which ideally
express ideas of entities that are assumed to be already active in the other
person’s thought, thus establishing them as given. That assumption depends on
the prior activation of the ideas in question, as when the idea of Jane was
assumed to be active in Fred’s thought because she was his conversational
 From a Thought to a Sound in English 71

thought

semantic
structure

Figure 9.3 From a thought to a semantic structure

partner, the idea of Molly because of her presence in Jane’s room, and the idea
of Steve’s apartment because of Jane’s earlier mention of it.
There were several features that created a distance between Jane’s thought
and what would have been directly conveyed by her words if they were taken
literally. The words hang and out in themselves bore no direct relation to
what Jane was thinking, which included neither hanging nor being out in any
literal sense. The word gonna is of course a way of expressing a future
intention, with its origin in the sequence going to, but the literal meaning of
those words also bore no direct relation to Jane’s thought. On a subtler level
the word go in the sequence go hang out was more closely related to her
thought – she was indeed planning to go somewhere – but in this case the go
expressed an “andative” orientation of the hanging out, and not the separate
event it would have been if she had said we’re gonna go there. The example
thus illustrates in several ways a separation of Jane’s words, interpreted
literally, from what she was thinking. These words, furthermore, were not
the only options available to her. With roughly the same thought she might
have said, among other possibilities:
We’re planning to mess around there.
The two of us are gonna spend some time over there.
Each such alternative, with others that can be easily imagined, would have
packaged Jane’s thought in a different way while retaining its essential nature.

From a Thought to a Semantic Structure

We can begin tracing the path from Jane’s thought to her sound by exploring
the way her ideas were organized into a semantic structure (Figure 9.3). Once
those choices were made, the thought was committed to a particular path of
verbalization whose representation in syntax and then in phonology was for
the most part already determined. Semantic choices, in other words, are
largely responsible for giving language its shape. These choices include
selection, categorization, orientation, and combination, each of which we
can examine in turn.
72 Verbalization Illustrated

thought
|selection|

semantic
structure

Figure 9.4 Selecting parts of a thought to be verbalized

Selection
A thought probably always contains more than is, or perhaps even can be
represented in language. We don’t say everything we think. An essential first
step in verbalization is thus to choose how much and what parts of a thought
will proceed to be represented with sound. On a large scale that choice can be
a commitment to an over-arching topic. In this case there had already been
discussion of Molly’s relation to Steve, a topic that was now “in the air.”
Speakers navigate through a topic with a succession of thoughts that are
often guided by interaction with an interlocutor. Figure 9.4 shows this selection
process as the first step toward a semantic structure.

Categorization
The anticipated event of hanging out at Steve’s was going to be particular in
time and space and its particularity called for the next step. It would obviously
be impossible for every particular idea to be associated with a particular sound.
Franz Boas recognized this problem a century ago: “Since the total range of
personal experience which language serves to express is infinitely varied, and
its whole scope must be expressed by a limited number of phonetic groups, it is
obvious that an extended classification of experiences must underlie all articu-
late speech” (Boas 1963 [1911]: 18).
Not only is the number of thinkable ideas vast if not infinite, but even if the
assignment of a particular idea to a particular sound were less problematic than
it is, there would be no way a listener could know which idea was associated
with which sound. Thought and language solve these problems by interpreting
particular ideas as instances of categories. A category serves two functions.
On the one hand it triggers expectations regarding an idea, expectations that are
typically associated with instances of that category. Perhaps we can lift it,
perhaps we can eat it, perhaps we can sit on it, perhaps it barks and chases
squirrels. But at the same time a category provides each of its instances with
a sound that can be used to verbalize it.
Edward Sapir emphasized the fact that a category can apply to an unlimited
range of particular instances:
 From a Thought to a Sound in English 73

We must cut to the bone of things, we must more or less arbitrarily throw whole
masses of experience together as similar enough to warrant their being looked upon –
mistakenly, but conveniently – as identical. This house and that house and thousands of
other phenomena of like character are thought of as having enough in common, in spite
of great and obvious differences of detail, to be classed under the same heading. In other
words, the speech element “house” is the symbol, first and foremost, not of a single
perception, nor even of the notion of a particular object, but of a “concept,” in other
words, of a convenient capsule of thought that embraces thousands of distinct
experiences and that is ready to take in thousands more. (Sapir 1921: 13)

Categorizing an idea lets us feel we understand it by placing it in the

company of other ideas that are categorized in the same way. “As soon as the
category is at hand, we instinctively feel, with something of a sigh of relief, that
the concept is ours for the handling. Not until we own the symbol do we feel
that we hold a key to the immediate knowledge or understanding of the
concept” (Sapir 1921: 17). A search for this feeling of satisfaction
might even sometimes lead us to carry with us a guidebook that will help us
experience it, for example with respect to a particular flower or bird.
Categories have various properties that have been described in numerous
other works (e.g. Rosch et al. 1976; Rosch 1978; Lakoff 1987). They may form
a hierarchy such as animal > dog > spaniel > springer spaniel. One of those
levels (probably in this case dog) may function as a basic or default level.
Certain categories may be favorite or typical representatives of that basic level,
as robins are typical birds whereas penguins are not. The fact that an idea may
be categorized differently on different occasions shows that categorizing does
not force an idea to adhere consistently to a single choice.
An idea may favor a particular categorization to a greater or lesser degree,
establishing its degree of “codability” (Brown 1958). At one extreme a person
may without hesitation characterize a particular idea as an instance of the dog
category. At the other extreme a person may have a perfectly clear idea of
something while nevertheless lacking a category that will provide it with
a sound. Most of us are familiar with the small plastic sheath on the end of
a shoelace that allows it to be passed through a small hole. A few people may
have learned the word aglet, but probably most of those who are completely
familiar with this object have never associated it with that or any other sound.
There is a resemblance here to the tip-of-the-tongue experience, but in this case
the sound is not just momentarily inhibited but unknown.
Categorizing illustrates repeatedly on a small scale the two components
of understanding that were discussed in Chapter 2: observing something
particular and inventing something general to accommodate it. We are
constantly observing particular events, states, and entities, but our under-
standing of them would be hopelessly limited if we stopped there.
The thousands of categories stored in our minds are thousands of little
74 Verbalization Illustrated

thought
selection
|categorization|

semantic
structure

Figure 9.5 Adding categorization to the creation of a semantic structure

thought
selection
categorization
|orientation|

semantic
structure

Figure 9.6 Adding orientation

theories, without which everyday experiences would be a mass of confu-

sion. We can now add categorization as a second step on the path from
a thought to a semantic structure (Figure 9.5).
Jane interpreted the event she had in mind as an instance of a category
whose semantic representation will be shown here as [hang out]. The square
brackets are a way of showing that a categorized idea does not, at this
semantic stage, consist, for example, of the word hang followed by the
word out but is a unitary semantic element directly related to an idea.
We have seen also that the idea of this event was associated with the idea
of an entity comprising two individuals, Jane and Molly, one of whom was
the speaker, and thus this idea was categorized as what can be shown as [first
person plural]. The third idea, the location of the hanging out, was categor-
ized in a way representable as [distal], capturing a location some distance
from where the speaker was at the moment.

Orientation
We saw in Chapter 5 that ideas are oriented within a multidimensional matrix of
time, space, epistemology, emotions, social interactions, and context. It was
mentioned that languages differ in the importance they assign to these dimen-
sions, with English devoting considerable attention to time. Figure 9.6 adds
orientation to the processes that lead to a semantic structure.
 From a Thought to a Sound in English 75

Table 9.1 Semantic choices in Jane’s utterance

Idea type Categorization Orientation

event [hang out] new
future
andative
participants [first person plural] given
contrastive
location [distal] given

How, then, were the ideas in Jane’s utterance oriented? Mentioned earlier in
this chapter was empathy, the ability of a speaker to imagine the thoughts of
a listener, and Jane was seen as assuming that two of her ideas – the idea of the
event’s participants [first person plural] and the idea of its location [distal] –
were already active in Fred’s consciousness because they had been activated in
the preceding talk. They were thus oriented as given. In addition, however, the
hanging-out event was to be performed, not by Molly alone, but jointly by
Molly and Jane, and thus Jane’s [first person plural] was oriented as contrastive
(contrasting Molly and Jane as a pair with Molly alone).
Jane assumed that her idea of hanging out was not already active in Fred’s
consciousness and that she was about to say something that would activate it;
thus it was oriented as new. Other orientations of this event included the fact
that it was intended for the future – here we can label it simply future – and that
it would require moving to a different location, or andative. How Jane categor-
ized and oriented her ideas is summarized in Table 9.1, showing the idea types,
how they were categorized, and how they were oriented.

Combination
The ideas of the event, its participants, and its location were not floating
independently in semantic space but were combined in a hierarchy of immedi-
ate constituents, as explored years ago by Rulon Wells (1947) and subsequently
incorporated in some fashion in many theories of language structure.
The central idea of hanging out was amplified with [distal], and its participants
were added as [first person plural] as shown in Figure 9.7. These three ideas
were not temporally ordered, either in thought or in semantic structure, and
Figure 9.7 can be imagined as a mobile with parts that swing freely in semantic
space.
Figure 9.8 now adds combination as the last step in the creation of a semantic
structure from a thought:
76 Verbalization Illustrated

[first person plural]

given
contrastive [hang out] [distal]
new given
andative
future

Figure 9.7 Combining ideas

thought
selection
categorization
orientation
|combination|

semantic
structure

Figure 9.8 The processes contributing to a semantic structure

From Semantics to Syntax

Semantic choices are seldom passed on directly to phonology but are almost
always reformulated in ways that replace semantic structures with what we
know as syntax (Figure 9.9).
If one views language within the framework developed here, one may well
ask why this syntactic stage intrudes itself at this point. Why aren’t semantic
structures represented by sounds directly? The answer lies in the fact that
languages are always changing, and that innovations in the semantic realm
often occur without corresponding changes in already established representa-
tions in sound. Those changes fall under the headings of “idiomaticization” and
“grammaticalization.” These similar processes create situations in which newly
derived semantic elements continue to be represented by the structures from
which they were derived. Those earlier structures are now left stranded with
“quasi-semantic” elements whose direct relation to thought has been distorted.
It is these quasi-semantic elements that proceed to find a representation in
sound.
Syntax has something in common with a sailor suit, the traditional uniform
of enlisted men in the United States Navy as pictured in Figure 9.10. Parts of
this uniform were and still are directly functional in keeping the sailor warm
and covering parts of his body that are meant to be covered. But at least two
 From a Thought to a Sound in English 77

thought

semantic
structure

syntactic
structure

Figure 9.9 From semantics to syntax

Jumper flap

Bell bottoms

Figure 9.10 Parts of a sailor suit

parts of the uniform are no longer directly functional. The jumper flap is
there because in earlier times sailors greased their hair and this flap kept the
grease from soiling their jumpers. The origin of bell bottom trousers is more
obscure. One explanation is that they made it easier for sailors to pull up their
trousers from the bottom when they swabbed the deck or climbed the rigging.
“Whatever the real reason, reliable documentation validating any assertion
was lost long ago. What is true, however, is that the bell bottom trouser has
consistently remained as a part of the identifiable occupational dress of
sailors throughout modern history” (Dervis 2000). A sailor suit is thus
a mixture of functional and no longer functional elements and syntax is
like that too, combining elements that are semantically relevant and thus
bear a direct relation to thought with elements whose connection to thought
has been distorted and whose explanation can only be found in a language’s
history.
How then did it happen that the semantically unitary element [hang out]
came to be represented syntactically with the words hang out? The English
language is notorious for its phrasal verbs: constructions in which a verb is
followed by a particle or preposition. The words hang and out are promis-
cuous in their availability to such constructions: hang on, hang up, hang
78 Verbalization Illustrated

thought

semantic
structure
|literalization|

syntactic
structure

Figure 9.11 Literalization as a contributor to syntax

around, hang in there, along with bring out, look out, carry out, not to
mention nouns like hangout, hangup, and hangover. Jane’s use of hang out
seems to have originated in twentieth-century teenage slang as a way of
expressing the idea of spending time somewhere without a clearly defined
purpose.
The distance between an idiomatic (semantic) meaning and a literal (quasi-
semantic) meaning can vary considerably. Jane’s thought lacked the idea of
vertical suspension conveyed by hang as well as the container-related meaning
of out, but the relation of syntax to thought can be closer, as with let it all hang
out. It has been observed that when the relation is relatively close, people may
remain peripherally aware of the literal meaning (e.g. Gibbs 1994), or what
I have called a “shadow meaning” (Chafe 2008: 265). The idiom spill the beans
has become a canonical example. It expresses the idea of prematurely disclos-
ing a secret, but while that idea occupies center stage one may at the same time
experience a visual and perhaps even an auditory image of beans being spilled
(Lakoff 1980). The presence and nature of a shadow meaning may vary with
the idiom and from person to person, but the point is that quasi-semantic
syntactic elements sometimes feed back into the way a thought is experienced
without replacing its primary meaning, in this case the idea of disclosing
a secret.
We can begin to form a syntactic structure by replacing a semantically
categorized idea with the literal form of an idiom, a process called “literaliza-
tion” in Figure 9.11. The result of this step can be shown by removing the
square brackets around [hang out], replacing this unitary semantic element with
the words hang and out, and by replacing [first person plural] with the pro-
noun we.
Grammaticalization is a process that literalizes the orientation of an
idea. It applies here to the orientation of [hang out] as future and andative.
Future has by far the more complex history and it provides an excellent
example of how far syntax can be separated from semantics as a result of
language change. The construction be going to has been discussed
 From a Thought to a Sound in English 79

frequently as an illustration. Hopper and Traugott (1993) cite the fifteenth-

century attestation found in the Oxford English Dictionary: “Thys onhappy
sowle . . . was goyng to be broughte into helle for the synne and onleful
lustys of her body.”
This construction, however, could not have arisen if the language had
not already acquired the use of be . . . -ing as a way of expressing the so-
called progressive aspect. There is a useful discussion of the history of
this construction in A Middle English Grammar by Mustanoja (1960).
In Old English there were evidently two constructions. One was
a participial construction with the -end ending, the use of which was
discussed in Nickel (1966). The other was a gerundial construction with
the -ing ending accompanied by the preposition on. In the thirteenth
century the -end ending merged with the -ing ending and the distinction
between the two constructions was blurred. Meanwhile the on was
reduced to a-, as in he went a-hunting. The form without this a-, simply
he went hunting, gradually replaced the other, leaving a-hunting as the
quaint relic it is today.
With be . . . -ing now in place as the way of verbalizing the progressive
aspect, statements like I’m góing, to buy bréad (the purpose of my going is
to buy bread, expressed in two intonation units) were reinterpreted as
a future orientation that was expressed in a single intonation unit without
the accent on going. “The change depended on an inference of futurity
from purposives” (Hopper and Traugott 1993: 3). If I am going somewhere
to buy bread, the buying will necessarily take place in the future. English
already possessed a future orientation that was realized syntactically with
the auxiliary will, as in we will hang out there. The change just described
introduced a new and different future orientation with the purposive flavor
observable in we’re going to hang out there. The syntactic realization of
this now very common future orientation arose in the ways just described,
illustrating how syntactic structures can result from a series of semantic
innovations.
The syntactic representation of the orientation andative is easier to describe:
simply the word go contracted from the sequence go and (Jane might have said
we’re gonna go and hang out there). For this andative meaning the English
language offers another alternative: go . . . -ing as in go bowling and go skiing,
which appears to be largely restricted to recreational activities that include go
shopping. The remaining orientations in Jane’s utterance – given, new, and
contrastive – were realized prosodically in ways to be described later.
The syntactic structure of Jane’s utterance can now be represented as shown
in Figure 9.12, where the andative and future orientations have been literalized
as discussed above.
80 Verbalization Illustrated

we
given
contrastive are going to
go
hang out there
new given

Figure 9.12 The syntactic structure of Jane’s utterance

thought

semantic
structure

syntactic abstract
symbolization
structure phonology

Figure 9.13 From syntax to abstract phonology

From Syntax to Phonology

We have now reached the point in verbalization where a syntactic structure
on the thought side of language comes to be represented by a phonological
structure on the sound side (Figure 9.13). The bottom tier of this figure is
the place where thought-based and sound-based representations are linked
through symbolization (Chafe 1967; Langacker 1990). The importance
that is so often assigned to syntax may stem in part from its location at
this point where the thought–sound divide is bridged. The bottom tier is
also the place where a direct relation to both thought and sound has been
distorted.
Here a word or construction relates to both a thought and a sound: to
a thought on the left and a sound on the right. For those of us who have grown
up in a literate society it may be virtually impossible to think of a word in
isolation from its sound or even its spelling. In Chapter 3 Earl could hardly be
blamed for associating the word plinth with the way it sounded or was
spelled. His conception of that word placed it entirely in the right-hand
column of Figure 9.13. Very differently, Harry and Lotte Webster conceived
of the Onondaga word onų́ :dó:da’ entirely in the left-hand column, with no
clear idea of how to spell it and with minimal interest in how it sounded.
It was, to be sure, the sound that activated the idea in their minds, but they
 From a Thought to a Sound in English 81

thought

semantic overt
structure phonology

syntactic abstract
structure symbolization phonology

Figure 9.14 From abstract to overt phonology

were no more conscious of that fact than Earl was conscious of the meaning
of plinth.

From Abstract to Overt Phonology

Proceeding further along the verbalization path, we can see why it is
necessary to distinguish overt from abstract phonology (Figure 9.14).
In our example the distinction is realized in the fact that we are is
expressed in sound as we’re and going to as gonna. These differences
can be attributed again to language change, in this case the contraction of
longer sequences into shorter ones. The historical basis of distinguishing
semantics from syntax is thus mirrored on the sound side in distinguishing
overt from abstract phonology. It is also apparent that sound changes fail
to respect the thought side of language. The contraction we’re combines
the idea of the participants in the event, we, with are, the first word of are
going to, the verbalization of future. The contraction gonna combines only
the second and third words of are going to. This independence of sound
change from thought-based structures will be demonstrated again in
Chapter 10.

From Overt Phonology to Sound

Figure 9.15 brings us to the end of the verbalization path: the sound that
expressed Jane’s thought. Figure 9.16 repeats Figure 9.1 by tracking the
fundamental frequency of her sound, where the rising pitch at the end of
hang out and the conspicuous fall on there are easily visible. An overall
decline in pitch is shown by the dashed line. It can also be seen that Jane’s
utterance gradually decelerated, the rapidly spoken syllables of we’re
gonna followed by a slower go, a still slower hang out, and finally
a much prolonged there. This overall decline in pitch combined with
82 Verbalization Illustrated

thought sound

semantic overt
structure phonology

syntactic abstract
structure symbolization phonology

Figure 9.15 From overt phonology to sound

yeah we’re
go
out
gonna
hang
there

Figure 9.16 The fundamental frequency of Jane’s sound

thought sound

semantic prosody and

symbolization
structure gesture

Figure 9.17 Semantic features symbolized directly by prosody and gesture

deceleration is characteristic of many, although certainly not all intonation

units.
Figure 9.17 makes the more radical suggestion that prosody and
gesture can bypass the bottom tier of Figure 9.15, allowing semantic
elements such as given, new, and contrastive to be symbolized
directly by prosody, and perhaps these or other semantic elements to be
symbolized directly by gesture (Bolinger 1983; McNeill 2005). In other
 From a Thought to a Sound in English 83

words, a separate syntactic structure may be irrelevant for certain aspects

of prosody and gesture, with gesture resembling prosody in a different
form.

Summary
The brief utterance that was discussed in detail in this chapter illustrated the
stages that intervene between a thought and a sound: semantic structure,
syntactic structure, and both abstract and overt phonology. We turn now in
Chapter 10 to the way a very different type of language would verbalize a very
similar thought.
10 From a Thought to a Sound in a Polysynthetic
Language

They don’t have articles, and they wouldn’t know how to compensate for this
lack of articles either with case or with prepositions, which they also lack.
Nevertheless, they have other ways of establishing and maintaining clarity of
discourse . . . They have only a few adverbs and conjunctions, but in fact they
have an astonishing wealth of verbs. In their language almost everything is
a verb, or can become one. Jean-André Cuoq (1866: 87)

Cuoq was a French priest who was describing the Mohawk language, which he
knew from his missionary work in Canada. Comparing it with his native French
and with other languages familiar to him, he found it lacking in the grammatical
elements a European linguist would expect a language to have. He was at the same
time impressed by the fact that its speakers had no problem, in spite of what might
have seemed major handicaps, in “establishing and maintaining clarity of dis-
course,” noting that they compensated for what they lacked with “an astonishing
wealth of verbs.” No better illustration could be found of the different paths
languages may follow as they move from a thought to a sound. In this chapter we
trace that path in the Seneca language, a close relative of Mohawk.
Exact translations are never possible, and the thought expressed by Jane in
Chapter 9 as we’re gonna go hang out there cannot be reproduced exactly in
Seneca. Discussed in this chapter, however, is a reasonable approximation.
Seneca of course lacks the idiom hang out, which is replaced here with
a Seneca idiom that can be translated visit. Unlike Jane’s utterance this example
was not taken from spontaneous speech but it is a realistic thing to say, even
though in actual speech it would almost certainly be accompanied by several
particles: short words that would position it within the ongoing discourse. Let
us suppose that it was uttered by a person named Andrea, who said:
Hëyagyajö́ ’se:nö’.5
“We will go visit there.”

5
Seneca spelling follows an orthography that has been in use since the 1970s. The dieresis in ë and ö
shows nasalization of those two vowels, jokingly referred to as nostrils. The two apostrophes
represent glottal stops, and the colon after the e marks it as a lengthened vowel. The acute accent
mark over the fourth vowel shows a syllabic prominence that is realized especially in higher pitch.

84
 From a Thought to a Sound in a Polysynthetic Language 85

thought sound

semantic surface
structure phonology

morphosyntactic reconstructed
symbolization
structure phonology

Figure 10.1 From a thought to a sound by way of morphosyntax

Immediately striking is the contrast between this one Seneca word and the five
words of the English. It has often been said that a polysynthetic language expresses
in one word what a language like English would express in an entire sentence.
It thus becomes necessary in this chapter to trace the path from a semantic structure
to a morphological structure, the internal structure of a word. The term “morpho-
syntactic” in Figure 10.1 leaves room for both morphology and syntax, and later in
this chapter we will see how both may be relevant.

Andrea’s Thought
We saw in the previous chapter that Jane’s thought included the idea of two
participants in the hanging-out event, Jane and Molly, verbalized with the word
we. Chapter 5 mentioned how ideas of particular events necessarily include
ideas of entities that initiate them, are affected by them, or otherwise participate
in them. That observation applies to all languages, but it is realized directly in
a polysynthetic language, where the participants in an event are tightly pack-
aged together with the event itself within a single word.
If a Seneca verb corresponds to an English sentence, one might at first
suppose that the range of thoughts expressed by such a verb would rival the
range of thoughts that can be expressed in an English sentence. But that is not
the case. A Seneca verb does make it possible to verbalize a number of things
that are relevant to an event, including not only its participants but also its
aspectual properties, its location in space and time, whether it is distributed
across several locations, its relation to reality, whether it is a cause or an
instrument, whether it benefits someone, whether it involves going some-
where to do something, whether it is performed while walking, whether it
happens easily, or whether it is the culmination of a series of other events.
Selections from these numerous possibilities are packaged tightly within the
verb, where each element has a form and position determined by the
language’s morphological patterning.
But this list of morphological possibilities is only a small subset of the
open-ended opportunities that are offered by combining words in sentences,
and later in this chapter we will see how the nearly infinite variety of possible
86 Verbalization Illustrated

{event} {event} {location} {participants}

{participants}
{mode and aspect}

Figure 10.2 Andrea’s thought compared with Jane’s

thoughts can lead speakers of a polysynthetic language to supplement their

morphology with options provided by their syntax. But it is still worth asking
whether the thoughts of a Seneca speaker have at least a tendency to conform to
the options offered by the language’s morphology as listed in the previous
paragraph. Does the easy availability of those options bias a speaker toward
thinking in those terms? And do those options feed back into the ways Seneca
speakers think? The larger question of whether speakers of different languages
think differently will be confronted in Chapter 13, but for now we can focus on
the present example.
Most Seneca words are built on a base that offers a large number of lexical
possibilities. This base is usually accompanied by one or more participants and
also by a mode and aspect. In our example Andrea’s thought is likely to have
contained at least the three ideas on the left in Figure 10.2, where placing them
in a single column shows their simultaneous presence within the word, as
compared with Jane’s thought on the right as copied from Figure 9.2, where
they were distributed across three words.

Andrea’s Semantic Structure

Andrea’s semantic structure had as a base her categorization of the event as
[atyö’se]. This notation follows the convention introduced in Chapter 9 where
Jane’s semantic category was shown as [hang out], the square brackets showing
here that [atyö’se] is a unitary semantic category and not at this stage composed
of the separate parts described below. The participants in this event, corre-
sponding to Jane’s [first person plural], can be shown as [exclusive dual agent].
But whereas [first person plural] was assigned in English to the separate word
we, the Seneca participants were included in the same word as the event and
will thus be represented directly below it, as in Figure 10.3. “Exclusive dual”
means that the participants excluded the listener (Fred) and that there were two
of them, Andrea and Molly. “Agent” means that those two would be the
instigators of the event. Seneca also requires that an event be specified in
terms of mode and aspect, shown here as [predictable event].
The idea of this event was semantically oriented as shown in Figure 10.3 by
placing andative and distal without brackets below the three ideas.
The andative orientation functions as it does in English: the event required
moving to a different location. The event was also oriented as distal, or located
 From a Thought to a Sound in a Polysynthetic Language 87

[atyö’se]
[exclusive dual agent]
[predictable event]
andative
distal

Figure 10.3 Elements of Andrea’s semantic structure

at a distance from the speaker. This orientation contrasted with the separate idea
that was expressed in English with the word there.
The properties just described – a base, its participants, and its mode and
aspect – are obligatory for most Seneca verbs, and in this case they were
supplemented with the andative and distal orientations.
An important question arises at this point. English does not say explicitly that
there were two people and that the listener was excluded, but to what extent are
people conscious of knowledge that is not directly verbalized in their language?
In this case Jane obviously knew that the participants in the hanging-out event
were two people and that neither of them was Fred, even though that knowledge
was not represented in the sound she uttered, whereas it was represented in
Andrea’s sound hëyagyajö́ ’se:nö’. We saw in Chapter 9 that a thought nearly
always includes more than is expressed in language, but the question now is
whether Jane’s knowledge that there were two people going to Steve’s, neither
of whom was Fred, occupied a less active state in her mind than the corre-
sponding knowledge in Andrea’s mind, where it was expressed overtly in the
sound she uttered.
In my own imperfect attempts to speak Seneca I have sometimes found
myself using a plural form of a Seneca verb when I should have used a dual,
even though I knew that only two people were involved. Introspection suggests
that it would have required an extra cognitive effort for me to pay attention to
the fact that there were only two people. It was something I knew, but that
knowledge was not as active in my consciousness as it would have been for
a fluent Seneca speaker. Observations like these suggest the value of distin-
guishing Jane’s tacit knowledge from Andrea’s overt knowledge (cf. Polanyi
2009). How the brain distinguishes overt, verbalized knowledge from tacit
knowledge is a question that clearly deserves further study.

Andrea’s Morphological Structure

How were these semantic elements represented morphologically? Just as the
semantic orientation future acquired its syntactic representation as be going to
in the historical stages described in Chapter 9, the morphological representa-
tion of [atyö’se] can also be traced to its history. First it came to be represented
88 Verbalization Illustrated

arrive + benefactive + middle voice

exclusive agent
dual
event
predictable
andative
distal

Figure 10.4 Elements of Andrea’s morphology

Table 10.1 A verb template with morphology above and phonology below
the line

Exclusive
Distal Future agent Dual Middle Arrive Benefactive Andative Eventive
h ë yak y at yö ’se hn a’

with a verb root meaning “arrive” accompanied by a “benefactive” derivation.

In other words “visit” was expressed literally as “arrive for someone’s benefit.”
That idea, however, was transitive (visit somebody) and in our example the
visiting was intransitive, a fact that was expressed by adding a marker of
“middle voice” (e.g. Kemmer 1993). The meaning “exclusive dual agent”
was represented by an element meaning “exclusive agent” to which “dual”
was added separately. These changes led to a morphological structure that
contained the elements shown in Figure 10.4.
What was then passed on to phonology is shown in the linear template above
the horizontal line in Table 10.1. The items above the line belong to the
thought-based face of morphology, those below the line to the sound-based
face. Removing the spaces from the elements below the line gives us the
reconstructed phonological form hëyakyatyö’sehna’.

