Download as pdf or txt
Download as pdf or txt
You are on page 1of 227

Contributions To Phenomenology 96

Patrizia Pedrini · Julie Kirsch Editors

Third-Person
Self-Knowledge,
Self-Interpretation,
and Narrative
Contributions To Phenomenology

In Cooperation with The Center


for Advanced Research in Phenomenology

Volume 96

Series Editors
Nicolas de Warren, KU Leuven, Belgium
Ted Toadvine, Pennsylvania State University, PA, USA

Editorial Board
Lilian Alweiss, Trinity College Dublin, Ireland
Elizabeth Behnke, Ferndale, WA, USA
Rudolf Bernet, Husserl Archive, KU Leuven, Belgium
David Carr, Emory University, GA, USA
Chan-Fai Cheung, Chinese University Hong Kong, China
James Dodd, New School University, NY, USA
Lester Embree, Florida Atlantic University, FL, USA
Alfredo Ferrarin, Università di Pisa, Italy
Burt Hopkins, University of Lille, France
José Huertas-Jourda, Wilfrid Laurier University, Canada
Kwok-Ying Lau, Chinese University Hong Kong, China
Nam-In Lee, Seoul National University, Korea
Rosemary R.P. Lerner, Pontificia Universidad Católica del Perú, Peru
Dieter Lohmar, University of Cologne, Germany
William R. McKenna, Miami University, OH, USA
Algis Mickunas, Ohio University, OH, USA
J.N. Mohanty, Temple University, PA, USA
Junichi Murata, University of Tokyo, Japan
Thomas Nenon, The University of Memphis, TN, USA
Thomas M. Seebohm, Johannes Gutenberg-Universität, Germany
Gail Soffer, Rome, Italy
Anthony Steinbock, Southern Illinois University at Carbondale, IL, USA
Shigeru Taguchi, Hokkaido University, Japan
Dan Zahavi, University of Copenhagen, Denmark
Richard M. Zaner, Vanderbilt University, TN, USA
Scope
The purpose of the series is to serve as a vehicle for the pursuit of phenomenological
research across a broad spectrum, including cross-over developments with other
fields of inquiry such as the social sciences and cognitive science. Since its
establishment in 1987, Contributions to Phenomenology has published more than
80 titles on diverse themes of phenomenological philosophy. In addition to
welcoming monographs and collections of papers in established areas of scholarship,
the series encourages original work in phenomenology. The breadth and depth of
the Series reflects the rich and varied significance of phenomenological thinking for
seminal questions of human inquiry as well as the increasingly international reach
of phenomenological research.
The series is published in cooperation with The Center for Advanced Research in
Phenomenology.

More information about this series at http://www.springer.com/series/5811


Patrizia Pedrini  •  Julie Kirsch
Editors

Third-Person
Self-Knowledge,
Self-Interpretation,
and Narrative
Editors
Patrizia Pedrini Julie Kirsch
Department of Letters and Philosophy Department of Liberal Arts
University of Florence D’Youville College
Florence, Italy Buffalo, New York, USA

ISSN 0923-9545     ISSN 2215-1915 (electronic)


Contributions To Phenomenology
ISBN 978-3-319-98644-9    ISBN 978-3-319-98646-3 (eBook)
https://doi.org/10.1007/978-3-319-98646-3

Library of Congress Control Number: 2018958002

© Springer Nature Switzerland AG 2018


This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, express or implied, with respect to the material contained herein or for any errors
or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims
in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG.
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface

 elf-Knowledge: A Plea for Closer Dialogue


S
Between the Analytic Tradition and Phenomenology1

This is a book about a much thrown-around topic in analytic philosophy of mind


and epistemology: self-knowledge. In particular, it focuses on third-person self-­
knowledge, that is, the kind of self-knowledge that a subject can achieve by taking
into account the evidence offered by his or her verbal and/or non-verbal behavior,
and his or her other mental states. Such evidence can help the subject reach a knowl-
edge that he or she could not access directly, without inference from the evidence.
When a subject is able to access directly his or her own mental states, he or she is
said to have first-person self-knowledge. Our project is to investigate third-person
self-­knowledge rather than first-person self-knowledge because the literature has
devoted comparatively less attention to the former. Furthermore, third-person self-
knowledge shows interesting connections with self-interpretation and narrative that
should be explored. For when a subject has to assess the evidence offered by his or
her verbal and/or nonverbal behavior, and his or her other mental states, in order to
get to know what mental states he or she really has, a certain amount of interpreta-
tive work has to be done, and relying on narratives is part of the rationalizing opera-
tion to which evidence can be subjected. All this will be explained at length in Chap.
1, and more specifically throughout the contributions assembled in the book.
However, although this is a book primarily inspired by the methodology with
which analytic philosophy of mind and epistemology address the topic, our explora-
tion of the connection between third-person self-knowledge, self-interpretation, and
narrative brings to light important overlaps with analogous topics discussed in other
philosophical traditions — phenomenology, in particular. None of the chapters
engage directly with it, so a systematic comparison of that overlap is not part of this
project, even if some chapters, as the reader will see, appeal to traditions that are not

1
 The preface’s author is Patrizia Pedrini. Julie Kirsch offered useful comments and approved it
as coeditor of the book.

v
vi Preface

simply contemporary analytic philosophy of mind and epistemology, but rather


hermeneutics, cognitive science, and psychoanalysis. Virtually all contributions,
however, more or less tacitly assume the analytic methodology. Yet all readers who
are familiar with the phenomenological tradition will find that the topics addressed
in the book can be of the utmost interest for the scholarship of phenomenology. A
few words on how related studies in phenomenology address some of the general
issues that this book tackles are thus in order. While doing this, we will try to offer
some thematic and terminological clarifications, which may be useful for seeing
exactly what it is that phenomenology tends to study, what the analytic tradition
may be taken to add or, more specifically, focus on, and how and why both disci-
plines might benefit from more direct cooperation with a view to reaching a better
understanding of the phenomenon of self-knowledge as a whole.
Prominent in the phenomenological tradition is the discussion of phenomena
referred to, respectively, as consciousness and self-consciousness. Although it may
be immediately apparent that consciousness and self-consciousness importantly
bear on self-knowledge, we should not assume without argument that the focus is
exactly the same, even if most theories regarding consciousness and self-­
consciousness that are offered in the phenomenological tradition are highly instruc-
tive and can easily become very fruitful in terms of advancing analytic philosophy
of mind.2 First, what in phenomenology goes under the name of consciousness is
often the so-called pre-reflective consciousness, while the so-called reflective self-­
consciousness is a more advanced state of mind, involving more complex psycho-
logical capacities.3 Pre-reflective consciousness is analyzed as an intrinsic feature of
any felt experience a subject may go through. Among them there may not be states
to which analytic philosophy of mind tends to attach phenomenal qualities, the
“what it is like” to have them, such as sensations. In phenomenology, the quality of
being “experienced” seems to be attached to much more than sensations. Be that as
it may, the felt experience is given to the subject in an already structured form, in
ways to be qualified, including intentionality,4 or directedness at an object. The pre-­
reflective experience is not, however, typically reflected upon — that is, it is not
typically thematized by the subject having it, it is not “objectual” to itself. It is a
very primary form of consciousness, and there need not be any higher-order mental
state (be it perceptual or thought) making it conscious as such.5

2
 A valuable attempt to show the fruitfulness of dialogue between philosophy of mind and phenom-
enology is offered by Gallagher & Zahavi, 2007. In Chap. 3 they develop a detailed overview of
the topic of consciousness and self-consciousness in phenomenology.
3
 Recently, Moran (2017) has discussed Husserl’s project of transcendental self-knowledge, with
clarifications about the bearing of such discussion on the phenomenon of pre-reflective and reflec-
tive awareness.
4
 It should be noted that when the term “intentionality” is used in philosophy of mind, it can have
two meanings depending on the context: either it refers to the capacity of a subject to conceive
voluntary actions or it can refer to the directedness at content that a mental state can have. The
latter meaning is close to what phenomenology calls “intentionality.”
5
 Cf. Gallagher & Zahavi, 2007, Chap. 3.
Preface vii

Yet consciousness can be said to qualify as self-conscious in a weak sense in so


far as the subject’s experience occurs within the first-person perspective. The first-­
person perspective that is relevant here is itself a weak first-person perspective, as it
does not involve the capacity of the subject to attribute said experience to him- or
herself via a conceptual, linguistic, and/or reflective act. But it counts as a first-­
person form of consciousness because the subject experiences it as his or her own
(see Gallagher & Zahavi 2007; see also Pedrini 2015a, 2015b for failures of the
sense of “mine-ness” in abnormal conditions of otherwise conscious thoughts).
In order for a subject to be able to attribute the experience to him- or herself in
conceptual form, reflection must come in: whenever a subject takes his or her felt
experience as the object of his or her reflection, and tries to interpret or ponder it,
the subject is engaging in reflective self-consciousness, this time within a strong
first-person perspective (see Baker 2000, 2007, 2013).
As we have said, according to phenomenology, pre-reflective consciousness is
not unstructured. Rather, it proves to be already significant, that is, it proves to have
an internal structure that is conferred upon it by the very subject’s interaction with
the physical and/or social world, as well as by his or her being an incarnate subject,
a subject with a highly specific psychological constitution and a narrative identity.
This directly stems from a conviction widely voiced by phenomenology, according
to which experience displays essential characteristics that can be investigated by a
very precise philosophical methodology.6
Of course, the subject may not be thematically aware of such structure at the time
of its occurrence. And it is not clear that, while making it the object of the subject’s
reflection, the structure can be reflected upon without distortion or loss. Husserl
(1966a, 1966b, 1973, 1984) and other leading phenomenologists (Sartre 1956,
1957; Merleau-Ponty 1962) discuss this at length, showing an awareness of topics
that analytic philosophy of mind has come to discuss much later. In particular, it will
strike the reader that some of the frontier topics discussed in this book have their
precedents in the phenomenological tradition, such as the analysis of the knowledge
we can obtain of our own “internal evidence,” which amounts to conscious states
that become the object of our own self-interpretative endeavors (they are discussed
by Coliva, Chap. 2, and Pedrini Chap. 4), or the knowledge we can obtain of other
conscious mental states we may have, such as intuition (see McGahhey and
Leeuwen, Chap. 5), memories (see Kirsch, Chap. 6 ), and so on.
Husserl and others were also interested in understanding how self-consciousness
stems from pre-reflective consciousness, but overall it might not be unfair to say that
analytic philosophy of mind and epistemology have raised very troublesome episte-
mological questions connected with how we get to know what we feel, think, desire,
intend, and so on. This focus on epistemology may not be immediately apparent
when one pays heed to the phenomenological literature on self-consciousness.

6
 It would be interesting to explain further the details of phenomenological analysis, including the
methodology of the “epochè,” etc. Addressing this issue, however, would lead us too far from our
present purposes.
viii Preface

Admittedly, most phenomenologists think that pre-reflective consciousness is a


necessary, though not sufficient condition of self-consciousness, and that it can
hardly amount to a form of knowledge. Some have clearly maintained that it is not
transparent (see Ricour 1966; Sartre 1956). What seems to be agreed upon, how-
ever, is that the endeavor of reflectively articulating our pre-reflective conscious
experience requires interpretation. This is one of the most interesting overlaps with
the focus of this book. When philosophy of mind and epistemology discuss third-­
person self-knowledge, as the reader will see, they try precisely to explain how we
can get to know some of our mental states that are not transparent to us, or not
immediately identified and conceptualized by the subject having them, so that self-­
interpretation comes in.
Third-person self-knowledge of already conscious mental states, which requires
interpretation, is a somehow quite new issue in philosophy in mind. For several
decades, third-person self-knowledge was mainly considered to have its main focus
on verbal and non-verbal behavior, or on unconscious mental states. On this specific
aspect we think that closer dialogue between the two traditions should be initiated,
as it could be most fruitful to both. Although this book does not develop this dia-
logue as such, we strongly encourage its opening.
An invitation to cooperate also comes from other projects and studies. Some are
derived from historically minded researches that have shown that phenomenology
and analytic philosophy have a common root. Famously, Michael Dummett (1993)
suggested that it was time for scholars to overcome any ideological barrier viewing
analytic philosophy and phenomenology as strongly opposed and conceived as sep-
arate traditions. The idea that philosophy of mind, in particular, has little in common
with a tradition, sometimes referred to as the “continental” tradition, that started
with Franz Brentano, was then developed by Edmund Husserl, and later on followed
by Martin Heidegger, Jean-Paul Sartre, and Maurice Merleau-Ponty,7 was mainly
guided by unilateral interpretations of the history, leading projects, methodologies,
and concepts used in both traditions. However, Husserl has also inspired leading
analytic philosophers such as Rudolf Carnap, Gilbert Ryle, Wilfrid Sellars, Hilary
Putnam, John Searle, and Jerry Fodor.8 One of the reasons that may have led schol-
ars to neglect this influence is the fact that phenomenology is often — wrongly —
associated with “introspectionism,” whose rejection was the central intellectual
program of analytic philosophy at its outset. But all those acquainted with phenom-
enology are aware that phenomenology is not the heir of the introspectionist
approach. This prejudice has been an impediment to a correct reading of both tradi-
tions and an obstacle to cooperation between them as well.

7
 In the Italian edition of The Phenomenological Mind by Gallagher and Zahavi, which I translated
(2009), I wrote an “Afterword” in which I drew an outline of those historiographical studies, cf.
Pedrini 2009a.
8
 For the relation between Husserl and Carnap, see M. Friedman 1999; between Husserl and Ryle,
see Livingston 2005, Thomasson 2002, pp. 115–142; between Sellars and Husserl, see Thomasson
2005; between Husserl and Putnam, see Putnam 1981, 1983, 1987; between Husserl and Searle,
see Mulligan 2003, pp. 261–286, and Dreyfus 2000, pp. 287–302. See also Fodor 1982, pp. 277–
303 for the relation between Fodor and Husserl.
Preface ix

In recent years several scholars have attempted to remedy this by encouraging


integration and collaboration between the two traditions.9 Some of these attempts
emphasized how “anti-psychologism,” that is, the independence of logic from the
psychic sphere, was a goal common to both traditions. Both Husserl and Frege took
this idea from Hermann Lotze, under whom Frege studied in Göttingen and to
whom Husserl dedicated his Logical Investigations (2001). Lotze, inspired by Kant
and elaborating further on Plato’s theories of ideas, was engaged in a struggle
against widespread nineteenth-century naturalism. He explained how several
notions, including the notion of a priori, fully escape a purely psychological treat-
ment. While Frege elected language as the medium between mind and logical
objects, Husserl developed the idea that judgments and logical structures are already
prepared by experience itself, that is, contents of thought and their logical relations
are objectively present in experience before linguistic thought and propositional
attitudes capture them.10 Thus, since its very beginning, phenomenology attempted
to study the mind by approaching it with a non-psychologistic and non-naturalistic
method, with the same spirit with which analytic philosophy rejected introspection-
ism. Well before the Vienna Circle, logical empiricism and positivism, Franz
Brentano had tried to open up a new approach to psychology. In his Psychology
from an Empirical Standpoint (1973), Brentano expressed dissatisfaction with the
idea that the mind was to be studied through a sort of “inner observation.”11 For this
reason, he drew an important distinction between what he dubbed “genetic psychol-
ogy” and “descriptive psychology”: while the former is described as the empirical
study of mental phenomena through experiments and statistical methodology, which
make it possible to elaborate laws and causal explanations, the latter is not aimed at
finding any cause-effect laws, or at any psychological episode in particular. Rather,
it tries to distinguish and classify basic kinds of mental phenomena by finding their
essential characteristics and mutual relations. Thus, a central focus of Husserl’s
Logical Investigations (2001), directly stemming from Brentano’s ideas, is the cata-
loguing of mental states and the answering of questions such as “What is a percep-
tion, a judgment, an emotion?”, “What is the relation between emotion and
judgment?”, “What is it that makes an emotion regret?”, and so on. In this way,
Brentano’s descriptive psychology precedes genetic psychology, in so far as study-
ing the causes of a perception, an emotion, a memory, etc. presupposes that we
understand what seeing, remembering, feeling, etc. are. Husserl made it clear that
this approach has nothing to do with introspection. Rather, it is the study of the
essences of mental states.