From Reconstructed to Surface Phonology

Through both internal and comparative reconstruction but also through the
fortuitous existence of French–Seneca and Seneca–French dictionaries from
about 1700, we know quite a bit about the state of this language 300 years ago.
During the centuries since then the language underwent more than fifty sound
changes, most of them during the eighteenth century. As a result the difference
between a reconstructed phonological structure and the sound of a word today
can be considerable. The following changes include only those that are relevant
to this example:
 From a Thought to a Sound in a Polysynthetic Language 89

Reconstructed form: hëyakyatyö’sehna’

Accent the fourth syllable: hëyakyatyö́ ’sehna’
Delete h before n and lengthen the preceding vowel: hëyakyatyö́ ’se:na’
Change k to g and t to d, before y: hëyagyadyö́ ’se:na’
Change dy to j: hëyagyajö́ ’se:na’
Change a to ö after n: hëyagyajö́ ’se:nö’

The last line contains the form of this word that is now realized in sound, and it
is the way Andrea expressed her thought.
The stages described above show how a linguist might account for the
steps that led from Andrea’s thought to the sound she uttered, but there is no
reason to suppose that Andrea herself followed those steps. To a large extent
Seneca speakers remember words as wholes, not as the morpheme sequences
identified by linguists. They are aware of the meanings that are contained
within a word, but they are often unable to associate parts of a word with
those meanings. If that is the case, how do they manage to create a word they
have never before spoken or heard? In fact they do that less often than one
might suppose. Verbal creativity is an ability that varies with different
speakers, and creative eloquence is admired. A creative speaker almost
certainly does not follow the stages described above but relies instead on
what might be termed a “folk morphology” consisting of ad hoc patterns
established individually and analogically on the basis of words that are
already familiar.

Establishing and Maintaining Clarity of Discourse

At the head of this chapter Father Cuoq was quoted as saying that a language of
this type was able to establish “clarity of discourse” by taking advantage of its
“astonishing wealth of verbs.” We can end this chapter by looking at one of the
ways that is accomplished. Despite the wealth of resources offered by Seneca
morphology, speakers often find them inadequate to express what they are
thinking. In that case Seneca syntax makes use of a pattern that allows the
information within a verb to be amplified in order to convey more of what is
included in a thought. As an example we can start with a thought that might
have been expressed in English in this way:
There were a father and son who came to visit.
We can compare that English sentence with the following Seneca:

Né:ne:’ wa:ya:jö’s neh, yadátawak.

those they (masculine dual) visited namely a father and his son
“They visited: a father and his son.”
90 Verbalization Illustrated

The verb wa:ya:jö’s captured the idea of visiting by two males. But that verb
failed to convey everything the speaker was thinking, so he went on to add the
idea that the two visitors were a father and his son. The third Seneca word, the
particle neh, signaled that the information in the verb was about to be amplified by
what would follow. The translation “namely” captures this function reasonably
well, although in writing one might simply use a colon, as followed the word
visited in the free translation here. Most Seneca kinship terms are verbs, and the
final word, yadátawak, can be translated literally “they were father and son to each
other.” This amplification repeats a reference to the participants who were
included in wa:ya:jö’s. The y- in both verbs showed that the participants in both
the visiting event and the father–son relationship were two males, thus explicitly
linking the two ideas. Amplification constructions like this are a common feature
of Seneca syntax (Chafe 2012, 2015), a common way of establishing “clarity of
discourse.”
One sometimes reads that children learn approximately 14,000 words by the
age of six. Many of those statements are traceable to studies that were compiled
in a single source (Templin 1957). Those studies mentioned the problem of
deciding whether, for example, look, looks, looked, and looking should count as
one word or four. Presumably children learn the meaning and form of the
lexical item look, but also learn the inflections that can be applied to a large
number of such items. Irregular items like go and went can of course only be
learned separately. With Seneca the number of possible inflections of a single
verb base is greatly magnified, reaching into the hundreds, and irregularities are
extensive, with cases resembling go and went on a larger scale. For a child
learning Seneca, much depends on learning a very large number of individual
words at an early age and deriving from them ad hoc patterns of a folk
morphology, each such pattern being applicable to a restricted number of cases.

Summary
Seneca served here as an example of the polysynthetic language type, packa-
ging within a single word a variety of elements resembling those which in
English would be assigned to separate words. Beginning with a thought that
combines what in English would be separate ideas, it categorizes and orients
that thought in its own ways, which leads to a morphological structure in which
those elements are arranged in a linear verb template. The result is then subject
to a number of phonological changes that reflect the language’s history during
the past 300 years. Although a variety of options are available for inclusion
within the meaning of a verb, they are often not enough to cover everything
a speaker is thinking, in which case recourse can be had to a syntactic
construction that amplifies the meanings within a verb in order to express
more of what is contained in a thought.
Part IV

Related Issues
11 The Translation Paradox

If different languages organize thoughts in different ways, how it is possible for

thoughts that are expressed in one language to be expressed in another language
without distorting those thoughts? But we do find people translating from one
language to another all the time, apparently with considerable success. Some
find little if any problem here. Ray Jackendoff wrote that “pretty much anything
we can say in one language can be translated into any other, preserving the
thought that the original language conveys” (Jackendoff 1994: 185). Those
who practice translation may disagree, and yet the prevalence of translation in
today’s world suggests that the statement has some validity. Therein lies the
paradox.
It would be a mistake to take the extreme position that the differences
between languages are so great that they make translation impossible,
or at the opposite extreme to suppose that the differences are so
trivial that there is no problem at all. The question should really be the
extent to which translation is possible. It is an important question, for
practical reasons of course, but also because it forces us to confront
head-on the nature of language differences. There may be no question in
linguistics that highlights so well the need to understand what those
differences are.
Figure 11.1 places thoughts at the heart of translation. The symbol ≈ is meant
to suggest that the thoughts expressed in the source language may be replicated
in the target language to varying degrees. Although it may be impossible for
either thoughts or semantic structures to be totally congruent across languages,
a translator may strive to bring them into as close an agreement as possible.
Syntactic structures are likely to diverge significantly from one language to
another, but syntactic differences themselves need not seriously affect the
success of a translation, whereas semantic differences can be more of
a problem. What is captured in a translation can often be the essentials of
thoughts that were embellished in the source language in ways that fail to
survive the transfer. The essentials that do survive may be enough to serve
the practical requirements of a translation, even when lacking subtleties that
contributed to the total effect of the original. Some literary translations may

93
94 Related Issues

Source Language Target Language

thoughts ≈ thoughts

semantic structure semantic structure

syntactic structure syntactic structure

Figure 11.1 Translation and the design of language

abandon any attempt at equivalence and aim for a target language creation that
has its own values that are more or less loosely connected to the source. That
approach will not concern us here.
There is a difference between interpreting, where the source and target
languages are spoken, and translating, where they are written. Interpreting
may be done on the fly with little or no time to weigh alternatives, whereas
translating gives the translator leisure to explore possibilities. Sometimes,
however, the source language may have begun in a spoken form that is recorded
and then transcribed in writing, after which it is the written transcription that is
translated into the target language. The example below followed a mixed path
of that kind.
Although I could have used examples from more familiar European
languages, this illustration is an excerpt from attempts to translate from
the Seneca language into English. Because the two languages are so differ-
ent in a variety of ways (see Chapter 10), those attempts highlight the kinds
of problems a translation may face. I could also have chosen an example in
which translation was hindered by the fact that certain Seneca ideas are
difficult to express in English simply because English speakers have never
entertained those ideas. There are events and entities in Seneca culture that
are unfamiliar to most English speakers. Problems of that kind, however, can
be overcome either by borrowing words from Seneca or by inventing new
words in English. The example that follows includes differences of another
sort.
For peculiar historical reasons the city of Salamanca in western New York
State is located entirely within the Allegany Seneca Reservation. Since the
nineteenth century white people living in Salamanca have leased their land
from the Senecas, but the amounts they paid for those leases were minimal.
The leases expired near the end of the twentieth century and the Senecas
negotiated for larger payments. As can be imagined, the negotiations aroused
strong feelings on both sides. The speaker quoted here summarized his own
views as follows:
 The Translation Paradox 95

(1) Da:h ne:’ nö:h hë:né:h, And I guess they thought

(2) ëyágwatgá’ negë’ neh yöëdzá’, we would give up the land,
(3) ögyöëdza:dé’ ne’hoh, which was our land,
(4) næ:h da’áöh ne’hóh nö:yawëh. but that couldn’t happen.
(5) Da:h ne:’ gáíö:níh në:gë:h hodínö’kwë’öh. And that was why they got mad.
(6) Næ:h da:digwe:göh, Not all of them,
(7) ne:’ shö:h neh, it was just those
(8) ne:’ neh da:diyëde:íh. who didn’t understand.
(9) Nö:h gaya:söh neh, I guess you could say
(10) honöhsigwé:ót ne’hoh. they were stirring up trouble there.
(11) Da:h negë’ nö:h næ:h, And I guess
(12) ëgáiwíyoak nö:h neh ae’ wëdöshö’öh. things will settle down eventually.
(13) Næ:h ní:’ a:yë:’ sö:gá:’ dé:gë’seh neh, As for me I don’t think less of anyone,
(14) hadinöge:nyö’ neh hadí:nyö’öh. among the white people that live
there.
(15) Ögwádéó’shö’ honötga’de’. A lot of them are my friends.
The speaker was equally at home in both cultures and both languages, and
there was little in the essence of his thoughts that was peculiar to Seneca.
The passage, however, categorized some of his ideas in ways that are not
easily reproduced in English. One example can be found in line 2, where
English provides no exact parallel to the verb ëyágwatga’. I translated it “we
would give it up,” but in other contexts this verb might be translated “supply”
or “provide.” In religious contexts it is used to refer to the Creator’s provi-
sion of essentials for life on earth, where “give up” would hardly be appro-
priate. The general notion is the idea of making something available that will
benefit the recipient. In line 2 it would involve giving up Seneca land, but
English “give up” adds a connotation of being victimized that was not
present in the original.
Line 12 might be more literally translated “the matter will continue to be
good again sometime.” The word ëgáiwíyoak predicted a future state in
which good feelings would prevail, and the -k at the end of the word said that
this state would continue indefinitely. The idea conveyed by “things will
settle down eventually” does not seriously distort what the speaker was
thinking, but settling down eventually is not the same as continuing to be
good.
The most problematic sequence appears in lines 9 and 11. The word
honöhsigwé:ót at the beginning of 10 translates literally as “they are stand-
ing up a fork in it,” a Seneca idiom expressing an idea that is similar to the
English “they are stirring up trouble,” another and very different idiom.
Idioms always involve both a literal and an idiomatic idea: in this case the
idea of sticking a fork in something (literal) and the idea of stirring up
trouble (idiomatic). The thought focused on causing trouble, but the use of
the idiom may have activated at the same time a shadow meaning of sticking
96 Related Issues

a fork in something. In English the idiomatic meaning of stirring up trouble

comes close to what the speaker had in mind, but in English the shadow
meaning of stirring something is very different.
In line 9 the word gaya:söh means literally “it is called,” a metalinguistic
comment on what the speaker was about to say in 11. He was saying in effect
that the idea he was about to express could be captured with the sticking-a-fork
idiom. The translation of 9 as “I guess you could say” might suggest that the
speaker was hedging his evaluation of the situation, but in fact he was hedging
his choice of an idiomatic way to categorize it.
The categorization expressed by the word dé:gë’seh in line 13 illustrates
another problem. The verb root -gë- in the middle of this word can be translated
“see.” Here, however, it is followed by -’s indicating that the seeing adversely
affected another person. The de- at the beginning of this word is a negative
prefix, and thus the literal meaning is approximately “I don’t see it in a way that
adversely affects him.” The translation “I don’t think less of him” approaches
a similar idea from quite a different angle.
In the Seneca there is an absence of past tense marking. The passage as it
stands is ambiguous with respect to its temporal orientation. The speaker
could just as well be describing a current situation and it is only the larger
context, including knowledge of when the lease controversy took place,
that would lead a listener to interpret what was said as describing some-
thing from the past. Without that knowledge the translation could equally
well have begun “And I guess they think we will give up the land.” But
language is always created in a context and the absence of past tense
marking is seldom a problem in practice.
The particle næ:h appears in lines 4, 6, 11, and 13, and like many Seneca
particles it is difficult to translate directly. It emphasizes or highlights an idea,
an orientation that is likely to be captured in English with a raised pitch on the
target word, as indicated below by italicizing those words. In line 13 the force
of the Seneca words næ:h ní:’ was captured more or less accurately with the
translation “as for me.”
but it was impossible for that to happen
not all of them
and I guess
As for me I don’t think less of anybody
As we saw in Chapter 10, the manner in which ideas and their orientations
are combined in Seneca is different from what we find in English. To choose an
example here, the word da:digwe:göh in line 6 contains the following roughly
translated elements in this order: negative + masculine plural + all + stative
aspect. The English translation “not all of them” captures a similar thought, but
fails to specify that there are at least three of them and that they were males.
 The Translation Paradox 97

The same masculine plural designation appears in line 14, referring to the white
people living in Salamanca.
In general, despite the different ways of combining ideas and orientations
that are characteristic of a polysynthetic language, such differences need not in
themselves be a serious impediment to translation. Problems arise from incom-
patible ways of categorizing ideas, from different shadow meanings when they
are present, and from ways of orienting ideas that have no easy equivalences.

Translations over Time

It is of some interest that those who are equally fluent in more than one
language may not always remember for very long the language they used to
express a thought on a particular occasion. Whereas some version of the
thought itself may persist in memory, the way it was verbalized may not.
We are thus left with the following question. Although translations may not
capture all the categorizations and orientations that were chosen to verba-
lize one’s thoughts at a certain time, if those choices were ephemeral and if
the translation succeeds in approximating them in a roughly satisfactory
way, does that mean that, in the end, whatever differences existed between
the original thoughts and the thoughts expressed by the translation tend to
fade, with more or less identical thoughts remaining in the minds of the
original speaker and the recipient of the translation? Does the answer
depend on how well the translation succeeded in conveying the ideas
expressed by the original: the ideas themselves and not the manner in
which they happen to have been verbalized?
To make the question more concrete, would the fact that a bilingual
Seneca–English speaker was exposed either to the Seneca on the left side of
the above example or to the English on the right make any difference in the long
run? Would such a person remember only the ideas that had been expressed, the
top level of Figure 11.1? If specific verbalizations are ephemeral, to what extent
does the memory for thoughts continue to be influenced in the long run by
whatever language they happen to have been expressed in?
This question leads to another. Introspective and anecdotal evidence suggest
that the thoughts conveyed by language may sometimes pass through the three
stages in the minds of language producers and receivers that are shown in
Figure 11.2. At Stage 1 – the moment when language is produced and received
and for a short time thereafter – there is immediate and richly embellished
awareness of the ideas that were expressed and of the particular ways those
ideas were categorized and oriented, including any emotions associated with
them.
Before long at least some of those categorizations and orientations will have
faded from memory, whereas the ideas and any emotions associated with them
98 Related Issues

Source Language Target Language

original partially different
Stage 1 ≈
thoughts thoughts

gradually attenuated thoughts closer

Stage 2 thoughts to the original

Stage 3 emotions only emotions only

Figure 11.2 Possible stages in a translation over time

will have been retained. That is Stage 2 and it lasts much longer, perhaps with
a gradual degradation of detail. It is during that period that we might say that
a translation has succeeded well, because most or all of the ideas that were in
the mind of the source language creator have been successfully transmitted
through the target language. The adequacy of a translation may depend on this
stage, the long-term memory stage, where ideas have been successfully trans-
ferred across the languages.
But occasionally, though certainly not always, there may be a Stage 3.
If Stage 2 retains the ideas and emotions of the original thoughts, Stage 3
retains only the emotions. When enough time has passed after thoughts have
been experienced and verbalized, all that remains may be a memory of the
emotions and attitudes that were at first only one component of the thoughts.
People may remember only how they felt about something, no longer
remembering what that something was. The remarks on the Salamanca lease
controversy provide a possible example. Eventually one might remember only
that the speaker’s attitude toward the controversy was a charitable one and that
he was willing to forget the animosity that had been generated by it. One might
be left with nothing more than the memory that the speaker was “a nice guy,”
without remembering exactly what led to that conclusion.
If this description of Stage 3 has any validity, we are left with a final question.
To what extent are emotions, apart from the ideas with which they are
associated, influenced by language differences? For example, is the attitude
expressed in the Seneca word ëgáiwíyoak, translatable literally as “the matter
will continue to be good,” adequately communicated with the English “things
will settle down”? Although I can hardly demonstrate it here, it could well be
the case that languages differ in the ways they express and communicate
emotions and that this kind of difference is in the final analysis the hardest
problem for a translation to overcome. It may in fact be inseparable from the
aesthetic component of language discussed in Chapter 21, and the aesthetic
dimension is likely in the long run to be the most intractable of translation
problems.
 The Translation Paradox 99

Summary
For thoughts to be verbalized, they must be adjusted to language in four ways,
which in Chapter 9 were termed selection, categorization, orientation, and
combination. For each of the four I gave brief examples here of differences
between the Seneca and English languages, each of those differences
a potential source of distortion in translation. I then speculated on whether
the ephemeral nature of specific choices of linguistic expression may in the
long run leave the recipient of a translation with a closer approximation to the
original ideas, since the specific language chosen to express those ideas will
have faded. More speculative is the possibility that memory sometimes retains
nothing more than the emotional component of the original.
12 Repeated Verbalizations of the Same Thought

A thought may be experienced any number of times, and in fact it is impossible

to conceive of a mental life in which every thought was a brand new one. Every
thought has its own history within the consciousness of a particular individual,
some thoughts being activated again and again while others may enter con-
sciousness only once. The unique history of a particular thought is governed by
the vicissitudes of an individual’s life. Many thoughts are of course experi-
enced silently, and it is only when thoughts are turned into overt sounds that an
outside observer has any access to them.
While Chapter 11 was concerned with thoughts expressed in more than one
language, this chapter concerns thoughts that are verbalized more than once in
the same language (Chafe 1998). The question is the extent to which verbaliza-
tion is the same each time. We can look at examples in which a speaker intended
to express the same thought or thoughts on different occasions, but where the
language was never precisely identical. My interest in this question goes back
to a time when I made multiple recordings of a speech that is delivered at the
beginning of almost every Seneca longhouse ceremony. Those who listened to
this speech believed it to be the same speech every time they heard it, but while
much of the wording was indeed the same, recordings showed that there were
always differences.
To identify places where differences may arise, Figure 12.1 repeats
Figure 9.9, listing the four processes that lead from a thought to a semantic
structure. When a thought is verbalized more than once, although the
thought itself may remain more or less the same, one or more of these
processes may be applied differently, allowing the thought to be associated
with partially different semantic structures and thus with partially different
sounds.
During a dinner-table conversation the speaker, Chris, told how he, his wife,
and their dog had just come back from the California mountains where, during a
blizzard, the dog was hit by a car. They needed to drive some distance to see a
veterinarian, and the snow made it necessary to put chains on their car’s tires.
On the way to the vet’s one of the chains broke, and Chris told how he had
received help in fixing it:
100
 Repeated Verbalizations of the Same Thought 101

thought
selection
categorization
orientation
combination

semantic
structure

Figure 12.1 The path from a thought to a semantic structure

Version 1
Next door to the vet’s was this service station and I got to be friendly with the guy
there, borrowing his pliers for the chains all the time? It was a special tool for fixing
chains.
A short time later another person joined the group and the story was repeated
for his benefit. This time Chris said:
Version 2
I got to be friends with the guys in the service station next door. They didn’t, let’s
see, they had, well okay, so I borrowed their chain pliers, you know, to fix the chain
thing.
The idea of the service station was categorized the same way each time.
The singular guy was changed to the plural guys. But there was another idea
that presented a problem in categorization. In Version 1 it was pliers, but
because this was not a typical instance of pliers Chris went on to explain that
it was a special tool for fixing chains. In Version 2 he hesitated with the words
let’s see, they had, well okay before finally settling on chain pliers and then
adding the explanation you know, to fix the chain thing. The object that
required fixing also presented a problem, leading to the ad hoc expression
the chain thing.
In a project described in Chafe and Danielewicz (1987) a number of dinner-
table conversations were recorded at different times. In two parallel narratives
recorded fifteen months apart, Pamela, who had recently lived in Japan, told
how an intruder had appeared outside the back window of her apartment. The
first time she described this incident she said:
Version 1
I noticed there was a guy, who walked by. And no one ever walked by back there, it
was only a bamboo grove you know. And so I thought hm that’s strange you know,
someone must be lost.
Fifteen months later she recalled the same incident during a conversation with a
different audience, saying:
102 Related Issues

Version 2
And I noticed someone go past my back window there, by this bamboo grove. I
thought that’s strange, no one’s ever back there unless they’re lost.
The intruder’s action was categorized first as walk by and later as go past. But
there was also a difference in the way items were combined. The idea of the
intruder was introduced in Version 1 as prominent new information: a guy.
Because intonation units are normally restricted to a single new idea, the idea of
his walking by appeared in a separate intonation unit: who walked by. In
Version 2 the vague reference to someone was combined in a single intonation
unit with his going past: I noticed someone go past.
Cases like these suggested that it would be interesting to collect more
examples of thoughts that were verbalized on different occasions, and during
my work with the Caddo language a Caddo speaker was asked to retell a story
that dealt with a wildcat and a turkey. As in most Caddo stories its protagonists
were animals who behaved in very human ways. The following are edited
translations of Versions 1 and 2, told three weeks apart (Chafe 1977: 29–31).
Version 1 was noticeably less fluent, with false starts and repetitions. Kishwah
is the name of a popular corn preparation that is usually called in English
parched corn. There is a racier version of this story in which Turkey had sex
with Mrs. Wildcat, disturbing Mr. Wildcat even more!
Version 1
A long time ago it is said, Wildcat was digging roots. Soon Turkey came along.
Turkey said, “My goodness, you’re busy.” “Yes,” he said, and soon he caught him and
plucked all his feathers. He told him, “Go over there where the woman is at my house.
Tell her to cook you for lunch and I’ll eat you.” Turkey said, “All right, I’ll go and tell
her.” He went and when he arrived there he said to her, “You husband is digging roots
over there. He said I should come and tell you to make me kishwah.” “Oh all right.”
She quickly had the kishwah ready. He left and arrived where a fawn was lying. The
fawn said, “I’m sleepy nu.” “My goodness he knows my name.” Soon when Mr.
Wildcat came home he said, “Did you cook the turkey I sent?” Then he followed him.
That is the reason now he eats him raw.
Version 2
A long time ago it is said, Mr. Wildcat was stooping over to dig up roots. He was
making a garden. Soon someone said, “My goodness, you’re busy.” “Yes,” he was
told. “I’m digging. What are you doing?” “Nothing. I was just looking around.” He
said, “You’ll be in my way.” He grabbed hold of him, plucked him, and said, “My
house is over there. My woman is there. Go tell her to cook you for lunch. Then I will
eat you.” Turkey trotted along and arrived where Mrs. Wildcat was sitting, rocking a
child in a hammock. She was singing as she rocked. [Mrs. Wildcat’s cradle song.] Mr.
Turkey said, “He sent me and said you should make me kishwah. When you have
made it I’ll go.” Quickly she made the kishwah and he left. Over there there was a big
log. A tree had fallen over. A fawn was lying there. Mr. Turkey climbed up and stood
on top. Then the fawn yawned. He said, “I’m tired nu.” Mr. Turkey said, “My
 Repeated Verbalizations of the Same Thought 103

goodness he knows my name.” He kept repeating when he yawned, “I’m sleepy nu’.”
Right at that time, when it was noon, Mr. Wildcat arrived at his house and said, “Did
you cook the turkey?” “No,” she said. “He came here and said you told him I should
make him kishwah. After I made it he put it on his back and went along.” Then he
followed him and caught up with him over there. That is why now he eats him raw.

Version 2 selected more thoughts for verbalization than Version 1. Only

Version 2 said that Wildcat was making a garden, and the conversation between
Wildcat and Turkey was more detailed:
Version 1
Turkey said, “My goodness, you’re busy.” “Yes,” he said and kept talking.
Version 2
Soon someone said, “My goodness, you’re busy.” “Yes,” he was told. “I’m digging.
What are you doing?” “Nothing. I was just looking around.” He said, “You’ll be in
my way.”
Version 1 completely omitted the description of what Mrs. Wildcat was doing,
including her song:
Version 2
Turkey trotted along and arrived where Mrs. Wildcat was sitting, rocking a child in a
hammock. She was singing as she rocked. [Followed by Mrs. Wildcat’s cradle song.]
Also truncated was the conversation between Mr. and Mrs. Wildcat:
Version 1
Soon when Mr. Wildcat came home he said, “Did you cook the turkey I sent?”

Version 2
Right at that time, when it was noon, Mr. Wildcat arrived at his house and said, “Did
you cook the turkey?” “No,” she said. “He came here and said you told him I should
make him kishwah.” After I made it he put it on his back and went along.”
The second version contained more detail than the first and it was a more
coherently organized story, illustrating what might be called “the second telling
principle” to the effect that when a topic is verbalized a second time it is likely
to be more complete and more coherent than the first time. Between 1903 and
1905 an anthropologist named George A. Dorsey wrote down English-lan-
guage versions of seventy Caddo stories (Dorsey 1997 [1905]). It is a valuable
collection although we can regret the absence of Caddo-language versions. But
it is also interesting to speculate that the published versions were probably the
first tellings of stories that might have been more complete if they had been told
again.
This story illustrates another feature that is found in other stories in this
tradition. Inserted within the main story was a substory that told of the fawn’s
104 Related Issues

apparent knowledge of Turkey’s name, based on the accidental convergence of

the last syllable of a word meaning “I’m sleepy” (kakkudikí:nu’) with the word
for turkey (nu’). This substory was only loosely related to the main story and
may once have been told separately, later being incorporated as part of this
longer story where it served to create suspense before Turkey was finally
caught and eaten.

Summary
When thoughts are verbalized on more than one occasion they are almost
always verbalized in ways that are partially different each time. The differences
can be located in the semantic structures that are imposed on thoughts through
selection, categorization, orientation, or combination, any of which may apply
differently from one verbalization to the next. When thoughts are verbalized a
second time, they are likely to exhibit more detail and greater coherence. The
larger lessons here are, first, that thoughts themselves are at least partially
independent of any particular way in which they happen to be turned into
sounds, and, second, that different tellings can illuminate different aspects of
the same larger thought.
13 Rethinking Whorf

During one of my earliest days of working with the Seneca language in 1956,
I was asking a Seneca speaker for a few more or less random words and for
some reason I asked how he would say “uncle.” Just at that moment the door
opened and in walked Floyd Lounsbury, my professor in Iroquois studies but
also a leading scholar of kinship systems, who was stopping by the Seneca
reservation on his way to visit his parents in Wisconsin. After my Seneca helper
had left for the day I was treated to an hour’s lecture on Seneca kinship – the
way one refers to and addresses one’s relatives – which differs significantly
from the English way.
Suppose, for example, as a Seneca speaker I wanted to refer to a relative who
was one generation older than me. In English I could have called that person
“mother,” “father,” “aunt,” or “uncle.” Seneca has what might at first look like
a parallel set:
mother no’yëh
father ha’nih
aunt age:hak
uncle haknó’sëh
The Seneca words, however, would apply to partially different sets of people.
Whereas English speakers usually have only one mother and one father (ignor-
ing step-parents and adoptions), Seneca no’yëh refers not only to my biological
mother but also to any woman of her generation who is related through her: her
sister(s) but also her female cousin(s). Similarly, ha’nih refers not only to my
biological father but also to any man of his generation who is related through
him: his brother(s) and his male cousin(s). The aunt and uncle words require
that the sex of the relative in question be different from the sex of the linking
relative. For example, age:hak refers to any woman of my father’s generation
related through him: his sister(s) and his female cousin(s). And haknó’sëh
refers to any man of my mother’s generation related through her: her brother(s)
and her male cousin(s).
I describe these differences because they illustrate how different languages
can divide the world of relatives in different ways and in so doing can affect the

105
106 Related Issues

way one thinks of them. As an English speaker, for example, there is only one
man I call my father and I think of him differently from his brothers.
As a Seneca speaker I would call all of those men ha’nih and think of them
in similar ways. And whereas as an English speaker I combine my father’s
brothers with my mother’s brothers, calling all of them my uncle, as a Seneca
speaker I would call members of the first set ha’nih and those of the second set
haknó’sëh and think of them differently.6
The question of whether people who speak different languages think
differently is an old one. It goes back at least to the German scholars
Johann Gottfried von Herder (1744–1803), Wilhelm von Humboldt
(1767–1835), and Heymann Steinthal (1823–1899). It received renewed
attention with the writings of Benjamin Lee Whorf from 1927 to 1941
(Carroll 1956). Strong opinions continue to be expressed on both sides of
this question. Those who are convinced that Whorf was hopelessly wrong –
and they are many – often express their opinion emotionally and ad
hominem. Geoffrey Pullum dismissed Whorf as a “Connecticut fire preven-
tion inspector and weekend language-fancier” (Pullum 1991: 163), and
Pieter Seuren wrote of “an amateur linguist who, as a mature student,
took courses with Sapir during his years at Yale and, for some time, gained
popularity with the American anthropological establishment” (Seuren
1998: 189). Steven Pinker wrote that Whorf’s proposal “is wrong, all
wrong! The idea that thought is the same thing as language is an example
of what can be called a conventional absurdity: a statement that goes
against all common sense but that everyone believes because they dimly
recall having heard it somewhere and because it is so pregnant with
implications” (Pinker 1994: 56–57).
Despite these strongly negative evaluations the question refuses to go
away and Whorf keeps resurfacing. Those who knew him took him
seriously. For a time some of his writings were distributed to American
State Department employees to sensitize them to different ways of think-
ing they might encounter in employment abroad where people spoke other
languages. More recent academic studies have suggested that Whorf was
onto something important (e.g. Lucy 1992, 1996; Gumperz and Levinson
1996). The fact that he was employed outside of academia is hardly
a reason to dismiss him so completely, and to say as Pinker did that
those who take him seriously are so foolish that they believe something
that “goes against all common sense” is offensive. To say that language
influences thought is not at all to say that “thought is the same thing as
language.” Reasons for distinguishing thought and language were

6
This traditional Seneca system is currently giving way to the English system as the Seneca
language is giving way to English.
 Rethinking Whorf 107

thought thought

semantic semantic
structure structure

syntactic syntactic
structure structure

Figure 13.1 The wrong question (on the left) and the right one (on the right)

reviewed here in Chapter 6. How language and thought are related

has been a central concern of this book, and the view of language advo-
cated here suggests that Whorf’s critics misunderstood where the relation
should be sought. A better way to view the question is pictured in
Figure 13.1.
Almost by definition, syntax is removed from a direct relation to thought.
We have seen how language change inevitably increases the distance
between the two. Asking whether syntax shapes thought is as pointless as
asking whether apples grow on orange trees. On the other hand it should be
obvious that different languages provide their speakers with different
semantic resources. No one doubts that different languages organize sounds
in different ways, and presumably they can differ at least as much in the
ways they organize thoughts. As Dan Slobin wrote, “The language or
languages that we learn in childhood are not neutral coding systems of an
objective reality. Rather, each one is a subjective orientation to the world of
human experience, and this orientation affects the ways in which we think
while we are speaking” (Slobin 1996: 91). If we accept the presence of
semantic differences across languages, the question we are left with is
whether or to what extent the different semantic resources of different
languages influence the thoughts of the people who speak those languages.
Answering that question leads us to take account of other aspects of the
relation between language and thought.
To return to differences between the English and Seneca kinship systems,
they seem to provide a clear case where the two languages do shape thoughts
in different ways. But the question remains whether those differences origi-
nated in language itself. Are the English and Seneca languages responsible
for the ways people think of their relatives, or do those differences stem from
the different ways people interact with those relatives in the world outside of
language? It is difficult to imagine how the ways we think about relatives
could have been determined initially by language. What could there possibly
be in language itself that would lead us, for example, to link the various
individuals the Senecas call ha’nih? On the other hand it is easy to imagine
108 Related Issues

Stage 1: the Emergence Stage 2: the Subsequent

of a Category Use of the Category

ways of behaving behavior

extralinguistic
toward certain influenced
experience:
individuals by that idea

the idea of such the idea evoked

thought:
an individual by that category

semantic its categorization the category

structure: as [ha’nih] [ha’nih]

Figure 13.2 How extralinguistic experience shapes a semantic category

how Seneca ways of interacting with those relatives would affect the ways
they are thought of, and that ways of thinking about men called ha’nih were
initially established by one’s experiences with men who were categorized in
that way.
But that is not the end of the story. Once a language has acquired
a category, from then on a category that was at first established in the
world outside of language becomes one more element in the huge inven-
tory of categories that are available to those who use the language.
A language is thus a repository for countless ways of thinking whose
origins lay outside of language, but when subsequently a category is
employed as the language is used, its relation to thought is activated in
combination with the relation of that thought to extralinguistic experience,
and from then on it is reasonable to say that language does play a role in
shaping a thought.
The two-stage view in Figure 13.2 may help to explain why people view
the so-called Whorf hypothesis in such opposite ways: they are looking at
different stages in the emergence of categories. At the first stage both
thoughts and semantic structures are shaped by the external environment,
but once a category has become part of a language it feeds back into a way of
thinking and acting. The first stage is independent of Whorf’s proposal while
the second stage confirms it. Those who are so violently opposed to Whorf
may be thinking of the stage at which ways of thinking are shaped by
experiences outside of language. Those who believe just as strongly that
Whorf had something valuable to say are thinking of the subsequent stage,
where linguistic categories have become established. Here are additional
examples.
 Rethinking Whorf 109

Visiting
Chapters 9 and 10 introduced ideas that were categorized in English as [hang
out] and in Seneca as [atyö’se]. The Seneca category [atyö’se] is approxi-
mately translatable with the English category [visit], but a Seneca speaker’s
idea of visiting differs from an English speaker’s and it is not irrelevant that
the literal meaning of [atyö’se] is “arrive for someone’s benefit.” Both the
English idea and the Seneca idea involve one or more people going to a place
occupied by one or more other people where there will be conversation and
perhaps other activities. That overlap in ideas is what makes the Seneca
category a reasonable translation of the English. In much of Native America,
however, visiting has always had a special significance as a highly valued
and frequently practiced activity that is often undertaken without prior
warning or preparation, and that functions to cement social relations
and exchange knowledge in ways that have now been appropriated by
newspapers and television. In the nineteenth century, government officials
who wanted to force Indians into the white man’s way of thinking often
complained about their incessant visiting.
The point here is that the thought associated with [atyö’se] was established
by events in the world outside of language, where the experiences of English
and Seneca speakers were different. Once established, however, this category
became part of the Seneca language’s semantic resources and fed back into
the way one thought of visiting. Figure 13.3 suggests the origin and subse-
quent fate of this category.

Thanking
Another Seneca idea is often translated “thank,” although, especially among
older speakers, that was by no means the only choice:
When confronted with the Seneca words involved, some speakers balk at any attempt to
give an English equivalent. Others translate, to some extent according to context, as
“thank, be thankful or grateful to or for, rejoice in, bless, greet.” The trouble is that the
Seneca concept is broader than that expressed by any simple English term, and covers
not only the conventionalized amenities of both thanking and greeting but also a more
general feeling of happiness over the existence of something or someone. One result is
that the English distinction between “give thanks to” and “give thanks for” has no
relevance. (Chafe 1961: 1)

This semantic category can be represented as [nöhönyö]. One of the many

words in which it appears is o’tsedwanö́ :nyö:’, roughly translatable as “we
hereby greet, rejoice in, give thanks to or for him [referring to Our Creator].”
These translations are attempts, none of them fully satisfactory, to capture the
Seneca idea. There is also an unanalyzable Seneca word nya:wëh that is usually
110 Related Issues

Stage 1 Stage 2

extralinguistic behavior influenced by

experience: that idea

thought: the idea the idea evoked by

of visiting that category

semantic categorization the category

structure: as [atyö’se] [atyö’se]

Figure 13.3 How outside reality shapes a category, which then affects thought

translated “thank you.” But that translation seems inappropriate as part of the
common Seneca greeting nya:wë́ h sgë:nö’, which translates literally as (first
word) “thank you” or “rejoice,” (second word) “peace, well-being, health,
contentment.” As can be seen, the particle sgë:nö’ also lacks a single satisfac-
tory translation. The point is that these Seneca words capture a thought of
appreciating the existence of something that does not coincide exactly with
anything in English.

Differences in Orientation
In addition to expressing ideas, as illustrated so far, the Seneca language often
expresses the orientation of an idea in a way that lacks a single clear equiva-
lence in English. The following words all contain a suffix -aje’ that is tradi-
tionally called the “progressive,” although its meaning is not at all like that of
the English aspect of that name. It signals that an event or state is distributed
along some continuum, either in time (an event) or in space (a state).
A translation sometimes approaches that meaning with the word “along”:
gëhö́ kdaje’ “along next to the creek”
oadaje’ “the road passing by”
satödaje’ “you’re listening along”
 Rethinking Whorf 111

hóíwa:nodaje’ “he’s preaching” (literally “piling up words as he goes along”)

Another Seneca orientation is provided by the “stative-distributive” suffix,

whose form is -’s. When added to a verb of motion it signals that the motion
takes place randomly across several locations. A translation may include the
word “around” or “about” as in the following examples. The third word was
used to describe a deer that was seen moving about randomly from one location
to another.
yeda:ke’s “she’s running around”
ye:awi’s “she’s carrying it about”
i:we’s “it’s around, it’s present, it’s there”
Orientations are often expressed with unanalyzable particles: short words
that function to orient the idea of an event or state. One of them is nö:h,
sometimes translated “I guess” but also “I think,” “perhaps,” or “probably,”
showing that the speaker is not completely certain about something:
Osdëöjö́ :h nö:h.
it’s raining I guess
“I think it’s raining.”

Another particle is gë:s, translatable as “repeatedly,” “commonly,” “habi-

tually,” “multiple times,” or as “keep doing something”:
Wá:ga:ödö:’ gë:s.
“He asked me repeatedly” or “he kept asking me.”

Seneca speakers often have trouble explaining the use of these particles and
may omit them in translating. They were added to the semantic resources of
Seneca at various times in the language’s history and they contribute, often in
subtle ways, to the manner in which people think in that language. The above
examples illustrated ways in which the Seneca language feeds back into the
thinking of its speakers, and how those ways lead to differences from the
thinking of English speakers.

Summary
Taking as a point of departure the strong opinions that have been expressed
with regard to the writings of Benjamin Lee Whorf, this chapter questioned
how language itself could possibly be the place where thoughts are initially
shaped. It was seen as more plausible that the shaping of thoughts is initially
112 Related Issues

determined by interactions with the world outside of language. However, as

that outside world is processed, first in thought and then as a part of
a language’s semantic resources, those resources become a repository of
ways in which aspects of the outside world are linguistically interpreted.
Once established in that way, subsequent uses of the language feed back into
ways of thinking that are indeed affected in different ways by different
languages.
14 Lessons from Literature

One of the properties of conscious thought that were listed in Chapter 2 was the
fact that it is continuous during all our waking hours. We are conscious of
thinking something as long as we are awake. For any normal person conscious
thought occupies the largest portion of their lives. But what are we thinking
about? Obviously we can’t list here the myriad things a person might think
about, but we can at least identify some of the directions a person’s thoughts
may take.
During most of human history the thoughts of hunter-gatherers must have
been focused part of the time on the details of hunting and gathering, and the
thoughts of agriculturalists must have been focused on details of planting and
harvesting. But those practical matters could not have occupied their thoughts
during all of those conscious hours. What were people thinking about the rest of
the time? Michael Corballis (2015) has called attention to the fact that our
minds often wander. It may be impossible to prevent what some might deni-
grate as lapses of attention that intrude on whatever task (such as reading this
book) we are “supposed to be” focusing on. Corballis shows that these lapses,
far from being something we should regret, are essential to our mental health, to
creativity, and to constructing our knowledge of ourselves and the world
around us. Mind wandering, however, is not always random in its content but
may follow a story of some kind. It is from these organized mind wanderings
that an “oral literature” develops, an oxymoronic label that combines qualities
shared by the oral dissemination of wisdom, entertainment, and beauty with the
ways such knowledge is expressed in writing.
A familiar manifestation of oral literature has been the folktale, known to
many of us through European “fairy tales,” but more widely through the
folktale traditions that contribute to the legacy of most if not all cultures.
These stories are usually characterized by a unidimensionality that focuses
largely if not exclusively on a character’s overt actions with little attention paid
to what that character is thinking. “The fairy tale indicates the action and does
not get lost in the portrayal of scenes and characters . . . Its hero moves and acts,
he does not stand still in astonishment, contemplation, or meditation” (Lüthi

113
114 Related Issues

1976: 50). To the extent that folktales are restricted to overt actions they
provide us with minimal guidance to the nature of thought.
As mentioned in Chapter 7, the invention of writing, a very recent develop-
ment in human history, has made it easy for language to reach an audience that
is located at a distant time and place and unknown to the language’s creator,
who also enjoys the benefit of leisure to reflect on and mold the shape and
purpose of language before it reaches its audience. This detachment from a
known audience combined with this contemplative leisure gives a writer the
freedom to probe a character’s innermost thoughts, thus opening unique win-
dows into the nature of thought itself (Chafe 2010).
If the real world becomes fiction only by revealing the hidden side of the human beings
who inhabit it, the reverse is equally true: the most real, the “roundest” characters of
fiction are those we know most intimately, precisely in ways we could never know
people in real life. (Cohn 1978)
To illustrate ways in which written literature can shed a special light on the
nature of thought this chapter draws examples from a short story and two
novels, but what is said of them applies, mutatis mutandis, to plays, movies,
operas, and whatever other creations allow their audiences to share vicariously
in the experiences of fictional others (cf. Clark and van der Wege 2015). A
writer can do this in two ways: by presenting a character’s thoughts from the
inside as the character is thinking them, an approach that has been called
“mimesis,” or by adopting an omniscient stance that describes those thoughts
from the outside while interpreting and commenting on them, an approach
called “diegesis.” We can examine each of those approaches in turn.

Mimesis
As an illustration we can look at James Joyce’s story “Eveline” from the
collection Dubliners (Joyce 1993 [1904]). Most of the story puts the reader
inside Eveline’s thoughts as she agonizes over whether she should abandon the
comfort of her home in Dublin and begin a new life with her boyfriend Frank in
Buenos Ayres. It begins:
She sat at the window watching the evening invade the avenue. Her head was leaned
against the window curtains and in her nostrils was the odour of dusty cretonne. She was
tired.
Few people passed. The man out of the last house passed on his way home; she heard
his footsteps clacking along the concrete pavement and afterwards crunching on the
cinder path before the new red houses.

Before discussing Eveline’s thoughts we might ask why, since we are being
inserted into her own stream of thought, she was referred to in the third person
and her experiences were reported in the past tense: she sat, she heard, etc. It is
 Lessons from Literature 115

in fact more often than not the case in both written and oral literature that
fictional events are reported in this way. Chafe (1994) distinguished the “repre-
senting” from the “represented” consciousness. The representing conscious-
ness is the one responsible for producing the language – in effect the author’s
consciousness – whereas the represented consciousness belongs to the char-
acter being portrayed. The first provides an outer frame within which the events
of a story unfold. This distinction pervades both written and oral literature,
where third person and past tense are established from the narrator’s point of
view.
Turning now to Eveline’s thoughts, we can distinguish those that were
immediate from those that were displaced – between what for her was here
and now and what she remembered or imagined. Immediate access to her
thoughts was signaled from the beginning by the fact that the story begins in
medias res. We are not told who or where she is but are placed at some arbitrary
point within her ongoing stream of thought. Another indication of immediacy is
a richness of detail that captures a character’s anoetic consciousness. The
reader experiences vicariously Eveline’s visual experience (watching), her
posture (sitting and leaning her head), her olfactory experience (smelling the
dusty cretonne), and her auditory experience (hearing the neighbor’s footsteps).
Emotion is captured in the reference to her being tired, but more powerfully
later when “she felt her cheek pale and cold and, out of a maze of distress, she
prayed to God to direct her, to show her what was her duty.” Emotion was
supplemented with action as “she gripped with both hands at the iron railing”
and “her hands clutched the iron in frenzy.”
But a shift from immediacy to displacement occurs as she recalls events from
her earlier life:
One time there used to be a field there in which they used to play every evening with
other people’s children. Then a man from Belfast bought the field and built houses in it –
not like their little brown houses but bright brick houses with shining roofs. The children
of the avenue used to play together in that field – the Devines, the Waters, the Dunns,
little Keogh the cripple, she and her brothers and sisters. Ernest, however, never played:
he was too grown up. Her father used often to hunt them in out of the field with his
blackthorn stick; but usually little Keogh used to keep nix and call out when he saw her
father coming. Still they seemed to have been rather happy then. Her father was not so
bad then; and besides, her mother was alive.
A consistent property of memory is its island-like nature, as seen here in
Eveline’s recall, first of the field, then of the man from Belfast, then of the
children’s play, then of her brother Ernest, then of her father’s threatening
appearance on the scene, and then of the family’s relative happiness. The same
island-like quality appears in a subsequent passage in which she remembers
how she got to know Frank:
116 Related Issues

How well she remembered the first time she had seen him; he was lodging in a house on
the main road where she used to visit. It seemed a few weeks ago. He was standing at the
gate, his peaked cap pushed back on his head and his hair tumbled forward over a face of
bronze. Then they had come to know each other. He used to meet her outside the Stores
every evening and see her home. He took her to see The Bohemian Girl and she felt
elated as she sat in an unaccustomed part of the theatre with him. He was awfully fond of
music and sang a little. People knew that they were courting and, when he sang about the
lass that loves a sailor, she always felt pleasantly confused. He used to call her Poppens
out of fun.

She remembered her first seeing Frank, then his meeting her after work, then
their visit to the theatre, then his singing, and then his nickname for her. In
addition to its island-like nature, memory lacks the richness of detail that
characterizes immediacy. Some of these selected memories were generic in
nature and not of particular events, as was signaled explicitly by the words used
to. Memory often combines multiple events into a single generic memory in
this way (Chafe 1994: 203–204).
In the next paragraph Eveline shifts from remembering the past to imagining
the future. Anticipation shares with memory its island-like quality, as seen first
in the thought of shelter and food in her home, then of how hard she worked,
then of what would be said about her, and then of the predictable reaction of
Miss Gavan:
She had consented to go away, to leave her home. Was that wise? She tried to weigh each
side of the question. In her home anyway she had shelter and food; she had those whom
she had known all her life about her. Of course she had to work hard, both in the house
and at business. What would they say of her in the Stores when they found out that she
had run away with a fellow? Say she was a fool, perhaps; and her place would be filled
up by advertisement. Miss Gavan would be glad. She had always had an edge on her,
especially whenever there were people listening.
The distinction between immediate and displaced thought is easy to relate to
ordinary experience. As we live through each day we constantly choose
between focusing our attention on what is present in our immediate environ-
ment or alternatively on remembering events from our past, events that were
conveyed to us by others or events we imagine for the future. There is no need
to draw a strict line between remembering and imagining, since both involve
creative interpretations of events outside our immediate environment.
Properties that distinguish immediate from displaced thought are summar-
ized in Table 14.1, where typical means of linguistic expression are included in
parentheses. With respect to continuity, immediate thoughts are experienced as
a continuous flow that may be realized by beginning a work in medias res,
contrasting with the island-like nature of displaced thought. The second row,
resolution, characterizes immediate thought as richly detailed, contrasting with
the attenuated detail of displaced thought. The third row, space and time,
 Lessons from Literature 117

Table 14.1 Properties of immediate and displaced consciousness

Immediate Displaced
Continuity Continuous (in medias res) Discontinuous (island-like)
Resolution High (rich detail) Low (attenuated detail)
Space and time Proximal (here, now) Distal (there, then)

distinguishes the proximal quality of immediate thought, which may be rea-

lized with deictic adverbs like here and now, from the distal quality of displaced
thought, realized with words like there and then.
The two examples that follow are taken from Edith Wharton’s novel The
House of Mirth, first published in 1905 (Wharton 1987 [1905]). The author’s
point of view is again recognized with third person and past tense, but the
richness of detail would be available only to an immediate consciousness:
Seating herself on the upper step of the terrace, Lily leaned her head against the
honeysuckles wreathing the balustrade. The fragrance of the late blossoms seemed an
emanation of the tranquil scene, a landscape tutored to the last degree of rural elegance.
In the foreground glowed the warm tints of the gardens. Beyond the lawn, with its
pyramidal pale-gold maples and velvety firs, sloped pastures dotted with cattle; and
through a long glade the river widened like a lake under the silver light of September.
(p. 79)
The reader shares vicariously what is passing through Lily’s thoughts as she sits
and leans her head and enjoys a succession of olfactory and visual experiences.
These thoughts are supplemented with emotion-laden judgments and
comparisons:
Lily smiled at her classification of her friends. How different they had seemed to her a
few hours ago! Then they had symbolized what she was gaining, now they stood for
what she was giving up. That very afternoon they had seemed full of brilliant qualities;
now she saw that they were merely dull in a loud way. Under the glitter of their
opportunities she saw the poverty of their achievement. It was not that she wanted
them to be more disinterested; but she would have liked them to be more picturesque.
And she had a shamed recollection of the way in which, a few hours since, she had felt
the centripetal force of their standards. (pp. 88–89)

The passage shows clearly how temporal expressions like a few hours ago,
then, now, that very afternoon have their center in a character’s immediate
thought, whereas third person and past tense remain to be determined by the
author – Lily leaned her head, Lily smiled.
Figure 14.1 summarizes the three levels of conscious thought that were
illustrated in these examples of mimesis. Thought may go no further than
anoetic experiences that have not reached the stage of reflective organization.
118 Related Issues

thought

anoetic reflective

unverbalized spoken

Figure 14.1 Levels of conscious thought

Or it may involve experiences that are more reflectively organized, in which

case they may be either unverbalized or fully clothed in language.

Diegesis
As an alternative to the mimetic options illustrated above, a writer may portray
a character’s thoughts as known from the outside by an omniscient observer.
Dorrit Cohn pointed out the advantages that are offered by what she called
psycho-narration:
One of the most important advantages of psycho-narration over the other modes of
rendering consciousness lies in its verbal independence from self-articulation. Not only
can it order and explain a character’s conscious thoughts better than the character
himself, it can also effectively articulate a psychic life that remains unverbalized,
penumbral, or obscure. Accordingly psycho-narration often renders, in a narrator’s
knowing words, what a character “knows”, without knowing how to put it into words.
(Cohn 1978: 46)

Psycho-narration thus lets narrators present their own interpretations of a

character’s thoughts, whether they are anoetic, unverbalized, or fully verba-
lized. For illustrations of all three we can look at excerpts from George Eliot’s
Middlemarch (1871), a novel that takes full advantage of the diegetic option.
Quoted here are portions of Chapter 47 of that work, which focuses entirely on
events that pass through the consciousness of Will Ladislaw as he attends
church in order to catch sight of Dorothea Brooke, the object of his affection,
thereby flouting the wishes of her husband, Edward Casaubon. The following is
a typical example of the manner in which the narrator conveys Will’s thoughts:
It may seem strange, but it is the fact, that the ordinary vulgar vision of which Mr
Casaubon suspected him – namely, that Dorothea might become a widow, and that the
interest he had established in her mind might turn into acceptance of him as a husband –
had no tempting, arresting power over him; he did not live in the scenery of such an
event, and follow it out, as we all do with the imagined “otherwise” which is our
practical heaven. It was not only that he was unwilling to entertain thoughts which
 Lessons from Literature 119

could be accused of baseness, and was already uneasy in the sense that he had to justify
himself from the charge of ingratitude – the latent consciousness of many other barriers
between himself and Dorothea besides the existence of her husband, had helped to turn
away his imagination from speculating on which might befall Mr Casaubon.
The author tells us about, rather than showing us, Will’s internal experiences.
She adds her own interpretations, commenting on his thoughts with statements
like “as we all do with the imagined ‘otherwise’ which is our practical heaven.”
Comments on a character’s thoughts occur frequently throughout the novel, as
in the second sentence of the following.
Will easily felt happy when nothing crossed his humour, and by this time the thought of
vexing Mr Casaubon had become rather amusing to him, making his face break into its
merry smile, pleasant to see as the breaking of sunshine on the water – though the
occasion was not exemplary. But most of us are apt to settle with ourselves that the man
who blocks our way is odious, and not to mind causing him a little of the disgust which
his personality excites in ourselves.
Anoetic experiences of all three types – perceptions, actions, and emotions –
are included in the following:
To his own surprise Will felt suddenly uncomfortable, and dared not look at her after
they had bowed to each other. Two minutes later, when Mr Casaubon came out of the
vestry, and, entering the pew, seated himself in face of Dorothea, Will felt his paralysis
more complete. He could look nowhere except at the quire in the little gallery over the
vestry-door . . .
There are references to where Will was looking and not looking, to his action of
bowing to Dorothea, and to his discomfort and feeling of paralysis.
Unverbalized thought was captured when Will realized that his visit to the
church was not having the results he had hoped for:
Dorothea was perhaps pained, and he had made a wretched blunder.

These need not have been words that passed through Will’s consciousness, but
simply one way of expressing a less committed unverbalized thought.
In contrast the italicized portions of the following two excerpts, with appro-
priate changes of person and tense, may have captured thoughts that were fully
verbalized:
It was no longer amusing to vex Mr Casaubon, who had the advantage probably of
watching him and seeing that he dared not turn his head. Why had he not imagined this
beforehand?
There was no delivering himself from his cage, however; and Will found his places
and looked at his book as if he had been a school-mistress, feeling that the morning
service had never been so immeasurably long before, that he was utterly ridiculous,
out of temper, and miserable. This was what a man got by worshipping the sight of a
woman!
120 Related Issues

Summary
This chapter began by suggesting how mind wandering is sometimes funneled
into the creation of stories. It then turned to ways in which fictional literature
can shed light on the nature of thought, which an author may present either
mimetically or diegetically – from inside or outside a character’s own mind.
The mimetic approach highlights a distinction between immediate and dis-
placed thoughts: between what is here and now and what is remembered or
imagined. Immediacy is characterized by continuity and rich detail, displace-
ment by islands of experience and attenuated detail. The diegetic approach lets
an author evaluate and comment on what a character is thinking. With either
approach the thoughts themselves may be anoetic, unverbalized, or fully
verbalized.
Part V

Common Ways of Orienting Thoughts

15 Small Numbers and Subitizing

This chapter shows how different approaches to a particular area of

investigation can make their own separate contributions which together
can augment the understanding provided by any one approach. The
findings discussed here have not been aware of each other, as happens
often in a world where separate disciplines follow their own traditions.
The chapter can thus be seen as a plea for interdisciplinary cross-
fertilization.
It happens frequently as people are speaking or writing that their language
forces them to make a distinction within a range of very small numbers. People
using the English language, for example, are required to distinguish one item from
more than one, in the latter case choosing the plural form of a noun, most often
with the suffix spelled -s. We say face when there is one of them and faces when
there are more than one. With pronouns we say he, she, or it when there is one, but
they regardless of gender when there are at least two. This distinction between one
and more than one is often obligatory in English. Whenever we use a so-called
count noun like face as opposed to a mass noun like sand we are obliged to use
a singular or plural word as the occasion demands. This distinction is often
imposed on verbs as well, as can be seen with is and are in the third sentence of
this paragraph.

The Typology of Number Marking

Other languages may treat number differently. Such differences were docu-
mented in a comprehensive study of number marking by Greville Corbett
(2000). At the low end are languages like Japanese that may use the same
form of a noun regardless of whether there is one or more than one of some-
thing, although such languages usually provide a way of marking plurality
when it is salient. This way of talking does not mean, of course, that Japanese
speakers are unaware of the difference between one and more than one, but
only that their language does not demand that they express that difference
overtly. The element tachi can be added when it is important to specify plurality
for humans:

123
124 Common Ways of Orienting Thoughts

inu “dog” or “dogs”

otokotachi “men”

Many languages use a special form when there are precisely two, thus
distinguishing one (singular), two (dual), and three or more (plural).
The Seneca language requires its verbs to mark a singular–dual–plural distinc-
tion when talking about participants in events or states:
yenö́ e’s “she likes it”
ginö́ e’s “they [two women] like it”
wadínöe’s “they [three or more women] like it”
English lacks the obligatory dual category, but it does make frequent use of
words expressing twoness like both, pair, twin, and double. In the Seneca
language, nouns, as opposed to verbs, do not make an obligatory number
distinction, although they do offer distinctions like these:
ji:yäh “dog”
jiyä́ shö’öh “dogs”

However, the second word differs from the translation “dogs” in two ways.
First, it is not obligatory when talking about more than one dog. Second, it not
only marks plurality but is a “distributive” marker, implying that the multiple
entities differ from each other in some way. Perhaps they belong to different
breeds, or the dogs are scattered in different places, or they lived at different
times. Distributives thus include plurality but something more as well (Mithun
1999: 79–94; Corbett 2000: 111–117).
Beyond this singular–dual–plural distinction a smaller number of
languages offer a “trial” category that is used when there are specifically
three, thus distinguishing singular–dual–trial–plural where “plural” means
more than three. Corbett (2000: 21–22) mentions several aboriginal
languages of Australia that make this four-way distinction. Instead of
a trial, other languages offer a “paucal,” whose use corresponds to the
English expression “a few,” not marking a specific number of items but
covering a range of low-number possibilities. Paucals are found, for exam-
ple, in various Oceanic languages (Corbett 2000: 22–26). The presence of
duals and trials raises the question of whether any languages offer
a “quadral” for specifically four entities. Corbett (2000: 26–30) examined
the evidence for a quadral in several languages of the Austronesian family
and concluded that in each case the form in question was not restricted to
precisely four but was a type of paucal.
In brief, many languages make either an obligatory or an optional distinction
between one and more than one, quite a few languages also give special
attention to two, and a smaller number of languages also distinguish three.
 Small Numbers and Subitizing 125

Figure 15.1 How many faces?

Whether any language gives attention to precisely four is doubtful, but recourse
may be had to a form that covers a few. These observations arise from a large
and varied sample of the world’s languages.