9
 See Føllesdal 1958; Monthly 1982, 1985; Willard 1984; Sommer 1985; Cobb-Stevens 1990;
Smith 1982.
10
 In the analytic tradition, there seems to be a primacy of the predicative form of thought over the
non-predicative form. Tugendhat (1976, pp. 94) claims that this is the major difference between the
analytic tradition and phenomenology.
11
 In Pedrini (2009b) I analyzed the Cartesian model of self-knowledge vis-à-vis the observational
perceptual models of it. I tried to explain to what extent the perceptual model of self-knowledge
could be derived from the Cartesian model of self-knowledge, although the latter has been gener-
ally taken to be opposed to it.
x Preface

By considering this profound goal of phenomenology, Paul Livingston (2005)


has shown how Husserl influenced Ryle and his idea of conceptual analysis. In his
The Concept of the Mind (1949), Ryle discussed why it was necessary to analyze
mental concepts and how they are used in ordinary language. He was not using
Husserl’s style, he was not talking about “intuiting” and “analyzing essences,” yet
his proposal has close affinities with Brentano’s and Husserl’s in so far as he studied
the fundamental types of mental states, their functioning, their intentionality or
directness at objects, their logical status, their structures, and their mutual relation-
ships. Ryle at some point said that his work “could be described as a sustained essay
in phenomenology, if you are at home with that label” (Ryle 1971, p. 188). From the
point of view of analytic philosophy, there should be nothing mysterious about the
Husserlian project of studying essences of mental phenomena. Those essences
would be just the “concepts” of experiences, what is necessary for an experience to
count as a perception, a judgment, a memory, and so on. The study of essences
would thus lead us back to conceptual analysis, which is characteristic to at least
one of the major trends of analytic philosophy. Of course, Ryle tried to develop an
approach to the mind where any discourse on the mental should be reformulated as
a discourse about observable behavior, which led to logical behaviorism. Yet, his
book remains a proof of the close relationship between the two traditions.
Despite this common origin, another reason why philosophy of mind and phe-
nomenology have long been separate and developed in parallel, without touching
upon each other, may certainly have been the primacy of consciousness in phenom-
enology, in contrast with the crisis of the concept in philosophy of mind during the
“cognitivist revolution,” as it is known, which has influenced the future develop-
ment of philosophy of mind for some decades. The cognitivist revolution has it that
consciousness is a minor aspect of the mental, and also highly misleading as to the
correct methodology, i.e., scientific methodology, that philosophy of mind should
adopt. This is because consciousness is held to be inessential, or epiphenomenal, in
so far as it does not play any cognitive and causal role, so that cognition that counts
could be implemented even if consciousness did not exist at all.12
Later on scholars realized that overlooking consciousness led to discarding one
crucial aspect of cognition, that of phenomenological and conscious data, that cog-
nitive sciences simply would not explain, thus being incomplete as a theory of the
mind.13 An overreaction to introspectionism had led philosophers of mind to jettison

12
 The cognitive tradition has had a long history. The attempt to elaborate a scientific theory of the
mind has been long prepared by the discussions that followed the birth of modern science after
Galileo Galilei about how natural sciences are separate from the science of the spirit. By way of
introducing a discontinuity with the western metaphysics stemming from Aristotle, Galileo had
brought to the fore the problem of how the mind is recalcitrant to a reduction to the kind of mecha-
nistic paradigm that informs biology. For a reconstruction of such development from Galileo to
cognitive sciences, see Petitot, Varela, Pachoud & Roy 1999, pp. 1–80.
13
 For this line of criticism against cognitive sciences see Nagel 1979, pp. 162–175; Jackendoff
1987; Dennett 1991; Searle 1992; Flanagan 1992; Chalmers 1996. This gap has helped encourage
skepticism toward science and its underlying naturalism: whereas objectivity in science has it that
“a view from nowhere” (see. Nagel 1986) should be adopted to study the world scientifically,
Preface xi

such a central fact of the mind as consciousness. From that moment onward, we
have witnessed the “consciousness boom,” that is, the flourishing of studies aimed
at filling this programmatic and explanatory gap.14 And that gap has prompted the
necessity of cooperation with phenomenology and an integration between cognitive
sciences and phenomenology.15
If consciousness has become a central locus of encounter for comparative and
collaborative investigation, self-knowledge should be even more cooperatively
tackled. We hope that this book will call on scholars to forge bridge questions with
a view to mutually enriching and deepening our perspectives of the topic, and that
it may at least count as one step to opening up the conceptual space for discussing
the prospects of a fruitful interaction between phenomenologists, philosophers of
mind, and epistemologists in the attempt to clarify further the questions posed by
self-knowledge.

References

Baker, L.R. 2000. Persons and body. Cambridge: Cambridge University Press.
Baker, L.R. 2007. The metaphysics of everyday life. An essay in practical realism. Cambridge:
Cambridge University Press.
Baker, L.R. 2013. Naturalism and the first-person perspective. Oxford: Oxford University Press.
Brentano, F. 1973. Psychology from an empirical standpoint. London: Routledge and Kegan Paul.
Trans. A.C. Rancurello, D.B. Terrell, and L.L. McAlister.
Chalmers, C. 1995. The conscious mind. Oxford University Press.
Cobb-Stevens, R. 1990. Husserl and analytic philosophy. The Hauge: Kluwer Academic
Publishers.
Dennett, D. 1991. Consciousness explained. Boston: Little & Brown.
Dreyfus, H.L. 2000. “A Merleau-Pontyian Critique of Husserl’s and Searle’s Representationalist
Accounts of Action”, in Proceedings of the Aristotelian Society, New Series, vol. 100.
Dummett, M. 1993. Origins of analytic philosophy. London: Duckworth.
Flanagan, O. 1992. Consciousness reconsidered. Cambridge, MA: MIT Press.
Friedman, M. 1999. Reconsidering logical positivism. Cambridge and New  York: Cambridge
University Press.
Fodor, J. 1982. Methodological solipsism considered as a research strategy in cognitive psychol-
ogy. In Husserl, Intentionality and cognitive science, ed. H.L. Dreyfus. Cambridge, MA: MIT
Press.
Føllesdal, D. 1958. Husserl und Frege: Ein Beitrag zur Beleuchtung der Entstehungder phänome-
nologischen Philosophie, Ascheoung, Oslo.
Gallagher, S., and Zahavi, D. 2007. The phenomenological mind. London: Routledge.

phenomenological data are intrinsic to a point of view, so they would fall squarely outside the
domain of science. Furthermore, such data can hardly be naturalized.
14
 “Explanatory gap” is the felicitous phrase used by Levine 1983, pp. 354–361.
15
 Neurophenomenology works on the hypothesis that the functioning of consciousness could be
closer to the functioning of the brain than it is generally held to be in traditional cognitive sciences.
Phenomenological differences could correspond to neurological differences. See Varela 1996,
pp. 330–350.
xii Preface

Gallagher, S., and Zahavi, D. 2009. La mente fenomenologica. Milano: Raffaello Cortina Editore.
Trans. P. Pedrini.
Husserl, E. 1966a. Zur Phänomenologie des inneren Zeitbewußtseins (1893–1917). Husserliana
X. Den Haag: Martinus Nijhoff; On the phenomenology of the consciousness of internal time
(1893–1917). Trans. J. Brough. Dordrecht: Kluwer Academic, 1991.
Husserl, E. 1966b. Analysen zur passiven synthesis. Husserliana XI. Den Haag: Martinus Nijhoff.
Husserl, E. 1973. Zur Phänomenologie der Intersubjektivität III. Husserliana XV.  Den Haag:
Martinus Nijhoff.
Husserl, E. 1984. Einleitung in die Logik und Erkenntnistheorie. Husserliana XXIV. Den Haag:
Martinus Nijhoff.
Husserl, E. 2001. Logical Investigations I-II. Trans. J. F. Findlay. London: Routledge.
Jackendoff, R. 1987. Consciousness and the Computational Mind. Cambridge: MIT Press.
Levine, J. 1983. “Materialism and qualia: The explanatory gap”, in Pacific Philosophical Quarterly,
64.
Livingston, P. 2005. Functionalism and logical analysis. In Phenomenology and philosophy of
mind, ed. D.W. Smith e A.L. Thomasson.
Merleau-Ponty, M. 1962. Phenomenology of perception. Trans. C. Smith. London: Routledge &
Kegan Paul.
Monthany, J.N. 1982. Husserl and Frege, Bloomington: Indiana University Press.
Monthany, J.N. 1985, Husserl, Frege and the overcoming of psychologism. in The possibility of
transcendental philosophy. Boston: Kluwer.
Moran, D. 2017, Husserl’s phenomenology and the project of transcendental self-­knowledge. In
Self-knowledge. A history, ed. U. Renz. Oxford: Oxford University Press.
Mulligan, K. 2003, Searle, Derrida and the end of phenomenology. In B.  Smith, John Searle,
Cambridge: Cambridge University Press.
Nagel, T. 1986. The view from nowhere. Oxford: Oxford University Press.
Nagel, T. 1979. What it is like to be a bat? Mortal questions. Cambridge: Cambridge University
Press.
Pedrini, P. 2009a. Postfazione, in Gallagher & Zahavi 2009. Trans. P. Pedrini.
Pedrini, P. 2009b. Prima persona. Epistemologia dell’autoconoscenza. Pisa, ETS Edizioni.
Pedrini, P. 2015a. Rescuing the ‘Loss-of-Agency’ account of thought insertion. Philosophy,
Psychiatry, & Psychology 22(3): 221–233.
Pedrini, P. 2015b. Reply to commentaries: On the pre-reflective perplexity of a Schizophrenic
thinker. Philosophy, Psychiatry, & Psychology 22(3): 243–245.
Petitot, J., Varela, F.J., Pachoud, B., Roy, J.-M. 1999. Beyond the gap: An introduction to natural-
izing phenomenology. In Naturalizing phenomenology: Issues in contemporary phenomenol-
ogy and cognitive science, eds. J. Petitot et al. Stanford: Stanford University Press.
Putnam, H. 1981. Reason, truth and history. New York: Cambridge University Press, 1981.
Putnam, H. 1983. Representation and reality, Cambridge: MIT Press.
Putnam, H. 1987. The many faces of realism. Open Court, La Salle.
Ricoeur, P. 1966. Freedom and nature: The voluntary and the involuntary. Evanston: Northwestern
University Press.
Ryle, G. 1949. The concept of the mind. New York: Barnes and Noble.
Ryle, G. 1971. Phenomenology versus the concept of mind, in Collected papers, London:
Hutchinson.
Sartre, J.-P. 1956. Being and nothingness. Trans. H.E. Barnes. New York: Philosophical Library.
Sartre, J.-P. 1957. The transcendence of the ego. Trans. F. Williams & R. Kirkpatrick. New York:
The Noonday Press.
Searle. J. 1992. The Rediscovery of the Mind. Cambridge, MA.: MIT Press.
Smith, D.W. 1982. R. McIntyre, Husserl and intentionality: A study of mind, meaning, and lan-
guage. New York: Springer.
Sommer, H. 1985. Husserl und der frühe Positivismus. Frankfurt am Main: Klostermann.
Preface xiii

Thomasson, A.L. 2002, Phenomenology and the development of analytic philosophy. Southern
Journal of Philosophy 40, Supplement.
Thomasson, A.L. 2005. First person knowledge in phenomenology. In Phenomenology and phi-
losophy of mind, eds. D.W. Smith e A.L. Thomasson, . Oxford: Oxford University Press.
Tugendhat, E. 1976. Einfürhung in die sprachanalytische Philosophie. Frankfurt am Main:
Suhrkamp.
Varela, F. 1996. Neurophenomenology: A methodological remedy to the hard problem. Journal of
Consciousness Studies 3.
Willard, D. 1984. Logic and objectivity of knowledge: A study in Husserl’s early philosophy.
Athens: Ohio University Press.
Contents

I ntroduction: Getting to Know Our Own Minds������������������������������������������    1


Julie Kirsch and Patrizia Pedrini
Self-Knowing Interpreters������������������������������������������������������������������������������   13
Annalisa Coliva
Extended Self-Knowledge ������������������������������������������������������������������������������   31
J. Adam Carter and Duncan Pritchard
 he ‘Crux’ of Internal Promptings����������������������������������������������������������������   51
T
Patrizia Pedrini
Interpreting Intuitions������������������������������������������������������������������������������������   73
Marcus McGahhey and Neil Van Leeuwen
Interpreting Things Past����������������������������������������������������������������������������������   99
Julie Kirsch
 elf-Interpretation as Software: Toward a New Understanding
S
of Why False Self-Conceptions Persist����������������������������������������������������������  115
Tadeusz Wiesław Zawidzki
Self-Interpretation and Social Cognition������������������������������������������������������  145
Shaun Gallagher
 ermeneutics, Self-Knowledge and Self-Interpretation������������������������������  159
H
Bruce B. Janz
Identification and Self-Knowledge ����������������������������������������������������������������  177
Luca Malatesti and Filip Čeč
 ausal Inference in the Clinical Setting: Why the Cognitive
C
Science of Folk Psychology Matters ��������������������������������������������������������������  191
Andrew Sims

xv
Contributors

J. Adam Carter  University of Glasgow, Glasgow, UK


Filip Čeč  Department of Philosophy, Faculty of Humanities and Social Sciences,
University of Rijeka, Rijeka, Croatia
Annalisa Coliva  Department of Philosophy, University of California, Irvine, CA,
USA
Shaun Gallagher  Department of Philosophy, University of Memphis, Memphis,
TN, USA
Faculty of Law, Humanities and the Arts, University of Wollongong (AU),
Wollongong, NSW, Australia
Bruce B. Janz  Department of Philosophy, The Center for Humanities & Digital
Research, and Core faculty in the Texts and Technology Ph.D. Program, University
of Central Florida, Orlando, FL, USA
Julie Kirsch  Department of Liberal Arts, D’Youville College, Buffalo, New York,
USA
Luca  Malatesti  Department of Philosophy, Faculty of Humanities and Social
Sciences, University of Rijeka, Rijeka, Croatia
Marcus McGahhey  Department of Philosophy, University of California, La Jolla,
CA, USA
Patrizia  Pedrini  Department of Letters and Philosophy, University of Florence,
Florence, Italy
Duncan Pritchard  University of California, Irvine, CA, USA
University of Edinburgh, Edinburgh, UK

xvii
xviii Contributors

Andrew Sims  Institut supérieur de philosophy, Université Catholique de Louvain,


Louvain-la-Neuve, Belgium
Neil Van Leeuwen  Department of Philosophy and Neuroscience Institute, Georgia
State University, Atlanta, GA, USA
Tadeusz Wiesław Zawidzki  Department of Philosophy and Mind-Brain Institute,
George Washington University, Washington, DC, USA
Introduction: Getting to Know Our Own
Minds

Julie Kirsch and Patrizia Pedrini

Abstract  Our project is to answer questions that lead to a clearer picture of third-
person self-knowledge, the self-interpretation it embeds, and its narrative structure.
Bringing together current research on third-person self-knowledge and self-inter-
pretation, our authors focus on third-person self-knowledge, and the role that narra-
tive and interpretation play in acquiring it. We regard the third-personal epistemic
approach to oneself as a problem worthy of investigation in its own right, and our
contributors try to clarify the relation between third-person self-knowledge, self-
interpretation, and narrative capacities. We also invite cooperation between analytic
philosophy and phenomenology about issues regarding self-knowledge.

Keywords  Self-knowledge · First-person self-knowledge · Third-person self-­


knowledge · Self-interpretation · Narrative

In recent years, the idea that each person is in a privileged position to acquire knowl-
edge about her own mental states has come under attack. A growing body of empiri-
cal research has cast doubt upon the existence of what philosophers call ‘first-person
self-knowledge’, i.e., knowledge about our mental states that is often thought to be
immediate, transparent, and authoritative. This line of thought, along with certain
theoretical commitments, has led some philosophers to claim that what seems to be
‘first-person self-knowledge’ is really just ‘third-person self-knowledge,’ i.e.,
knowledge about our mental states that is inferential and fallible (cf. Carruthers
2009, 2011; Gopnik 1993; Lawlor 2009; Dennett 1987). Other philosophers, uncon-
vinced by this research, have tried to uphold the asymmetry between first-person
and third-person self-knowledge (cf. Moran 2001; Bilgrami 2006; Coliva 2012a, b).

Authorship of this chapter is divided equally.


J. Kirsch (*)
Department of Liberal Arts, D’Youville College, Buffalo, New York, USA
e-mail: kirschj@dyc.edu
P. Pedrini
Department of Letters and Philosophy, University of Florence, Firenze, Italy

© Springer Nature Switzerland AG 2018 1


P. Pedrini, J. Kirsch (eds.), Third-Person Self-Knowledge, Self-Interpretation,
and Narrative, Contributions To Phenomenology 96,
https://doi.org/10.1007/978-3-319-98646-3_1
2 J. Kirsch and P. Pedrini

The debate is then divided up between those who argue that all self-knowledge
worthy of the name is actually third-person self-knowledge and those who argue
that there are two distinct kinds of self-knowledge, first-person self-knowledge and
third-person self-knowledge. Against the state of the art, this volume focuses instead
on the clarification of the very nature of third-person self-knowledge, whether or not
it is all the self-knowledge there is.
We are personally convinced that there is self-knowledge of both types, and that
the existence and importance of first-person self-knowledge is worth defending.
Nonetheless, in the last few decades, there has been a significant lack of research
about third-person self-knowledge as such, which has long been in need of
­clarification of its very nature and implications. Comparatively, in the effort to resist
the assimilation of first-person self-knowledge with third-person self-knowledge,
defenders of first-person self-knowledge have offered important clarifications of the
former. But many important aspects of third-person self-knowledge have been
largely unexplored. Recently, Cassam (2014) has tried to fill this gap laying down
an interesting theory of third-person self-knowledge, which he claims is quite far
extended and able to cover instances of self-knowledge previously considered by a
wide number of others as indisputable candidates of first-person self-knowledge.
The book has had the merit of inspiring several commentaries which explicitly take
a stand on the topic (Boyle 2015; Coliva 2015, 2016). Most of them, however, try to
show that the scope of third-person self-knowledge is not after all as extended as
Cassam thinks it is by finding fault with the arguments Cassam gives to defend such
extension. Coliva 2016, in particular, argues in favor of a novel, pluralistic view of
self-knowledge, which gives ample citizenship to both kinds.
Our concern in this book is orthogonal to the debate about the principled differ-
ences between first- and third-person self-knowledge. This volume aims to fill the
gap about third-person self-knowledge further, and explain important features of
third-person self-knowledge that will prove to be of interest for both those who
believe that it is all the self-knowledge there is and for those who think it is just one
of two kinds. In doing this, we offer a series of contributions which try to illuminate
how third-person self-knowledge is connected to interesting forms of self-­
interpretation and narrative, which can have as their object not only our own verbal
or non-verbal behavior, but also other mental states, some of which may be either
conscious or unconscious, and to our narrative capacities.
One of the most important features of third-person self-knowledge is its “medi-
ated” character, its being inferential and then fallible. When a subject gains third-­
person self-knowledge, she infers a truth about her mental states or her psychological
experience from evidence provided by her behavior, testimony of this behavior, or
other mental states. However, the very machinery of the inference in question has
not yet been fully unpacked. The hypothesis that guides this volume is that when we
perform such inference, we apply a certain mental concept, or cluster of concepts,
to the available evidence (how we behave, verbally and/or non-verbally; what we
experience; how we feel; the general theory of the mind we subscribe to; the cultural
patterns in which we are immersed; the behavior and reactions of others towards
our own verbal and/or non verbal behavior; and in general all of the implications
Introduction: Getting to Know Our Own Minds 3

and “implicatures” of our social interaction). Important factors that need to be clari-
fied seem to determine the choice of a certain concept, or cluster of concepts, to
capture what we are going through. The subject’s individual culture and identity, the
very conceptual vocabulary that she possesses, the degree of reasoning sophistica-
tion she can deploy in getting to novel concepts that are appropriate to capture her
inner life, and the exposure to conceptual sources by others she trusts are factors
that need to be explored even more than they have been by the recent attempts to
analyze third-person self-knowledge. It is agreed by Cassam and Coliva, for exam-
ple, that when we perform inferences to the concept, or cluster of concepts, that best
capture our experience and behavior, we interpret our experience and behavior to a
certain degree. It is the richness and peculiarity of the interpretation of a subject’s
experience and behavior that is thus the second focus of our current interest and we
wish to explore the role played by self-interpretation during third-person attempts
at gaining knowledge of one’s own psychological states.
There is a diverse body of empirical research that suggests that human beings
engage in self-interpretation at least some of the time. Take, for example, Michael
Gazzaniga’s famous studies involving split-brain patients (cf. Gazzaniga 1995a, b).
Gazzaniga spent years studying patients whose corpus callosum was severed for the
purpose of relieving epileptic seizures. Among other things, his research shows that
the left and right hemispheres of the brain have different specializations: the left
hemisphere is superior for language and interpretation, whereas the right is better
able to carry out visual-spatial tasks. Interestingly, when a patient’s corpus callo-
sum is severed, the right hemisphere is unable to communicate what it ‘sees’ with
the left brain. In one of several well-known experiments, Gazzaniga presented the
command ‘walk’ to a patient’s left brain. After viewing the prompt, the patient stood
up and approached the door of the testing van to leave. When asked where she was
going, she replied, “I’m going into the house to get a Coke”. Unaware of what her
right brain ‘saw’, the patient’s left brain developed a faulty interpretation of her
behavior. The patients involved in Gazzaniga’s research were clearly confabulating;
their left brains were constructing remarkably inventive interpretations of their
behavior based upon an incomplete understanding of things. But what, if anything,
do Gazzaniga’s patients reveal about ordinary human beings whose brains are
intact? Are our own self-interpretations as wildly off the mark as theirs? An optimist
would argue that self-interpretation can be a useful and reliable tool for acquiring
knowledge about our own mental states. We might gain new and valuable insight
into our mental states by reflecting upon them from the perspective of an external
observer. But a pessimist would be quick to point out that exercises of self-­
interpretation are often tainted by ‘the dear self’ (as Immanuel Kant might put it).
We all-too-often use self- interpretation as a way of flattering our egos and enhanc-
ing our virtue.
This leads us to other questions that can be asked about third-person self-­
knowledge, and that are still waiting for an answer. Some of these questions are
genuinely epistemological: what are the necessary and sufficient conditions for
genuine knowledge from the third-person perspective? How can we be sure that we
are not just getting (inner) things wrong? In particular, when we try to understand
4 J. Kirsch and P. Pedrini

our mental states or our behavior on a sufficiently large scale (e.g., when we try to
make sense of our life, or of a bit of it, whether or not such operation is legitimate),
it makes a huge difference to the outcome of our attempt if we possess a sufficiently
nuanced psychological vocabulary and the correct sense of the concatenation of its
terms, or if instead we work under a poor conceptual scheme. Some have the intu-
ition that in the second case we can be badly wrong and don’t in fact acquire knowl-
edge of what’s going on inside. But the skeptical threat is even more worrisome:
how do we adjudicate between different conceptual schemes that cultures offer us
to interpret ourselves? It seems that some narratives are more adequate than others
to understand ourselves, but it’s a difficult task to rank those narratives in terms of
adherence to how things actually stand. Philosophy of psychology, of psychiatry,
and of psychoanalysis dispute on such issue, in so far as different conceptual para-
digms compete for the correct understanding of our psychology. But that is not
simply a theoretical problem. It easily gets individual, as depending on the para-
digm that one uses to look at oneself, one can reach very different interpretations
about oneself.
Some of these interpretations can create psychological tensions that seem to
reveal that the subject is self-deceiving, or at least “innocently” interpreting incor-
rectly what he or she goes through. Be that as it may, it seems that inner experience
can tolerate a certain degree of variation in self-interpretation, without us feeling
excessively uncomfortable with it on many occasions. A suspicion looms that inner
experience is somehow under-determined with respect to psychological theories, to
put it in Quinean terms, to the extent that the very same inner state can tolerate
vastly different interpretations by different individuals, or by the same individual at
different times. In all, third-person self-knowledge is based largely upon self-­
interpretation, which is a complex activity that goes hand in hand with the concep-
tual resources of the subject performing it. And there may, in principle, be a criterion,
or set of criteria, to distinguish between knowledge-conducive self-interpretations,
on the one hand, and self-interpretations that instead put us off track. Or at least, it
would be desirable to come up with such criterion, or set of criteria, as a guidance
to action and thinking. But whether or not there is such criterion, or set of criteria,
at all, it seems to be a constraint of a theory of it that we can adjudicate between
available conceptual vocabularies.
Finally, the third focus of our interest is the narrative structure that self-­
interpretation often takes. Human beings have long been recognized to be storytell-
ers (Dennett 1991) and that is said to matter to the constitution of a sense of the self
(cf. Goldie 2012; see also Schechtman 1996). This volume further investigates how
and why self-interpretation often gets incorporated into a self-told story of one’s
life, or pieces of one’s life. It looks at the distinctive contribution that narrative
explanation can make to self-understanding. Shaun Gallagher (Chap. 6), for exam-
ple, contrasts a theory of mind approach with narrative approach to self understand-
ing and argues that the latter, unlike the former, “involves a rich knowledge of one’s
own embodied comportments and skills, one’s affective life, one’s autobiography,
and a knowledge that derives from one’s relations with others”  (this volume).
Similarly, Janz describes narrative as a “major way that language structures and
Introduction: Getting to Know Our Own Minds 5

makes available a world” (this volume). In this view, narrative plays an important
role in the way that we understand and make sense of ourselves and others. However,
given that there is often more than one way to construct a narrative, skeptical wor-
ries about interpretation and misinterpretation reappear here. How can we be sure
that the cultural forces that influence our narratives (see Janz, Zawidzki, and
Gallagher, this volume) do not distort or corrupt them? Can we ever be confident
that the particular narrative that we have settled upon is not delusional or self-­
deceptive? In trying to understand the contribution that narrative makes to third-­
person self-knowledge, we cannot afford to neglect these important skeptical
questions.
The three themes explored in this volume — third-person self-knowledge, inter-
pretation, and narrative — combine in novel and sometimes unexpected ways. The
authors featured in this collection bring together a broad range of theoretical per-
spectives and areas of expertise.
In Chap. 2, Coliva investigates the asymmetry between first- and third-person
self-knowledge in the pluralistic spirit she has also defended elsewhere (2016), and
offers a detailed account of the machinery of third-person self-interpretation. She
presents a set of characteristics of first-person self-knowledge which she then uses
to single out crucial aspects of third-personal self-knowledge. She prepares the logi-
cal map for all those interested in third-person self-knowledge, and the self-­
interpretation it involves. As a defender of first-person self-knowledge, she explains
how first-person self-knowledge is neither inferential nor observational. Rather, it is
direct or immediate, groundless, and authoritative. It’s also the basis for further
inferences to the presence of other mental states that can’t be known first-­personally.
Such inferences, to get started at all, and on pain of an infinite regress, do count on
such first-personal self-knowledge as contents of the premises of reasoning. Also,
it’s not observational, as it would be troubling to assume that most mental states are
simply there to be observed, even if some can just be grasped by observing one’s
overt behavior and other mental evidence one might have. Even if groundless, that
is, even if not reached by evidence as other cognitive achievements are, first-­personal
self-knowledge is not based on “anything” either, as some instead have maintained,
by subtracting favorable prospects to the possibility of giving a credible epistemo-
logical account of it. Coliva argues that it is possible to redeem an epistemological
account of first-person self-knowledge, following the track of a priori reflection on
a set of interrelated concepts, including transparency and authority, and involving,
crucially, concepts possession, cognitive well-functioning, alertness, attentiveness,
and excluding unconscious and purely dispositional mental states.
The theoretical apparatus Coliva uses allows her to individuate all the cases in
which first-person self-knowledge doesn’t hold, thus refining the scope of first-­
person self-knowledge, as well as singling out cases in which third-person self-­
knowledge is in order. She presents all of the troubling epistemic features of the
latter, including fallibility, and she illustrates with high precision the machinery of
self-interpretation that third-person self-knowledge requires. Cases of vague states,
self-deception, uncertain or ambivalent emotions, affective states and affective fore-
6 J. Kirsch and P. Pedrini

casting, inductive reasoning from the past, and  counterfactual imagination show
how the mischaracterization of our inner states and behavior is always looming.
Coliva also excellently characterizes self-interpretation by invoking the activity
of subsuming patterns of behavior or inner elements under concepts newly acquired
and applied, and which can also be finessed by experience and culture. She explains
that new concepts can render salient aspects of our mind previously unknown to us,
or to which we were simply blind, and that we can also improve, becoming more
proficient in concepts application.
Interestingly, she also argues that third-person self-knowledge is not simply on a
par with knowledge of other minds, or knowledge of the external world, as some
claim, as the kind of evidence we can use in the case of third-person self-knowledge
can be inner, and supplied by first-person self-knowledge itself. She thus defends a
plurality of methods which can help us reach third-personal self-knowledge, which
pair with an ontological plurality of mental states which should not be underesti-
mated or overlooked.
Chapter 3 is contributed by leading epistemologists Pritchard and Carter, who
offer an important advancement in the general epistemology of self-knowledge, tra-
ditionally under attack from content externalism, and more recently from cognitive
science. The authors take an original stand on the contrast between self-knowledge
and content externalism, with a view to establishing an interesting expansion of
self-knowledge, contrary to recent objections which seem to jeopardize self-­
knowledge further. The possibility of knowing oneself has been long challenged by
content externalism, according to which physical and social environment play a role
in the individuation of mental content, and hence in determining what our thoughts
are thoughts about. This has created a much investigated philosophical tension, as it
seems obvious that if content externalism is true, then one can’t know one’s own
mental states directing one’s attention “inwards”. One has to look at the world, if
one wants to know what one’s thoughts are thoughts about. The extended cognition
thesis and situationism seem to have created even more tension: according to the
first, cognitive process can supervene on parts of the world that are external to the
biological agents, whilst the second says that cognitive success depends to a surpris-
ing extent on various extra-agential factors specific to certain situations. Pritchard
and Carter argue that, on closer inspection, such challenges do not diminish self-­
knowledge; rather, they augment it, in so far as extended knowledge is not incom-
patible with self-knowledge at all. On the contrary, they are an important alley of it.
While acknowledging all the threats to first-person self-knowledge, including non-­
inferential biases and what they call “negative epistemic dependence”, they make a
tactic move which capitalizes “positive epistemic dependence” to defuse the chal-
lenge from extended cognition. Among the many interesting results they reach —
including separating the level of integration on the information from its storage
provenance, and importantly blocking the identification between privacy and intro-
spection — they also offer a general lesson that it is worth considering: we should
be careful to judge instances of self-knowledge, as it may happen that what is taken
to be third-person self-knowledge, such as in many examples given of extended
cognition, in fact it is not. Not all external factors and supports involve inference or
Introduction: Getting to Know Our Own Minds 7

observation. From a descriptive point a view, Pritchard and Carter reshape the
domain of first-person self-knowledge by widening it. They also give an important
contribution to general epistemology, by facing and brilliantly defusing the chal-
lenges for externalism, extended cognition, and situationism.
In Chap. 4, Pedrini focuses on a more specific application of third-person self-­
interpretation: “raw” internal promptings. She deals directly with the theory of
third-person self-knowledge that Quassim Cassam has offered in Self-Knowledge
for Humans (2014). Cassam defends a quite broad inferentialist theory of substan-
tive third-person self-knowledge, which promises to extend virtually all mental
states, including the so called “internal promptings”. “Internal prompting” is a tech-
nical phrase Cassam borrows from Lawlor (2009) to indicate experiential episodes
which present themselves spontaneously. They may not always be phenomenologi-
cally salient, nor may they be conceptually clearly subsumed. According to Cassam,
however, their spontaneous manifestation does not preclude that our access to them
is in fact inferential. Pedrini offers reasons to question the claim that internal
promptings can really be covered by an inferentialist theory of self-knowledge.
Even if Cassam is right to claim that it is by inference (possibly to the best explana-
tion) that we can adjudicate the epistemological question regarding what they are,
their initial surfacing in our psychological experience as raw unconceptualized
internal promptings tells us that the mere capacity we have to notice them must
amount to some kind of self-knowledge nonetheless. She thus distinguishes between
two epistemological accounts of self-knowledge: self-knowledge by discrimination,
or by identification, and self-knowledge without discrimination, or identification.
She argues that while the former may not apply to raw internal promptings, the lat-
ter does. While self-knowledge without discrimination, or identification, may be
perceived as quite modest, as a kind of self-knowledge, it has the advantage of
respecting the elusive, opaque, phenomenologically inexplicit nature of an impor-
tant class of internal promptings. She also discusses to what extent internal prompt-
ings can function as a basis for inferences to discriminative self-knowledge, in case
their vagueness reaches a considerable extent, and she addresses the issue of the risk
of their under-determination with respect to the conceptual schemes used to identify
them. By making use of the notion of “psychic residual tension after concepts appli-
cation” to elusive promptings she individuates a possibly reliable track to the sound-
ness of a conceptual scheme application.
In Chap. 5, we have another instance of application of self-interpretation to an
interesting and specific class of mental states: intuitions. We often associate self-­
interpretation with beliefs, desires, and other propositional attitudes, but McGahhey
and Van Leeuwen show that it has a crucial role to play in the way that we under-
stand and formulate judgments about theoretical intuitions. They argue, against the
mainstream view (or ‘Mainstream Propositionalism’), that many intuitions that we
have in response to philosophical thought experiments, such as Gettier cases, are
not propositional in nature; instead, they lack propositional content and are more
like “hunches, urgings, murky feelings, and twinges.” According to McGahhey and
Van Leeuwen’s own positive view, ‘Interpretationalism,’ we must first interpret
intuitions in light of background beliefs before forming propositional judgments in
8 J. Kirsch and P. Pedrini

response to them. When we first encounter a Gettier case, for example, we may have
an initial feeling that something is not right, but we must interpret this initial
response against our background beliefs before we can formulate a propositional
judgment about it. Although McGahhey and Van Leeuwen focus upon the role that
interpretation plays in formulating judgments about thought-experimental intu-
itions, something similar may be said about our intuitions in a variety of other
contexts. Indeed, McGahhey and Van Leeuwen argue that we have intuitions of a
similar form  — “spontaneous, conscious, non-theoretical, and non-proposi-
tional”  — in response to fictional sentences. Their analysis, and emphasis upon
non-­propositional mental content, is also compatible with the findings of contempo-
rary cognitive science and its recognition of sub-doxastic mental systems and
information.
In Chap. 6, Kirsch continues the discussion of the epistemic risks associated with
self-interpretation. In particular, she is interested in biases affecting our self-­
interpretation of the past, while also analysing the possibility of redemption from
the third-person perspective. She examines the role that interpretation plays in the
way that we remember and make sense of the past. There is often more than one
way to interpret past events, and our memories, which are often distorted and inac-
curate, reflect this interpretative flexibility. The narratives that we construct about
our lives  — the stories that we tell ourselves  — shape the way that we see and
understand the past. Kirsch argues that our ability to interpret the past in more than
one way gives rise to a mixed set of consequences. On the one hand, this ability may
empower us to view a personal setback or failure in a more positive light, thus
prompting us to respond to it in a productive, non-defeatist way. But, on the other,
it may encourage self-deception and enable us to interpret away, and in some cases
forget altogether, the harm that we have caused to others. Kirsch nevertheless
believes that we have some limited control over our memories and interpretations of
the past. We can, for example, guard against potential well-documented biases and
make an effort to compare our interpretations with the interpretations of others.
In Chap. 7, Tad Zawidzki explores a puzzle that arises within the domain of self-­
interpretation: many of our action-relevant self-conceptions persist even though
they are false and deeply misleading, while in other domains false action-relevant
beliefs, such as beliefs about food location or environmental dangers, would disap-
pear because the beliefs, as well as the mechanisms that generate them, would be
selected against. Zawidzki argues that misleading self-conceptions persists because
they do not “aim to represent independently constituted facts” (this  volume).
Instead, such beliefs aim to create facts and transform human beings into the kinds
of persons who accept the constraints of these beliefs for the purpose of social coor-
dination. Thus, in his view, false self-conceptions exist not because they accurately
represent facts about ourselves, but because they allow us to coordinate with others
relative to our social environment. Zawidzki’s chapter highlights the fact that our
self-interpretations may be motivated by considerations other than truth. Even
though a given self-interpretation is false, it may persist because it is socially useful
and benefits us in other ways. Towards the end of the chapter, he comments on the
way in which social and cultural forces influence how we interpret ourselves. For
Introduction: Getting to Know Our Own Minds 9

example, Zawidzki speculates that “the availability of cheap and safe birth control
and the pop-cultural revolution of the 1960s” had some effect on women’s self-­
conceptualizations (this volume).
There are, however, also other influences on self-interpretations, and not all of
them need be malign, but rather simply constitutively present in the social environ-
ments we all live in. In Chap. 8, Gallagher explores the social and cultural con-
straints on self-interpretation which are embedded in the narratives we are exposed
to. He defends a narrative approach to self-understanding and believes that a narra-
tive approach provides a better account of self-understanding precisely because it
provides a more parsimonious account of other-understanding than its main com-
petitor, the theory-of-mind (ToM) approach. Whereas ToM approaches “emphasize
mindreading by inference or simulation,” a narrative approach looks at the way in
which social and cultural norms affect our narratives (this volume). It is therefore
better able to account for the narrative competency that we have when making sense
of ourselves and others. According to Gallagher, self-understanding, on a narrative
approach, “involves a rich knowledge of one’s own embodied comportments and
skills, one’s affective life, one’s autobiography, and a knowledge that derives from
one’s relations with others” (this volume). Gallagher’s chapter sheds light upon the
social and cultural forces that influence our self-interpretations. He shows that the
process of constructing a narrative is never solipsistic; instead, it plays out interper-
sonally and under the influence of social and cultural norms and institutions. From
an early age, children are exposed to the narratives of others through both personal
interactions and larger cultural experiences. They draw upon these narratives, and
their responses to them, later in life to interpret themselves and others.
In Chap. 9 Bruce Janz continues the discourse about the role of narratives and
conceptual schemes in self-interpreting, and he does so from the point of view of
hermeneutics. In particular, he explores the way in which hermeneutics can help us
understanding third-person self-interpretation. Hermeneutics has a long history of
placing the subject in an already meaningful world, which means that the task of
self-interpretation and self-understanding (these two may be different, as he argues)
will always be a continuous one between various modes of reference and perspec-
tive. If the task of self-knowledge is one of getting things right about the self, this
will have to be understood in the context of being part of an already meaningful
world. That is to say, knowledge about the self is not the primitive building-block of
understanding the self — we do not first get the facts about the self (or about our
self) right and only after that come to an understanding of the self. Janz’s reflections
cover an investigation about how we use the rational capacity he calls “forensics” to
navigate diverse and incommensurable forms of reason. Hermeneutics also enables
a second-person capability, that of being open to the voice of the other, which aids
in forensically sorting and interpreting between third-person accounts. While
Cartesian versions of the self solve the problem of multiple narrative sources for the
self by giving one version regulatory power, this pride of place for the first-person
has been questioned since at least the middle of the nineteenth century. Marx, Freud,
Saussure, Lacan, Derrida, and a host of others showed the limits of thinking about
the first-person as reliable or regulatory over other frames of reference. Hermeneutics
10 J. Kirsch and P. Pedrini