Introspection
How many happy faces do you see in the left box in Figure 15.1? How many in
the right box? You almost certainly followed very different procedures in
answering these questions. On the left you must have known immediately
and confidently that there were three. On the right you must have counted
before you knew there were eleven.
Evidently we process very small numbers of items quickly, accurately, and
confidently, whereas larger numbers leave us no choice but to count or guess.
This difference is experienced constantly in daily life, no matter what language
we speak, and recognizing it requires no special investigative techniques. Can
we conclude that our brains are constructed in such a way that we know
immediately when there are two or three of something, but not when there
are eleven or seventeen?
Things become less clear as we add successively more items as in
Figure 15.2. Perhaps we know immediately when there are four, but is that
because we see two groups of two and apply arithmetic, or do we see four all at
once as we did three? How is it with five? Do we add two and three, or do we
see all five together? Do we see six as two groups of three, three groups of two,
or six all at once? By the time we come to seven, seeing them all at once is
unlikely. Either we divide the group into smaller subsets or we count.
With larger numbers counting may be our only choice unless the items
appear in an array that fosters the application of arithmetic to subsets.
126 Common Ways of Orienting Thoughts

Figure 15.2 Successively larger numbers

Figure 15.3 Unpatterned and patterned displays

Compare the different ways in which sixteen items are displayed in Figure 15.3.
Probably it was easier to identify sixteen in the bottom array than in the
undifferentiated sequence on top, although in neither case was the experience
like that of identifying the three in Figure 15.1.
These observations are neither idiosyncratic nor new. In the middle of the
nineteenth century Sir William Hamilton wrote in his published lectures on
metaphysics:
 Small Numbers and Subitizing 127

Consciousness will thus be at its maximum of intensity when attention is concentrated

on a single object; and the question comes to be, how many several objects can the mind
simultaneously survey, not with vivacity, but without absolute confusion? . . . You can
easily make the experiment for yourselves, but you must beware of grouping the objects
into classes. If you throw a handful of marbles on the floor, you will find it difficult to
view at once more than six, or seven at most, without confusion; but if you group them
into twos, or threes, or fives, you can comprehend as many groups as you can units;
because the mind considers these groups only as units – it views them as wholes, and
throws their parts out of consideration. You may perform the experiment also by an act
of imagination. (Hamilton 1859: 176–177)
To summarize, introspection shows that when it comes to recognizing the
number of items in a group, recognition is immediate when there are one, two,
or three. With four, five, or six, either we see the number all at once or we
arrange the items in subgroups of two or three. Beyond about six we can do
nothing but count, unless we are aided by a pattern that encourages grouping
into subsets and applying arithmetic.
Does this special way of processing very small numbers relate to human
needs? It seems that humans frequently need to think or behave differently
depending on whether there are one, two, or three of something, whereas we
seldom need to think or behave differently depending on whether there are
twenty-six or twenty-seven. The dominance of “two” can be seen as reflecting
a multiplicity of real-world circumstances favoring that number: the symmetry
of the human body (two eyes, two hands, etc.), the frequent salience of dyadic
relations in society (men and women, parents and children, Democrats and
Republicans), as well as abstract oppositions like good and bad, large and
small, up and down. When it comes to three, the special salience of that number
in folklore and in music has been discussed, for example, by Alan Dundes
(1980), and for music by Paul Rozin et al. (2006).

Subitizing
The observation that we have a special way of processing very small numbers
has intrigued psychologists for a long time. W. Stanley Jevons wrote in Nature
in 1871 of 1,017 trials during each of which he threw some black beans toward
a white box in such a way that anywhere between one and fifteen beans landed
randomly in the box. Each time, he immediately guessed at the number in the
box and found he was always accurate up to four, but that with five he already
made an occasional error. Somewhat later, working with two subjects besides
himself and with a more complex apparatus, Howard C. Warren concluded that
“except under special stress of attention, or with subjects especially apt in this
direction, the function of perceptive counting [as he called it] is limited to the
numbers One, Two and Three” (Warren 1897: 589). The term “subitize” was
128 Common Ways of Orienting Thoughts

introduced for this phenomenon in the 1940s by psychologists at Mount

Holyoke College (Kaufman et al. 1949), with terminological advice from
Cornelia Coulter, a classicist colleague who attached the Greek suffix -ize to
the stem of the medieval Latin verb subitare “to arrive suddenly.” This word
has in the meantime come into widespread use among psychologists, although
it is little known elsewhere.
It appears, then, that subitizing and counting utilize different mental pro-
cesses, and psychologists have tried to identify what those processes are and
how they relate to more general mental abilities. A useful review of relevant
research was provided by Lana Trick and Zenon Pylyshyn (1993, 1994; Trick
2005), relating it to a theory of vision (e.g. Pylyshyn 1989). Visual processing
was thought to include two stages: first, a spatially parallel “preattentive” stage,
where analyses of visual input occur holistically; second, a spatially serial
“attentive” stage, where analyses are performed one item and one location at
a time with a moving focus. Visual analysis was thought to take place either all
at once (preattentively) or one item at a time (attentively). The evidence of
subitizing led Trick and Pylyshyn to hypothesize an intermediate but still
preattentive stage of very limited capacity where very small numbers are
processed differently from larger numbers. At this intermediate stage the visual
process individuates each of a small number of items in parallel, but only when
the number does not exceed its highly limited capacity. When that capacity is
exceeded, the only recourse is to move the attentional focus serially from one
item to another, as in counting. This explanation may now need to take account
of the fact that Riggs et al. (2006) found subitizing in tactile as well as visual
perception.

Back to Language
Languages suggest another way of understanding this phenomenon. A clue was
present in the statement of William Hamilton quoted above, where he asked
how many objects the mind can “simultaneously survey,” implying a limit on
what this book calls focal consciousness. Language gives evidence that con-
sciousness moves continually and rapidly from one focus to the next as people
are talking (Chapter 8). Subitizing may be evidence for a similar pace of
consciousness when people are not talking, with a single focus of conscious-
ness processing the items in a group simultaneously.
Language also gives evidence of the limited capacity of each focus:
a restriction to one new item at a time – one item that is assumed not to be
already active in the listener’s consciousness. What, then, constitutes a single
item? Here we can invoke categorization: interpreting a particular mental input
as an instance of a familiar and unitary category as discussed in Chapter 9.
We interpret our perception of a particular animal as an instance of a category
 Small Numbers and Subitizing 129

Figure 15.4 Instances of the categories “dog” and “three”

that allows us in English to use the word dog. Do we also possess categories of
twoness and threeness that lead us to interpret particular groups of items as
instances of those categories? In other words categorization may apply to both
the examples in Figure 15.4: on the left an instance of the dog category, on the
right an instance of the three category. The latter does not, of course, lead
necessarily to the marking of “trial” in a language’s morphology, but only to its
inclusion as an immediate element of thought.
If we do in fact operate with small number categories, why should they be
limited to three? Might not some language somewhere, for example, make
available a seventeenness category that would create immediate recognition
of that many items? Such a category is implausible, of course, and we might
see it as excluded by capacity limitations on focal consciousness. Not only is
a focus restricted to a group constituting one new item, it may also be
restricted to no more than three or four separable items within a group.
A category of seventeen would far exceed that limitation. If a display of
seventeen were categorized at all, it might be as an instance of a texture and
not of individuated items.
Subitizing does require that the mind individuate the items that constitute
the group. Trick and Pylyshyn found their subjects subitizing the first two
displays in Figure 15.5 but not the third. We immediately interpret the first
two as instances of threeness whereas the third, with rectangles inside
rectangles, is not immediately interpretable in that way. Identifying the
number of rectangles requires an extra process of individuation, and that
takes time.
To understand subitizing in terms of a limited focus of consciousness,
categorization, and individuation need not be inconsistent with a theory of
preattentive parallel processing but only an elaboration of that theory. But
restrictions on the flow of focal consciousness that are necessary to explain
130 Common Ways of Orienting Thoughts

Same size Different sizes Concentric

Figure 15.5 Different groupings of rectangles (after Trick and Pylyshyn

1993: 338)

what is found in language may help account for the reaction times that have
been consistently identified in subitizing experiments. The fact that number
marking by languages may not extend beyond three provides evidence for this
upper limit. These avenues converge on an understanding that is richer than
anything offered by a single discipline.
We have seen here a relatively simple example of how introspection can be
augmented with evidence from both linguistics and psychology to bring us to
a fuller understanding of a frequently encountered and easily observable mental
phenomenon. Introspection alone is suggestive and may ultimately be deter-
mining, but it invites support from publicly accumulated evidence. Linguistics
shows how this phenomenon affects the ways languages organize thought and
communication, while psychology suggests its relation to processes of greater
generality.
Other chapters in this work illustrate other aspects of human experience that
are manifested in some way in language. Understanding their nature can be
enhanced by psychological research and sometimes by research in other fields.
If this support is sporadic and uneven, that only shows the need for more work
in all relevant areas, including the construction of more comprehensive theories
within which evidence of all kinds finds a comfortable fit.

Summary
There is abundant evidence that we recognize the number of at least three
individuated items immediately, accurately, and confidently. This ability bears
an obvious relation to the restriction of number marking in languages to
singular, dual, and trial. It was suggested that we process number in terms of
categories such as oneness, twoness, and threeness, and that these categories
have special relevance to human needs.
16 Thoughts and Gender

Gender is a way of classifying entities that is found in some form in many if not
all languages. Corbett (1991) provides a comprehensive survey of the various
ways gender is manifested. It is familiar in languages like French and German,
where it is regarded as a property of nouns, whose gender may be masculine or
feminine as in French, or masculine, feminine, or neuter as in German. It affects
several areas of syntax including the form of the definite article, as when a noun
in French is preceded by le if it is masculine and la if it is feminine. Those are
the singular forms and the distinction is neutralized with les if the noun is
plural. German distinguishes masculine, feminine, and neuter with der, die, and
das, which are neutralized with die in the plural. Gender is also manifested in
pronouns, as with French il (masculine) and elle (feminine), a distinction that is
maintained in the plural, at least in spelling, with ils and elles. In the nominative
case German has the pronouns er (masculine), sie (feminine), and es (neuter),
all three neutralized with sie in the plural.
When the pronouns refer to actual people or sometimes to pets, the distinc-
tion between masculine and feminine has a basis in how one thinks of the sex of
those people or animals. But otherwise why, for example, should German use
die See for the sea (feminine) and der See for the lake (masculine)? Or why
should it be der Tisch (masculine) in German but la table (feminine) in French?
In some cases the gender of a noun is influenced by its phonological shape as
when French nouns ending in -ion like question are feminine, but the influence
of phonology on gender is at best limited.
The rest of this chapter describes the gender system of the so-called
Lake Iroquoian languages (spoken in the area of the Great Lakes), which
include all of the currently spoken Northern Iroquoian languages except
Tuscarora, i.e. Mohawk, Oneida, Onondaga, Cayuga, and Seneca, but also
Huron or Wendat, which was once spoken in several dialects in Canada
but recently only in a version called Wyandot that survived in Oklahoma
until the mid-twentieth century. The examples that follow are from Seneca
(cf. Chafe 2002b).
Of interest here is the way the determination of gender in these languages is
affected by certain cultural practices that affect how one thinks of men and

131
132 Common Ways of Orienting Thoughts

Table 16.1 Intransitive agent and patient prefixes in Seneca

Agents Patients
Singular Dual Plural Singular Nonsingular
Masculine ha- hni- hati- ho- hoti-
Feminine ye- kni- wati- yako- yoti-
Neuter ka- yo-

women. As illustrated in Chapter 10, pronominal reference in Iroquoian lan-

guages is realized in prefixes that are attached to a polysynthetic verb.
Table 16.1 shows the Seneca forms for intransitive agent and patient prefixes.
The agent–patient distinction was discussed in Chapter 6. To summarize
briefly, agents do things whereas patients either have things done to them or
are in some kind of state.
If we look first at the singular forms, the division into masculine, feminine,
and neuter suggests a system like that of German. There is more to say about
feminine singular, but first we can note that whereas masculine is treated
everywhere as a separate category, feminine is treated separately only in the
singular, outside of which it is merged with neuter. In fact, however, its separate
identity is blurred in the singular as well. The forms ye- and yako- are used not
only for females but also for people in general, where it might be translated
“one” or “they” (in a general sense), in that case resembling French on and
German man. This usage has been labeled with terms like “indefinite” or
“nonspecific.” In short, whereas masculine maintains its separate identity
everywhere, in the singular feminine merges with nonspecific and in the dual
and plural it merges with neuter.
It is also worth noting that in the dual or plural, if at least one member of
a group is male then the masculine form is used, even if the group consists
entirely of women except for one man. Furthermore, in Mohawk, Oneida, and
Onondaga a certain girl or woman may be referred to as ye- or yako- while
another girl or woman may be referred to as ka- or yo-. Thus the merger of
feminine and neuter extends in those cases to the singular as well. As explained
for Oneida by Abbott (1984: 127), the feminine prefix ye- “is used to convey
the impression that the female referred to is small, graceful, or petite,” while the
neuter prefix ka- “is used to convey the impression that the female is large,
awkward, or aggressive.” Abbott quotes Karin Michelson as saying that in
Canadian Oneida the neuter prefix sometimes conveys “an indifferent or
detached attitude, sometimes a lack of respect, and sometimes an acknowl-
edgment that the referent has authority or power over the speaker” (quoted in
Abbott 1984: 128).
 Thoughts and Gender 133

Table 16.2 First and second person agents with third person patients

Third person patient

Masculine singular Everything else
First singular he- khe-
Exclusive dual shakni- yakhi-
Exclusive plural shakwa-
Agents Inclusive dual shetni- yethi-
Inclusive plural shetwa-
Second singular hehs- she-
Second dual shesni- yetshi-
Second plural sheswa-

Table 16.3 Third person agents with first and second person patients

Third person agent

Masculine Feminine
singular singular Nonsingular
First singular hak- yök- hök-
First dual shökni- yökhi-
Patients First plural shökwa-
Second singular hya- yesa- hösa-
Second dual shesni- yetshi-
Second plural sheswa-

In short, with intransitive prefixes males play a dominant role and females
are often merged with either neuter or nonspecific reference. The same situa-
tion can be seen in transitive prefixes too: prefixes in which agent and patient
are combined in a single form. Table 16.2 shows agent prefixes with
a masculine third person singular patient and with all other third person
patients. It can be seen that masculine singular patients distinguish more
possibilities.
Table 16.3 shows the reverse: patient prefixes with masculine singular and
other third person agents, where again the masculine singular dominance is
evident.
Does this conspicuous role of masculine singular in the pronominal prefix
system reflect the way speakers of Iroquoian languages think about men and
their role in society? To what extent can we extrapolate from these morpholo-
gical patterns to ways of thinking? This question was raised in Chapter 13,
where it was suggested that thought is shaped initially by interactions with the
world outside of language, but that those externally imposed patterns of
134 Common Ways of Orienting Thoughts

thought are then incorporated into a language’s semantic resources and are
realized in its morphosyntactic structures. The question in this chapter is
whether and how the conspicuous role of masculine singular in Iroquoian
prefixes relates to ways of thinking that characterize (or once characterized)
Iroquois society.

Men and Women in Iroquois Society

The most detailed knowledge we have of Iroquois culture at an
earlier time involves the Hurons, who lived south of Georgian Bay in
what is now the province of Ontario, Canada. In 1649 the Hurons were
overrun and dispersed by Iroquois warriors from further south. What is
known about Huron culture before the dispersal comes from reports by
Samuel de Champlain (Biggar 1929) and Gabriel Sagard (Wrong 1939),
and the many reports by Jesuit missionaries that were compiled in what
are known as the Jesuit Relations (Thwaites 1896–1901). Modern
descriptions of Huron culture based on these materials have been
provided by Elisabeth Tooker (1964) and Bruce Trigger (1976). Trigger
wrote that:
the most basic distinction in Huron society was that made between the sexes. Almost
every task was considered to be either exclusively men’s work or exclusively women’s
work, and every Huron was expected to be familiar with all or most of the tasks
appropriate to his or her sex. For the most part men engaged in tasks that required
considerable physical strength, or which took them away from home for long periods.
Women performed tasks of a more routine nature that kept them in or close to their
villages. (Trigger 1976: 34)
Council meetings were attended only by men, where:
decisions were made that influenced many aspects of village life . . . The council
arranged public feasts, dances, and lacrosse matches, and decided for whom special
curing rites requiring village participation would be performed. The council also under-
took to see that no one was in need and coordinated communal projects, such as
constructing palisades and relocating the village. All legal disputes arising between
members of the different clan segments that lived in the village were adjudicated by the
council. (Trigger 1976: 56–57)

Tooker wrote that “the ‘old men’ in a village decided all matters within the
village and their advice was tantamount to an order” (Tooker 1964: 42). She
noted that “despite the extensive characterization of Iroquois society as being
a matriarchy, women usually had no voice in council meetings. They exercised
their influence behind the scenes, not directly by speaking or attending council
meetings. Even today women do not speak in the Longhouse” (Tooker
1964: 48).
 Thoughts and Gender 135

I recorded a story told by a Seneca man named Roy Jimerson regarding an

athletic contest between the Senecas and their neighbors the Kahkwas, a test of
male prowess that was viewed by an entire Seneca town. Of interest is the way
the audience was divided: “They formed a circle; in front stood the chiefs and
the warriors. And behind them stood the mothers. And different still were the
children.” In other words it was the men who had the front row seats, with the
women behind them. Even today it is only men who perform ceremonial
speeches in a Seneca longhouse. Some women know those speeches just as
well as the men, and a woman may even prompt a man who has forgotten part
of a speech. I was told, however, that for a woman to perform the speech herself
would devalue the role of men.
The roles of men in war and hunting parties have changed as opportunities
for demonstrating masculinity have changed, but Morris Freilich (1958) wrote
that the conspicuous nature of men’s roles has been retained, especially among
Mohawk men, in ways that arose with changing circumstances:
When the ancient hunting and warfare pattern could no longer be maintained,
Mohawk men turned to alternative occupations that also took them far from home
and offered a significant element of danger: the fur trade, timber rafting, river
boating, dock work, and participation in circuses. But fortuitously in 1886, as
a result of the construction of a St. Lawrence River bridge that abutted on the
Caughnawaga Reserve, Mohawk men took up structural high steel work with
enthusiasm and in ever-increasing numbers, and many have continued in that
employment to the present. It is work that also involves long periods away from
home and returning periodically as conquering heroes (now with considerable sums
of money rather than captives or game), working in an all-male group with
a minimum of hierarchical authority, and “having chances to display daring and
courage and thereby gain personal prestige both from the whole community and from
the group one fought or worked with.” (Freilich 1958: 478)
This retention of a culture pattern whose function has changed considerably –
from providing meat and wreaking vengeance on one’s enemies to building the
frameworks of enormous bridges and buildings – shows the same process that
was seen in Chapter 9 as responsible for idiomaticization, where newly derived
semantic elements continue to be represented by the structures from which they
were derived. Here the continuous functioning of a culture pattern appeared
against changing ways in which that pattern was realized.
In short, gender roles were (and to some extent still are) distributed in
Iroquois culture in such a way that men were conspicuous, even flamboyant,
with decision-making powers, while women remained in the background,
a position from which they nevertheless exerted considerable influence on
what men did. Women were far from unimportant. They were responsible for
keeping life going from day to day and from generation to generation.
The importance of women was described by the anthropologist Cara
136 Common Ways of Orienting Thoughts

Richards, who went so far as to write, “If you must be born a woman, try to be
an Onondaga.” She noted a continuation of female roles from the past into
the present:
Historically, the relatively high status of Iroquois women is well documented. Descent
was traditionally traced through the women, land belongs to the women, and chiefs are
appointed by the women – who theoretically can depose them at will. This means that
female children are important to a family, the opinions of women carry weight, and
women have some legal and political power. Even after four hundred years of contact
with the male-oriented European culture and its American descendant, Iroquois women
still maintain their high status and may even have improved it somewhat. (Richards
1974: 401)
How justified would we be in hypothesizing that the Northern Iroquoian
pronominal prefix system developed as it did, with its skewing in favor of
masculine singular, because of the conspicuous nature of men’s role in
Iroquois society? In favor of this hypothesis is the striking nature of the
correlation. Without question men have played a more conspicuous role than
women, and without question the masculine singular category plays a more
conspicuous role in the prefix system. It is certainly more plausible to
conclude that the linguistic pattern arose out of the culture pattern than
vice versa. A scenario in which the high visibility of men arose through
complex interactions of both genetic and cultural forces is not hard to
imagine, as opposed to an unlikely scenario in which the linguistic pattern
arose first for no apparent reason and then became responsible for men
attaining the status described here.
It is true that many other societies in North America and elsewhere share
similar traits of male visibility and yet have not developed analogous morpholo-
gical traits in their languages. One cannot predict that a language spoken in
a society where men play similarly conspicuous roles will develop a morphology
that is skewed toward masculine singular. It is thus preferable to speak of
a culture pattern as motivating rather than causing a linguistic pattern.

Summary
Gender differences were seen as thought-based in those cases where they relate
to a person’s actual sex, but in other cases they are introduced syntactically or
occasionally phonologically. Most of this chapter described a situation in
which the conspicuous role of men in Iroquois society, and hence in the way
men were thought of, led to the dominance of masculine singular in the
pronominal prefix systems of Iroquois languages.
17 Time, Tense, Memory, and Imagination

Time is not something we can see, hear, touch, or perceive in any direct way.
It is not an entity like a house or dog or an event like eating or barking. The very
existence of time can only be inferred from experiences we can perceive.
Linguists have remarked on the fact that metaphors for time often exploit
analogies of space (Traugott 1978). We talk about a “short” or “long” time,
the time “behind” us or “in front of” us, and in many other ways we call on
space to give us a handle with which time can be grasped.
The fact that languages recognize time by marking past, present, and future
tense is familiar to anyone who has attended what used to be called a grammar
school. Even a superficial look at the English language, however, shows that
things are not as simple as that three-way division suggests. English expresses
past tense either with a simple suffix (walk–walked), a change of vowel (sing–
sang), or a completely different form of the verb (go–went). Future tense may
be expressed with an auxiliary verb (will sing), but more often with a more
complex grammatical construction (is going to sing). Present tense may be
expressed by the bare verb (I sing) or with a suffixed -s (he sings), but the same
words may be used more often in speaking generically (he sings perhaps
professionally). If he is doing it right now, English prefers the “progressive”
construction (he’s singing), with the understanding that he started to sing before
the present moment and will continue to sing after it, but probably also that he
will eventually stop. (It may be a little odd to say, “the moon is orbiting the
earth,” because we expect it to keep doing that indefinitely.)
This progressive construction is usually regarded as an “aspect” rather than
a tense because it shows, not a simple location in time, but a more complex
temporal structure whose center may be the present moment (he’s singing), or
a past moment (he was singing), or a future one (he’s going to be singing).
To complicate things further, the progressive is sometimes used for an event
that is scheduled for the future: he’s singing tomorrow at eight, while it also
allows generic alternatives: he’s singing a lot these days, or he’s singing
next year in Milan (perhaps once, perhaps repeatedly). Another aspect, tradi-
tionally called the “perfect,” is used to express the current relevance of an event
that took place earlier (he has sung already), with its own generic option (he

137
138 Common Ways of Orienting Thoughts

has sung often). There is a “past perfect” in he had sung and a “future perfect”
in he will have sung. These are some of the ways the English language reflects
divisions of time. Other languages have ways of their own.
Bernard Comrie (1976, 1981, 1985) and Östen Dahl (1984, 1985) surveyed
tense and aspect in various languages around the world, and their work was
later extended by contributors to Binnick (2012). Like what we saw in
Chapter 14 with respect to number, some languages may not require the overt
marking of tense at all; Comrie mentioned Burmese. But that does not mean
that speakers of Burmese are unaware of the passage of time, which they may
capture with adverbs that have meanings like “yesterday” or “tomorrow,” or in
other ways. It is just that tense is not an element in their language’s obligatory
grammatical system.
Linguistic studies of tense have often made use of a scheme that was
introduced by the logical empiricist philosopher Hans Reichenbach in a book
on symbolic logic (Reichenbach 1947; cf. the Reichenbachian approach men-
tioned in Hornstein 1990). Comrie found certain shortcomings in
Reichenbach’s scheme and suggested the use of semiformal representations
like those in Table 17.1, where S represents the time of speaking, E the time of
the event being talked about, and R a contextually established reference time.
These three locations in time may be related by overlap, including simultaneity,
or one may precede or follow another.
Formulas like these imply the existence of a “time line” pictured as running
from left (earlier) to right (later), where S, E, and R are positioned either as
points or as spans. Figure 17.1 shows how the past, past perfect, present
progressive, future, and future perfect tenses and aspects can be represented
on such a line.
Representations like these, whether verbal formulas like “E precedes S” or
positions on a time line, ignore the roles of memory and imagination in
establishing the ways in which E, S, and R are experienced. Obviously, how-
ever, when E precedes S it must be something that is remembered or imagined
as having occurred, and when E follows S it must be something imagined for

Table 17.1 Ways of representing tense and aspect

Tense/aspect Structure Examples

past E precedes S sang
past perfect E precedes R precedes S had sung
present progressive E overlaps S is singing
future E follows S will sing
future perfect E precedes R follows S will have sung
 Time, Tense, Memory, and Imagination 139

E S E R S S S E S E R

E
past past perfect present future future perfect
progressive

Figure 17.1 Tense and aspect on a time line

the future. The way we experience these temporal relations depends on the way
we experience the present moment as it relates to the ways we experience
events or states that are displaced from the present moment through remember-
ing or imagining.

Languages That Mark Multiple Distances from the Present

Of special interest in this regard is the fact that languages in various parts of the
world have ways of marking tense that distinguish degrees of distance from the
present (Dahl 1984; Comrie 1985: 83–101; Botne 2012). The term “metrical
tense” has sometimes been used for systems of this kind (Chung and
Timberlake 1985: 207–209; Timberlake 2007: 307–308). One such language
is Washo, a language spoken in the Lake Tahoe region of California and
Nevada as described in a dissertation by William Jacobsen (1964). Washo
verbs offer four suffixes that all locate an event or state at a time before the
present, but make the following distinctions (pp. 631–636):
(1) Earlier on the same day or during the preceding night.
(2) Earlier than the same day but not in the more distant past.
(3) In the more distant past but within the lifetime of the speaker.
(4) Before the lifetime of the speaker.
These differences are partially mirrored in events imagined for the future:
(1) From the immediate future to perhaps an hour or so later (p. 589).
(2) Not immediately, but after the lapse of a short interval of time, usually in
a later portion of the same day, although if one is speaking late at night it
may be the following morning (p. 638).
(3) On the following day or at any later time (p. 647).
In a very different part of the world, the Dschang dialect of Bamileke,
a Bantu language of Cameroon, distinguishes five past tenses described as
follows (Hyman 1980):
(1) Immediately before the time of speaking.
(2) Earlier the same day.
140 Common Ways of Orienting Thoughts

(3) The day before.

(4) Some days before.
(5) A long time before.
These distinctions are mirrored symmetrically in the future:
(1) Immediately after the time of speaking.
(2) Later the same day.
(3) The day after.
(4) Some days after.
(5) A long time after.
Multiple past tenses are common in South America. The Yagua language of
Peru, for example, has been described as making the following five distinctions
(Payne and Payne 1990: 384–388):
(1) Within the past few hours.
(2) Yesterday.
(3) Up to about a month ago.
(4) Several months ago.
(5) Distant or legendary.
Five past tenses are also distinguished in Creek, a language native to the
southeastern United States, as recorded by Jack Martin (2010), who described
their meanings as follows:
(1) Today up to last night.
(2) Yesterday to several weeks ago.
(3) Several weeks to a year or so ago.
(4) Long ago by at least several years.
(5) Very long ago.
These distinctions are marked with verb suffixes. The “today” suffix (1) is
restricted to events that are incomplete or repeated, although it can also be
applied to states and negative events. A different morphological structure
(showing aspect rather than tense) is used for events that occurred only once
and were successfully completed. In other words, very recent events may be
marked in two different ways, either by tense or by aspect. Martin noted that the
above list describes the usage of older speakers, whereas Creek speakers born
later than about 1940 no longer recognize option 4 and have expanded the
ranges of 3 and 5 to cover that interval.
Martin illustrated ways in which these tenses are used, as well as the effect of
allowing tenses to “creep forward” from 5 to 4 or 3 in the course of a single
narrative. The following translation of a historical account began with 5, then
 Time, Tense, Memory, and Imagination 141

moved to 4 before settling on 3 at the climax, with a return to 5 at the end

(Martin 2010: 62–64):
A man named Ispahihcha opposed the Muskogee [Creek] constitution, gathered many of
his people, made war against the supporters of the constitution, and divided (Past 5) the
Muskogee Nation . . . So the lawyer accompanied by two people went, and we went
about (Past 4) guarding the old people that were there. And though many were afraid to
guard the old people, they couldn’t (Past 3) do anything. They had heard stories from
long ago about the old ones, and so they were afraid (Past 3). “Never again disturb the
peace or conduct meetings opposing the law,” the judge warned them, and they
dispersed (Past 3). It was about fifty years ago (Past 5).