(and phenomenology more generally), however, does not start with that assumption
or the prioritizing of the individual. All self-understanding occurs within a context
where all the forms of referentiality, perspective, and relationality exist
simultaneously.
Also, importantly, Janz tackles an issue that lies at the core of our interests: not
only is it important to address the central questions of what an epistemological
theory of the third-person is, how we can adjudicate between competing third-­
person accounts, and how we can account for less adequate ones, but it is also impor-
tant to embrace the equally crucial question, both for the hermeneuticists and the
analytic philosophers, about how we can account for the construction of the self and
the world amidst the competing forms of reason that the self deploys. By explicitly
focusing on whether there is a way of recognizing misinterpretation of the self and
the world, while not ruling out the possibility that a minority of  third-person
accounts or an unpopular account might be correct after all, Janz makes use of the
idea of “travelling concepts” (Bal 2001), that is, concepts which are developed in
one conceptual ecology, and then travelling to another. Janz also leads us to ques-
tions about the production of the self in third-person narratives. Hermeneutics
emphasizes the importance of language, and the world that language creates. In
other words, it is not that language is a vehicle for truth-values, as in the epistemo-
logical approach. It is that language itself is important in creating a space in which
truth or falsity makes sense at all. But this does not mean that any third-person nar-
rative is as good as any other, and that it is merely an act of the will or, alternatively,
an act of social construction that determines the self — hermeneutics is not volun-
tarism, nor is it social determinism.
An issue connected with self-interpretation is whether one can be a fallible self-­
interpreter, and yet be credited with having a “real self” after all, which be sufficient
to warrant autonomy. This analysis is offered in Chap. 10 by Malatesti and Čeč.
They respond to Peter Carruthers and Matt King’s criticism of Real Self accounts of
moral responsibility and autonomy. Carruthers and King have argued that such
accounts are under threat because they rely upon an implausible conception of self-­
knowledge  — one that is incompatible with contemporary cognitive science.
Specifically, Real Self accounts of moral responsibility and autonomy assume that
we have transparent self-knowledge of propositional attitudes, such as beliefs and
desires. Malatesti and Čeč try to rescue such accounts, not by showing that we have
transparent access to our propositional attitudes, but by showing that we have other
forms of self-knowledge that meet the requirements of the Real Self view. Malatesti
and Čeč maintain that the transparent access to sensory and perceptive data that we
do have is enough to support the types of self-knowledge that are required by Real
Self views. Thus, Malatesti and Čeč offer a version of the Real Self view that is
compatible with contemporary cognitive science and the work of King and
Carruthers. They nevertheless stress that their account does not rule out the possibil-
ity that we may be mistaken about our mental states. Like Carruthers and King,
Malatesti and Čeč believe that we acquire knowledge about ourselves through the
same process that we use when acquiring knowledge about others (although we have
a richer set of data to draw upon in the former than in the latter case). Just as we may
Introduction: Getting to Know Our Own Minds 11

be mistaken about the mental states of others, we may be mistaken about our own
mental states. Although they retain the idea of the ‘Real Self’, they admit that we
lack “direct and infallible knowledge of the mental states that we identify with or
integrate in our real self” (this volume). Given that, on their view, self-­knowledge is
based largely upon interpretation, there is considerable conceptual space left open
for confabulation, self-deception, and other errors of self-ascription.
In Chap. 11, Andrew Sims offers a discussion of self-interpretation in the clinical
setting, where third-person self-knowledge is reached both by the analyst and the
patient. By focusing on the psychoanalytic clinical setting, Sims investigates the
dynamics and warrants of inferences in the clinical setting, with a particular attention
to psychoanalytic practice, but with the wider intent of extending the results to other
talking cures which equally use folk-psychological concepts in order to reach their
conclusions and to make patients’ behavior and mentality intelligible. His analysis
focuses on whether, and to what extent, psychoanalytic explanations can be consid-
ered an extension of folk-psychology. In doing so, he offers an interesting picture of
how our ordinary psychological concepts are put in the service of getting to know the
minds of others both in everyday life and in the context of psychological therapy. His
inquiry about how the causal inferences are performed, and on which possibly legiti-
mate basis, is instructive to reflect on the functioning of interpretation when faced
with initially unintelligible behavior and mental states. Even if Sims’s focus is
mainly directed to the clinical setting and the interpretation a clinician gives to the
patient’s mind and behavior, his considerations are relevant to any individual attempt
to make sense of oneself and to get to know one’s own mind through the practice of
individual third-person self-knowledge and self-­interpretation, which is nowadays
largely imbued with the concepts made widely available by psychoanalysis itself.
Collectively, the chapters contained in this volume help to establish a conceptual
framework for important future research on third-person self-knowledge, interpre-
tation, and narrative. It is our hope that this work will help to inspire and promote
additional research in these areas across a broad range of disciplines. Philosophers,
psychologists, cognitive scientists and others cannot provide a comprehensive anal-
ysis of self-knowledge without treating third-person self-knowledge, and the pro-
cesses that contribute to it, as a problem worthy of investigation in its own right.

References

Bilgrami, A. 2006. Self-knowledge and resentment. Boston: Harvard University Press.


Boyle, M. 2015. Critical study: Cassam on self-knowledge for humans. European Journal of
Philosophy 23 (2): 337–348.
Carruthers, P. 2009. How we know our own minds: The relationship between mindreading and
metacognition. Behavioral and Brain Sciences 32: 1–18.
———. 2011. The opacity of mind: An integrative theory of self-knowledge. Oxford: Oxford
University Press.
12 J. Kirsch and P. Pedrini

Cassam, Q. 2014. Self-knowledge for humans. Oxford: Oxford University Press.


Coliva, A. 2012a. In One variety of self-knowledge: Constitutivism as constructivism, ed. A. Coliva
2012b.
———., ed. 2012b. The self and self-knowledge. Oxford: Oxford University Press.
———. 2016. The varieties of self-knowledge. London: Palgrave Macmillan.
Dennett, D. 1987. The intentional stance. Cambridge, MA: MIT Press.
Dennett, D.C. 1991. Consciousness explained. Boston: Little, Brown and Company.
Gazzaniga, M. 1995a. Consciousness and the cerebral hemispheres. In, ed. M.  Gazzaniga, vol.
1995b, 1391–1400.
———., ed. 1995b. The cognitive neurosciences. Cambridge: MIT Press.
Goldie, P. 2012. The mess inside. Narrative, emotion, and the mind. Oxford: Oxford University
Press.
Gopnik, A. 1993. How we know our own minds: The illusion of first-person knowledge of inten-
tionality. Behavioral and Brain Science XVI: 1–14.
Lawlor, K. 2009. Knowing what one wants. Philosophy and Phenomenological Research 79:
47–75.
Moran, R. 2001. Authority and estrangement: An essay on self-knowledge. Princeton: Princeton
University Press.
Schechtman, M. 1996. The constitution of selves. Ithaca: Cornell University Press.
Self-Knowing Interpreters

Annalisa Coliva

Abstract  The paper illustrates and clarifies the distinction between first-personal
and third-personal self-knowledge. It is argued that the characteristic traits of first-­
personal self-knowledge are groundlessness, transparency and authority. It is main-
tained that each of these characteristics is a necessary and a priori aspect of
first-personal self-knowledge.
The thesis that there is first-personal self-knowledge is then defended from chal-
lenges coming from cognitive science, which are taken to show that we may fail to
know dispositional (sometimes causal) elements of our mental states. Yet, this does
not impugn the fact that, regarding the non-dispositional aspects of our mental life,
we can and do have first-personal knowledge of them.
It is then argued that third-personal self-knowledge is achieved through a plural-
ity of methods, which vary from inference to the best explanation, to induction,
simulation, testimony and inferential conceptual deployment. The ensuing self-­
ascriptions are therefore neither groundless, nor transparent or authoritative and are
the result of substantive cognitive achievements.

Keywords  First-personal self-knowledge · Third-personal self-knowledge ·


Constitutivism · Pluralism · Cognitive psychology

In this paper, I first present the characteristic features of what may be regarded as
truly first-personal self-knowledge. Namely, groundlessness, transparency and
authority (§1). The key idea defended in the following is that they are necessary and
a priori aspects of what goes by the name of “(first-personal) self-knowledge”. For
massive failures at self-knowledge would display either the lack of the relevant
psychological concepts or failures at rationality, understood in a “thick” sense,
which is accordingly specified. These characteristic traits of first-personal self-­
knowledge are then defended against possible objections stemming from skepticism

A. Coliva (*)
Department of Philosophy, University of California, Irvine, CA, USA
e-mail: a.coliva@uci.edu

© Springer Nature Switzerland AG 2018 13


P. Pedrini, J. Kirsch (eds.), Third-Person Self-Knowledge, Self-Interpretation,
and Narrative, Contributions To Phenomenology 96,
https://doi.org/10.1007/978-3-319-98646-3_2
14 A. Coliva

regarding self-knowledge deriving from recent findings in cognitive sciences (§2).


It is argued that none of this shows that we never have essentially first-personal self-­
knowledge. Rather, it shows that the scope of first-personal self-knowledge is lim-
ited and does not extend to our deep seated and future dispositions, to the
dispositional elements of our feelings and emotions, and to the causal relations
among our various mental states, which are known, if and when they are, in a third-­
personal way. Various modes of third-personal self-knowledge are then presented
and discussed (§3). Yet, all this is compatible with the fact that we have essentially
first-personal knowledge of a wide range of mental states.

1  First-Person Self-Knowledge

First-personal self-knowledge, which consists in the ability correctly to self-ascribe


mental states, like “I am in pain”, “I see a canary”, “I believe it will rain tomorrow”,
“I am scared of that dog”, is puzzling, from an epistemological point of view. For,
on the one hand, it regards contingent states of affairs. On the other, however, it does
not seem to be obtained by the same means as the ones we employ when we form
knowledge regarding contingent states of affairs concerning the physical world or
other minds. In particular, first-personal self-knowledge seems neither inferential
nor observational. For, of course, we sometimes find out our own mental states by
means of inference starting with the observation of our behvior and further mental
states. However, for the inference to get started at all, we must already have knowl-
edge of those mental states of ours that figure as contents of the premises of our
reasoning. On pain of an infinite regress, there must be knowledge of our own men-
tal states that is not inferential. Somewhere down the line, we must be able to know
our own mental states directly.1
Moreover, quite intuitively, first-personal self-knowledge cannot be a matter of
observation either: first, mental states are just not the kind of thing one could
observe. Secondly, the Cartesian picture of an inner eye, which is supposed to
observe mental states that are luminously presented in the mental arena, is more a
recipe for trouble than a viable explanation of the sense in which self-knowledge
could be observational. Very briefly: it is a recipe for trouble because it would lead
to solipsism—the idea that each of us is caught up in her own world insulated from
anyone else, being unable to know whether others have mental states and are, there-
fore, full-fledged persons. After all—the train of thought would go—how would I
know that other people have mental states at all, if those mental states are intrinsi-
cally private to them and foreclosed to me? It may really become a recipe for catas-
trophe when taken to involve conceiving of our psychological language as private.
Since Wittgenstein (1953), however, private language has become synonymous
with incoherence. The idea is this: any language is a rule-governed practice and it is
essential to it that there be a distinction between correct and incorrect applications

 Cf. Wright (1998), p. 16.


1
Self-Knowing Interpreters 15

of the terms. Still, if the meanings of our psychological vocabulary are constituted
by mental states that are private to each individual, then whatever seems to me the
right application of the psychological term “S” is right. This just means that the
distinction between being right/seeming right, which is much the same as the dis-
tinction between correct/incorrect applications of a word, has vanished. Yet, as we
have just seen, that distinction is essential to there being a (psychological) language
at all. Thus, Cartesianism about self-knowledge may take us to a conception of our
psychological language, which turns it into a non-language. However, we do have a
psychological language—after all we tell each other about our own thoughts and
feelings, sensations and emotions all the time and we seem to understand each other
well. Therefore, it seems safe to conclude that Cartesianism will not do. If
Cartesianism has to go, so has the idea that self-knowledge could be observational.
First-personal self-knowledge is therefore groundless. That is to say, it is based
( 1) neither on observation
(2) nor on inference.
If groundlessness so characterized consists in the idea that (first-personal) self-­
knowledge is not the result of any substantial cognitive achievement, such as observ-
ing or inferring from a symptom to its likely cause, it does not necessarily involve
the idea that one’s psychological self-ascriptions are not based on anything, such as
the very experience one is undergoing when, for instance, one is in pain and avows
it.
That said, it has to be registered that many philosophers think that the on-going
experience, which would allegedly ground one’s psychological self-ascription,
would not play the role of genuine evidence, or of a reason, for one’s self-ascription.
Accordingly, self-ascribing pain, while undergoing a painful sensation, would sim-
ply amount to identifying pain as pain, or even to giving immediate expression to it
through conceptual and/or linguistic means. It would not amount to having an inde-
pendent mental state in view which could, as such, corroborate the truth of one’s
self-ascription. For having that mental state “in view”, in such a way as to ground
one’s self-ascription, would, presumably, be identical, or at least extremely close to
already recognizing it as the mental state it is. Still, recognizing a given mental state
as an instance of pain is just—or is very close to—judging “I am in pain”. The jus-
tification of the relevant self-ascription that should be underwritten by the mental
state would thus be either straightforwardly circular, or based on a mental state too
tightly connected to the self-ascription one would like to justify thereby. These theo-
rists would therefore subscribe to a stronger notion of groundlessness, according to
which mental self-ascriptions, characteristic of first-personal self-knowledge, not
only fail to be the product of observation and inference, but are not based—in any
epistemologically relevant way—on one’s previous awareness of on-going mental
states.
Groundlessnessweak  first-personal self-knowledge is neither observational nor
inferential.
16 A. Coliva

Groundlessnessstrong  first-personal self-knowledge is neither observational


nor inferential, or epistemologically based on one’s previous awareness of one’s
on-going mental states.
For our purposes, it is enough to register these different conceptions and to point
out that, in a minimal sense, groundlessness amounts to the idea that first-personal
self-knowledge is not the result of any substantial cognitive achievement such as
inner perception or inference to the best explanation.
Now, suppose that a subject, endowed with the relevant concepts, and quite pro-
ficient in applying them to herself and others, did so, in her own case, by always
observing her behavior and inferring to its likely cause. She would not know that
she is in pain because she feels it, but because she would see herself scream and
moan, after getting injured, and would infer that she must be in pain. She would not
know that she intends to do some groceries because she is aware of that intention,
but because she would realize that she has gone to the shops and done her shopping.
Such a subject would strike us as someone incapable of having genuine first-­
personal self-knowledge and in some sense as alienated from herself. The word
“alienation” may sound grandiose in this connection, but, minimally, it would mean
that the subject would bear to herself—that is, to her mental states—the kind of rela-
tion she bears to those of others. However, what makes a mental state of ours, ours,
beside its origin in our brain activity, is the fact that it is known to us in a way which
differs from how we know of other people’s mental states, and in which no one else
can know of it, so as to figure among the subject’s motivations for her further
actions (as opposed to something she would know of after those actions). The fail-
ure of groundlessness would therefore result in a lack of “ownership”, or “miness”,
as we might put it, over the mental states one would be inferring to have. Or, again,
in the case of sensations and other mental states with a distinctive phenomenology,
not knowing them directly, but by having to infer to them, starting from the observa-
tion of one’s overt behavior, by application of a little theory, therefore hypothesizing
having them, would cast doubt on the fact that one is capable of enjoying sensations
and other phenomenologically salient mental states at all. Moreover, when inten-
tions (and other propositional attitudes relevant to action) are at stake, a subject who
would know of them only inferentially, from reflection on her behavior and by infer-
ence to its best explanation, besides being “alienated” from her actions, would actu-
ally strike us as not fully responsible for them. For she would know of her motives
only afterwards, after having witnessed herself act in certain ways. Yet, at the time
when the action was accomplished she would not know either what action she was
in fact doing, or why. She would not know her action, because, prior to having inter-
preted it, it would just be a movement or a piece of overt behavior to her. Moreover,
failing to have immediate access to her intentions, she would not know why, for
instance, she happens to be pushing a cart around in a supermarket—as she wit-
nesses herself doing. She would not act but be acted upon and at any instance of the
question “Why did you do/are you doing that?” she would be entitled to respond “I
do not know. Let me find out” and she would then be entitled to preface whatever
answer she might wish to give, based on inference to the best explanation, with
Self-Knowing Interpreters 17

“Probably I did/am doing that because…” or “It is my hypothesis that I did/am


doing this because…”. Now, these responses, though understandable from the point
of view of a subject like the one just described, would actually strike us extremely
odd. They would strike us as reports on another subject, who is the subject of one’s
actions, which would happen to live in one’s body.2 On a “thick” notion of being a
rational agent, who can be considered rationally responsible for (at least many of)
her actions, knowing of them and of their motivations only through inference to the
best explanation and in the way of a hypothesis would actually represent a major
departure from rationality.3 Hence, groundlessness seems to be an element of genu-
ine first-personal self-knowledge, which is in turn a constitutive element of being a
responsible agent.
To claim that groundlessnessweak is constitutive of first-personal self-knowledge
and to redeem that claim by means of a priori considerations having to do, as we
have seen, with the notion of sensation (and possibly other phenomenologically
salient mental states), and with that of being rational agents who can be held respon-
sible for their actions, does not mean to prejudge the issue of how a satisfactory
account of first-personal self-knowledge can go about trying to meet this desidera-
tum. In particular, it is still possible to propose a fully epistemic account of first-­
personal self-knowledge that aims to respect this requirement. It is only by the lights
of non-epistemic theories of first-personal self-knowledge that groundlessness, just
like the other features of first-personal self-knowledge we will examine in the fol-
lowing, cannot be satisfied by epistemic accounts and has to be considered an a
priori feature of what goes by the name of first-personal self-knowledge, which is
not underwritten by any epistemic achievement on a subject’s part, but can only be
redeemed on the basis of a priori reflection on a set of mutually interdependent
concepts. We will not have to address this issue here. Still, it is worth bearing in
mind that defending groundlessness on a priori grounds is compatible with different
accounts of first-personal self-knowledge.
Consider a sensation of cold or the deliberation to cook risotto for dinner. If
someone asked you “Are you feeling cold?” or “What do you intend to cook for
dinner?”, you would be in a position to answer their question immediately. Our
mental states seem to be directly, or transparently known to us. More precisely, their
occurrence is usually of a piece with one’s awareness of them. According to some
theorists (Shoemaker 1996), they are self-intimating. If one has them one would be
ipso facto aware of them. Similarly, the occurrent painful sensation, or fear one
might have is of a piece with one’s awareness of them. Again, one’s seeing a PC in
front of one is of a piece with one’s awareness of it. Surely, however, the occurrence
of states of affairs out there—even the most banal ones—is not of a piece with one’s
awareness of them. The trivial fact that there is some pigeon flapping its wings right
now outside the window, for instance, is not of a piece with my awareness of it. I

2
 Similar considerations are advanced in Shoemaker (1996).
3
 To be rational agents does not mean acting necessarily for good reasons. It means, however, to
have knowledge of the actions one is performing and of their motivations, at least for the most part,
so as to be in a position to be held responsible for them.
18 A. Coliva

will have to look out of the window and recognize the bird as a pigeon and so on.
Nor is the occurrence of other people’s mental states of a piece with one’s awareness
of them. Hence, for example, no matter how good one might be at figuring out what
is crossing someone else’s mind right now, that someone else is feeling bored, per-
plexed or annoyed by what they are reading is not something one can be immedi-
ately aware of. As said, it is something one will have to figure out by taking into
account facial expressions, sighs, bodily movements, by connecting them with gen-
eral knowledge of what those reactions are an expression of and finally by inferring
that they are bored, perplexed or annoyed. Transparency, then, is one of the features
that set first-personal self-knowledge apart from all other kinds of knowledge.
Some theorists are unconvinced that transparency is a characteristic feature of
our own mental states (see Snowdon 2012 for a criticism of transparency with
respect to sensations, emotions, and passing thoughts). They point out, for instance,
that higher-order mammals and infants do not seem to be capable of self-ascribing
sensations and basic emotions, or even perceptions, while they can enjoy them. To
such a worry, one might respond by noticing that we have been talking of transpar-
ency mostly in terms of awareness and it seems safe to hold that at least in the case
of sensations, occurrent emotions with a characteristic on-going phenomenology,
and conscious perceptions, their occurrence seems to coincide with one’s awareness
of them, even if one does not possess the concepts necessary to self-ascribe them.
More precisely:
Transparencyweak  if one has a given mental state M, one is aware of it. That is
to say, the mental state M is phenomenologically salient to the subject.
Even so, it is clear that transparency does not hold unconditionally, but only for
those mental states which have a distinctive phenomenology—leaving aside for
present purposes the width of this class—and are not purely dispositional.
Furthermore, subjects undergoing these mental states will have to be cognitively
lucid, attentive and alert.
However, one might object to this characterization by pointing out that it cap-
tures self-awareness rather than self-knowledge. For the latter, contrary to the for-
mer, consists in the ability correctly to make the relevant psychological
self-ascriptions. Furthermore, one might want to capture the idea that also proposi-
tional attitudes that may lack a distinctive phenomenology can be transparently self-­
known. As we saw at the very beginning, having formed the intention to cook risotto
for dinner, one may be immediately in a position to self-ascribe that mental state,
even if, arguably, there is no distinctive phenomenology to forming intentions. In
such a scenario, the likely move would be to conditionalize transparency to the
obtaining of the relevant (C-)conditions. Namely, subjects should be conceptually
endowed, besides being cognitively lucid, attentive and alert. By means of such a
conditionalization, then, usual counterexamples to transparency, often presented in
the literature on self-knowledge, would easily be dispensed with. The case of ani-
mals and infants would be countered by the requirement that subjects should pos-
sess the relevant concepts. If, in addition, a subject is under the effect of drugs, has
her attention occupied by a pressing task, or is distracted, she may be having
Self-Knowing Interpreters 19

s­ ensations, emotions, perceptions, and yet not be aware of them. Given the condi-
tionalization just proposed, these cases would not represent counter-examples to
transparency.
A separate remark is apposite in relation to the Freudian idea that we may have
many unconscious propositional attitudes, such as beliefs, desires, and intentions, of
which, by definition, we are totally unaware. Obviously, there is no denying that this
might be the case. However, Freudian mental states are mental dispositions. Hence,
one may add to the C-conditions that when propositional attitudes are at stake, they
should be beliefs, desires and intentions as commitments, not as dispositions. I can-
not go into details here (but see Coliva 2015a, 2015b, 2016). The basic distinction
is that propositional attitudes as commitments are intrinsically normative4 and in
fact necessarily self-known, while dispositions are not. Hence, transparency can be
characterized as follows.
Transparencystrong  Given C-Conditions (including concepts’ possession, cog-
nitive well-functioning, alertness and attentiveness and to the exclusion of uncon-
scious and purely dispositional mental states) if one has a given mental state M, one
will be in a position to judge and/or believe that one has it.
Now we should consider whether transparency is a contingent feature of first-­
personal self-knowledge or, rather, a necessary and a priori aspect of it. Let us start
with sensations, perceptions and basic emotions. One may distinguish between sen-
sory states, with no representational content and therefore without correctness con-
ditions, which simply register proximal stimuli, and sensations properly so regarded.
Only the latter are necessarily conscious—that is, such that having them coincides
with being aware of them. That is to say, transparency is, on this reading, a neces-
sary and a priori aspect of what goes by the name of “sensation”. Similar consider-
ations may be put forward in the case of at least those emotions that have a distinctive
phenomenology to them. However, the claim that transparency holds a priori, at
least for some kinds of mental state, would then seem to depend on a terminological
choice. Furthermore, one may want to insist that for instance the soldier who has
been marching for days does feel pain, even if he is not aware of it because his
fatigue and stress are so conspicuous that he is oblivious to them and just keeps
walking. It is at this stage that the C-conditions we have introduced in the character-
ization of transparencystrong are helpful to delimit those cases in which it makes sense
to think that our knowledge of our on-going sensations meets the transparency
requirement. If a subject were always unable, while feeling pain and being attentive
and alert, to judge that she is, this would cast doubt on the fact that she does possess
the relevant concepts; alternatively, if we had reasons to think that she does have
those concepts, her inability to self-ascribe an on-going pain would cast doubt or on
her cognitive well-functioning to the point of impairing the idea that we are dealing
with a rational subject. Furthermore, she would not be in a position to take respon-
sibility for her actions. She would try to cure her wound, for instance. If asked why,

4
 For instance, in the case of belief one cannot believe as a commitment that P while also knowingly
and willingly assenting to its negation or remaining agnostic about P.
20 A. Coliva

however, she would not know, even after screaming and moaning and while having,
ex hypothesi, the relevant psychological concepts.
The case of perceptions is clearly different, for, along  with conscious percep-
tions, whose occurrence would be of a piece with one’s awareness of them, we have
unconscious ones. Blind-sight would be a case in point. Hence, there would be
perceptions whose occurrence would not be of a piece with one’s awareness of
them. To such a challenge, we can respond by noticing that clearly unconscious
perceptions fall out of first-personal self-knowledge. For we would know of them
by observation and inference, by having witnessed ourselves act in ways which can
be made sense of only by presupposing that we did have those perceptions. By con-
trast, if we were dealing with conscious perceptions, then, their occurrence would
be of a piece with one’s awareness of them. Furthermore, if endowed with the rele-
vant concepts, while being cognitively alert and attentive, we would be in a position
immediately to judge (or believe) that we are having them.
It is worth noting that, given the strong link between perception and action, the
actions we would perform based on unconscious perceptions would fall out of
responsible agency. The subject affected by blind-sight does not know either that
she is catching a ball as she does it or why. Hence, she cannot be held responsible
for it. Transparency is therefore a necessary and a priori element only of our knowl-
edge of conscious perceptions, which are, in turn, constitutive of responsible outer
actions.
Similarly, a subject who, having formed an intention or other propositional atti-
tudes as commitments, were systematically unable to self-ascribe them, would
either be lacking the relevant concepts; or else, if there were reasons to think she
possessed them, she would seem to be somehow mentally deranged. For she would
systematically respond “I don’t know” to the question “Do you believe/desire that
P/intend to φ?”, after in fact having asserted (or judged) “P”, “P is worth-doing” and
“I will φ”, while having the relevant psychological concepts.
Once more, claiming that transparency is a constitutive feature of first-personal
self-knowledge and that it can be redeemed on the basis of a priori considerations
concerning a range of interconnected concepts, such as the concept of responsible
agency, rationality and first-personal self-knowledge, does not preclude the possi-
bility of giving an epistemic account of it. It is only by the lights of non-epistemic
accounts of first-personal self-knowledge that this requirement is not underwritten
by any kind of epistemic performance on a subject’s part and is in fact just a concep-
tual truth concerning a set of interlocking concepts. Once again, we do not have to
address this issue here, but it is worth keeping in mind that defending transparency
on a priori grounds is compatible with different accounts of first-personal
self-knowledge.
Let us now consider authority. If you are sincere and competent with respect to
the concepts you use to express your mental states, nobody can—rationally—cast
any doubt on your avowals. If you answer my question “What are you thinking?” by
saying “I think that summers in Italy are really too hot” and you are sincere and
know how to use ‘I’, ‘think’, ‘summers’, ‘Italy’, …, then nobody could challenge
you by saying “Are you sure that this is what you are thinking?”. Of course, one can
Self-Knowing Interpreters 21

challenge the subject’s grounds for believing that summers in Italy are too hot, by
pointing out that it is August 15, you are in Italy and there are only 15 °C. Yet, one
cannot challenge the subject’s own believing that she believes it. By challenging the
grounds in favor of that content, one can lead a subject to revise her belief, but this
possibility depends on the (implicit) acknowledgement that the subject does
believe—erroneously, from your point of view—that summers in Italy are too hot
and that she knows that much about herself. Similarly, if I sincerely avow a painful
sensation, it would be pointless for someone to challenge me by saying “Are you
sure that is what you are feeling?”. To the restatement of my avowal, I could only
add that I am being sincere and that I can speak English. Therefore, subjects are
authoritative with respect to their own mental states: if they say (or judge) that they
have a certain mental state M, then they have it.
Surely, however, if someone asked you “What is the weather like?” or “Is Elly
sad since her sudden loss?”, from your sincere and conceptually competent answers
it wouldn’t follow at all that it would be inappropriate for someone to challenge
your claims. Therefore, authority is another feature that sets self-knowledge apart
from all other kinds of knowledge.
Again, some theorists are wary of authority (see again Snowdon 2012). With
respect to sensations, for instance, they point out that one can envisage a situation
where a subject is asked to locate her sensation and is uncertain about that. This,
however, would merely show that we are not authoritative with respect to the bodily
location of our on-going sensations. Yet, it would not show that we are not authorita-
tive with respect to the fact that we are undergoing them.
Another case often discussed in the literature is the one of a subject who, by
dreading the pain a dentist will produce by messing about her mouth, claims to be
feeling pain already when the dentist has not even touched her. This case can be
accommodated by specifying the C-conditions under which authority is supposed to
hold. In fact, the subject should be cognitively lucid. Arguably, however, intense
fear can impair one’s cognitive functions. Once again, the possibility, on specific
occasions, of mistaken self-ascriptions of sensations, should not be taken as a coun-
terexample to authority, but only as imposing the need of properly characterising its
extent.
Another prima facie difficult case to handle is the one of confused or vague sen-
sations, such as itches that border pain. A subject could then be unclear whether she
is undergoing one or the other kind of sensation and make avowals she would then
like to retract. This, however, would not be a counterexample to authority, because
we are in fact dealing with a confused or vague sensation, which escapes definite
categorization. Hence, we vacillate between two concepts we do possess, when both
of them seem to apply to it to some extent. In this sense, our self-ascriptions may be
imprecise and retractable. Yet, this does not show that, at least when sensations are
not of a confused nature, we would not be authoritative with respect to them. More
generally, the ubiquitous presence of vagueness in our concepts should not be
taken—by itself—as a serious challenge to the possibility of first-personal self-­
knowledge. No more than the ubiquitous presence of vagueness in the domain of
22 A. Coliva

colour is normally taken as a challenge to the possibility of correct colour percep-


tions and judgments.
Another possibility is that one might be wrong about the content of one’s percep-
tual appearances. For instance, one can say that one’s after image contains five red
dots, when in fact it contains only four. However, authority can be maintained even
in this case because determining the number of dots involves counting and one may
go astray in doing it. Still, one would be authoritative with respect to the fact of hav-
ing an after image.
Interestingly, Eric Schwitzgebel (2008) has recently argued that we are not
authoritative with respect to our inner feelings and emotions. The idea is that we
may be bad at judging that the feelings we are currently experiencing are feelings of
anger, say, when in fact this would be clear to a third party. It should be admitted
that sometimes we do not realize what the feelings we may experience are symp-
toms of. This, however, could be due to poor performance in connecting one’s on-­
going feelings to the dispositional elements of the relevant emotion. Hence, it would
be a case of mistaken self-interpretation. That is to say, it would be a case of lacking
or of going astray in one’s third-personal knowledge of one’s own mental states,
rather than a case of failure at first-personal knowledge of them.
Finally, moving on to propositional attitudes, it may happen that one says that P,
or that P would be good to have, thus manifesting the belief or the desire that P, and
yet behaves in ways that run contrary to one’s professed beliefs and desires. This
phenomenon, known as self-deception, is taken to be the most powerful counterex-
ample to authority. There are three possible answers on behalf of the supporters of
authority. First, one may complicate the C-conditions to exclude cases of self-­
deception from the range of propositional attitudes one is authoritative about. This
move, however, seems quite ad hoc, hence it would be preferable to avoid it. Second,
one may propose an alternative account of self-deception, which is compatible with
the retention of authority over one’s propositional attitudes (see Bilgrami 2006 and
Coliva 2016). Finally, one may hold that authority can admit of exceptions, yet
argue that it is present in a significant amount of self-ascriptions about one’s own
intentional mental states.
Hence, we face the possibility of two different readings of authority, which differ
in strength. A minimal one has it that we are mostly—though not always—authori-
tative with respect to a specified class of mental states we can enjoy, and, in particu-
lar, that self-deception may lead us to wrong psychological self-ascriptions. A
stronger reading, in contrast, has it that, as a matter of fact, there are no relevant
exceptions to authority, once we have appropriately specified the class of mental
states for which it holds and have independently accounted for self-deception in
ways which explain away the impression that it should pose a challenge to authority.
Accordingly, we would have the following two possible readings of authority:
Authorityweak  Given C-Conditions (including concepts’ possession, cognitive
well-functioning, alertness and attentiveness) if one judges to have a mental state M
(save for dispositional ones or for the dispositional elements of some mental states),
one will usually have it.
Self-Knowing Interpreters 23

Authoritystrong  Given C-Conditions (including concepts’ possession, cognitive


well-functioning, alertness and attentiveness) if one judges to have a mental state M
(save for dispositional ones or for the dispositional elements of some mental states)
one will always have it.
Be that as it may, authority too would seem to be a necessary and a priori feature
of first-personal self-knowledge. For, if one were systematically proved wrong in
one’s psychological self-ascriptions, doubt would be cast upon one’s possession of
the relevant concepts. Alternatively, if one wished to maintain that a constantly mis-
taken subject (with respect to her own mental states) could still be said to have the
relevant psychological concepts, perhaps because she retains the ability to apply
them to other people’s mental states, her persistent self-deception would impair her
rationality. For she would avow certain mental states, while she would systemati-
cally behave in ways which run contrary to them. Hence, we could no longer make
sense of her linguistic and non-linguistic behavior.
Once again, to claim that authority is constitutive of first-personal self-­knowledge
and that it is a priori connected with other notions, such as the one of being a ratio-
nal agent who can be held responsible for her actions, does not preclude the possi-
bility of accounting for it in epistemic terms. It is only in the perspective of
non-epistemic theorists that those attempts fail and that authority is not underwrit-
ten by any epistemic achievement on a subject’s part but is purely a fall-out of the
necessary and a priori links holding between some interconnected concepts.
Groundlessness, transparency and authority, therefore, are constitutive features
of first-personal self-knowledge, and are constitutively tied to what it means to pos-
sess psychological concepts, as well as to what it means for someone to be a rational
subject, who can be held responsible for her own actions.5

2  Counterexamples from Cognitive Science?

Recent studies in cognitive psychology tend to challenge the idea that we are
authoritative with respect to our own mental states. In what follows, I will draw
extensively on Brie Gertler’s (2011) excellent discussion of the topic.6 First, there
are studies that impugn the idea that we have privileged access to our character
traits, such as jealousy or loyalty. People tend to deny being jealous and to affirm to
be loyal, even if they actually behave in ways which run contrary to their psycho-
logical self-ascriptions. Second, it has been shown that we are bad at “affective
forecasting”.7 That is to say, we predict that a life-changing event, such as winning