Martin noted a parallel in the use of the so-called historical present in English:
“That is, just as English speakers will sometimes describe past events with the
present tense, Creek speakers will sometimes slip into past 4 or even past 3
when vividly remembering remote circumstances” (p. 62).
Comparing a variety of languages that distinguish multiple past tenses,
Comrie (1985: 87–88) identified the most common distinction as that between
today and before today. Also commonly marked is the difference between
recent events and events of a more distant past. Sometimes yesterday is
distinguished from before yesterday, or a few days ago is distinguished from
more than a few days ago. Sometimes there is a special way of marking
a remote or legendary past. As far as the future is concerned, it may be treated
as symmetric with the past as in Bamileke, or with fewer distinctions as in
Washo. Common distinctions in the future are between today or tomorrow
versus later than tomorrow, or between the next few days and later than the next
few days.
Figure 17.2 summarizes the kinds of time periods that are distinguished in
various languages. The borders between them may be vague and different
languages may define them in different ways. Comrie noted that in English,
day before yesterday

day after tomorrow

before my lifetime

about to happen
some time ago

in the future
later today
yesterday

tomorrow

someday
just now
today

the present moment

Figure 17.2 Some linguistically distinguished divisions of time

142 Common Ways of Orienting Thoughts

even though “today” begins technically at midnight, in ordinary usage it begins

when one wakes up, whereas there are other languages in which it begins with
the preceding sunset so that the preceding night also belongs to today (Comrie
1985: 89).
These divisions show that locations on a time line are not arbitrary but are
governed by qualitative differences in the ways people think of events at
different degrees of removal from the present, as well as by partially parallel
differences in the ways people imagine events for the future. Introspection may
confirm that things that happened a few minutes ago are experienced differently
from things that happened yesterday or last month or ten years ago. Similarly,
what we anticipate happening a few minutes from now or later today differs in
quality from what we expect months or years from now. There may be no
reason to believe that people experience a time line as such. It seems more
likely that they experience memories of past events and anticipations of future
events with qualitative differences that lead to these distinctions.

The Present Moment

The fulcrum of Figure 17.2 is the present moment. The present cannot be
a point in any literal sense; how would one experience a point? To quote
William James once again:
The practically cognized present is no knife-edge, but a saddle-back, with a certain
breadth of its own on which we sit perched, and from which we look in two directions
into time. The unit of composition of our perception of time is a duration, with a bow
and a stern, as it were – a rearward- and a forward-looking end. (James 1890: 609)
James was anticipating the focus of consciousness described in Chapter 8:
whatever passes through a speaker’s consciousness during a momentary perch-
ing on a brief segment of experience before consciousness moves restlessly on
to the next. The time occupied by each perching defines the present moment,
the “specious present,” which is available to be verbalized as an intonation unit.
With a different metaphor we can picture the present moment as a small
window of focal consciousness, in front of which focused segments of
experience are continually passing, usually at one- to two-second intervals.
Focal consciousness has something in common with foveal vision in the
sense that both are surrounded by a larger periphery that provides a context
for them (Chafe 1994: 140). The events of our lives are in constant change,
and experiencing this constant change as it passes before the window of
consciousness may very well be responsible for the way we experience
time itself.
Figure 17.3 reduces time and space to a contrast between the present time
and place and everything that lies outside it. Immediate experience focuses on
 Time, Tense, Memory, and Imagination 143

displaced

displaced
the present moment

Figure 17.3 Immediate versus displaced experience

the here and now, but the input to it consists in part of ideas that occupied
immediate consciousness at an earlier time and are now reactivated. That is
what we call remembering. Other ideas are anticipated for a later time.
Thoughts, furthermore, are not limited to our own immediate experiences,
memories, and anticipations but often include things we have learned from
others. And a significant proportion of our conscious life contains ideas of
events that were never experienced by anyone but are products of imagination,
our own or someone else’s.
Focal consciousness, in short, includes a mixture of experiences immedi-
ately perceived, remembered, or anticipated, experiences learned from others,
and experiences that are only imagined. The boundaries separating these
sources of conscious experience are not rigid. Even immediate experiences
are likely to be interpretations to some degree, and not veridical reproductions
of what is out there. The brain is a creative organ, even when its creativity is
constrained by input from outside.
Research of a different kind has suggested that different areas of the brain are
involved in immediate (“stimulus-driven”) thought and displaced (“stimulus-
independent”) thought (e.g. Spreng and Grady 2009). Displaced thought may be
associated with the “default mode network,” a set of brain structures that are
activated together during the experiencing of autobiographical memory, anticipa-
tion (“prospection”), and empathy (curiously labeled “theory of mind”).
In contrast, immediate experience is associated with a different set of brain areas
dubbed the “task-positive network”; e.g. Fox et al. (2005). One can hope that
further brain studies will harmonize with the linguistic evidence presented here.

The Order and Prosody of Temporal Adverbs

Years ago I explored certain linguistic evidence for different levels of memory
as they relate to the order and prosody of temporal adverbs (Chafe 1973). From
evidence gathered at that time I identified what I called surface, shallow, and
deep memory with parallels in surface, shallow, and deep anticipation of the
future. Although I would now conduct that study differently, its findings need
not be entirely discarded.
144 Common Ways of Orienting Thoughts

Temporal relations and corresponding levels of memory are signaled by tense,

but also by adverbial expressions like “yesterday” or “three weeks ago.”
I hypothesized that experiences from the most recent past, although they are
verbalized with past tense, are not marked with adverbs. A suggestive piece of
evidence was provided by the comic strip Dennis the Menace, where the
following conversation took place between Dennis and his parents. Humor
arose from the fact that Dennis omitted a temporal adverb in spite of the fact
that he was talking about something that happened many years before. His
parents interpreted the absence of an adverb as signaling a time very close to
the present, residing in what I then called surface memory.
father: Hi, Dennis! What’s the news?
dennis: Somethin’ terrible! Did ya know Mr. Wilson broke his arm?
father: No!
mother: How awful!
dennis: He fell down his cellar stairs!
mother: The poor man!
father: When did this happen, Dennis?
dennis: When he was a little kid my age. He jus’ told me about it today!

I also hypothesized that experiences in shallow memory, perhaps extending

back through yesterday in Figure 17.2, are marked with adverbial expressions
in sentence-final position and weak prosody, as with the word yesterday in he
fell down the stáirs yesterday, while experiences in deep memory are marked
with adverbial expressions in sentence-initial position and prosodic promi-
nence, as in lást yéar he fell down the stáirs. If a similar study were conducted
today, it would profit from access to corpora of natural speech and techniques of
electronic sound processing. A sample of what might be achieved is offered by
the following recorded sentence:
Day before yesterday was the most weird day I’ve ever seen in my entire adult life.
Figure 17.4 traces the fundamental frequency of that sentence. The prominence
of the initial adverbial phrase day before yesterday is evident. The depth of
memory communicated by that phrase corresponds to the location of “day
before yesterday” or earlier in Figure 17.2. The extent to which such a finding
might be repeated remains to be seen.

Figure 17.4 F0 showing prominence on an initial temporal adverb

 Time, Tense, Memory, and Imagination 145

A decade after that study was published the linguist Östen Dahl wrote the
following (note the date):
I think it would be a bit rash to say that the existence of morphological categories
marking temporal distance provides definite evidence for the kind of structuring of
human memory that Chafe proposes. What can be said, though, is that it would be rather
strange if these categories did not reflect any general properties of human cognition and
that the uniform character of remoteness systems from different parts of the world
constitutes a challenge for cognitive psychology.
It should certainly be possible for an experimental psycholinguist with interest for the
structure of memory and access to a sufficient number of informants for languages with
remoteness distinctions to construct suitable experiments to find any possible connections.
It seems that psychologists have paid relatively little attention to the question whether
there is any qualitative differentiation of information within long-term memory. As far
as I know, there has been no reaction to Chafe’s paper in the ten years that have elapsed
since it was written. Time may now be ripe. (Dahl 1984: 120)

Marigold Linton
Despite the attention that has been given to the psychology of both time and
memory, a restriction to laboratory studies has sidetracked the investigation of
questions like those raised here. Marigold Linton, unusual among psycholo-
gists in her interest in interactions between the mind and the real world,
lamented, “What do we know about the contents of human memory?
A preliminary listing might be obtained by surveying the literature. Such
a review shows, however, that we know precious little about the denizens of
that natural habitat, the mind . . . Why has cognitive psychology not begun with
a basic understanding of what inhabits the mind?” (Linton 1986: 50).
During a six-year study she conducted in the 1970s, Linton wrote down
every day brief descriptions of at least two events that took place in her life
during that day. Once a month she drew pairs of items from this growing pool
of events, tried to date them, and tried to reconstruct their relative chronology.
Several findings are of interest here. One was that specific chronology tends
soon to be lost, although it may be preserved when there is a relation of an
event to another event that was salient enough to carry a date. (I remember
very well what I was doing on September 11, 2001.) She also found that
repeated exposure to events that are similar “makes specific episodic knowl-
edge increasingly confusible, and ultimately episodes cannot be distin-
guished” (Linton 1982: 79). In short, as time passes (1) events tend to lose
their place in chronology, and (2) similar events tend to lose their uniqueness.
It may well be that a subconscious awareness of these changes in the quality
of memories influences at least some of the distinctions captured in
Figure 17.2.
146 Common Ways of Orienting Thoughts

Other Evidence for the Changing Quality of Memories

The observation that some languages distinguish events earlier today from
events before today raises the question of the extent to which those
distinctions are influenced by sleep. How is memory affected by a single
night of sleep, and to what extent are multiple instances of sleep respon-
sible for other divisions in Figure 17.2? Sleep researchers have generally
focused on how well people remember an artificial stimulus rather than
differences in the quality of memories for naturally occurring events.
Evidence of the sort presented in this chapter suggests that once an
experience has lost its immediacy (1) it loses detail progressively as more
and more instances of sleep intervene, and (2) the details that remain are
gradually confused with other knowledge.
We can look to the British psychologist Frederic Bartlett (1932) for relevant
observations. In what he called The Method of Repeated Reproduction, Bartlett
asked people to read a Native American (Kathlamet) folktale originally recorded
by Franz Boas (1901), and then to write what they remembered of it after
different periods of time. The results suggested (1) a gradual loss of detail and
(2) the gradual importation of other knowledge. The story began as follows:
One night two young men from Egulac went down to the river to hunt seals, and while
they were there it became foggy and calm. Then they heard war-cries, and they thought:
“Maybe this is a war-party.” They escaped to the shore, and hid behind a log. Now
canoes came up, and they heard the noise of paddles, and saw one canoe coming up to
them. There were five men in the canoe, and they said:
“What do you think? We wish to take you along. We are going up the river to make
war on the people.”
People were asked to write their memories of this story after varying intervals
of time. One person began his first written version as follows (p. 72):
Two youths were standing by a river about to start seal-catching, when a boat appeared
with five men in it. They were all armed for war. The youths were at first frightened, but
they were asked by the men to come and help them fight some enemies on the other
bank.
The loss of detail is evident. As for the importation of other knowledge, the
original said nothing about the five men being armed, although it could be
inferred from their plan to make war. The idea that the young men were
frightened, which was also absent from the original, could be inferred from
their escaping and hiding. “Up the river” was reinterpreted as “on the other
bank,” and the targets of the attack were now described as enemies. Two weeks
later the same man wrote (p. 73):
There were two young men who once went out in the afternoon to catch seals. They were
about to begin when a boat appeared on the river and in it were five warriors. These
 Time, Tense, Memory, and Imagination 147

looked so fierce that the men thought they were going to attack them. But they were
reassured when they asked the youths to enter the boat and help them to fight some
enemies.

Introduced here was the location in the afternoon as well as the fierceness of the
five men. After another month this man wrote (p. 73):
Two youths went down to the river to fish for seals. They perceived, soon, coming down
the river, a canoe with five warriors in it, and they were alarmed. But the warriors said:
“We are friends. Come with us, for we are going to fight a battle.”
Added was the statement that the five men were coming down the river and that
they were friends. After two more months (p. 74):
Two youths went down to the river to hunt for seals. They were hiding behind a rock
when a boat with some warriors in it came up to them. The warriors, however, said they
were friends, and invited them to help them to fight an enemy over the river.

The original said they hid behind a log, which had now become a rock. After
two years and six months the story was still vaguely remembered, but its
beginning was reduced to the following (p. 75):
Some warriors went to wage war against the ghosts.
Among his conclusions Bartlett mentioned that “in long-distance remember-
ing, elaboration becomes rather more common in some cases; and there may be
increasing importation, or invention, aided . . . by the use of visual images”
(p. 93). Here we can note the creativity of incorporating other material, but also
the attenuated detail that was evident even after several months, but more
strikingly after several years. Chafe (1986) extended Bartlett’s study to spoken
language.

The Pear Film

When we made the Pear Film we wanted among other things to explore how
thoughts were related to reality. We weighed various possibilities for recording
different people reacting to the same event, including bees swarming outside
the Berkeley campus library and a devastating church fire in San Francisco. But
because we were also interested in seeing how speakers of different languages
would interpret a shared reality, we settled on producing a film that could be
shown in different locations around the world (Chafe 1980). The film began
with a man picking pears in a tree, then another man who walked past with
a goat, and after that a boy who approached on a bicycle. It is not uncommon to
find people’s thoughts multiplying events or the number of participants in
them:
148 Common Ways of Orienting Thoughts

And you see passerbyers on bicycles and stuff go by. (Speaker 8, p. 307)
A number of people are going by. (Speaker 9, p. 308)
There was variability in the way people viewed the pear picker:
It’s like he’s been doing this all day, and it’s just a monotonous kind of thing for him.
(Speaker 2, p. 303)
You got the feeling that he pretty much liked his pears, because he was so gentle with
them. (Speaker 12, p. 311)
He’s . . . sort of not inclined to do a lot of work, and uh sort of a little bit slow mentally.
(Speaker 18, pp. 317–318)
One member of our group, Deborah Tannen, showed the film in Athens and
recorded the following descriptions of this man from two different people
(edited from Tannen’s translations from the Greek):
It showed how a person gathered the pears, and insisted that that which he did he
lived. In other words the fact that he was cultivating the earth, that he was gathering the
harvest, was for him something special. It was worth something, he lived that which he
did, he liked it. (Speaker G12, p. 68)
His movements basically gathering the fruits don’t show a person who loves them
very much, he pulls them very I don’t know. I didn’t like generally the way he was
pulling them. (Speaker G15, p. 63)
The Greeks were especially prone to assigning their own interpretations to what
they saw. Tannen mentioned “the overriding effect of the Greek narratives as
‘good stories’” (p. 80).

Eyewitness Testimony
Perhaps nothing shows better the creativity of memory than the fallibility of
eyewitness testimony, as famously explored and described by Elizabeth Loftus
(e.g. Loftus 1996, 2003). The real-world consequences of the creative remem-
bering of crime scenes can be devastating for those who are falsely accused.
Loftus divided the stages at which creativity may take place into (1) the initial
perception of events, (2) the retention of events in memory, and (3) their
subsequent retrieval when witnesses are questioned. Each stage is open to
distortions of “what really happened.” Perception can be distorted by expecta-
tions: “We can identify four different sorts of expectations that will affect
perception: cultural expectations or stereotypes, expectations from past experi-
ence, personal prejudices, and momentary or temporary expectations. When
any of these are present, they can distort perception; the perceptual material that
enters stored memory will accordingly be distorted in a manner consistent with
the expectation” (p. 37). Retention in memory can also be creative:
 Time, Tense, Memory, and Imagination 149

During the time between an event and a witness’s recollection of that event – a period
often called the “retention interval” – the bits and pieces of information that were
acquired through perception do not passively reside in memory waiting to be pulled
out like fish from water. Rather, they are subject to numerous influences. External
information provided from the outside can intrude into the witness’s memory, as can
his own thoughts, and both can cause dramatic changes in his recollection.
(1996, pp. 86–87)

And retrieval can introduce further distortions:

Most people, including eyewitnesses, are motivated by a desire to be correct, to be
observant, and to avoid looking foolish. People want to give an answer, to be helpful,
and many will do this at the risk of being incorrect. People want to see crimes solved and
justice done, and this desire may motivate them to volunteer more than is warranted by
their meager memory. The line between valid retrieval and unconscious fabrication is
easily crossed. (1996, p. 109)
In short, while thoughts are obviously influenced by “the real world” outside
the thinker, distortions are already introduced as reality is perceived, and these
distortions only multiply as thoughts are retained in memory and later
retrieved.

Summary
We have seen how observations derived from introspection and from language
converge on a view in which the immediacy of the present moment is char-
acterized by:
(1) an unbroken continuity in the flow of consciousness,
(2) a wealth of detail, and
(3) direct involvement in a person’s emotional experiences.
In contrast, displacement from the present moment is characterized by:
(1) isolated islands of experience,
(2) attenuation of detail, and
(3) an absence of direct emotional involvement.
Displaced experiences that arise from memory pass through approximately the
following stages:
(1) Earlier the same day: less detailed and less continuous than the present.
(2) Yesterday, after sleep has intervened: more attenuation and already an
island-like quality.
(3) Earlier than yesterday: still more attenuation, the importation of extraneous
material, the coalescing of similar events into a single event type, and the
loss of temporal ordering.
150 Common Ways of Orienting Thoughts

(4) Remote time, with experiences often known from outside sources rather
than memories of one’s own experiences.
Imagining the future may loosely follow a similar course.
Toward the end of this chapter there was mention of other sources of
evidence for the quality of immediate and displaced experiences, including
the order and prosody of temporal adverbs (Chafe 1973), long-term diary
studies (Linton 1982), and memory for artificially presented materials in the
manner of Bartlett (1932) and its extension to spoken language in Chafe (1986).
Eyewitness testimony provides further evidence for the creativity of memory
(Loftus 1996, 2003).
In Chapter 15 psychological findings dovetailed nicely with those in linguis-
tics, the only regret being an absence of interdisciplinary awareness.
Unfortunately the same cannot be said of research on memory, where the use
of artificial data has produced a huge literature on “working memory,” “episo-
dic memory,” “semantic memory,” and related topics. In a welcome collection
of articles on “remembering in natural contexts” Ulrich Neisser wrote, “If X is
an interesting or socially significant aspect of memory, then psychologists have
hardly ever studied X” (Neisser 1982: 4). Although Neisser subsequently found
the situation to have improved, the concerns of this chapter show little overlap
with the psychological tradition.
18 Relating Ideas to Reality

There appears to be a nearly universal belief in the existence of a real world that
is known in part through sensory contact with it, but is also assumed to exist
beyond any direct experiencing of it. At the same time, however, everyone
realizes that thought includes not only what belongs in that real world but also
much that does not and is known only through imagination. Different lan-
guages structure ideas of this second kind in different ways, and this chapter
samples a few of them.
Franz Boas worked extensively with the Kwakiutl (or Kwak’wala) language
in British Columbia. In his Introduction to the 1911 Handbook of American
Indian Languages Boas mentioned that if a Kwakiutl speaker wanted to say
that a certain man was sick, “in case the speaker had not seen the sick person
himself, he would have to express whether he knows by hearsay or by evidence
that the person is sick, or whether he has dreamed it” (Boas 1963 [1911]: 43).
In his posthumously published Kwakiutl Grammar he mentioned how “a small
group of suffixes expresses source and certainty of knowledge, such as ‘accord-
ing to hearsay’, ‘experienced in a dream’, ‘evidently’, ‘probably’; also ‘con-
trary to fact’” (Boas 1947: 206). Later Edward Sapir noted “how frequently the
form [of a grammatical category] expresses the source or nature of the speak-
er’s knowledge (known by actual experience, by hearsay, by inference)” (Sapir
1921: 114–115).
These were early mentions of a concern for epistemology, which in some
languages may overshadow the concern for time that characterizes languages
of Europe and was discussed in Chapter 17. Tense has been a recognized
grammatical category for several thousand years, but epistemology has only
recently been given the attention it deserves. If linguistics had been invented by
speakers of Native American languages the emphasis might have been
reversed. As noted by Alexandra Aikhenvald:
Up until the late nineteenth century only those linguistic categories which were found in
classical Indo-European languages were accorded due status and investigated in some
depth. Since these language have not grammaticalized information source, the concept
of evidentiality had not made its way into linguistics until “exotic” languages started

151
152 Common Ways of Orienting Thoughts

being described in terms of categories relevant for them, rather than from a limited Indo-
European perspective. (Aikhenvald 2004: 11)
Martha Hardman, who worked with South American languages in which
epistemology is fundamental, noted how earlier investigators called it nothing
more than a decoration: “ornate particles” that “serve no other function than to
adorn the sentence” (Hardman 1986: 113). A survey of other early references in
this area is available in William H. Jacobsen (1986).
A concern for evidentiality or knowledge source is conspicuous in many
languages of the western hemisphere. It attracted scholarly attention when it
was described for the Wintu language of northern California by Dorothy
Demetracopoulou Lee. She wrote (1938: 92) that if I wanted to translate the
English sentence Harry is chopping wood into Wintu, I would have five
options:
(1) Harry kupake (if I knew this by hearsay)
(2) Harry kupabe (if I see or have seen Harry chopping)
(3) Harry kupante (if I hear him, or if a chip flies off and hits me)
(4) Harry kupare (if I have gone to his cabin to find him absent and his axe gone)
(5) Harry kupael (if I know that Harry has a job chopping wood every day
at this hour, that he is a dependable employee, and, perhaps, that he is not in
his cabin)
The article in which these observations appeared was titled “Conceptual
Implications of an Indian Language.” Lee began as follows:
It has been said that a language will delineate and limit the logical concepts of the
individual who speaks it. Conversely, a language is an organ for the expression of
thought, of concepts and principles of classification. True enough, the thought of the
individual must run along its grooves; but these grooves, themselves, are a heritage from
individuals who laid them down in an unconscious effort to express their attitude toward
the world. Grammar contains in crystallized form the accumulated and accumulating
experience, the Weltanschauung of a people. The study which I propose to present is an
attempt to understand, through a study of grammar, the unformulated philosophy of the
Wintu tribe of California. (Lee 1938: 89)
Lee was expressing the view discussed in Chapter 13 that the relation
between language and thought is established by a people’s “accumulated
and accumulating experience,” making language “an organ for the expres-
sion of thought.” Speakers of Wintu, it seems, paid special attention to
sources of their knowledge.
For the most part we can translate her examples into English:
(1) I understand that Harry is chopping wood (hearsay)
(2) I see that Harry is chopping wood (visual evidence)
(3) It sounds like Harry is chopping wood (auditory evidence)
(4) Harry must be chopping wood (inference)
 Relating Ideas to Reality 153

English differs from Wintu, however, in the fact that these expressions are not
part of a commonly employed or even obligatory verb structure. We can, if we
wish, use these longer phrases to convey similar meanings in English, but they
are not imposed on us by English grammar. It is worth noting too that, whereas
Wintu has a straightforward way of marking a piece of knowledge as learned
from someone else (the “hearsay” option), the use in English of “I understand
that” is only one among a variety of possibilities including “it appears that,”
“I hear that,” “supposedly,” etc. Wintu makes one distinction that is not as easy
to reproduce in English, distinguishing inference based on perceptible evi-
dence, as in 4 in the Wintu list, from inference based on familiar patterns of
behavior as in 5. The use of “must” in 4 of the English list can have either
origin.
What may have been the first conference devoted exclusively to this topic
took place in Berkeley in 1981. The resulting publication (Chafe and Nichols
1986) was titled Evidentiality: The Linguistic Coding of Epistemology, a title
from which one might have inferred that “evidentiality” and “epistemology”
were names for the same phenomenon. In the present chapter I follow
Aikhenvald in restricting “evidentiality” to knowledge source, where “evi-
dence” is the main issue. For closeness to reality I borrow Boas’s term
“certainty” and reserve “epistemology” to cover both. Evidentiality and cer-
tainty are not unrelated, for the source of one’s knowledge may very well
influence one’s judgment of its degree of certainty.

Evidentiality
A wide-ranging study of evidentiality has been provided by Aikhenvald
(2004), who classified the types of knowledge sources that are variously
recognized by languages across the world in a way that closely mirrors Lee’s
Wintu categories but divides hearsay into hearsay and quotation (Aikhenvald
2004: 63–64):
(1) Visual: covers information acquired through seeing.
(2) Non-visual sensory: covers information acquired through hearing, and is
typically extended to smell and taste, and sometimes also to touch.
(3) Inference: based on visible or tangible evidence, or result.
(4) Assumption: based on evidence other than visible results; this may include
logical reasoning, assumption, or simply general knowledge.
(5) Hearsay: for reported information with no reference to those by whom it
was reported.
(6) Quotation: for reported information with an overt reference to the quoted
source.
154 Common Ways of Orienting Thoughts

It can be seen that 1 and 2 depend on knowledge acquired through the senses,
with vision given special place, while 3 and 4 involve some sort of reasoning,
based either on specific evidence or on general knowledge, the latter called
“assumption.” As for 5 and 6, both involve knowledge acquired from others.
Hearsay does not identify those others, nor does it pretend to reproduce specific
language through which the knowledge was acquired. Quotation does identify
the verbal source, and pretends to repeat what was said (direct quotation), or
settles for the gist of it (indirect quotation). Lee’s Wintu examples included all
but 6.
Hearsay may be the most ubiquitous of these knowledge sources. Seneca and
Caddo and many other languages have a special way of marking knowledge
that was not acquired through the speaker’s firsthand experience but was
acquired secondhand, often but not necessarily through language. In Seneca
the hearsay marker has the form gyö’öh, which is often translated “it is said.”
This word occurs frequently in stories, which almost always begin with a fixed
phrase translatable as “long ago it is said.” Compare English “once upon
a time” with its temporal rather than evidential focus. The word gyö’öh is
often inserted repeatedly within the body of a story; in one narrative it con-
stituted as much as 14 percent of the total word occurrences.
Gyö’öh is not confined to stories but also appears in conversations.
As a man was describing a nearby river he said something that can be
translated “they used to go down the river on rafts,” but he added gyö’öh to
show that he had not seen them doing that himself. Telling about a bear she
herself had not seen, a woman said, “it was big gyö’öh” to make it clear that
the bear’s size was secondhand information. Another woman who was
talking about a man who was sick qualified her statement with gyö’öh to
suggest that she had no direct experience of his sickness and implying that it
might be feigned. The need to identify knowledge as hearsay is so ingrained
in the minds of Seneca speakers that they sometimes insert gyö’öh into their
English as well.
The Caddo language uses the word bah’nah in virtually the same way, but
whereas Seneca gyö’öh has no obvious etymology, the Caddo word is trans-
parently derived from a verb meaning “say.” Hearsay status in Caddo is in
addition signaled by a verb prefix kan-, which combines with past tense in the
form kín- that appears at the beginning of the fourth word of the following
(Chafe 1977: 27). Caddo often marks hearsay evidence redundantly, using both
bah’nah and kan- or kín-, just as English may mark occurrence in the past with
both an adverb and a past tense verb.
Ahya’ tiki: bah’nah, kínkambašuh ’wan’ti’ tikiy.
in the past far hearsay past.hearsay water dried up all far
“Long ago it is said water is said to have dried up far and wide.”
 Relating Ideas to Reality 155

Certainty
Probably all languages have ways of recognizing whether an event or state is or
is not “real.” People know that their thoughts are not always occupied with
things they regard as the way things really are but that many of them are
speculative or even in conflict with reality. In what follows we look at two
different ways in which languages recognize degrees of certainty.

A Three-Part Scale
In Seneca a verb that expresses the idea of an event necessarily begins with one
of three prefixes. The following are three possibilities offered by the idea of
hanging something up. At issue are the epistemological implications of the
prefixes o’-, ë-, and a::
(1) o’kniyö:dë’ “I hung it up”
(2) ëkniyö:dë’ “I’ll hang it up”
(3) a:kniyö:dë’ “I might hang it up”
The first two examples might suggest that the prefixes o’- and ë- mark past and
future tense. Several observations, however, cast doubt on that conclusion.
Most conspicuous is the fact that the third prefix, a:-, does not mark tense.
The translation “might” is a way of capturing a speaker’s judgment that an
event has not occurred and that its occurrence is only a possibility, not a known
fact. That is a judgment of certainty, and the prefix can be labeled “hypothe-
tical.” From that perspective we can look back at the first two prefixes and ask
whether they might also involve certainty.
The first of them, o’-, was translated with a past tense, but viewing it
epistemologically suggests that an event remembered from the past is one
whose closeness to reality is derived from the fact that it actually happened.
The past-tense translation selects one property of such an event, its relation to
time, but the Seneca can be understood as marking its factuality instead, for
there are other uses in which this prefix does not convey a past event at all, but
one that is simultaneous with the act of speaking. Thus the verb o’ge:gë’, which
can sometimes be translated “I saw it,” is equally appropriate if I am looking
out the window and see something at this very moment: “I see it.” Its agreement
with reality stems in that case, not from its prior occurrence but from my
immediate perception of it. The o’- prefix can thus be appropriately labeled
“factual.”
The prefix ë- was translated with a future tense: “I’ll hang it up.” An event
predicted for the future cannot be factual but has a higher probability of
actually occurring than an event marked as hypothetical. Furthermore, just as
the o’- prefix does not always signal past tense, ë- is frequently found in polite
156 Common Ways of Orienting Thoughts

imperatives, as in the instruction ëhsa:wa:k “you will sift it” said during
a cooking class. In some contexts it also implies only a possibility of occurring:
ëgéyë’he’t “I will learn it” may have the force of “I can learn it,” something
more likely to occur than a similar event conveyed by the hypothetical prefix in
a:geyë́ ’he’t “I might learn it.” An appropriate label is “predictive.” Thus we can
arrange these three prefixes on a continuum from factual to predictive to
hypothetical:
(1) o’- “factual”
(2) ë- “predictive”
(3) a:- “hypothetical”
Mithun (2016) provides a detailed account of the ways languages of the
Iroquoian family deal with the areas traditionally called mood and modality.
A useful survey of meanings in this area is provided by Jan Nuyts (2016).