5
 Whether, in turn, they are also constitutive elements of what it means to be a self or a subject at
all, insofar as they are constitutive of being critical reasoners and subjects of moral norms as Burge
(2011) maintains, is a further issue, which would need a separate treatment.
6
 See Gertler (2011), pp. 70–86.
7
 Gilbert (2006).
24 A. Coliva

the lottery, will permanently affect our psychological attitude towards life. Studies
have shown, however, that after a short period in which one’s moods are elated,
subjects revert to their “happiness baseline”. In both cases, we may have privileged
access to our resolutions or intentions and expectations, such as the intention to be
loyal, or the expectation that winning the lottery will make us permanently happy,
and yet we are not authoritative with respect to whether we will actually behave in
the way we intend or expect we will behave. Gertler rightly notices that both char-
acter traits and affective forecasting involve access to our own psychological dispo-
sitions. Dispositions involve counterfactual conditions and it may well be the case
that we are bad at determining whether we will satisfy them or not. This clearly
limits the scope of first-personal knowledge of our own mental states, but does not
rule it out.
Other studies show that we are bad not only at determining our psychological
dispositions, but also at indicating the causes or motivations of our actions.8 In fact,
they often remain unconscious, by operating at the subpersonal level. Hence, we do
have first-personal knowledge of certain occurrent thoughts, such as one’s desire to
have ice cream, which we take to be the cause of our action of going towards the
fridge to have some, but in fact that thought may not be the real cause of our action.
Again, Gertler rightly notices that, since Hume, causes have been shown not to be
directly observable and rather be the result of theorizing about established correla-
tions between different types of events. Therefore, it should not be surprising that
we lack authoritative access to the causes of our actions. Still, we are under the illu-
sion of having a distinctively first-personal access to them because we do have such
privileged access to the thoughts that we take to be the causes of our actions. Once
more, the scope of first-personal knowledge of our own mental states appears to be
limited, yet unassailable.
Similar results have been achieved by studies concerning our moods, such as
pessimism, like the tendency to regard the future as bleak, and emotions, such as
anger. Gertler correctly points out that moods are dispositional mental states and
that emotions involve, besides a certain feeling, with respect to which we do have
first-personal access, certain dispositional elements. This idea chimes well with the
borderline conception of the emotions defended in Coliva (2016, ch. 2), according
to which emotions cannot be reduced to either feelings or dispositions. It may well
be, then, that while we have privileged access to their phenomenological aspects,
we do not have any to their dispositional elements.
Finally, if we do not have first-personal knowledge of dispositional mental states,
that entails that we are not authoritative, nor do we have immediate and groundless
access to a lot of propositional attitudes, such as beliefs and desires, which are dis-
positional. Indeed, this is only to be expected. Yet we do also have the impression of
having first-personal access to our beliefs, desires and intentions. The distinction
between propositional attitudes as dispositions and as commitments comes in handy
at this stage (see also Coliva 2015a, 2016, ch. 2). For we may say that while we lack

8
 Nisbett and Wilson (1977); Libet (1985); Wegner (2002); Wilson (2002); Wegner and Wheatley
(1999).
Self-Knowing Interpreters 25

first-personal knowledge of the former, we have it of the latter. Yet, while this
imposes a further limitation to first-personal self-knowledge, it does not show that
we lack it altogether. Rather, it merely shows that first-personal self-knowledge is
limited to occurrent sensations, including those characteristic of at least basic emo-
tions, to perceptions, current thoughts and propositional attitudes as commitments.

3  Third-Person Self-Knowledge

So far, we have examined the characteristic aspects of first-personal self-­knowledge.


That is to say, those aspects, which set our knowledge of our own mental states apart
from other kinds of knowledge, such as knowledge of truths about the external
world, usually achieved through perception, or about other people’s mental states,
customarily obtained through observation and inference to the best explanation,
starting from the observation of their overt behavior. The various limitations to first-­
personal self-knowledge we considered, however, show that ample room for error is
provided in figuring out our own mental states. This can be explained by noticing
that we often acquire knowledge of our mental states in a third-personal way, even
in our own case, as a result of observation of our behavior and of inference to the
best explanation; or else, thanks to inference to the best explanation starting from
one’s first-personal knowledge of one’s on-going sensations, feelings, perceptions
and propositional attitudes. We therefore possess not only first-personal but also
third-personal self-knowledge. In general, that is, we are often self-knowing inter-
preters, as we may put it.
When third-personal self-knowledge is at stake, however, none of the character-
istic features of first-personal self-knowledge is present. There is no groundless-
ness, for indeed our knowledge of our own mental states is either based on observing
our behavior and inferring to its likely cause, or it is based on our awareness of our
inner phenomenology and to inferring to its probable explanation. Two avenues for
error are open, though. For we may mischaracterize our own behavior and hence
start out our inferences with mistaken premises. Or else, we may be wrong in iden-
tifying a certain mental state as the likely cause of our on-going feelings or overt
behavior. Thus, authority does not hold either. Nor does transparency, as the occur-
rence of one’s mental states is characteristically not of a piece with one’s awareness
of them or indeed with being in a position to self-ascribe the relevant mental state.
Indeed, to gain knowledge of them, we need to engage in observation and inference
and we may fail to perform them, thus remaining blind to ourselves, as it were.
Still, there is an element of third-personal self-knowledge that sets it apart from
knowledge of other people’s mental states. Namely, sometimes the inference can
start on the basis of other mental states one has knowledge of in a first-personal way,
such as one’s sensations, immediate feelings, propositional attitudes as commit-
ments, etc.9 This obviously cannot be the case when we acquire knowledge of other

 I was pleased to find a similar claim in Cassam (2014), reviewed in Coliva (2015b).
9
26 A. Coliva

people’s mental states through inference to the best explanation. In that case, the
inference starts from observing their overt linguistic and non-linguistic behavior.
Let us therefore consider some examples of third-personal self-knowledge.
Cases of inferences starting from the observation of one’s own behavior can be
those that lead us to self-ascriptions of propositional attitudes as dispositions,
whether of a Freudian kind or not. Consider a mother who reflects on her behavior
towards her son and compares it with the one towards her daughter. She can judge
that she is more often benevolent towards the latter and therefore infer that she pre-
fers her daughter to her son. Of course, she may have gone wrong in judging that
she is more munificent with her daughter in the first place, as she may under-­estimate
the number of times in which she has been open-handed towards her son. Hence, her
self-ascription is neither groundless, nor transparent or authoritative.
Sometimes we figure out our complex emotions, such as love towards someone,
by reflecting on our behavior and inner phenomenology in their presence or at the
prospect of some event involving them. Obviously, we can go wrong in our self-­
interpretation. Our behavior and feelings may not be symptoms of love but of being
possessive and thus dreading the prospect that a good friend should inevitably stop
being so close to us as the result of getting married to another person. Alternatively,
we can mistake our behavior and feelings as symptoms of love, when they are in fact
signs of our vanity and complacent attitude towards those people who are usually
very kind to us. Sometimes, in contrast, we are self-blind in the sense that despite
showing many of the characteristic symptoms of love, say, towards someone, we do
not realize that we are in love with them. Such blindness can have different causes,
both cultural and psychological.
Sometimes, we gain knowledge of our own minds, and in particular of our char-
acter traits, by means of testimony. That is to say, we form the belief “I am F”,
where “F” is a psychological concept, because someone tells us we are F. This kind
of third-personal self-knowledge, therefore, has an essentially social aspect to it and
exceeds by far the case of the highly institutionalized communication that takes
place between a patient and a psychological therapist. More specifically, it involves
taking the other person at her word, in the absence of reasons to doubt of her com-
petence in the application of the relevant psychological concepts and of her sincer-
ity; or else, in light of positive reasons in favour of her competence and sincerity.10
A conversation with a friend, or a relative we have no reason to distrust, or that we
have positive reasons to trust and consider competent both in her mastery of the
relevant psychological concept and in its application in our own case, can go a long
way in giving us knowledge of our own minds. Still, we may be led astray and there-
fore err in our eventual psychological self-ascription.
Furthermore, there are cases in which we gain knowledge of our minds by means
of inductive evidence. For instance, by knowing how we have reacted to certain situ-
ations in the past, we can inductively predict we will feel thus-and-so in the future,
if the same conditions should obtain. Once more, this method may, on occasion,

 It depends on which theory of testimonial justification and knowledge one adopts, whether
10

Humean in kind or Reidian.


Self-Knowing Interpreters 27

lead us astray, or we can fail to draw the relevant conclusions even if we have previ-
ous evidence which should lead us to form the relevant belief. Thus, whenever self-­
knowledge is based on induction, groundlessness obviously fails, and so do
transparency and authority.
We have also seen how alleged counter-examples to first-personal self-­knowledge
drawing on recent findings in cognitive studies can in fact be interpreted as interest-
ing examples of third-personal knowledge of our own mental states. Therefore, for
instance, we think we have certain character traits, or that we will have certain
emotional responses in the light of alleged life-changing events, which we actually
lack. The interesting aspect of this failure at self-knowledge is that the reason why
we think we would have these mental dispositions is that we presumably engage in
imagining how we would behave on a given situation. For whatever reason, which
may or may not be biased, we think we would behave in ways that would depend on
having a certain mental state. Yet, when faced with a real life situation, we do not
behave as we imagined we would do. The kind of procedure followed in these cases,
in order to try to gain knowledge of our mental dispositions, is indeed similar to the
one we would apply if we were to figure out someone else’s mental states by means
of simulation and its correct outcome is obviously not guaranteed. It is therefore
another example of third-personal self-knowledge.
Another intermediate case, which does not immediately coincide with knowl-
edge of our own mental states either through inference to the best explanation or
through mental simulation, is the one in which we suddenly notice an aspect of
ourselves we had been previously blind or oblivious to; or indeed reconceptualize
an aspect of ourselves by suddenly seeing it in a different way. We may all of a sud-
den realize that we are not nice to other people—we do not actually care about them
that much—but are only polite; or we may suddenly see a character trait of ours not
as a symptom of strength but of weakness. Seeing-as, switches of aspect and notic-
ing aspects are complex phenomena even when they clearly concern perceivable
objects and properties and involve the operation of perceptual faculties.11 Obviously,
in the case of our own mental states, these notions find only mediated or indirect
employment, for no real perceptual faculty is involved and yet it does not seem that
we engage in any complex inferential thinking, or in mental simulation. Furthermore,
the relevant psychological concepts are necessarily needed, while, arguably,12
purely perceptual switches of aspects do not necessarily require concepts (although
they may and often do involve them). We may say that we suddenly take a pattern
of overt behavior, and possibly other inner aspects of our psychology, as instantiat-
ing this or that psychological property either because that pattern is manifest to us
for the first time, or because the relevant concept is available to us at last and a pat-
tern of behavior and inner elements we had possibly noticed before seems immedi-
ately to fit the newly acquired concept. Sometimes these new conceptualizations or
reconceptualizations contradict previous ones and that gives rise to a switch of
aspect regarding ourselves and our personalities. Still, in these cases, inference is

11
 Cf. Wittgenstein (1953, II, xi).
12
 See Coliva (2012) for a discussion of seeing and seeing-as and the role of concepts in it.
28 A. Coliva

involved at least in the following sense. The characteristic notes of the newly
acquired (or applied) concept guide the process of subsuming the observed pattern
of behavior and inner elements under the concept. That may happen very rapidly,
and so, phenomenologically, we may not even seem to be engaging in inferences,
yet epistemically that is what grounds our judgment “I am F” (where F is a psycho-
logical concept). It is one possible case of self-interpretation and it can go wrong
because, after all, the pattern observed may not fit the concept or we may have
ignored other features, which would, if taken into account, lead to a different psy-
chological self-ascription. Furthermore, self-interpretation can be wrong for moti-
vated reasons, often unconscious in nature, and this can explain at least some forms
of self-deception.
What is important to notice is that third-personal self-knowledge exhibits all the
typical aspects of any substantive cognitive achievement, which usually results in
knowledge. We base our psychological self-ascriptions on various kinds of evi-
dence, we engage in inferences and we deploy imagination to simulate a situation
and predict our own actions, therefore determining their psychological motivations.
All these procedures admit for the possibility of error and for not being deployed at
all, thus resulting in a form of self-blindness. Furthermore, the kind of abilities
called upon can be variously distributed among subjects, who are otherwise equal
with respect to cognitive functions, conceptual endowment and rationality.
Moreover, it can be finessed with experience, including exposure to literature, mov-
ies and drama, which very often affect us precisely because we identify with the
characters and thereby get to know something about ourselves. Study too can aug-
ment our third-personal self-knowledge  – especially when new concepts are
acquired and make salient to us aspects of our minds we were previously blind to.
We can also improve over time by exercising these judgmental abilities and by
learning to pay heed to relevant feelings and/or behavioural patterns and by becom-
ing more and more proficient in the application of the relevant (dispositional) psy-
chological concepts. Third-personal self-knowledge, therefore, is not any
different – epistemically speaking – from knowledge of truths about physical objects
around us, or about other people’s mental states in these respects. Yet, it is special in
that the kind of evidence it may derive from may indeed be uniquely available to us.
For, as we have seen, it would depend on our first-personal knowledge of occurrent
sensations, feelings, perceptual experiences and propositional attitudes as commit-
ments. Yet, it is only for third-personal self-knowledge that the usual traits of knowl-
edge are clearly present. For there is cognitive achievement, there is ample possibility
of error, or self-blindness, without resulting in failures at rationality or ipso facto in
lack of concepts’ possession, and there is room for expertise and for improving
one’s skills. None of this, in contrast, seems to characterize first-personal self-­
knowledge. For some theorists, this means that first-personal self-knowledge is not,
after all, real knowledge. Rather, it is a set of constitutive claims, which can be vari-
ously grounded and redeemed. I have explored the issue at length in Coliva (2016).
For our purposes here, there is no need to take a stance on that.
What matters, rather, is to recognize that a powerful case can be made in favour
of a pluralist conception of self-knowledge, which, as we have seen, comprises both
Self-Knowing Interpreters 29

first- and third-personal elements to it; and, in turn, for the view that third-personal
self-knowledge is gained through a plurality of methods, which should be thor-
oughly investigated.

References

Bilgrami, A. 2006. Self-knowledge and resentment. Cambridge: Harvard University Press.


Burge, T. 2011. Self and self-understanding. The Dewey lectures (2007–2011). The Journal of
Philosophy CVIII/6-7. Reprinted in Cognition through understanding (pp. 140–228). Oxford:
Oxford University Press.
Cassam, Q. 2014. Self-knowledge for humans. Oxford: Oxford University Press.
Coliva, A. 2012. Human diagrammatic reasoning and seeing-as. Synthese 186 (1): 121–148.
———. 2015a. How to commit Moore’s paradox. The Journal of Philosophy 112 (4): 169–192.
———. 2015b. Review of Cassam’s self-knowledge for humans. Analysis 1 (1): 1–6.
———. 2016. The varieties of self-knowledge. London: Palgrave Macmillan.
Gertler, B. 2011. Self-knowledge. London/New York: Routledge.
Gilbert, D. 2006. Stumbling on happiness. New York: Knopf.
Libet, B. 1985. Unconscious cerebral initiative and the role of conscious will in voluntary action.
Behavioral and Brain Sciences 8: 529–566.
Nisbett, R., and T. Wilson. 1977. Telling more than we can know: Verbal reports on mental pro-
cesses. Psychological Review 84: 231–259.
Schwitzgebel, E. 2008. The unreliability of naïve introspection. Philosophical Review 117:
245–273.
Shoemaker, S. 1996. The first person perspective and other essays. Cambridge: Cambridge
University Press.
Snowdon, P. 2012. How to think about phenomenal self-knowledge. In The self and self-­knowledge,
ed. A. Coliva, 243–262. Oxford: Oxford University Press.
Wegner, D. 2002. The illusion of conscious will. Cambridge: MIT Press.
Wegner, D., and T. Wheatley. 1999. Apparent mental causation: Sources of the experience of will.
American Psychologist 54: 480–492.
Wilson, T. 2002. Strangers to ourselves. Discovering the adaptive unconscious. Cambridge: MIT
Press.
Wittgenstein, L. 1953. Philosophical investigations. Oxford: Blackwell.
Wright, C. 1998. Self-knowledge. The Wittgensteinian legacy. In Knowing our own minds, ed.
C. Wright, B. Smith, and C. Macdonald, 15–45. Oxford: Clarendon Press.
Extended Self-Knowledge

J. Adam Carter and Duncan Pritchard

Abstract  We aim to move the externalism and self-knowledge debate forward by


exploring two novel sceptical challenges to the prospects of self-knowledge of a
paradigmatic sort, both of which result from ways in which our thought content,
cognitive processes and cognitive successes depend crucially on our external envi-
ronments. In particular, it is shown how arguments from extended cognition (e.g.,
Clark A, Chalmers D.  Analysis 58:7–19 (1998); Clark A.  Supersizing the mind:
Embodiment, action, and cognitive extension. Oxford: Oxford University Press
(2008)) and situationism (e.g., Alfano M. The Philosophical Quarterly 62:223–249
(2012), Alfano M. Expanding the situationist challenge to reliabilism about infer-
ence. In Fairweather A (ed) Virtue epistemology naturalized, Springer, Dordrecht,
pp 103–122 (2014); Doris JM. Noûs 32:504–530 (1998), Doris JM. Lack of charac-
ter: Personality and moral behavior. Cambridge University Press, Cambridge
(2002); Harman G.  Proceedings of the Aristotelian Society. 99:315–331 (1999),
Harman G. Proceedings of the Aristotelian Society 100:223–226 (2000)) pose hith-
erto unexplored challenges to the prospects of self-knowledge as it is traditionally
conceived. It is shown, however, that, suitably understood, these apparent chal-
lenges in fact only demonstrate two ways in which our cognitive lives can be depen-
dent on our environment. As such, rather than undermining our prospects for
attaining self-knowledge, they instead illustrate how self-knowledge can be extended
and expanded.