A Binary Distinction with Subtypes

A very different way of evaluating the certainty of knowledge is illustrated by
the Caddo language. Like Seneca and many other polysynthetic languages,
Caddo verbs contain prefixes that refer to the participants in events and states.
A special property of these Caddo prefixes is their division into two subsets that
have been labeled “realis” and “irrealis” (Chafe 1995). The choice between the
two is governed by whether the speaker judges an event or state to conform to
what one believes to be real, or whether it belongs to a world that is only
imagined. Usages in the irrealis category include the following (Melnar 2004:
49–55):
(1) yes–no questions sày:báwnah “Have you seen him?”
(2) negations kúyt’áybáwnah “I haven’t seen him.”
(3) prohibitions kaššáy’bah “Don’t look at it!”
(4) obligations kassánáy’aw “He ought to sing.”
(5) conditions hít’áybah “If I see it.”
(6) simulations dúyt’áybah “As if I saw it.”
(7) infrequentatives wást’áybah “I seldom see it.”
(8) surprises húsba:sáy’k’awihsa’ “He knows my name!”

With the possible exception of 8, it can be seen that the event in question is
one that is judged to be in less than complete accord with reality, but one that is
imagined as only possibly occurring or even not occurring at all. Differing
prefixes distinguish irrealis events as (1) yes–no questions (where the answer is
not yet known), (2) negations (where the event does not occur), (3) prohibitions
(instructions not to do something), (4) obligations (so-called deontic modality),
(5) conditions (imagined events that may cause other events), (6) simulations
(imagined analogies), or (7) events that occur only infrequently. The usage in
 Relating Ideas to Reality 157

(8) illustrates what has been labeled “mirative” (Aikhenvald 2004: 195–215).
An event or state that is surprising departs from expected reality. This word
appeared in in both versions of the Caddo story in Chapter 12, where the turkey
exclaimed, “My goodness he knows my name!”

Summary
This chapter illustrated a few ways in which different languages distinguish
two major properties of thought: evidentiality (the source of a thought) and
certainty (its closeness to what is regarded as real). Evidentiality was illustrated
with examples from Wintu, Seneca, and Caddo, certainty with examples from
Seneca and Caddo. These examples are a small sampling of the epistemological
diversity displayed by different languages.
Part VI

The Emotional Component of Thoughts

19 Emotional Involvement in a Conversation

In Chapter 8 we saw how the flow of thought is reflected in prosody – in pitch,

loudness, timing, and voice quality. Prosody performs the four functions listed
at the bottom of Figure 19.1 (Chafe 2000a). Two of these functions involve the
way thoughts are organized and two of them involve the way they are eval-
uated. In signaling the organization of discourse, prosody delimits units such as
words, phrases, sentences, and topics. At the same time it signals relations
between those units and their larger contexts, as when a rising pitch at the end
of a phrase shows there is more to come and a falling pitch shows closure. With
respect to the way thoughts are evaluated, some elements stand out more
prominently than others, expressing new information, contrast, or emphasis.
Our concern in this chapter, however, is the role of prosody in expressing
emotions. The first three items are realized in some fashion in any sample of
speech, which is always segmented into units that are related in various ways
and which show various degrees of prominence. While emotions are nearly
always signaled with prosody, the absence of such marking may itself commu-
nicate an attitude of emotional detachment.
One need only listen to any sample of speech to appreciate the continually
varying contributions of pitch, volume, timing, and voice quality in commu-
nicating a speaker’s emotions or their absence. They may range from detach-
ment at one extreme to strong involvement at the other, and specific prosodic
patterns convey a variety of emotional states. Few studies have examined in
detail this relation of emotion to the prosody of naturally occurring speech, but
observations in that area have the potential to offer important insights into
varieties of emotional experience.
Previous work in this area has been largely experimental. Some of it has used
actors to simulate standard emotions like anger, sadness, or joy, testing whether
and how well subjects can identify those emotions from their acoustic signals
(e.g. Bezooyen 1984). The most detailed and ramified experiments were those
conducted by Klaus Scherer and his collaborators during the 1970s and 1980s.
A summary of that work is available in Goldbeck, Tolkmitt, and Scherer
(1988), which mentioned that “a close collaboration between psychology and
the language sciences will be required to disentangle the complex web of

161
162 The Emotional Component of Thoughts

functions
of prosody

organization evaluation

units relations prominence emotions

Figure 19.1 The functions of prosody

factors that determine human vocal expression” (Goldbeck et al. 1988: 137).
This chapter offers a linguistic contribution to that effort, focusing on emo-
tional expression in the course of a conversation.
Prosody often expresses a generalized heightening of emotional involve-
ment that is associated with a variety of experiences. Much of the discussion in
this chapter focuses on fundamental frequency and its perceptual correlate in
pitch, but timing, volume, and voice quality may enter the picture as well. In the
course of one recorded conversation the discussion turned to the way an
acquaintance dressed, and particularly to the shoes he wore. Someone said:
That was the ugliest set of shoes I ever saw in my life!

As written, this statement can be read as a relatively straightforward negative

evaluation of part of the man’s attire. The exclamation point at the end may
suggest something of its prosody, but the full range of involvement is
shown more adequately in the display of fundamental frequency and timing
in Figure 19.2. Noteworthy are the prolonged rise–fall pitch contours on the
first syllable of ugliest and, less exaggerated, on the word life. Rise–fall
contours, sometimes supplemented with lengthening as here, are common
signals of emotional involvement.
Figure 19.3 and those that follow are taken from a conversation in which the
prosody of the principal speaker, Kay, fluctuated considerably with degrees of
her involvement. One of the topics in this conversation was initiated by Sue as
follows.7
(1) Sue: What I’d like to find out more about.. is.. your operátion.
(2) Sue: That you’re having (laugh).
(3) Kay: My nô=se operation.
(4) Sue: Yes.. your nose operation.

7
Accent marks are inserted where they are relevant to the discussion. An acute accent shows
heightened pitch, a grave accent a fall, and a circumflex accent a rise–fall. The equals sign shows
lengthening of the preceding vowel.
 Emotional Involvement in a Conversation 163

u– life
was the
– gli est set of shoes I
ever saw in my

Figure 19.2 Extreme involvement

If one were to read line 3 in an uninvolved way, one could imagine it being
spoken with either of two prosodic contours. One of them would have heigh-
tened pitch on nose and a further rise at the end. That contour is suggested by
the question mark here:
My nóse operation?
Spoken in that way, line 3 would have requested confirmation of the operation
Sue had in mind. With a different prosody, the peak on nose would be followed
by a fall that continued to the end, as suggested by the period here:
My nóse operation.
Spoken in that way, line 3 would have expressed an emotionally neutral
confirmation of the topic Sue had just introduced.
What Kay actually said differed from both of these, although it had more in
common with the second. The word nose was spoken with a rise–fall contour,
the peak of which reached 396 Hz, well above Kay’s range when she was
speaking in a more neutral way. Figure 19.3 traces the fundamental frequency
(F0) on the words my nose. The word nose was prolonged, occupying more than
half a second, as compared with Kay’s average speaking rate of about 150
milliseconds per syllable.
The idea of Kay’s nose had not been activated previously in the conversation
and was thus a new idea. We would expect it, then, to be uttered with a higher
pitch than my, which expressed the given idea of Kay herself. The same can be
said of the word operation, which expressed an idea that had already been
activated by Sue. In less involved utterances Kay expressed new ideas with F0
levels no higher than about 260 Hz. Why, then, did nose rise to almost 400?
Kay was conveying her acceptance of a newly introduced topic, but it was a
topic that was emotionally loaded. The operation promised to be a major event
in her life, and Sue’s introduction of it aroused in Kay a feeling of high
involvement. By way of comparison the following example and Figure 19.4
illustrate Kay’s less involved speech. The maximum frequency was 264 Hz on
the first syllable of “maybe,” with a minimum 163 Hz at the end of the entire
sequence. Both maybe and why showed minor rise–fall contours that expressed
mild, brief, and localized involvement. The word maybe qualified the factuality
of what Kay was saying – there was no certainty that she should preface this
164 The Emotional Component of Thoughts

400 nose

340

280
my
220

Figure 19.3 F0 of “my nose” in line 3

400

320
maybe I should preface this thing
240 ... why I want it done.
Well,
160

Figure 19.4 Minor involvement

400
great!

340

280 That was

220

Figure 19.5 Exciting news

thing as she did. The word why signaled a contrast with Sue’s request, which
focused on the operation itself rather than the reason for it. A brief rise–fall
contour can serve to localize minor involvement on a single word in this way.
(1) Well,
(2) . . . mâybe I should preface this thing whŷ I want it done.
For some time Kay had entertained the thought that she could modify the
shape of her nose, but because of the unwillingness of her insurance company
to pay for cosmetic surgery she had not acted on this wish. Then came the
exciting discovery that she actually had a medical condition that her insurance
would cover. She reacted with the highly positive evaluation of this news
shown in Figure 19.5.
 Emotional Involvement in a Conversation 165

460

390 Sure,

320
it’s a medical problem
250
She said,
180

Figure 19.6 Quoting the nurse

Her nurse proceeded to confirm this finding as shown in Figure 19.6.

Quoting another person’s speech often leads to a raised pitch level, but quoting
the nurse here led to a peak F0 as high as 432 Hz on the word sure, evidently
conveying excitement on the part of Kay herself rather than the nurse.
(1) . . . She said . . . súre.
(2) . . . It’s a médical problem.
In what followed, similar rise–fall peaks extending to nearly 500 Hz
occurred in the two instances of the word no that answered Sue’s ques-
tions in lines 1 and 3, reaching 490 Hz in 2 and 496 Hz in 4. Here we
can see also that, whereas ideas tend to be replaced at intervals of one or
two seconds, an emotion dissipates more slowly, often extending across a
sequence of ideas:
(1) Sue: Didn’t you notice you had a deviated septum?
(2) Kay: . . . Nô.
(3) Sue: You couldn’t breathe through one side of your nose?
(4) Kay: . . . Nô.
Later in this conversation Kay mused as follows:
(1) Kay: . . . Well,
(2) . . . imagine this.
(3) .. I mean all those years you’ve been keeping yourself away from muggers,
(4) . . . rapists,
(5) . . . things like that,
(6) and then all at once,
(7) . . . you díe.
(8) . . . Because you’ve got a nose op.
(9) Sue: Choked on a piece of food.
(10) Kay: I mean that’s a real . . . sort of a . . . disgusting way to go.
166 The Emotional Component of Thoughts

you die

Figure 19.7 F0 and intensity in line 7

The climax of this scenario came in line 7. Figure 19.7 shows F0 at the top
and intensity at the bottom. F0 on the word die reached no higher than
260 Hz, and the emotion associated with this phrase was reflected chiefly
in the greater loudness of the word die as can be seen at the bottom of this
figure.
For an emotion that was triggered in a very different way we can move to a
segment later in this conversation when Kay described her frustration with
computers, and specifically the tendency of her word processor to delete
material against her will:
(1) . . . Or. . when the paragraph goes . . . off the page,
(2) when you’re just trying to. . type in that . . . little bit of information,
(3) and I’m screaming to Jim
(4) . . it just léft.
(5) . . . I didn’t dó it.
(6) . . . I only wanted to put an A in there,
(7) instead of an E,
(8) . . . and he keeps saying,
(9) . . you . . are the person in charge of this computer,
(10) . . and . . it only listens to whatever you punch in.
This sequence began with a restrained, matter-of-fact prosody, but then came
the self-quote in lines 4–7, with an expanded range and a raised baseline that
are conspicuous in Figure 19.8. The peaks on “left,” “do,” and “A” reached
487, 466, and 411 Hz respectively:
(4) . . it just léft.
(5) . . . I didn’t dó it.
(6) . . . I only wanted to put an A in there,
(7) instead of an E,
Jim’s quoted remarks in lines 9 and 10 provide a significant contrast with the
F0 in Figure 19.8, as shown in Figure 19.9. By imitating Jim’s speech in this
way Kay conveyed his attitude of condescending instruction such as one might
express to a child – his careful explanation, one syllable at a time, of an
 Emotional Involvement in a Conversation 167

500 I didn’t do it.

400 It just left. wanted to put
... I only
300 an A in there,instead of an E,
200
100

Figure 19.8 F0 of lines 4–7

person of this
you
charge
are the in computer and

Figure 19.9 F0 of line 9

listens to
it only in
whatever punch
you

Figure 19.10 F0 of line 10

elementary fact. The accented syllables of person, charge, and computer in line
9 were evenly spaced approximately 700 ms apart:
(9) Yóu are the pérson in chárge of this compúter,
(10) and it only listens to whatever you punch in.
In parallel fashion, the words only, whatever, and in from line 10 were evenly
spaced at approximately 350 ms, as can be seen in Figure 19.10:

Summary
This chapter described and illustrated a few of the ways prosody communicates
emotion. Among the many factors that can stimulate emotional involvement
are the uptake of an emotionally charged new topic, the anticipation of a life-
changing experience, and frustration with recalcitrant equipment. A rise–fall
contour on a single element expresses localized involvement that may be
associated with contrast or a subjective evaluation. There are numerous specific
168 The Emotional Component of Thoughts

pitch contours that convey specific attitudes. They are expressed not only with
pitch but also with variations in intensity, timing, and voice quality. Emotions
are gradient and not categorical, and they change more slowly than the idea-
tional content of language. This chapter illustrated just a few of the ways
speakers use their voices to express their feelings. Many conversations need
to be studied in a similar way before more generally applicable conclusions can
be drawn.
20 The Feeling of Nonseriousness

Instructions for making macaroni: take a long thin piece of air and wrap pasta
around it. Robert Hetzron (1991: 89)

Reading these instructions may elicit a familiar emotional experience with

unique properties that affect the way one thinks. Chafe (2007) called this
emotion the feeling of nonseriousness and discussed its effect on thought as
well as the nature of the sound that is often associated with it, the sound we call
laughter. It shares properties of other emotions, including the fact that it is
experienced to varying degrees, that it persists longer in thought than is true of
unemotional ideas, and that it is contagious, universal, and difficult to describe
with language. Because it is a pleasant emotion, people have invented ways to
experience it, ways that fall under the heading “humor.” They may be jokes, but
humorous situations often arise spontaneously in other contexts. Furthermore,
this feeling is often elicited by experiences that are not at all humorous, so we
need to understand its relation to those experiences as well.

Humor
To begin with humor, it is generally agreed to depend on a conjunction of two
features, one of which is a recognition of absurdity. The absurdity of wrapping
pasta around a thin piece of air is obvious. Absurdity alone, however, is not
enough to elicit humor, which requires that the absurdity be accompanied by
a realization that the idea actually possesses a certain plausibility. One can at
least imagine making macaroni by wrapping pasta around something that is
long and thin, making these instructions in that respect “pseudo-plausible”
despite the absurdity of wrapping pasta around air. Elie Aubouin recognized
this requirement years ago: “If we strip off from our examples that justification,
whether real or apparent, then the contradiction, incongruity, or unsuitability
becomes simply absurd, ceasing thereby to be humorous” (Aubouin 1948).
Salvatore Attardo and Victor Raskin wrote more recently “It has been
frequently noted in humor research . . . that a joke must provide a logical or
pseudological justification of the absurdity or irreality it postulates”

169
170 The Emotional Component of Thoughts

Figure 20.1 Canine theorizing

© The New Yorker Collection 2005 Charles Barsotti from cartoonbank.com.
All Rights Reserved.

(Attardo and Raskin 1991). That is true not only of jokes but of humor in
all its forms. Pseudo-plausibility may be experienced in any number of
ways, but what is important is that the absurd idea have some partial basis
in reality.
There are multiple facets to the absurdity in Figure 20.1. Among other things
dogs do not formulate theories they express in language. And the idea that
opposable thumbs might have led to what dogs see as the crazy behavior of
humans is certainly an absurd idea. But if all that absurdity is set aside,
combining human behavior with the human possession of opposable thumbs
lends a pseudo-plausibility to the theory expressed.
Because jokes are deliberately designed to elicit the feeling of nonserious-
ness, they provide useful illustrations. Jokes have a bipartite structure consist-
ing of a “buildup” followed by a “punchline.” The buildup may be any length,
but the punchline is usually short and expresses something that comes as
a surprise while it also conflicts with reality and is thus absurd. Here is an
example (Chafe 2007: 105):
A man was driving a truckload of penguins to the zoo. On the way his truck broke down
and he was forced to park beside the road. Pretty soon a farmer came along with his own
truck, which was empty, and the first man flagged it down and asked, “If I give you
a hundred dollars, can you take these penguins to the zoo?”
 The Feeling of Nonseriousness 171

“Fine,” said the farmer. So they transferred the penguins to the second truck, the
farmer received the hundred dollars, and he drove off. Some time later, however, the first
man was still trying to fix his truck when the farmer came back with the penguins still
there.
“Hey,” said the first man. “I thought I gave you a hundred dollars to take these
penguins to the zoo.”
“Oh, I did,” said the farmer. “We had a great time and there was some money left over,
so now I’m taking them to a movie.”
The idea that penguins would enjoy visiting the zoo and then going to
a movie is patently absurd. But the request to “take these penguins to the
zoo” is ambiguous between delivering them to a zoo and taking them some-
where where they would be entertained, and it is this second interpretation that
makes the story pseudo-plausible. The buildup sometimes foreshadows the
humor to come, in this case by introducing penguins, whose upright posture
and what looks like formal dress are unexpected characteristics of birds.

Laughter
When the feeling of nonseriousness is strong enough it may be manifested
audibly with laughter. This sound may be produced by a speaker or a listener or
both, and it serves not only to signal the presence of this emotion but also to
affect the behavior of those who are experiencing it. Laughter consists of
sudden, spasmodic expulsions of air from the lungs, usually with greater
force than is found during normal breathing or speaking. These exhalations
are usually followed by a single, more prolonged inhalation that replenishes the
lost air. The sound wave in Figure 20.2 shows twelve exhalation pulses
followed by an inhalation.
The laugh pulses pass from the lungs into the larynx, where in many cases
they are voiced and are thus subject to variation in pitch. The pulses in this
illustration began at a frequency of 516 Hz and descended to 157 Hz, but rising
pitches are also found and some laughs are even in the falsetto range. Laugh

exhalation pulses inhalation

Figure 20.2 Components of a laugh

172 The Emotional Component of Thoughts

pulses then pass from the larynx into the mouth, where the tongue usually
occupies a relaxed position so that most laugh pulses resemble the vowel sound
of a schwa, leading them often to be spelled ha ha ha. Laughs seldom sound
like ho ho ho, unless at the North Pole.
The number of pulses can vary from one to many, although there are seldom
more than about a dozen, as in this example. Typically their rate is slightly
below five pulses per second, but sometimes there is an acceleration or decel-
eration. Although the pulses are most often voiced, voiceless laughs are by no
means rare. Most laughs allow the expelled air to pass relatively freely through
the mouth, but sometimes the lips are closed so that air can pass only through
the nose, resulting in a laugh with an M-like sound: hm hm hm.
People often laugh at the same time they are speaking. Speaking and laugh-
ing both begin with air from the lungs that passes through the larynx and exits
through the mouth, so these two competing uses of the same vocal channel must
be reconciled in some way. Sometimes the two activities are separated, with
segments of speech alternating with bursts of laughter. Sometimes speech is
superimposed on the laugh pulses, either with one spoken syllable per pulse or
with a more complex intermingling of the two. Particularly interesting is the
introduction of tremolo: rapid oscillations imposed on a single exhalation to
produce a machine-gun-like sound. Tremolo is relatively common when laugh-
ter is superimposed on speech, but it has not been observed elsewhere except
when a child is imitating a machine gun.
Laughter forcefully removes air from the lungs but the lost air is replenished
by the subsequent inhalation. This interference with normal breathing affects
the cardiovascular system, elevating the heart rate to a degree roughly propor-
tional to the duration and intensity of the laugh. Nevertheless, although both
speaking and laughing interfere with breathing, neither produces a drop in
blood-oxygen level (Fry and Stoft 1971), a fact which suggests that the human
body has evolved to retain stability in spite of the respiratory disturbances
produced by both speaking and laughing.
It is difficult to laugh without smiling. There appears to be a continuum in
which the feeling of nonseriousness in its weakest form is expressed by nothing
more than smiling, then more strongly by chuckling, and finally by full-
throated laughter as described above.

Nonseriousness as a Property of Thought

The physical properties of laughter that were sketched above are relatively easy
to record and measure, but the question of why people behave in this special
way has been open to debate. The question is related to speculations on the
nature of humor, but the relation between laughter and humor cannot provide
a complete answer because laughter often occurs without humor.
 The Feeling of Nonseriousness 173

Bachorowski, Smoski, and Owren (2001) suggested that a laugh aims at

eliciting some sort of emotional response in those who hear it, and that laughers
use the variable acoustic features of their laughs to shape the particular emo-
tional responses they want their listeners to experience.
Provine (2004) moved out of the laboratory to record more than a thousand
instances of spontaneous laughter in shopping malls and a university student
union. He found that speakers laughed more often than listeners, showing that
laughter is not just a reaction to things said by others but is in fact more often an
accompaniment to one’s own speech. He also found women laughing more
often than men, but men doing more to provoke laughter. He found, too, that
laughter occurred more often in social situations than in solitary environments.
And he found that things said just prior to a laugh were often not humorous. He
concluded that laughter functions to solidify friendship and bring people
together. He went further to speculate on the place of laughter in human
evolution, noting that the evolution of bipedality gave humans a freedom of
breath control that set the stage for both speech and laughter.
The view proposed in Chafe (2007) is that the feeling of nonseriousness,
whether or not it is expressed overtly with laughter, functions as a kind of safety
valve that keeps us from taking seriously experiences where seriousness would
be counterproductive. We are hindered from responding physically by the
spasmodic expulsions of air from our lungs that interfere with breathing and
make it difficult to perform physical tasks. (Try doing pushups while laughing.)
At the same time we are distracted from serious thought by the euphoria that
accompanies this emotion. We are thus for a time incapable of either perform-
ing serious acts or thinking serious thoughts.
The feeling of nonseriousness can also mitigate unpleasantness, as can be
observed in the instances where laughter occurs without humor. Chafe (2007:
73–87) describes a variety of nonhumorous experiences that elicit laughter,
including the use of profanity, talking of something disgusting or depressing,
uncertainty about a choice of words, interrupting another person’s talk, self-
deprecation, regret, bereavement, embarrassment, criticism, and talking of
subjects that are abnormal, anomalous, surprising, or awkward. As an example
of an attempt to mitigate depression, the following exchange took place during
a discussion of a book in which the author described how the world of nature
was being replaced by humans (Chafe 2007: 81):
roy: And then he goes on, for the rest of the book . . .
marilyn: Then it gets really depressing.
pete: [sarcastically] Oh good. [laughs]

While much depends on the people involved and the topic under discussion,
laughter may occur frequently in ordinary conversations as an overt manifesta-
tion of the feeling of nonseriousness, whether it is triggered by humor or by
174 The Emotional Component of Thoughts

experiences whose negative properties it serves to mitigate. The ability to

manipulate that feeling in conversational interaction is a frequently observable
human trait.
The question of whether laughter is restricted to humans or whether other
animals also laugh has been controversial, although it is difficult to see how
other animals could experience the clearly human cognitive basis of laughter.
It is true that chimpanzees and other primates emit a panting noise that sounds
a bit like laughing. However, each noisy exhalation is immediately followed by
an equally noisy inhalation, a pattern very different from the sequence of
exhalations followed by a single inhalation that is characteristic of laughter.
Jaak Panksepp and Jeff Burgdorf (2003) discovered that stroking rats in
sensitive areas caused the rats to emit a chirping sound above the range of
human hearing. The rats evidently enjoyed the experience and their chirping
communicated their pleasure to other rats. The chirping, however, is physiolo-
gically unlike laughter, and it is unlikely that rats emit those sounds in the same
way or for the same reasons as the laughter of humans. Despite these superficial
resemblances, it seems probable that laughter is an exclusively human trait.

Summary
The feeling of nonseriousness may for many people be as frequently if not more
frequently experienced than generally recognized emotions like anger, sadness,
or joy. It prevents people from interpreting seriously experiences with respect
to which it would be counterproductive to think or act in a serious way.
At a mild level it may be experienced with a silent smile, but it is often
expressed overtly with laughter, which not only signals the presence of this
emotion but also disrupts serious action by interfering with breathing while at
the same time it hinders serious thought with a distracting euphoria.
The pleasurable nature of this feeling has led people to invent ways of enjoying
it, ways that fall under the heading of humor, which depends on combining
absurdity with pseudo-plausibility.
21 How Language Can Be Beautiful

Poetics deals primarily with the question, “What makes a verbal message
a work of art?” . . . Poetics deals with problems of verbal structure, just as the
analysis of painting is concerned with pictorial structure. Since linguistics is
the global science of verbal structure, poetics may be regarded as an integral
part of linguistics. Roman Jakobson (1990: 70)

Jakobson’s use of the word “poetics” need not be understood as restricted to

poetry in any narrow sense but rather as applying to any use of language with an
aesthetic purpose or effect. Given the importance and variety of aesthetic
experiences in people’s lives and their frequent manifestations in language,
one might indeed ask why linguistics, the systematic study of language, has
given this area so little attention. For one thing, of course, investigating the
relation between language and beauty is hardly something linguists are trained
to do; they have been educated to channel their efforts in quite different
directions. Perhaps, too, it is not the sort of thing in which most linguists feel
they should take a professional interest, seeing it as lying outside their profes-
sional domain. One should not conclude that linguists are philistines, and there
are some who have responded to Jakobson’s challenge in ways that cross
disciplinary boundaries, profiting from the vigor and coping with the confusion
that interdisciplinary studies entail. There exists in fact a relatively new and
promising discipline known as “cognitive poetics” whose approach is, how-
ever, different from the approach followed in this chapter (e.g. Tsur 2008;
Burke and Troscianko 2017; Kukkonen 2017).
This chapter attempts to deal, in what can only be speculative ways, with
three fundamental, intriguing, and challenging questions:
(1) What is beauty?
(2) How is beauty created?
(3) In particular, how is beauty created by language?
These are obviously large questions with no simple answers, but that does not
mean we should avoid exploring areas where answers might be sought. It is the
third question that brings these questions within the scope of this book, where it
will be approached by looking at a few examples of language with a clear

175
176 The Emotional Component of Thoughts

aesthetic value, trying to identify some of the properties that contribute to that
value and in that way trying to provide at least some preliminary answers to the
last question.
But we need to begin with the first question. I suggest that beauty can be
understood as a property of anything in our experience that elicits a certain
identifiable emotional response, a response we recognize when we feel it and
one that is pleasurable and often sought after. As with other emotions,
experiencing beauty can vary in intensity, and English sometimes recog-
nizes degrees of it with sets of words like pretty, beautiful, and gorgeous.
To some extent beauty is in the eye of the beholder. What you find beautiful
may not be what I find beautiful. And although many experiences may be
considered beautiful everywhere, there may also be cultural differences.
Individuals, furthermore, may differ in their sensitivity to beauty, some
being overwhelmed by experiences to which others may react only mildly
if at all. Any conclusions we may reach need to allow for variations like
these.
If anyone would question whether we are dealing here with an emotion,
it may be instructive to begin with an experience entirely outside of
language. Ever since I first visited it many years ago I have found
Yosemite Valley in California to be extraordinarily beautiful. Its beauty
was immediately apparent to one of the first white men to look upon it.
A soldier named Lafayette Bunnell belonged to the Mariposa Battalion
whose assignment was to rid this valley of the Indians who had long
inhabited it, but who were proving to be a nuisance to white people in the
surrounding area. Bunnell wrote, “None but those who have visited this
most wonderful valley, can even imagine the feelings with which I looked
upon the view that was there presented. . .. As I looked, a peculiar exalted
sensation seemed to fill my whole being, and I found my eyes in tears
with emotion” (Bunnell 1892: 59).
Reports like this suggest that beauty has the power to elicit “a peculiar
exalted sensation” which, when it is strong enough, may be felt throughout
one’s body and even elicit tears. Can we identify properties of the scene
confronting Bunnell that produced this feeling? These properties share
something that might be called “transcendence,” by which I mean departure
from the ordinary, mundane experiences of daily life. There are many ways
in which transcendence can be manifested, and this chapter tries to identify
some of them. Yosemite Valley contrasts with ordinariness in various ways,
perhaps the most obvious being its magnitude. Things there are just big.
Their transcendent size is accompanied by a repetition of shapes that is
apparent in an array of huge cliffs that extend into the distance. And there are
waterfalls whose sights and sounds far exceed our ordinary ways of experi-
encing water. But above all, as with other mountainous landscapes, there is
 How Language Can Be Beautiful 177

a radical departure from the usual ways we experience the earth on which we
walk. Bunnell continued: “The grandeur of the scene was but softened by the
haze that hung over the valley – light as gossamer – and by the clouds which
partially dimmed the higher cliffs and mountains. This obscurity of vision
but increased the awe with which I beheld it.”
Evidently the degree to which beauty is experienced can be augmented by
a partial obscuring of the very properties that elicit it. There may be two factors
at play here. For one thing an experience can in this way be reduced to its
essentials, freed of irrelevant and extraneous details. At the same time there is
an appeal to imagination, an invitation to augment our interpretation of a scene
in our own way.
To summarize up to this point, we have seen how transcendence from the
mundane may include such factors as unusual magnitude, repetition,
a reduction to essentials, and an appeal to imagination. We can look for these
and other factors as we shift our attention now to language. In doing so we can
recall findings from two earlier chapters. In Chapter 14 we saw how literature
can highlight the distinction between an immediate and a displaced thought,
whether that displaced thought is remembered or imagined. Immediacy is
characterized by continuity and rich detail, displacement by islands of experi-
ence and attenuated detail. In Chapter 19 we saw how emotions are expressed
with variations in prosody. In this chapter we can see how language may take
advantage of these and other factors as beauty is created.