Keywords  Self-knowledge · Extended cognition · Epistemic situationism ·


Content externalism · Memory

J. Adam Carter (*)
University of Glasgow, Glasgow, UK
e-mail: adam.Carter@glasgow.ac.uk
D. Pritchard
University of California, Irvine, CA, USA
University of Edinburgh, Edinburgh, UK
e-mail: dhpritch@uci.edu

© Springer Nature Switzerland AG 2018 31


P. Pedrini, J. Kirsch (eds.), Third-Person Self-Knowledge, Self-Interpretation,
and Narrative, Contributions To Phenomenology 96,
https://doi.org/10.1007/978-3-319-98646-3_3
32 J. Adam Carter and D. Pritchard

1  Introduction

Although much human knowledge is acquired by looking ‘outward’ to the world,


one especially important variety of knowledge, self-knowledge, is characteristically
attained by looking instead inward at oneself—i.e., toward one’s own mental states
(e.g., beliefs, desires), with the objective of representing these states aright. The
most famous challenge to the very possibility of self-knowledge adverts to a thesis
that is now orthodoxy in the philosophy of mind and language—viz., content exter-
nalism (e.g., Burge 1986; Putnam 1975). This is, broadly, the view that one’s physi-
cal or social environment can play a role in the individuation of mental contents, and
hence can play a role in determining what our thoughts are in fact thoughts about.
There is a clear tension between content externalism and self-knowledge, in that
if the content of one’s thoughts can be in part determined by environmental factors,
then how, in the relevant cases at least, can one come to know about one’s thoughts
by directing one’s attention inwards (and hence away from environmental factors)?
A great deal of recent philosophical discussions of self-knowledge have focused on
this tension, which is in effect a tension between, on the one hand, the privileged
‘first-person’ perspective that seems to be a hallmark of the kind of self-knowledge
that we claim of our own mental lives, and, on the other hand, the ‘third-person’
perspective which would, given content externalism, seem necessary to ascertain
with accuracy what, in the relevant cases, our thoughts are about.
Rather than retrace the contours of this well-worn ‘externalist’ objection to self-­
knowledge, we want to move the self-knowledge debate forward by exploring two
novel sceptical challenges to the prospects of self-knowledge—challenges which
exploit two entirely different ways in which knowledge seems to depend on ele-
ments of the world beyond the (biological) agent. One such challenge arises from
what has been called in the philosophy of mind the extended cognition thesis (e.g.,
Clark 2008; Clark and Chalmers 1998). This is the thesis that cognitive processes
can supervene on parts of the world that are external to the biological agent. The
other challenge—known as situationism (e.g., Doris 1998, 2002; Harman 1999,
2000)—has its roots in recent work in moral psychology. It insists that our cognitive
successes depend to a surprising extent on various extra-agential factors specific to
particular situations.
After setting out how these two challenges look initially very troublesome for the
prospects of self-knowledge, we argue that in fact the opposite is the case. Rather
than diminishing our self-knowledge, these two strands of argument in fact reveal
thus far unexplored ways that this kind of knowledge can be realised. In particular,
these two challenges highlight the importance of what we call extended knowledge,
which is knowledge that depends in significant ways on an epistemic contribution
from factors outwith the (biological) subject, but which is no less compatible, as a
result, with self-knowledge.
Here is the plan. In §2, we outline the challenge posed to self-knowledge on the
basis of extended (or ‘active’) cognition and articulate how this challenge differs
importantly from the more familiar challenge posed by content (or ‘passive’)
Extended Self-Knowledge 33

e­ xternalism. In §3, we highlight the key contours of the situationist critique and
outline how a recent attempt (e.g., Alfano 2012, 2014) to argue from situationism to
the falsity of virtue epistemology can be straightforwardly redeployed so as to
threaten the scope of self-knowledge. In §4 we show how, on closer inspection, both
of these novel challenges to self-knowledge fail to generate the sceptical conclu-
sions for the prospects of self-knowledge which they appear to threaten. Instead,
both challenges ultimately reveal hitherto unforeseen ways in which self-knowledge
can be extended and expanded. Finally, in §5, we offer some concluding remarks.

2  The Challenge from Extended Cognition

The sceptical challenge to self-knowledge posed by content externalism has puz-


zled epistemologists and philosophers of language, given that the following two
independently plausible claims seem prima facie incompatible: (i) that the content
of our thoughts is determined partly by the environment; and (ii) that a subject can
know by reflection alone (i.e., from the armchair) the content of her own thoughts.1
While (i) has been widely endorsed in mainstream philosophy of mind and lan-
guage since at least the 1980s, widespread endorsement for (ii) goes back much
further, at least to Descartes, who famously noted important disanalogies between
our knowledge of our own mental states, and our knowledge of other things in the
world—viz., the things which mental states concerning the world, such as percep-
tual beliefs, aim to represent. There is no general consensus regarding which dis-
analogies are the most important in characterising self-knowledge. However, three
important disanalogies which will be of interest to us in what follows involve direct-
ness, authority and method.2
First, regarding directness: we generally don’t regard ourselves as becoming
aware of our own mental states by being aware of something else first.3 In this way,
it looks as though our mental states are thus distinctively direct in a way that knowl-
edge of the world characteristically isn’t. Second, regarding method: we gain knowl-
edge of our own mental states by introspection, by looking inward rather than
outward.4 Introspection seems idiosyncratic of self-knowledge because it is not the
kind of method that we can use to gain knowledge of external things.5 Third,

1
 For a comprehensive overview, see Parent (2013). Some philosophers who have found (i) and (ii)
at tension take the incompatibility at face value, though a further popular way to defend the
(alleged) incompatibility is via appeal to slow-switching arguments (e.g., Boghossian 1989; Burge
1988; Ludlow 1995, 1997).
2
 There are a range of other claimed disanalogies. For discussion, see Gertler (2015).
3
 Gertler (2015, §3.1) notes that such awareness is also taken to be metaphysically and not just
epistemically direct ‘in that no event or process mediates between my awareness and the mental
state itself’.
4
 Though, this is disputed by proponents of transparency accounts of self-knowledge (e.g., Dretske
1994).
5
 See, however, Carter and Palermos (2015) for critical discussion on this point.
34 J. Adam Carter and D. Pritchard

­regarding authority: we generally should treat subjects as authoritative about their


own states. If you tell someone you believe something, it’s normal practice for that
individual to take you at your word, rather than to challenge you.
While these disanalogies have been deployed in the service of demonstrating
tensions between content externalism and self-knowledge, they can as well be put
to work in the service of demonstrating how a more radical form of externalism
poses threats to self-knowledge.
According to active externalism, parts of the world can do much more than indi-
viduate mental contents. Parts of the world—for example, a notebook, an iPhone,
Google Glass, etc., and sometimes even other individuals—can partly constitute
cognition.6 Active externalism is a bold thesis, though it is widely held in contem-
porary cognitive science, and it is quickly gaining traction in contemporary philoso-
phy of mind and cognitive science.
A commonly discussed variety of active externalism is the hypothesis of extended
cognition, according to which cognitive processes are realizable by brain, body and
world. To bring this idea into focus, consider the following widely referenced case-­
pair from Clark and Chalmers (1998):
Inga: Inga has a normally functioning biological memory. When she learns new informa-
tion, she stores it in her biological memory (as is typical and familiar) and it is her biologi-
cal memory which she consults when she needs old information.
Otto: Otto suffers from Alzheimer’s disease, and like many Alzheimer’s patients, he relies
on information in the environment to help structure his life. Otto carries a notebook around
with him everywhere he goes. When he learns new information, he writes it down. When he
needs some old information, he looks it up. For Otto, his notebook plays the role usually
played by a biological memory.

The salient similarity between Inga and Otto is functional; they are both relying
on something (a biological brain and a notebook, respectively) to play the functional
role of information storage and retrieval. A salient dissimilarity is of course that
what’s playing this functional role for Otto is something in the world which he is
physically interacting with.
One very natural response to what’s going on in this case would be, of course, to
grant that Otto is using his notebook in a way that is in fact very similar (perhaps
even functionally isomorphic) to the way in which Inga is relying on her biological
memory, but then to not draw any further conclusions. On this conservative diagno-
sis of the case, whatever functional similarities there are between Otto and Inga,
they don’t motivate any interesting metaphysical conclusions—viz., we should sim-
ply hold firm that whilst Inga’s biological memory is part of the cognitive process
she employs, Otto’s (non-biological) notebook can’t be part of his cognitive ­process.

6
 The three central varieties of active externalism are the extended cognition, extended mind, and
distributive cognition theses. For a recent survey of the differences between these versions of
active externalism, and how they interface with more familiar forms of externalism in epistemol-
ogy and the philosophy of mind, see Carter et al. (2014).
Extended Self-Knowledge 35

It is, after all, something in the world external to him, made of paper, not brain
matter.
Clark responds that this kind of conservative diagnosis reveals a kind of ‘bioprej-
udice’—i.e., an unprincipled inclination to regard the physical constitution of some-
thing as inherently interesting with respect to whether the thing in question is a
component part of a cognitive process. As Clark and Chalmers (1998) have famously
argued, our judgments about what counts as part of a cognitive process should be
guided not by traditional bioprejudice but rather by what they call the parity
principle:
Parity Principle: If, as we confront some task, a part of the world functions as a process
which, were it to go on in the head, we would have no hesitation in accepting as part of the
cognitive process, then that part of the world is part of the cognitive process. (Clark and
Chalmers 1998, 8)

If Clark and Chalmers are right about this, then notice that, with reference to the
parity principle, we should include Otto’s notebook as part of his cognitive process
insofar as we count Inga’s biological brain as part of hers.7 Put more forcefully:
what follows is that Otto’s memory (i.e., his process of memory storage and
retrieval) is located partly outside his head. And, furthermore, proponents of
extended cognition typically reason that, to the extent that we attribute in the default
case certain dispositional beliefs to Inga in virtue of having certain contents stored
in her biological memory, so we should attribute dispositional beliefs to Otto in
virtue of having certain contents stored in his notebook.
Extended cognition has, of course, its critics.8 But given that the view is becom-
ing increasingly mainstream, the time is ripe to consider whether and to what extent
this species of active externalism might itself pose challenges to self-knowledge
which mimic or even outstrip the challenges posed by content externalism.9
A natural starting point to this end will be to juxtapose the alleged tension
between content externalism (CE) and self-knowledge (SK) with the pairing of
extended cognition (EC) with self-knowledge.
Tension (CE + SK): (i) that the content of our thoughts is determined partly by the environ-
ment; and (ii) that a subject can know by reflection alone (i.e., from the armchair) the con-
tent of her own thoughts.
Tension (EC + SK): (iii) that the processes driving our thoughts are realized partly by the
environment; and (iv) that a subject can know by reflection alone (i.e., from the armchair)
the content of her own thoughts.10

7
 While ‘commonsense functionalism’ is the primary way in which extended cognition has been
philosophically motivated, the position also has been argued for by appeal to dynamical systems
theory. See, in particular, Palermos (2011; 2014).
8
 See, for instance, Adams and Aizawa (2001, 2008) and Rupert (2004).
9
 See Carter et al. (2018) for a collection of recent papers which highlight not only the recent popu-
larity of extended cognition in the philosophy of mind and cognitive science, but also how this
view is increasingly influencing positions in contemporary epistemology which have traditionally
proceeded against a background of epistemic individualism.
10
 Note that there is a parallel literature which engages with the apparent tension between content
externalism and internalist accounts of epistemic justification. Bonjour (1992, 136), for instance,
remarks in an oft-cited passage that characterises one line of motivation for incompatibilism:
36 J. Adam Carter and D. Pritchard

Note that part of the philosophical basis for the alleged tension between CE and
SK is that if content externalism is true, then it looks as though we lack the kind of
knowledge that we take to be distinguished from third-person knowledge on the
basis of the directness, authority and method disanalogies which are often used to
characterize self-knowledge. And so to the extent that the tension between EC and
SK is genuine, it will be helpful to ask: if extended cognition is true, then what
import does this have for the kind of knowledge that we take to be distinguished
from third-person knowledge on the basis of the directness, authority and method
disanalogies? Consider to this end a ‘self-knowledge’ twist on Clark and Chalmers’
(1998) original case of Otto.
Otto(SK): Otto* suffers from Alzheimer’s disease, and like many Alzheimer’s patients, he
relies on information in the environment to help structure his life. Unlike some Alzheimer’s
patients who have (in losing their biological memory) lost in the process significant self-­
knowledge (i.e., knowledge of their own mental states, beliefs, and feelings), Otto* carries
a notebook around with him everywhere he goes, and this notebook plays in important role
in Otto*’s preservation of his own mental narrative. For example, when he learns new infor-
mation about his own mental states (i.e., beliefs, feelings, desires, etc.)—information about
his mental states which would be lost in biological storage—he writes it down in the note-
book. Likewise, when he needs some old information about his mental life, he looks it up.
For Otto*, his notebook plays the role usually played by a biological memory in preserving
a mental narrative.

The relevant contrast is now between Otto* and Inga*, where the latter is a paral-
lel self-knowledge version of our original description of Inga.
An initial observation about this case is that Otto*’s beliefs about his own mental
states seem at least to be straightforwardly indirect, rather than direct, in that Otto*
is becoming aware of his own mental states by looking at something else first. This
indirectness seems, following Brie Gertler’s (2015) distinction, to be both epistemo-
logical and metaphysical. Otto* after all appears to be taking the evidence he obtains
via consulting his entries in the notebook as his epistemic grounds for his beliefs
about his mental states. Furthermore, the act of consulting the notebook comes
between his seeking knowledge of his mental states and his awareness of these
states.
Second, regarding authority: as Crispin Wright (1989, 630) remarks, drawing
inspiration from Wittgenstein, it seems very much that mental states are ‘subject to
groundless, authoritative self-ascription’.11 As Sanford Goldberg (1999, 169)

“The adoption of an externalist account of mental content would seem to support an externalist
account of justification in the following way: if part or all of the content of a belief is inaccessible
to the believer, then both the justifying status of other beliefs in relation to that content and the
status of that content as justifying further beliefs will be similarly inaccessible, thus contradicting
the internalist requirement for justification.”
Chase (2001), Vahid (2003), Kallestrup and Pritchard (2004), and Williamson (2007) have fol-
lowed Bonjour in defending versions of incompatibilism. See, however, Brueckner (2002), Gerken
(2008), Madison (2009), and Carter and Palermos (2016) for some arguments for compatibilism.
11
 See Goldberg (1999) for a criticism of Wright’s neo-Wittgenstenian view. Goldberg’s view, in
short, is that Wright’s project of articulating an account of the epistemology of self-knowledge
Extended Self-Knowledge 37

c­ ondenses the thrust of this line: ‘it is part of our belief-attribution practices, that we
treat as legitimate a person’s self-attributions of belief even when these self-­
attributions are formed in an effortless manner and without any evidence’. Consider
now Inga* and Otto*. Whereas it is part of our belief-forming practices that we treat
Inga*’s self-attributions as authoritative, it’s far less clear that we will be inclined to
treat Otto*’s self-attributions as equally authoritative, in the sense intimated by
Wright. To appreciate why, we can just run a ‘high-stakes’ case—i.e., where there
is a lot of practical significance to whether Inga* and Otto* are correct in their
respective self-belief reports. In such a circumstance, we will be disinclined to sim-
ply take Otto*’s self-attribution as a ‘groundless, authoritative self-ascription’, even
if we take Inga*‘s to be, in the very same practical environment. After all, one might
be tempted to point out that Otto*’s attribution does not appear to be groundless at
all, but rather, based on his consultation of the notebook. Rather than treat Otto*’s
self-report as authoritative in a high-stakes case—as we might do with regard to
Inga*’s self-report—we will thus be inclined to verify Otto*’s report specifically by
verifying the report’s basis (i.e., by looking in his notebook).
Thirdly, regarding method: Otto* hardly appears to be introspecting when ascer-
taining what he believes. After all, Inga can tell just as easily as Otto* can what
Otto* believes by looking in Otto*’s notebook. Yet, as Eric Schwitzgebel (2014)
puts it, introspection is generally viewed as a way of learning about your own mind
which no one else can.12
Putting these points about directness, authority and method together, a special
kind of threat to self-knowledge, borne from extended cognition, materialises. For
lack of a better term, let’s call the ensuing argument the argument from distinctive
knowability. On first blush, call a mental state, M, distinctively knowable if M is
knowable, for S, in a way that is at least one of the following: (i) direct, (ii) authori-
tative; and/or (iii) via an introspective method. We use the term ‘distinctively know-
able’ for the following reason: self-knowledge is supposed to be interestingly
different from garden-variety, third-person knowledge not because each of our men-
tal states is actually known in ways that third-person knowledge is not (i.e., directly,
authoritatively, or introspectively). Consider after all, Freudian suppression cases.
Rather, the idea is that if self-knowledge (as such) is interestingly different from
third-person knowledge, it’s because our mental states are distinctively knowable—
viz., mental states are the sort of things that can be known in ways that are direct,
authoritative and via an introspective method. Other kinds of knowledge simply
aren’t knowable in these ways.
We can now state the argument from distinctive knowability as follows: for any
mental state m, m is distinctively knowable only if there is a subject S such that m is
knowable for S in a way that is distinctive—i.e., direct, authoritative, or via an

(i.e., by connecting our epistemic justification for our own mental states with their authoritative
nature in socio-linguistic practice) cannot be reconciled with a plausible construal of the psychol-
ogy of self-knowledge.
12
 For a helpful overview on various theories of introspection, see Schwitzgebel (2014).
38 J. Adam Carter and D. Pritchard

i­ntrospective method. If self-knowledge is interesting as a kind of knowledge, then


necessarily, mental states are the sort of thing which are distinctively knowable
(regardless of whether they are known). However, proponents of extended cognition
will be inclined to regard the case of OttoSK as one where the contents of Otto*’s
notebook (which include entries about his beliefs, feelings, etc.,) are such that it’s
not the case that there is a subject for whom these states are distinctively knowable
(e.g., especially once the beliefs Otto has about his mental life have eroded entirely
from biological storage). Therefore, as the argument goes, if extended cognition is
true, then self-knowledge, as such, is not an interesting kind (or, at least, a much less
interesting kind) of knowledge.13

3  The Challenge from Epistemic Situationism

The challenge posed by extended cognition to self-knowledge was undercutting in


the sense that it calls into doubt that self-knowledge is as distinctive as we’d been
inclined to think. In this section, we want to consider a very different kind of chal-
lenge to our self-knowledge, one which does not take issue with self-knowledge as
distinctive, but rather which appeals to our dependence on our environments so as
to motivate a special sort of sceptical challenge about self-knowledge. This is not a
form of radical scepticism about self-knowledge, but rather a moderate sceptical
challenge according to which we plausibly have much less self-knowledge than we
ordinarily suppose.
The challenge here stems from epistemic situationism, the view that our judg-
ments are to a surprising extent influenced by factors in our environments, factors
that are external to our cognitive agency and mental lives. Epistemic situationism
can be organised into three key ‘steps’. The first step is descriptive: there is a grow-
ing body of empirical literature which demonstrates that ‘extra-agential’ factors
such as ambient light levels and sounds, ambient smells and mood depressors, the
presence of bystanders, hunger levels and so on can influence our actions and judg-
ments to a surprising extent.14 The second step is analogical. John Doris (1998,
2002) and Gilbert Harman (2000; 1999; 2003) have reasoned from these kinds of
empirical observations to the conclusion that virtue ethics is empirically inadequate,
insofar as the virtue ethicist’s postulation of stable character traits—i.e., moral vir-
tues and vices—in the service of explaining moral (and immoral) behaviour is
empirically undermined. This is because of how it is in fact extra-agential factors,
including ones of which the subject may be consciously unaware, that are playing