Poetry
Because poetry is a use of language that is intentionally designed to create
beauty it is a good place to begin. Examples here are from Robert Frost and
Emily Dickinson, two very different American poets who created beauty in
ways that overlap but also differ considerably.

Robert Frost
Frost’s poem “Stopping by Woods on a Snowy Evening” is widely known and
appreciated (Frost 1979: 224). It begins:
Whose woods these are I think I know.
His house is in the village though;
He will not see me stopping here
To watch his woods fill up with snow.
As Frost himself knew, the effects of such a poem can be divided into
features on the sound side of language and features on the thought side.
Obvious features on the sound side are meter and rhyme, to which Frost paid
178 The Emotional Component of Thoughts

a great deal of attention. Metrically the entire poem is organized into lines of
iambic tetrameter:
Whose wóods | these áre | I thínk | I knów.
Rhyme is centered on the final syllables of each line. Three of the four stanzas
follow the rhyme pattern AABA, a pattern that has been called the Rubaiyat
Quatrain because it was followed in Edward FitzGerald’s 1997 [1859] transla-
tion of The Rubaiyat of Omar Khayyam. Frost may or may not have known that,
but it suggests that adherence to an established tradition may contribute to
a pattern’s effectiveness.
In this poem the B of one stanza becomes the A of the next, or in other words:
AABA
BBCB
CCDC
The final stanza, however, preserves the same rhyme throughout: DDDD.
The woods are lovely, dark and deep.
But I have promises to keep,
And miles to go before I sleep,
And miles to go before I sleep.
Frost evidently believed that a different third line (DDED) would have pre-
dicted that another stanza would follow. “What it [the repeat of the last line]
does is save me from a third line promising another stanza . . . I considered for
a moment winding up with a three line stanza. The repetend was the only
logical way to end such a poem” (Gillespie 1994).
How then is beauty created by the repetition of iambic tetrameter and final-
syllable rhymes? Ordinary uses of language exhibit a diversity of prosodic
patterns that reflect the diversity of ordinary experience. Replacing that diver-
sity with regular patterns of meter and rhyme is one way of reducing ordinary
experience to essentials, a process already included on our list of transcendent
devices.
What can be said about the thought side of this poem? Its most obvious
property is its invitation to the reader or listener to inhabit the poet’s
consciousness and share with him his thoughts as he lingers by dark
woods on his way home. His (and our) thoughts focus first on stopping
to enjoy the beauty of the woods through which he is passing, while at the
same time thinking about the owner of those woods who won’t know he is
stopping there. Then he thinks how his horse won’t understand why he is
stopping and wants to keep moving. Finally his thoughts return to the
beauty of the woods, which are “lovely, dark and deep,” but then he
remembers that he is expected elsewhere and still has “miles to go.”
 How Language Can Be Beautiful 179

Snow, it is worth noting, has an inherent beauty that may be attributed to its
obliteration of extraneous detail, serving once more to reduce an experience
to its essentials. We will meet it again at the end of this chapter.
Those who analyze poetry often feel obliged to discover in it some deeper
meaning; for example:
The theme of “Stopping by Woods” – despite Frost’s disclaimer – is the temptation of death,
even suicide, symbolized by the woods that are filling up with snow on the darkest evening
of the year. The speaker is powerfully drawn to these woods and – like Hans Castorp in the
“Snow” Chapter of Mann’s Magic Mountain – wants to lie down and let the snow cover and
bury him. The third quatrain, with its drowsy, dream-like line: “Of easy wind and downy
flake,” opposes the horse’s instinctive urge for home with the man’s subconscious desire for
death in the dark, snowy woods. The speaker says, “The woods are lovely, dark and deep,”
but he resists their morbid attraction. (Jeffrey Meyers 1996)
Frost rejected such fanciful interpretations and recommended that we take his
poem to mean what it says.
To summarize, the beauty of this poem derives in part from its imposition of
a pattern of sounds that reduces the complex prosody of ordinary speech to
a few essentials, while at the same time inviting us to share in the poet’s
consciousness. As we move to a second illustration we can see how powerful
inhabiting another consciousness can be.

Emily Dickinson
Dickinson’s poems demand more effort on the part of a reader, and this one can
be challenging (Dickinson 1924: 293-294).
I started Early – Took my Dog –
And visited the Sea –
The Mermaids in the Basement
Came out to look at me –
And Frigates – in the Upper Floor
Extended Hempen Hands –
Presuming Me to be a Mouse –
Aground – upon the Sands –
But no Man moved Me – till the Tide
Went past my simple Shoe –
And past my Apron – and my Belt
And past my Bodice – too –
And made as He would eat me up –
As wholly as a Dew
Upon a Dandelion’s Sleeve –
And then – I started – too –
180 The Emotional Component of Thoughts

And He – He followed – close behind –

I felt His Silver Heel
Upon my Ankle – Then my Shoes
Would overflow with Pearl –
Until We met the Solid Town –
No One He seemed to know
And bowing – with a Mighty look –
At me – The Sea withdrew –
With one small exception meter and rhyme are just as regular here as they
were in Frost’s poem, although a bit more complex. Each stanza is composed of
two couplets, and in each couplet the first line is iambic tetrameter and
the second iambic trimeter:
I stárt | ed éar | ly – tóok | my dóg –
and ví | sitéd | the séa –
The tetrameter of the first line creates an expectation that the next line will
also contain four feet, but that expectation is thwarted by a second line of
only three. We see here another way of creating transcendence: creating an
expectation and then flouting it. There is a deviation from Dickinson’s
pattern here in the third line, which lacks the final accented syllable that
would have made it tetrameter:
the mér | maids ín | the báse | ment . . .
This exception to what is otherwise a very regular prosody might signal
a transition from the normal content of the first couplet into the fantastic
world that follows, or perhaps Dickinson was just being inconsistent. One
wonders why she couldn’t have said “the básement flóor.”
Rhyming is restricted to the second and fourth lines of each stanza. In the
fifth and sixth stanzas the rhymes are less complete: heel and pearl, know and
withdrew. There are several instances of alliteration: hempen hands, presuming
me to be a mouse, till the tide, simple shoe, belt and bodice, dew and dandelion.
Alliteration can be added to meter and rhyme as a third kind of poetic sound
patterning.
But what can we make of the puzzling thoughts expressed in this poem? Like
the Frost poem it also puts us inside the poet’s mind, in this case as the narrator
sets out with her dog to “visit the sea,” which initially welcomes her but then
threatens to engulf her. Finally she reaches the safety of “the solid town,” where
“bowing with a mighty look” at her, the sea withdraws.
Unlike many of Dickinson’s poems, this one has a plot – a plot that follows
a sequence of events in which a normal walk by the sea turns into a disturbingly
dangerous experience that finally runs its course and subsides. Does this plot
reflect anything in Dickinson’s life? Lyndall Gordon (2010: 114–136) makes
 How Language Can Be Beautiful 181

a convincing case that Dickinson suffered from periodic and unpredictable

epileptic seizures. Her fear of their recurrence may help to explain what
appeared to others as eccentric behavior: “What seemed eccentric was simply
dread” (p. 117). Interpreted as a fantasy that traces the course of a seizure the
poem makes a great deal of sense.
Although Gordon did not mention it, it is possible that Dickinson also
experienced episodes of a scintillating scotoma, a common problem that begins
with a small blind spot that gradually expands to the edges of the visual field
and then disappears. Although epilepsy and scotoma are different problems,
Dickinson and her doctors may very well have associated the two, both of
which begin without warning, are disturbing and debilitating at their climax,
and eventually subside, a pattern that is captured with powerful imagination in
this poem.
The poem is full of surprising images that depart radically from ordinary
experience: the mermaids, the frigates with their hempen hands, the dande-
lion’s sleeve, and the sea’s mighty look. These images stand in sharp contrast to
the normalcy of Frost’s experience by the woods, but each is effective in its own
way. Dickinson shows well the aesthetic value of imagery that transcends the
ordinary.

Prose
The beauty of prose writing is centered more on the thought side of language,
although sound-based features like alliteration and rhyme may occasionally
intrude. Chapter 14 showed how leading the reader to inhabit another con-
sciousness sheds light on the nature of consciousness itself. Here, as with the
poems discussed above, we can appreciate its contribution to the aesthetic
effect of a piece of writing, transcending ordinary solipsistic experience.

James Joyce
Discussed here are certain aspects of Joyce’s short story “The Dead,” which,
like “Eveline” in Chapter 14, belongs to the collection Dubliner’s. The story is
divided into two distinct episodes, the first describing a dance party at the home
of Aunt Julia and Aunt Kate. The second takes place almost entirely within the
consciousness of Gabriel, one of the guests at the party, as he returns to his hotel
with his wife Gretta. This section begins:
The morning was still dark. A dull yellow light brooded over the houses and the river;
and the sky seemed to be descending. It was slushy underfoot, and only streaks and
patches of snow lay on the roofs, on the parapets of the quay and on the area railings.
The lamps were still burning redly in the murky air and, across the river, the palace of the
Four Courts stood out menacingly against the heavy sky.
182 The Emotional Component of Thoughts

The reader shares Gabriel’s anoetic experiences as he begins walking

to the hotel. Aside from the value of inhabiting Gabriel’s consciousness,
one finds beauty in the description itself. We share visual detail like
the “dull yellow light” and tactile detail like “slushy underfoot,” as
well as the emotion conveyed by the words “brooded,” “murky,” and
“menacingly.”
From this point on, the story follows Gabriel’s changing emotions before he
is finally ready to retire. His attention shifts from the surrounding scene to
Gretta as he imagines things he might do:
She was walking on before him so lightly and so erect that he longed to run after her
noiselessly, catch her by the shoulders and say something foolish and affectionate into
her ear. She seemed to him so frail that he longed to defend her against something and
then to be alone with her. Moments of their secret life together burst like stars upon his
memory.
Then come three islands of memory, each of which Gabriel experiences
briefly:
A heliotrope envelope was lying beside his breakfast-cup and he was caressing it with
his hand.
They were standing on the crowded platform and he was placing a ticket inside the
warm palm of her glove.
He was standing with her in the cold, looking in through a grated window at a man
making bottles in a roaring furnace.
Once they are in their room he notices Gretta crying and there follows
a sequence of questions and answers that lead to a narrative climax.
The beauty of this passage resides in our sharing of Gabriel’s changing emo-
tions as the narrative unfolds:
—Gretta, dear, what are you thinking about?—
. . . —O, I am thinking about that song, The Lass of Aughrim—
. . . —What about the song? Why does that make you cry?—
She raised her head from her arms and dried her eyes with the back of her hand
like a child. A kinder note than he had intended went into his voice.
—Why, Gretta? he asked—
—I am thinking about a person long ago who used to sing that song—
—And who was the person long ago? asked Gabriel, smiling—
—It was a person I used to know in Galway when I was living with my grandmother,
she said—
The smile passed away from Gabriel’s face. A dull anger began to gather again
at the back of his mind and the dull fires of his lust began to glow angrily in his
veins.
 How Language Can Be Beautiful 183

—Someone you were in love with? he asked ironically—

—It was a young boy I used to know, she answered, named Michael Furey. He used to
sing that song, The Lass of Aughrim. He was very delicate—

Gabriel was silent. He did not wish her to think that he was interested in this
delicate boy.
. . . A thought flew across Gabriel’s mind.
—Perhaps that was why you wanted to go to Galway with that Ivors girl? he said
coldly—
She looked at him and asked in surprise:
—What for?—
Her eyes made Gabriel feel awkward. He shrugged his shoulders and said:
—How do I know? To see him perhaps—
She looked away from him along the shaft of light towards the window in
silence.
—He is dead, she said at length. He died when he was only seventeen. Isn’t it a terrible
thing to die so young as that?—
. . . —And what did he die of so young, Gretta? Consumption, was it?—
—I think he died for me, she answered—
Gabriel, leaning on his elbow, looked for a few moments unresentfully on her tangled
hair and half-open mouth, listening to her deep-drawn breath. So she had had that
romance in her life: a man had died for her sake.
Later, in a different location, Gabriel imagines Aunt Julia’s death:
Soon perhaps he would be sitting in that same drawing-room, dressed in black,
his silk hat on his knees. The blinds would be drawn down and Aunt Kate would
be sitting beside him, crying and blowing her nose and telling him how Julia had
died. He would cast about in his mind for some words that might console her,
and would find only lame and useless one. Yes, yes, that would happen very
soon.
The last sentence exemplifies what has been called “free indirect style” or
“erlebte Rede” (McHale 1978) as it applies not only to speech but also to
thought – to what I have called “verbatim indirect thought” (Chafe 1994:
247). The repeated “yes” is verbatim language, while the words “that would
happen very soon” are indirect thought that follows the pattern of indirect
speech.
The story approaches its end with a peaceful conclusion that rounds off the
emotional turmoil of what preceded as it exploits the beauty of snow and zooms
out to take in a much larger picture:
Yes, the newspapers were right: snow was general all over Ireland. It was falling on
every part of the dark central plain, on the treeless hills, falling softly upon the Bog of
Allen and, farther westward, softly falling into the dark mutinous Shannon waves. It was
184 The Emotional Component of Thoughts

falling, too, upon every part of the lonely churchyard on the hill where Michael Furey
lay buried. It lay thickly drifted on the crooked crosses and headstones, on the spears of
the little gate, on the barren thorns.

At the very end there is a return to Gabriel’s consciousness, with alliteration

in “his soul swooned slowly” and “falling faintly . . . faintly falling,” and even
the orthographic overlap of “swooned” with “snow”:
His soul swooned slowly as he heard the snow falling faintly through the universe and
faintly falling, like the descent of their last end, upon all the living and the dead.

Summary
This chapter explored certain features of writing that stimulate in a reader the
feeling of beauty. It was suggested that beauty arises from experiences that
transcend the ordinary while appealing to the imagination and transporting the
reader into another consciousness. In poetry an important role may be played
on the sound side of language by meter, rhyme, and alliteration. On the thought
side both poetry and prose may create beauty by inviting a reader to share with
another consciousness its perceptions, actions, and emotions. To follow
Jakobson’s recommendation by interpreting these and other considerations as
concerns of linguistics can present linguistics with new and exciting
challenges.
 Epilogue

This work hoped to uncover a few of the mysteries of the mind as viewed by
someone whose career has focused to a large extent on the complexities of
certain Native American languages, but who has at the same time puzzled over
what language can tell us about “how the mind works.” These concerns have
played themselves out against shifting trends in linguistics with none of which
the author has felt fully comfortable: first a behavioristically oriented “struc-
turalism” that was followed by the still more limited “generative” tradition.
The “cognitive” approach that followed still seemed lacking in deeper, wider-
ranging insights. In this work the focus has been on the importance of recog-
nizing thought as a fundamental part of language design.
There is an obvious need to explore further all the topics that were introduced
here, but it may be useful in conclusion to mention a few that were only touched
on and that call loudly for further investigation. There is, for example, the
question of how and to what extent personality type influences one’s choice of
research, the manner in which that choice is pursued, and what it manages to
accomplish. The difference between language-primary and language-
secondary researchers could turn out to have relevant and interesting conse-
quences, but might also turn out to be too simple a distinction.
There is the unresolved question of imageless thought that is independent not
only of language but also of imagery in any form. Is it frequently experienced,
and can it explain what William James’s intelligent friend knew of his breakfast
table? The existence and widespread occurrence of anoetic thought seems
beyond question, and it deserves an important place in our understanding.
The hierarchical structuring of smaller and larger elements that is shared by
both speech and music should be apparent from careful listening to both.
The role of prosodic contours in expressing a variety of emotions in speech is
easy to hear but less easy to characterize, and it calls for further study whose
results should be transferrable to our understanding of music as well as
language.
Behind all this lies the human brain, whose mysteries are exciting to con-
template. Its neglect in this work can be attributed in part to a lack of expertise
and in part to a belief that it is too soon to integrate it into a coherent

185
186 Epilogue

understanding of issues that were raised here. It would be valuable, for exam-
ple, to know more about the fragility of the thought–sound relation as evi-
denced by the tip-of-the-tongue experience. How the brain distinguishes tacit
from verbalized knowledge is another question worth pursuing.
While it may be too much to hope for, much could be gained by productively
combining the efforts of linguists, psychologists, literary scholars, and others.
This work tilted toward linguistics, but Chapter 15 showed how linguistics and
psychology can each shed its own light on a shared topic. In that case the two
fields were unaware of each other, but in the future a familiarity with and
sympathy for varying traditions could be fostered to their mutual benefit.
Chapter 1 began by noting that language and the mind offer an endless array
of exciting topics for investigation. I hope that this work has brought a few of
them into the foreground.
 References

Abbott, Clifford (1984). Two Feminine Genders in Oneida. Anthropological Linguistics

26, 125–137.
Adamson, Tim (2004). Is Cognitive Linguistics Our Best Phenomenology of Language?
A Philosophical Challenge. In Michel Achard and Suzanne Kemmer (eds.),
Language, Culture and Mind, 93–108. Palo Alto, CA: CSLI.
Aikhenvald, Alexandra Y. (2004). Evidentiality. Oxford: Oxford University Press.
Andresen, Julie Tetel (1990). Linguistics in America 1769–1924: A Critical History.
London: Routledge.
Attardo, Salvatore, and Victor Raskin (1991). Script Theory Revis(it)ed: Joke Similarity
and Joke Representation Model. Humor 4, 293–347.
Aubouin, Elie (1948). Technique et psychologie du comique. Marseille: OFEP.
Auden, Wystan Hugh (1965). About the House. New York, NY: Random House.
Bachorowski, Jo-Anne, Moria J. Smoski, and Michael J. Owren (2001). The Acoustic
Features of Human Laughter. Journal of the Acoustical Society of America 110,
158–197.
Bartlett, Frederic C. (1932). Remembering: A Study in Experimental and Social
Psychology. Cambridge: Cambridge University Press.
Bezooyen, Renee van (1984). Characteristics and Recognizability of Vocal Expressions
of Emotion. Dordrecht: Foris Publications.
Biggar, H. P. (ed.) (1929). The Works of Samuel de Champlain. Toronto: Publications of
the Champlain Society.
Binnick, Robert I. (ed.) (2012). The Oxford Handbook of Tense and Aspect. Oxford:
Oxford University Press.
Bischoff, Shannon, and Carmen Jany (2013). Functional Approaches to Language.
Berlin: Mouton de Gruyter.
Bloomfield, Leonard (1936). Language or Ideas? Language 12, 89–95.
Boas, Franz (1901). Kathalamet Texts (Bureau of American Ethnology Bulletin 26).
Washington, DC: Government Printing Office.
Boas, Franz (1947). Kwakiutl Grammar with a Glossary of the Suffixes. Transactions of
the American Philosophical Society 37(3), 201–377.
Boas, Franz (1963 [1911]). Handbook of American Indian Languages (Bureau of
American Ethnology Bulletin 10, Part 1). Washington, DC: Georgetown University
Press.
Bolinger, Dwight (1975). Aspects of Language, second edn. New York, NY: Harcourt
Brace Jovanovich.
Bolinger, Dwight (1983). Intonation and Gesture. American Speech 58, 156–174.

187
188 References

Botne, Robert (2012). Remoteness Distinctions. In Robert I. Binnick (ed.), The Oxford
Handbook of Tense and Aspect, 536–562. Oxford: Oxford University Press.
Brown, Gillian, and George Yule (1983). Discourse Analysis. Cambridge: Cambridge
University Press.
Brown, Roger (1958). Words and Things: An Introduction to Language. Glencoe, IL:
The Free Press.
Brown, Roger, and David McNeill (1966). The “Tip of the Tongue” Phenomenon.
Journal of Verbal Learning and Verbal Behavior 5, 325–337.
Bunnell, Lafayette (1911). Discovery of the Yosemite, and the Indian War of 1851 Which
Led to That Event, fourth edn, reprinted from third edn (1892), with new map and
illustrations. Los Angeles, CA: G.W. Gerlicher.
Burke, Michael, and Emily T. Troscianko (2017). Cognitive Literary Science:
Dialogues between Literature and Cognition. Oxford: Oxford University Press.
Buswell, Guy T. (1935). How People Look at Pictures: A Study of the Psychology of
Perception in Art. Chicago, IL: University of Chicago Press.
Carroll, John B. (ed.) (1956). Language, Thought, and Reality: Selected Writings of
Benjamin Lee Whorf. Cambridge, MA: MIT Press.
Chafe, Wallace (1961). Seneca Thanksgiving Rituals (Bureau of American Ethnology
Bulletin 183). Washington, DC: Government Printing Office.
Chafe, Wallace (1962). Phonetics, Semantics, and Language. Language 38, 335–344.
Chafe, Wallace (1964). Review of Sol Saporta (ed.), Psycholinguistics: A Book of
Readings. Romance Philology 17, 668–671.
Chafe, Wallace (1967). Language as Symbolization. Language 43, 57–91.
Chafe, Wallace (1970a). Meaning and the Structure of Language. Chicago, IL:
University of Chicago Press.
Chafe, Wallace (1970b). A Semantically Based Sketch of Onondaga. International
Journal of American Linguistics 43(2) (Memoir 25, Supplement).
Chafe, Wallace (1973). Language and Memory. Language 49, 261–281.
Chafe, Wallace (1977). Caddo Texts. In Douglas R. Parks (ed.), Caddoan Texts
(International Journal of American Linguistics Native American Text Series 2,
No. 1), 27–43. Chicago, IL: University of Chicago Press.
Chafe, Wallace (ed.) (1980). The Pear Stories: Cognitive, Cultural, and Linguistic
Aspects of Narrative Production. Norwood, NJ: Ablex.
Chafe, Wallace (1986). Beyond Bartlett: Narratives and Remembering. Narrative
Analysis: An Interdisciplinary Dialogue (Special issue of Poetics) 15, 139–151.
Chafe, Wallace (1987). Cognitive Constraints on Information Flow.
In Russell Tomlin (ed.), Coherence and Grounding in Discourse, 21–51.
Amsterdam: John Benjamins.
Chafe, Wallace (1991). Sources of Difficulty in the Processing of Written Language.
In Alan C. Purves (ed.), The Idea of Difficulty in Literature, 7–22. Albany, NY: State
University of New York Press.
Chafe, Wallace (1993). Prosodic and Functional Units of Language. In Jane A. Edwards
and Martin D. Lampert (eds.), Talking Data: Transcription and Coding in Discourse
Research, 33–43. Hillsdale, NJ: Lawrence Erlbaum.
Chafe, Wallace (1994). Discourse, Consciousness, and Time: The Flow and
Displacement of Conscious Experience in Speaking and Writing. Chicago, IL:
University of Chicago Press.
 References 189

Chafe, Wallace (1995). The Realis-Irrealis Distinction in Caddo, the Northern Iroquoian
Languages, and English. In Joan Bybee and Suzanne Fleischman (eds.), Modality in
Grammar and Discourse, 349–365. Amsterdam and Philadelphia, PA: John
Benjamins.
Chafe, Wallace (1998). Things We Can Learn from Repeated Tellings of the Same
Experience. Narrative Inquiry 8, 269–285.
Chafe, Wallace (2000a). The Interplay of Prosodic and Segmental Sounds in the
Expression of Thoughts. In Matthew L. Juge and Jeri L. Moxley (eds.),
Proceedings of the 23rd Annual Meeting of the Berkeley Linguistics Society, 1997,
389–401. Berkeley, CA: Berkeley Linguistics Society.
Chafe, Wallace (2000b). A Linguist’s Perspective on William James and the Stream of
Thought. Consciousness and Cognition 9, 618–628.
Chafe, Wallace (2002a). Searching for Meaning in Language: A Memoir.
Historiographia Linguistica 29, 245–261.
Chafe, Wallace (2002b). Masculine and Feminine in the Northern Iroquoian
Languages. In Nicholas J. Enfield (ed.), Ethnosyntax, 99–109. Oxford: Oxford
University Press.
Chafe, Wallace (2006). Reading Aloud. In Rebecca Hughes (ed.), Spoken English,
Applied Linguistics and TESOL: Challenges for Theory and Practice, 53–71.
London: Palgrave.
Chafe, Wallace (2007). The Importance of Not Being Earnest: The Feeling Behind
Laughter and Humor. Amsterdam and Philadelphia, PA: John Benjamins.
Chafe, Wallace (2008). Syntax as a Repository of Historical Relics. In Alex Bergs and
Gabriele Diewald (eds.), Constructions and Language Change, 259–266. Berlin:
Mouton de Gruyter.
Chafe, Wallace (2010). Literature as a Window to the Mind. Travaux du cercle linguis-
tique de Copenhague 34, 51–63. Also listed as Acta Linguistica Hafniensia 42
(Supplement 1).
Chafe, Wallace (2012). The Seneca Amplification Construction. Linguistic Discovery
10(1), 27–41.
Chafe, Wallace (2015). A Grammar of the Seneca Language (University of California
Publications in Linguistics 149). Oakland, CA: University of California Press.
Chafe, Wallace, and Johanna Nichols (1986). Evidentiality: The Linguistic Coding of
Epistemology. Norwood, NJ: Ablex.
Chafe, Wallace, and Jane Danielewicz (1987). Properties of Spoken and Written
Language. In Rosalind Horowitz and S. J. Samuels (eds.), Comprehending Oral
and Written Language, 83–113. Cambridge, MA: Academic Press.
Chafe, Wallace, and Deborah Tannen (1987). The Relation between Written and Spoken
Language. Annual Review of Anthropology 16, 383–407.
Chomsky, Noam (1957). Syntactic Structures. The Hague: Mouton.
Chomsky, Noam (1965). Aspects of the Theory of Syntax. Cambridge, MA: MIT Press.
Chomsky, Noam (2000). The Architecture of Language. Oxford: Oxford University
Press.
Chung, Sandra, and Alan Timerlake (1985). Tense, Aspect, and Mood.
In Timothy Shopen (ed.), Language Typology and Syntactic Description, first edn.
Volume III, Grammatical Categories and the Lexicon, 202–258. Cambridge:
Cambridge University Press.
190 References

Clark, Herbert H., and Mija M. van der Wege (2015). Imagination in Narratives.
In Deborah Tannen, Heidi E. Hamilton, and Deborah Schiffrin (eds.),
The Handbook of Discourse Analysis, second edn., 406–421 Chichester: John
Wiley and Sons.
Coetzee, John Maxwell (2007). Diary of a Bad Year. New York, NY: Random House.
Cohn, Dorrit (1978). Transparent Minds: Narrative Modes for Presenting
Consciousness in Fiction. Princeton, NJ: Princeton University Press.
Comrie, Bernard (1976). Aspect. Cambridge: Cambridge University Press.
Comrie, Bernard (1981). On Reichenbach’s Approach to Tense. In Roberta
A. Hendricks, Carrie S. Masek, and Mary Frances Miller (eds.), Papers from the
Seventeenth Regional Meeting, Chicago Linguistic Society, 24–30. Chicago, IL:
Chicago Linguistic Society.
Comrie, Bernard (1985). Tense. Cambridge: Cambridge University Press.
Corballis, Michael C. (2015). The Wandering Mind: What the Brain Does When You’re
Not Looking. Chicago, IL: University of Chicago Press.
Corbett, Greville G. (1991). Gender. Cambridge: Cambridge University Press.
Corbett, Greville G. (2000). Number. Cambridge: Cambridge University Press.
Croft, William (2001). Radical Construction Grammar: Syntactic Theory in
Typological Perspective. Oxford: Oxford University Press.
Cuoq, Jean-André (1866). Etudes philologique sur quelques langues sauvage de
l’Amerique. Montreal: Dawson Brothers. Reprinted 1966 by Johnson Reprint
Corporation.
Dahl, Östen (1984). Temporal Distance: Remoteness Distinctions in Tense-Aspect
Systems. In Brian Butterworth, Bernard Comrie, and Östen Dahl (eds.),
Explanations for Language Universals, 105–122. Berlin: Mouton.
Dahl, Östen (1985). Tense and Aspect Systems. Oxford: Blackwell.
Damasio, Antonio R. (1994). Descartes’ Error: Emotion, Reason, and the Human
Brain. New York, NY: Putnam.
Day, Ruth S. (1977). Systematic Individual Differences in Information Processing.
In P. G. Zimbardo and F. L. Ruch, Psychology and Life, 5A–5D. Glenview, IL:
Scott, Foresman.
Day, Ruth S. (1979). Verbal Fluency and the Language-Bound Effect. In Charles
J. Fillmore, D. Kempler, and William S-Y. Wang (eds.), Individual Differences in
Language Ability and Language Behavior, 57–84. New York, NY: Academic Press.
Dervis, Peter A. (2000). Bell Bottom Blues. Made to Measure Magazine, March 23.
Dickinson, Emily (1924). The Poems of Emily Dickinson. Reading Edition. Cambridge,
MA and London: The Belknap Press of Harvard University Press.
Dorsey, George A. (1997 [1905]). Traditions of the Caddo. Introduction by Wallace
L. Chafe. Lincoln, NE: University of Nebraska Press.
Du Bois, John W., Stephan Schuetze-Coburn, Susanna Cumming, and Danae Paolino
(1993). Outline of Discourse Transcription. In Jane A. Edwards and Martin
D. Lampert (eds.), Talking Data: Transcription and Coding in Discourse Research,
45–89. Hillsdale, NJ: Lawrence Erlbaum.
Dundes, Alan (1980). The Number Three in American Culture. In Alan Dundes,
Interpreting Folklore, 134–159. Bloomington, IN: Indiana University Press.
 References 191

Duponceau, Peter S. (1838). Mémoire sur le système grammatical des langues de

quelques nations indiennes de l’Amérique du Nord. Paris: A. Pihan de la
Forest.
Eliot, George (1871). Middlemarch: A Study of Provincial Life. Edinburgh and London:
William Blackwood.
Esser, Jürgen, and Andrzej Polomski (1987). Reading Intonation. Language and
Communication 7, 59–75.
Esser, Jürgen, and Andrzej Polomski (1988). Comparing Reading and Speaking
Intonation. Amsterdam: Rodopi.
FitzGerald, Edward (1997 [1859]). Rubaiyat of Omar Khayyam. Charlottesville, VA:
University Press of Virginia.
Fortescue, Michael, Marianne Mithun, and Nicholas Evans (2017). Oxford Handbook of
Polysynthesis. Oxford: Oxford University Press.
Fox, Michael D., Abraham Z. Snyder, Justin L. Vincent, Maurizio Corbetta, David
C. Van Essen, and Marcus E. Raichle (2005). The Human Brain is Intrinsically
Organized into Dynamic, Anticorrelated Functional Networks. Proceedings of the
National Academy of Sciences 102, 9673–9678.
Freilich, Morris (1958). Cultural Persistence among the Modern Iroquois. Anthropos 53,
473–483.
Frost, Robert (1979). The Poetry of Robert Frost. Edited by Edward Connery Lathem.
New York, NY: St. Martins Press.
Fry, William F., and P. E. Stoft (1971). Mirth and Oxygen Saturation Levels of
Peripheral Blood. Psychomatics 19, 76–84.
Gibbs, Raymond W. (1994). The Poetics of Mind: Figurative Thought, Language, and
Understanding. Cambridge: Cambridge University Press.
Gillespie, Patrick (2009). Interpreting Robert Frost’s “Stopping by Woods” [blog post].
Poemshape. Available at https://poemshape.wordpress.com/2009/05/24/interpreting-
robert-frosts-stopping-by-woods
Givón, T. (2001). Syntax: An Introduction. Volume 1. Amsterdam and Philadelphia, PA:
John Benjamins.
Goldbeck, Thomas, Frank Tolkmitt, and Klaus R. Scherer (1988). Experimental Studies
on Vocal Affect Communication. In Klaus R. Scherer (ed.), Facets of Emotion:
Recent Research, 119–137. Hillsdale, NJ: Lawrence Erlbaum.
Gordon, Lyndall (2010). Lives Like Loaded Guns: Emily Dickinson and Her Family’s
Feuds. New York, NY: Viking Penguin.
Gumperz, John J., and Stephen C. Levinson (eds.) (1996). Rethinking Linguistic
Relativity. Cambridge: Cambridge University Press.
Hamilton, William (1859). Lectures on Metaphysics and Logic. Volume 1, Metaphysics.
Edited by Henry L. Mansel and John Veitch. Boston, MA: Gould and Lincoln.
Hardman, Martha (1986). Data-Source Marking in the Jaqi Languages.
In Wallace Chafe and Johanna Nichols (eds.), Evidentiality: The Linguistic Coding
of Epistemology, 113–136. Norwood, NJ: Ablex.
Harris, Zellig (1951). Methods in Structural Linguistics. Chicago, IL: University of
Chicago Press.
Hartgenbusch, Hanns Georg (1933). Untersuchungen zur Psychologie der
Wiedererzählung und des Gerüchtes. Psychologische Forschung 18, 251–285.
Hetzron, Robert (1991). On the Structure of Punchlines. Humor 4: 61–108.
192 References

Hirsch, Eric Donald, Jr. (1977). The Philosophy of Composition. Chicago, IL:
University of Chicago Press.
Hockett, Charles (ed.) (1970). A Leonard Bloomfield Anthology. Bloomington, IN:
Indiana University Press.
Holland, V. Melissa (1981). Psycholinguistic Alternatives to Readability Formulas.
Technical Report Number 12. Washington, DC: American Institutes for
Research.
Hopper, Paul J., and Elizabeth Closs Traugott (1993). Grammaticalization. Cambridge:
Cambridge University Press.
Hornstein, Norbert (1990). As Time Goes By: Tense and Universal Grammar.
Cambridge, MA: The MIT Press.
Humphrey, George (1951). Thinking. New York, NY: Wiley.
Hyman, Larry M. (1980). Relative Time Reference in the Bamileke Tense System.
Studies in African Linguistics 11, 227–237.
Jackendoff, Ray (1994). Patterns in the Mind: Language and Human Nature.
New York, NY: Basic Books.
Jackendoff, Ray (2007). Language, Consciousness, Culture: Essays on Mental
Structure. Cambridge, MA: MIT Press.
Jacobsen, William H. (1964). A Grammar of the Washo Language. PhD dissertation,
University of California, Berkeley.
Jacobsen, William H. (1979). Noun and Verb in Nootkan. In Barbara S. Efrat (ed.),
The Victoria Conference on Northwestern Languages, 83–155. British Columbia
Provincial Museum Heritage Record No. 4. Victoria: British Columbia Provincial
Museum.
Jacobsen, William H. (1986). The Heterogeneity of Evidentials in Makah.
In Wallace Chafe and Johanna Nichols (eds.), Evidentiality: The Linguistic Coding
of Epistemology, 3–28. Norwood, NJ: Ablex.
Jacobson, Steven A. (1995). A Practical Grammar of the Central Alaskan Yup’ik
Eskimo Language. Fairbanks, AL: Alaska Native Language Center.
Jakobson, Roman (1990). On Language. Edited by Linda R. Waugh and
Monique Monville-Burston. Cambridge, MA: Harvard University Press.
James, William (1890). The Principles of Psychology. Volume 1. New York, NY: Henry
Holt. Reprinted 1950 by Dover Publications.
Jespersen, Otto (1924). The Philosophy of Grammar. London: George Allen and Unwin.
Jevons, W. Stanley (1871). The Power of Numerical Discrimination. Nature 3, 281–282.
Joyce, James (1993 [1904]). Dubliners. Edited by John Wyse Jackson and
Bernard McGinley. New York, NY: St. Martin’s Press.
Kaufman, E. L., M. W. Lord, T. W. Reese, and J. Volkmann (1949). The Discrimination
of Visual Number. American Journal of Psychology 62, 498–525.
Keenan, Edward L. (1984). Semantic Correlates of the Ergative/Absolutive Distinction.
Linguistics 22, 197–223.
Kemmer, Suzanne (1993). The Middle Voice. Amsterdam and Philadelphia, PA: John
Benjamins.
Kibrik, Aleksandr E. (1979). Canonical Ergativity and Daghestan Languages.
In Frans Plank (ed.), Ergativity: Towards a Theory of Grammatical Relations,
62–77. London: Academic Press.
 References 193

Klare, George R. (1974). Assessing Readability. Reading Research Quarterly 10,

62–102.
Kukkonen, Karin (2017). A Prehistory of Cognitive Poetics: Neoclassicism and the
Novel. Oxford: Oxford University Press.
Lakoff, George (1980). Getting the Whole Picture: The Role of Mental Images in
Semantics and Pragmatics. In Bruce R. Caron, Meredith A. B. Hoffman,
Marilyn Silva, et al. (eds.), Proceedings of the Sixth Annual Meeting of the
Berkeley Linguistics Society, 191–195. Berkeley, CA: Berkeley Linguistics
Society.
Lakoff, George (1987). Women, Fire, and Dangerous Things: What Categories Reveal
about the Mind. Chicago, IL: University of Chicago Press.
Lakoff, George, and Mark Johnson (1999). Philosophy in the Flesh: The Embodied
Mind and Its Challenge to Western Thought. New York, NY: Basic Books.
Lambrecht, Knud (1994). Information Structure and Sentence Form. Topic, Focus, and
the Mental Representations of Discourse Referents. Cambridge: Cambridge
University Press.
Langacker, Ronald (1990). Cognitive Grammar: The Symbolic Alternative. Studies in
the Linguistic Sciences 20(2), 3–30.
Langacker, Ronald (2008). Cognitive Grammar: A Basic Introduction. Oxford: Oxford
University Press.
Laver, John (2009). The Phonetic Description of Voice Quality. Cambridge: Cambridge
University Press.
Lee, Dorothy Demetracopoulou (1938). Implications of an Indian Language.
Philosophy of Science 5, 89–102.
Lengyel, Thomas E. (1988). On the Structure and Function of Semantic Couplets
in Mayan Languages. Anthropological Linguistics 30, 94–127.
Lieber, Francis (1837). Remarks on Some Subjects of Comparative Philology and the
Importance of the Study of Foreign Languages. Southern Literary Messenger 3,
161–172.
Linell, Per (1982). The Written Language Bias in Linguistics. Linköping: Linköping
University.
Linton, Marigold (1982). Transformations of Memory in Everyday Life.
In Ulric Neisser (ed.), Memory Observed: Remembering in Natural Contexts,
77–91. San Francisco, CA: W. H. Freeman.
Linton, Marigold (1986). Ways of Searching and the Contents of Memory. In David
C. Rubin (ed.), Autobiographical Memory, 50–67. Cambridge: Cambridge University
Press.
Loftus, Elizabeth F. (1996). Eyewitness Testimony: With a New Preface by the Author.
Cambridge, MA: Harvard University Press.
Loftus, Elizabeth F. (2003). Our Changeable Memories: Legal and Practical
Applications. Nature Reviews: Neuroscience 4, 231–234.
Lucy, John A. (1992). Language Diversity and Thought: A Reformulation of the
Linguistic Relativity Hypothesis. Cambridge: Cambridge University Press.
Lucy, John A. (1996). Grammatical Categories and Cognition: A Case Study of the
Linguistic Relativity Hypothesis. Cambridge: Cambridge University Press.
Lüthi, Max. (1976). Once Upon a Time: On the Nature of Fairy Tales. Bloomington, IN:
Indiana University Press.
194 References

Martin, Jack B. (2010). How to Tell a Creek Story in Five Past Tenses. International
Journal of American Linguistics 76, 43–70.
McHale, Brian (1978). Free Indirect Discourse: A Survey of Recent Accounts. PTL:
A Journal for Descriptive Poetics and Theory of Literature 3, 249–287.
McNeill, David (2005). Gesture and Thought. Chicago, IL: University of Chicago
Press.
Melnar, Lynette R. (2004). Caddo Verb Morphology. Lincoln, NE and London:
University of Nebraska Press.
Meyers, Jeffrey (1996). Robert Frost: A Biography. Boston, MA: Houghton Mifflin.
Mithun, Marianne (1991). Active/Agentive Case Marking and its Motivations.
Language 67, 510–546.
Mithun, Marianne (1994). The Implications of Ergativity for a Philippine Voice System.
In Barbara Fox and Paul Hopper (eds.), Voice: Form and Function, 247–277.
Amsterdam and Philadelphia, PA: John Benjamins.
Mithun, Marianne (1999). The Languages of Native North America. Cambridge:
Cambridge University Press.
Mithun, Marianne (2016). Modality and Mood in Iroquoian. In Jan Nuyts and Johan van
der Auwera (eds.), The Oxford Handbook of Modality and Mood, 223–257. Oxford:
Oxford University Press.
Mithun, Marianne, and Wallace Chafe (1999). What are S, A, and O? Studies in
Language 23, 579–606.
Mustanoja, Tauno F. (1960). A Middle English Grammar. Part 1. Parts of Speech.
Helsinki: Société Néophilologique.
Neisser, Ulrich (ed.) (1982). Memory Observed: Remembering in Natural Contexts. San
Francisco. CA: W. H. Freeman.
Nickel, Gerhard (1966). Die Expanded Form im Altenglischen: Vorkommon, Funktion
und Herkunft der Umschreibung beon/wean + Partizip Präsens. Neumünster: Karl
Wachholtz.
Nuyts, Jan (2016). Analyses of the Modal Meanings. In Jan Nuyts and Johan van der
Auwera (eds.), The Oxford Handbook of Modality and Mood, 31–49. Oxford: Oxford
University Press.
Panksepp, Jaak, and Jeff Burgdorf (2003). “Laughing” Rats and the Evolutionary
Antecedents of Human Joy? Physiology and Behavior 79, 533–547.
Payne, Thomas, and Doris Payne (1990). A Grammarical Sketch of Yagua. In Desmond
C. Derbyshire and Geoffrey Pullum (eds.), Handbook of Amazonian Languages,
Volume 2, 249–474. The Hague: Mouton.
Pinker, Steven (1994). The Language Instinct. New York, NY: Harper.
Plank, Frans (1979). Introduction. In Frans Plank (ed.), Ergativity: Towards a Theory of
Grammatical Relations, 3–36. London: Academic Press.
Polanyi, Michael (2009). The Tacit Dimension. Chicago, IL: University of Chicago
Press.
Poltrock, Steven E., and Polly Brown (1984). Individual Differences in Visual Imagery
and Spatial Ability. Intelligence 8, 93–138.
Popper, Karl (1959). The Logic of Scientific Discovery. New York, NY: Routledge.
Provine, Robert R. (2004). Laughter: A Scientific Investigation. New York, NY: Viking.
Pullum, Geoffrey (1991). The Great Eskimo Vocabulary Hoax and Other Irreverent
Essays on the Study of Language. Chicago, IL: University of Chicago Press.
 References 195

Pylyshyn, Zenon (1989). The Role of Location Indexes in Spatial Perception: A Sketch
of the FINST Spatial-Index Model. Cognition 32, 65–97.
Reichenbach, Hans (1947). Elements of Symbolic Logic. New York, NY: Macmillan.
Richards, Cara B. (1974). Onondaga Women: Among the Liberated.
In C. J. Matthiasson (ed.), Many Sisters: Women in Cross-Cultural Perspective,
401–419. New York, NY: Free Press.
Riggs, Kevin J., Ludovic Ferrand, Denis Lancelin, Laurent Fryziel, Gérard Dumur, and
Andrew Simpson (2006). Subitizing in Tactile Perception. Psychological Science 17,
271–272.
Rosch, Eleanor (1978). Principles of Categorization. In Eleanor Rosch and Barbara
B. Lloyd (eds.), Cognition and Categorization, 27–48. Hillsdale, NJ: Erlbaum.
Rosch, Eleanor, C. B. Mervis, W. D. Gray, D. M. Johnson, and P. Boyes-Braem (1976).
Basic Objects in Natural Categories. Cognitive Psychology 8, 382–439.
Rozin, Paul, Alexander Rozin, Brian Appel, and Charles Wachtel (2006). Documenting
and Explaining the Common AAB Pattern in Music and Humor: Establishing and
Breaking Expectations. Emotion 6, 349–355.
Sachs, Jacqueline Strunk (1967). Recognition Memory for Syntactic and Semantic
Aspects of Connected Discourse. Perception and Psychophysics 2, 437–442.
Sapir, Edward (1921). Language: An Introduction to the Study of Speech. New York,
NY: Harcourt, Brace.
Saporta, Sol (1961). Psycholinguistics: A Book of Readings. New York, NY: Holt,
Rinehart and Winston.
Schiffrin, Deborah (1987). Discourse Markers. Cambridge: Cambridge University Press.
Schwartz, Bennett L. (2002). Tip-of-the-Tongue States: Phenomenology, Mechanism,
and Lexical Retrieval. Mahwah, NJ: Lawrence Erlbaum.
Seuren, Pieter (1998). Western Linguistics: An Historical Introduction. Hoboken, NJ:
Wiley-Blackwell.
Slobin, Dan I. (1996). From “Thought and Language” to “Thinking for Speaking”.
In John J. Gumperz and Stephen C. Levinson (eds.), Rethinking Linguistic Relativity,
70–96. Cambridge: Cambridge University Press.
Spreng, R. Nathan, and Cheryl L. Grady (2009). Patterns of Brain Activity Supporting
Autobiographical Memory, Prospection, and Theory of Mind, and Their Relationship
to the Default Mode Network. Journal of Cognitive Neuroscience 22, 1112–1126.
Stamenov, Maxim I. (ed.) (1997). Language Structure, Discourse and the Access to
Consciousness. Amsterdam and Philadelphia, PA: John Benjamins.
Stelma, Juurd H., and Lynne J. Cameron (2007). Intonation Units in Spoken Interaction:
Developing Transcription Skills. Text and Talk 27, 361–393.
Templin, Mildred C. (1957). Certain Language Skills in Children: Their Development
and Interrelationships. Minneapolis, MN: The University of Minnesota Press.
Thomson, William (1883). Electrical Units of Measurement: A Lecture Delivered at the
Institution of Civil Engineers, May 3, 1883. In Popular Lectures and Addresses,
Volume 1, 73–136. London: Richard Clay and Sons, Limited.
Thwaites, Reuben Gold (1896–1901). The Jesuit Relations and Allied Documents:
Travels and Explorations of the Jesuit Missionaries in New France, 1610–1791.
Cleveland, OH: Burrows Brothers.
Tian, Xing, and David Poeppel (2012). Mental Imagery of Speech: Linking Motor and
Perceptual Systems through Internal Stimulation and Estimation. Frontiers in Human
Neuroscience 6(314), 1–11.
196 References

Timberlake, Alan (2007). Aspect, Tense, Mood. In Timothy Shopen (ed.), Language
Typology and Syntactic Description, second edn. Volume III: Grammatical
Categories and the Lexicon, 280–333. Cambridge: Cambridge University Press.
Tooker, Elisabeth (1964). An Ethnography of the Huron Indians, 1615–1649 (Bureau of
American Ethnology Bulletin 190). Washington, DC: Government Printing Office.
Traugott, Elizabeth Closs (1978). On the Expression of Spatio-Temporal Relations in
Language. In Joseph H. Greenberg, Charles A. Ferguson, and Edith A. Moravcsik
(eds.), Universals of Human Language. Volume 3, Word Structure, 369–400. Stanford,
CA: Stanford University Press.
Trick, Lana M. (2005). The Role of Working Memory in Spatial Enumeration: Patterns
of Selective Interference in Subitizing and Counting. Psychonomic Bulletin and
Review 12(4), 675–681.
Trick, Lana M., and Zenon W. Pylyshyn (1993). What Enumeration Studies Can Show
Us about Spatial Attention: Evidence for Limited Capacity Preattentive Processing.
Journal of Experimental Psychology: Human Perception and Performance 19,
331–351.
Trick, Lana M., and Zenon W. Pylyshyn (1994). Why Are Small and Large Numbers
Enumerated Differently? A Limited-Capacity Preattentive Stage in Vision.
Psychological Review 101, 80–102.
Trigger, Bruce (1976). The Children of Aataentsic: A History of the Huron People to
1660. Kingston and Montreal: McGill-Queen’s University Press.
Tsur, Reuven (2008). Toward a Theory of Cognitive Poetics, second expanded and
updated edn. Portland, OR: Sussex Academic Press.
Van Eijk, Jan P., and Thom Hess (1986). Noun and Verb in Salish. Lingua 69, 319–331.
Vandekerckhove, Marie, Luis Carlo Bulnes, and Jaak Panksepp (2014). The Emergence
of Primary Anoetic Consciousness in Episodic Memory. Frontiers in Behavioral
Neuroscience 7(210), 1–8.
Warren, Howard C. (1897). The Reaction Time of Counting. Psychological Review 4,
569–591.
Watson, John B. (1913). Psychology as the Behaviorist Views It. Psychological Review
20, 158–177.
Wells, Rulon S. (1947). Immediate Constituents. Language 23, 81–117.
Welty, Eudora (1983). One Writer’s Beginnings. New York, NY: Warner.
Wharton, Edith (1987 [1905]). The House of Mirth. New York, NY: Macmillan.
Wittgenstein, Ludwig (1955 [1922]). Tractatus Logico-Philosophicus. London:
Routledge and Kegan-Paul.
Woodbury, Hanni (1975). Noun Incorporation in Onondaga. PhD dissertation, Yale
University.
Wrong, G. M. (ed.) (1939). Gabriel Sagard: The Long Journey to the Country of the
Huron. Toronto: Champlain Society.
 Index

adjectives, 34 Corballis, Michael, 113

aesthetic component of language, 98 Corbett, Greville, 123
aesthetic language, 175 count nouns, 123
affect, 17, 20, 33, 47, 49 couplets, 45
agent and patient, 132 creaky voice, 57, 59
Aikhenvald, Alexandra, 151 Creek language, 140
Allegany Seneca Reservation, 94 Croft, William, 35
alliteration, 180, 184
American Psychological Association, 16 Dahl, Östen, 138
amplification constructions, 90 Day, Ruth, 42
andative, 79 Dickinson, Emily, 179
andative orientation, 71, 75, 86 diegesis, 114
anoetic consciousness, 45–47, 49, 115 discourse marker, 69
anoetic experience, 117, 119, 182 disfluencies, 44, 49
Association for Psychological dispositions, 61
Science, 16 distributives, 124
Auden, W.H., 42
Eliot, George, 118
Bachorowski, Jo-Anne, 173 empathy, 70, 75
Bamileke language, 139 entities, 33
Bartlett, Frederic, 146 epistemology, 151
Berkeley Department of Linguistics, 9 erlebte Rede, 183
Bloomfield, Leonard, 15, 192 events, 33–36, 40–41, 47, 57, 61, 149
Boas, Franz, 72, 146, 151 evidentiality, 153
Bolinger, Dwight, 51 eyewitness testimony, 148
Brown, Roger, 73
Bunnell, Lafayette, 176–177 FitzGerald, Edward, 178
Burmese language, 138 foci of consciousness, 58–59, 61–63, 65
folk morphology, 89–90
Caddo language, i, ix, 18–19, 102, 154, 156, folklore, 127
188–190, 194 folktales, 113
categories, 73 free indirect style, 183
categorization, 72, 128 Freilich, Morris, 135
center of interest, 62–63, 65 French language, 131
Chomsky, Noam, 8, 43, 189 Frost, Robert, 177
codability, 73 functionalism, 31
Coetzee, John Maxwell, 44
Cohn, Dorrit, 114, 118, 190 gender, 131
combining ideas, 75 generative linguistics, 185
Comrie, Bernard, 138 German language, 131
content, 33 gesture, 82
contrastive orientation, 75 Givón, T., 34

197
198 Index

Gordon, Lyndall, 180 mass nouns, 123

grammaticality judgments, 13 memory
grammaticalization, 76, 78 autobiographical, 143
long-term, 98
Hamilton, Sir William, 126 surface, shallow, deep, 143
Hardman, Martha, 152 memory, properties of, 115
Harris, Zellig, 8 Middle English, 79
Hartgenbusch, Hanns Georg, 47 mimesis, 114
Herder, Johann Gottfried von, 106 mind wandering, 120
historical present, 141 mirative mood, 157
Humboldt, Wilhelm von, 106 Mohawk language, 84
humor, 169 morphosyntactic structure, 85
Huron culture, 134 Mozart, 65
music, 65–66, 127, 185
ideas, 33, 61
idiomaticization, 76 Neisser, Ulrich, 150
imageless consciousness, 46 nouns, 34
imagery, 2, 7–8, 14, 29, 46–47, 49 number, grammatical, 123
immediate constituents, 75
indirect thought, 183 Old English, 79
interpreting, 94 Oneida language, 132
intonation units, 2, 51, 58–62, 66, 79, Onondaga
82 culture, 136
Iroquoian languages, 2, 18, 23, 131, Onondaga language, 23, 80
189, 194 orientation, 74
orientations, 34
Jakobson, Roman, 175 orientations of ideas, 33
James, William, 1, 15, 17, 46, 57, 64, 142,
189 Panksepp, Jaak, 174
Japanese language, 123 participant roles
Jesuit Relations, 134 agent patient, 40
jokes, 170 ergative and absolutive, 38
Joyce, James, 114, 181 subject and object, 36
paucals, 124
Kahkwa Indians, 135 Pear Film, the, 147
Kathlamet language, 146 Pear Stories, 19, 59, 188
Kelvin, Lord, 13 phonological structure, 30
kinship systems phrasal verbs, 77
Seneca, 105 poetics, 175
Kwak’wala (Kwakiutl) language, cognitive, 175
151 polysynthetic language, 97
polysynthetic languages, 18, 84–85,
Langacker, Ronald, 35 156
language change, 78 Popper, Karl, 14
language learning, 29 progressive aspect, 79
laughter, 169, 171 pronouns, 70
Lee, Dorothy Demetracopoulou, 152 prose, 181
limbic system, 58 prosody, 31, 33, 51–52, 54, 57, 161
Linell, Per, 25 morphological, i, 3, 23, 54, 83
Linton, Marigold, 145 Provine, Robert R., 173
listenability, 53 pseudo-plausibility, 170
literalization, 78 psycho-narration, 118
literature, 114
Loftus, Elizabeth, 148 readability, 53
Lounsbury, Floyd, 105 reading aloud, 52, 54, 56
 Index 199

realis and irrealis, 156 tense, metrical, 139

Richards, Cara, 136 thanking in Seneca, 109
theories
Sachs, Jacqueline, 48 folk, 12
Salamanca, New York, 94 scientific, 13
Salishan language family, 34 thoughts
Sapir, Edward, 72, 151 expanded, 62
Scherer, Klaus, 161 split, 61
segmental qualities of sound, 57 tip-of-the-tongue experience, 28–29,
selecting what to verbalize, 72 186
semantic structure, 30–31, 71–72, 74–76, 93 Tooker, Elisabeth, 134
Seneca language, i, ix, 18, 39, 65, 84, 105, 132, topicality, aboutness, and subjecthood, 36
154–155, 188–189 topics, 64–65, 70
sentences, 39, 44, 48–49, 52, 54, 57, 62, 65, 89 translation, 93
Seuren, 195 tremolo, 172
shadow meaning, 78, 95 Trigger, Bruce, 134
sign languages, 28, 50
silent language, 47 verbalizations
silent thinking, 50–51 repeated, 100
sleep, influence on memory, 146 verbatim indirect thought, 183
Slobin, Dan, 107 verbs, 34
smiling, 172 vision, theories of, 128
snow, the beauty of, 179, 183–184 visiting in Native America, 109
solipsism, 70
spectrogram, 58 Wakashan language family, 34
states, 15, 33, 36, 41, 61, 140 Watson, John, 15
Steinthal, Heymann, 106 Welty, Eudora, 51
structuralist linguistics, 185 Wharton, Edith, 117
subitizing, 127 Whorf, Benjamin Lee, 106
symbolization, 80 Wintu language, 152
syntactic structure, 8, 30, 83, 93 Wittgenstein, Ludwig, 44
syntactocentrism, 31 writing, i, 2, 7, 9, 17, 25, 43, 51–53, 56, 114
syntax, 57 writing and reading, 50
Wyandot language, 131
tacit knowledge, 87, 186
tactile perception, 128 Yagua language, 140
Tannen, Deborah, 148 Yosemite Valley, 176