13
 A further interesting implication of the argument is that to the extent that (in ordinary biological
cases) some individuals have direct, authoritative, introspective access to their mental states, this
will be merely an accidental feature of their self-knowledge, one that is an artifact of the material
constitution of the kind of cognitive process they happen to use to store their beliefs about their
mental life.
14
 For further discussion of these points, see Carter and Pritchard (2015; 2016)
Extended Self-Knowledge 39

much of the relevant explanatory in our ethical conduct, thereby precluding stable
character traits from playing this explanatory role. The third step is that if these
empirical observations can undermine virtue ethics by showing that our ethical vir-
tues do not play the explanatory role in our ethical behaviour that the view proposes,
then—as Mark Alfano (2012, 2014) has argued—in principle the very same situa-
tionist critique can be applied mutatis mutandis to virtue epistemology to show that
our epistemic virtues do not play the explanatory role in our epistemic behaviour
that the view proposes. Virtue epistemology holds that knowledge is acquired via
the manifestation of epistemic virtues, where this is a class of belief-forming pro-
cesses which includes cognitive abilities, cognitive faculties, and intellectual vir-
tues. It follows that epistemic situationism can force a dilemma: either we deny that
knowledge generally results from epistemic virtues as virtue epistemology claims,
or we concede that we have much less knowledge than we hitherto supposed.
Alfano (2014) himself has clarified that for those who embrace the virtue episte-
mologist’s core insight, the epistemic situationist threat is best understood as a
threat that targets inferential knowledge. This is not surprising, given that the most
widely studied cognitive biases—i.e., the availability heuristic and the representa-
tiveness heuristic—are ones that manifest at the level of inference. Interestingly,
though, there are a range of biases that look as though they have direct relevance to
the scope of our self-knowledge, even on the assumption that self-knowledge in
paradigmatic cases is not inferential knowledge. For even if self-knowledge is para-
digmatically direct, in that we don’t typically come to know the content of our
mental states by reasoning through any premises, we might well still form beliefs
about our mental lives in a way that manifests cognitive bias. Consider here a non-
epistemic analogy to preference. I might state my preference directly, in a way that
is not on the basis of anything else, and so state my preference directly, or non-
inferentially. Nonetheless, my preference might manifest bias (e.g., a knee-jerk
preference to hire a person of one ethnicity rather than another).
To appreciate the relevance of this point to self-knowledge, consider now three
example biases that have been well-studied—the hindsight, egocentric and confir-
mation biases:
Hindsight Bias: the filtering memory of past events through one’s present knowledge, so
that those events are recalled as having been predicted by one; also known as the ‘I-knew-­
it-all-along effect’.15
Egocentric Bias: recalling the past, including one’s own actions, intentions and beliefs, in a
self-serving manner.16

15
 For an influential meta-analysis on studies on the hindsight bias, see Christensen-Szalanski and
Willham (1991).
16
 As Taylor & Brown (1988, 193) note, ‘considerable research evidence suggests that overly posi-
tive self-evaluations, exaggerated perceptions of control or mastery, and unrealistic optimism are
characteristic of normal human thought.’
40 J. Adam Carter and D. Pritchard

Confirmation Bias: the tendency to search for or interpret information, including about
one’s own mental life, in a way that confirms one’s preconceptions.17

The hindsight, egocentric and confirmation biases are suggestive of various ways
in which our beliefs about the content of our own minds might well be subject a
range of biases. Here are three examples. First, the hindsight bias engenders in indi-
viduals false beliefs about what they knew. This is an error in self-knowledge; indi-
viduals who manifest the bias have an inaccurate picture of the past quality of their
beliefs. And, furthermore, the hindsight bias can also engender in individuals false
beliefs about what proposition the individual did in fact believe.18 Therefore, by the
hindsight bias, you might thus err in thinking both that certain beliefs of yours were
held when they were not, and that of certain beliefs that were held but not known,
that they were known. Second, in the case of egocentric bias, an individual will be
inclined to, for example, believe that they’ve had many praiseworthy beliefs and
desires which the individual in fact did not have, thus generating for one an inac-
curate (but self-serving) picture of one’s doxastic (and cognitive) history. Third,
when exhibiting confirmation bias, individuals will reflect on the contents of their
minds in ways that are distorted by their preconceptions. If, for example, one has a
self-serving preconceived impression that some description of one’s beliefs and atti-
tudes is correct, then when reflecting on which beliefs and attitudes one holds one
is likely to (mistakenly) accept the preconception as accurate.19
The hindsight, egocentric and confirmation biases are by no means the only
biases which can distort our self-knowledge, but their prevalence forms the basis for
two distinct challenges with respect to the received thinking about the nature and
scope of our self-knowledge. The first worry concerns scope and can be stated sim-
ply: if, as the empirical literature suggests, our judgments about the contents of our
own minds often manifest bias (i.e., biases of the sort noted in this section), then it
looks as though we have significantly less self-knowledge than we might have oth-
erwise supposed. Put another way, the challenge can be understood as highlighting
important ways in which even our non-inferential, direct beliefs about the contents
of our own minds might well manifest (undetected by us) cognitive biases which
undermine the epistemic status of these beliefs about the contents of our mind as
knowledge.
A second, and associated, problem is that whatever distinctiveness self-­
knowledge is regarded as having on the basis of epistemic privilege is called into
doubt. The directness, authority and introspective character of paradigmatic self-­
knowledge is generally taken to afford such knowledge with a kind of epistemically
security (e.g., Horgan 2012; Gertler 2012). However, the supposition that the epis-
temic status of paradigmatic self-knowledge can be undermined by the kinds of

17
 See Nickerson (1998) for a representative overview.
18
 In the former case, one might, for instance, falsely believe that one knew that a previous predic-
tion would come out true. In the latter case, one might, via the hindsight bias, falsely believe that
one in fact made a prediction one did not make.
19
 See, for instance, Forer (1949) for a classic study.
Extended Self-Knowledge 41

biases which afflict our ‘non-privileged’ (third-person) beliefs accordingly chal-


lenges the presumption of epistemic security.20

4  Self-Knowledge, Extended (and Expanded)

The aim of this section will be to show how, to the extent which the challenges to
self-knowledge outlined in §§2–3 appear to threaten the nature and scope of self-­
knowledge, this may well be symptomatic of a tacit commitment to a traditional
position in epistemological meta-theory called epistemic individualism. After out-
lining the view, and suggesting some problems for it, we will then suggest how,
through the lens of epistemic anti-individualism, we have a way of reinterpreting the
challenges so as to extend and expand rather than to limit self-knowledge.
Epistemic individualism is the widely embraced view within contemporary epis-
temology that the cognitive processes grounding knowledge are exclusively internal
to the biological subject—indeed, typically, they will be supposed to take place
exclusively within the brain (and possibly central nervous system) of the subject.21
While the epistemic individualist picture fits snugly with internalist approaches in
epistemology, the view has also received explicit and implicit endorsement from
paradigmatic epistemic externalists, who regard cognitive processes, the reliability
of which matters for knowledge, as themselves resident in the agent’s head.22
If epistemic individualism is true, then the challenges articulated in §§2–3 do
indeed appear to have negative import for the prospects of vindicating the presumed
extent of self-knowledge, at least as it’s construed in paradigmatic cases. For in both
cases what these arguments seem to show is that a class of apparent cases of self-­
knowledge in fact depend in substantive ways on factors which are significantly
external to the subject. Granted epistemic individualism, then, it seems to follow
that such apparent self-knowledge in such cases is not bona fide.
There are, however, good reasons to reject epistemic individualism. In particular,
there are independent reasons for supposing that knowledge exhibits a phenomenon
known as epistemic dependence, such that it can substantively depend on factors
external to the subject’s cognitive agency. The notion of epistemic dependence in
play here arises out of Duncan Pritchard’s critique of a form of virtue epistemol-
ogy—known as robust virtue epistemology—that attempts to exclusively under-
stand knowledge in terms of the appropriate manifestation of cognitive ability or
virtue.23 If such a proposal were correct, then epistemic dependence would be ruled

20
 For a detailed discussion of the challenge that the prevalence of epistemic bias poses to third-
person beliefs, see Carter and Pritchard (2016).
21
 See Kallestrup and Pritchard (2012) and Pritchard (2016a) for further discussion of epistemic
anti-individualism.
22
 See, for example, Goldman (1979, 13).
23
 For some of the key defences of robust virtue epistemology, see Sosa (1991, 2007, 2009, 2015),
Zagzebski (1996, 1999), and Greco (2010).
42 J. Adam Carter and D. Pritchard

out: insofar as one’s cognitive success manifests a sufficient level of cognitive


agency, then one thereby has knowledge. Conversely, Pritchard argues that the exis-
tence of epistemic dependence demonstrates that such a proposal is not correct.
More precisely, the phenomenon of epistemic dependence shows that epistemic
individualism is untenable, and hence that robust virtue epistemology, insofar as it
is wedded to this thesis (as it standardly is), is also untenable.
This rationale for rejecting epistemic individualism can be most straightfor-
wardly appreciated with reference to the ways in which it is due to factors entirely
outwith an individual’s skin and skull whether her true belief amounts to knowl-
edge—i.e., the ways in which knowledge exhibits epistemic dependence. In particu-
lar, epistemic dependence comes in two varieties, positive and negative. Positive
epistemic dependence features in cases where an agent manifests very little cogni-
tive agency—much less than would normally suffice for knowledge—but where her
cognitive success amounts to knowledge nonetheless because of factors in her envi-
ronment. A good example of this is epistemically friendly testimonial environments,
in which an agent can come to have knowledge by for the most part trusting the
word of another.24 In such cases the knowledge that is acquired is not primarily due
to the agent’s manifestation of cognitive agency, but rather substantially depends on
the informant’s manifestation of cognitive agency. Negative epistemic dependence,
in contrast, is when an agent manifests a high level of cognitive agency (i.e., of a
level that would ordinarily easily suffice for knowledge), but where the cognitive
success does not amount to knowledge because of environmental factors.
Paradigmatic examples here include barn façade cases.25 In such case, the subject’s
cognitive success manifests a great deal of cognitive agency, and yet nonetheless
fails to amount to knowledge on account of how purely environmental factors
ensure that the belief so formed is subject to high levels of epistemic luck/risk.26
Our goal is not to further defend epistemic dependence here, since that thesis has
been extensively argued for elsewhere. Rather, we want to show that with this thesis
in place, and thus with epistemic individualism rejected, we are in a position to
reevaluate the putative challenges posed for self-knowledge by extended cognition
and situationism.

24
 Note that the clause, ‘for the most part’, is important here. The claim is not that in the right epis-
temic conditions mere trust can suffice for testimonial knowledge (gullibility is never a route to
knowledge), but rather that in the right epistemic conditions relatively low levels of cognitive
agency can suffice for knowledge.
25
 See also the epistemic twin earth case offered by Kallestrup and Pritchard (2014).
26
 For Pritchard’s initial critique of robust virtue epistemology, which turns on a distinction between
intervening and environmental epistemic luck, see Pritchard (2009a, 2009b, 2012) and Pritchard,
Millar & Haddock (2010, chs. 2-4). For a development of this critique in terms of the specific
notions of epistemic dependence and epistemic anti-individualism, see Kallestrup and Pritchard
(2012, 2013, 2014) and Pritchard (2016a). For further discussion of the notion of epistemic risk
and its relevance in this context, see Pritchard (2016b).
Extended Self-Knowledge 43

4.1  Extended Self–Knowledge

Once we embrace epistemic anti-individualism, then the force of the argument from
distinctive knowability (outlined in §2) can be significantly mitigated. First, con-
sider the point about directness. It was supposed that since Otto* is becoming aware
of his own mental states by looking at something else (i.e., the notebook), hence his
mental states are not directly knowable, but rather mediated by the notebook. With
epistemic anti-individualism in the background, however, there is a different way to
interpret this case. So long as we take the proponent of extended cognition’s func-
tional analogy seriously, then Otto* will be consulting his notebook in a way that is
both automatic and entirely uncritical, much as we uncritically embrace the deliver-
ances of intracranial self-scanning. That is, in order for Otto*’s use of the notebook
to count as an extended cognitive process, it will need to be cognitively integrated
with his other cognitive processes to the same extent as his biological memory, and
thus to a similar extent embedded within his cognitive character.27 But if that is
right, then there should be no significant difference between Otto* and Inga* in
terms of the directness of the belief so formed—it is just that Otto* is making use of
external technology while Inga* is relying exclusively on her biological memory.
But isn’t that simply a form of positive epistemic dependence, such that Otto*’s
cognitive processes now in part depends upon external factors, albeit in ways that
are favourable to his acquisition of knowledge? Moreover, insofar as Otto* and
Inga* are forming their beliefs in analogously direct ways, then it follows that just
as we should not regard Inga*’s self-knowledge as mediated by her biological brain,
neither should we regard Otto*’s cognitive processes as mediated. Hence there is no
reason for supposing, on this front at least, that the knowledge acquired by Otto*
and Inga* is in any way different in kind.
Second, consider the point about authority. The crux of the matter here is that
once we take on board the issue of cognitive integration just noted, then it is hard to
see why authority in the relevant sense will not transfer to Otto*’s self-ascriptions
which are mediated by the use of the notebook. Put another way, the idea that
authority doesn’t transfer in this way seems to assume that there is a functional dif-
ference between Otto*’s use of the notebook and Inga*’s use of her biological mem-
ory, and yet by hypothesis no such difference can in fact be present. In short: if
Otto*’s use of the notebook in making self-ascriptions is as seamless and integrated
within his cognitive processes as Inga*’s use of her biological memory, then on
what basis would we treat the former as less authoritative than the latter?
Relatedly, note that once we grant that there is a genuine functional equivalence
in play here, then the related point about the groundlessness of self-ascriptions fails
to gain any purchase. This point supposes, after all, that Otto* is self-consciously
basing his belief on the deliverances of the notebook, while Inga* is simply report-
ing the content of her mental states directly. But as we noted above, taking the

27
 For a discussion of the various conditions that should plausibly be satisfied for cognitive integra-
tion, see Pritchard (2010) and Palermos (2014).
44 J. Adam Carter and D. Pritchard

functional equivalence in play here seriously means taking both self-ascriptions to


be equally direct, and hence equally unmediated by a conscious basing on rational
grounds.
Third, consider the point about method. The objection that the entries in Otto*’s
notebook fail to be recoverable by introspection is an objection with considerably
less import on an anti-individualistic reading. The thrust of the original worry was
that Otto*’s would-be self-knowledge is not introspectively recoverable because it
is also recoverable by others. The underlying idea here was that introspection is a
way of knowing the contents of your own mind which no one else has. Call this idea
epistemic privacy.
If epistemic privacy is a necessary condition for introspection, then Otto* fails to
be introspecting. Interestingly, however, the three most commonly embraced neces-
sary conditions on introspection in the literature, while of course compatible with
epistemic privacy, do not entail epistemic privacy. And, furthermore, Otto*’s situa-
tion can plausibly meet these other conditions. As Schwitzgebel (2014, §1.1) notes,
introspection ‘is aimed at generating, knowledge, judgments, or beliefs about men-
tal events, states, or processes’ (mentality condition), which are ‘about one’s own
mind only and no one else’s’ (first-person condition), and which are ‘part of one’s
currently ongoing mental life’ (temporal proximity condition). Note that, on the face
of it, neither the mentality, first-person nor the temporal proximity conditions on
introspection imply that a process is not introspective if another individual can (as
Inga can, for instance) have access to one’s mind. We thus conclude that there is a
perfectly respectable sense in which Otto* counts as satisfying the conditions for
introspection, and thereby evades the point about method, even though we grant that
he doesn’t satisfy the more demanding conditions for epistemic privacy.
The point of the foregoing is that the distinctive knowability argument lacks bite,
at least insofar as we are willing to embrace (positive) epistemic dependence, and
thus reject epistemic individualism. One upshot of this point is that the scope of
self-knowledge is now considerably widened.28 Indeed, given that we now occupy
an age in which we are increasingly embedded within technologically rich environ-
ments, the opportunities for substantially extending our self-knowledge on this view
are legion. For example, consider the ubiquitous use of smart-phone journal apps,
such as the smartphone app ‘Diaro’, which allows users to record their ‘experiences,
thoughts and ideas’ throughout the day, and in a way that syncs to all of the user’s
other devices.29 Diaro also tracks time and location, to better facilitate the user’s
grasp of the chronology of their thought contents. To the extent that our philosophi-
cal theorising treats certain kinds of interactions with technology (such as Diaro) on
a par with Otto*’s use of his notebook as cases of extended self-knowledge, it is
accordingly clear how the apparent challenge to orthodox thinking about self-­
knowledge posed by extended cognition opens up the door to more varieties of

28
 Relatedly, we think it also follows that the line separating self-knowledge from other kinds of
knowledge is now also much more blurred, though we will not be exploring this issue further here.
29
 See https://play.google.com/store/apps/details?id=com.pixelcrater.Diaro
Extended Self-Knowledge 45

s­elf-­
knowledge than can be accounted for within the epistemic individualist’s
framework.
The second key benefit regards the quality of our self-knowledge, and in particu-
lar, its stability. The point about stability can be made helpfully with reference to the
von Restorff effect and the reminiscence bump. The von Restorff effect is the ten-
dency for individuals, in recovering their own beliefs and attitudes, to more easily
recover those that ‘stand out’ than other beliefs, which are more easily forgotten.30
The reminiscence bump is the effect by which individuals are more inclined to
recall the beliefs one has which are of a personal nature from adolescence and early
adulthood than from other life periods.31 In each case, the cost of biological memory
storage is stability; via effects such as these, some of our self-knowledge is lost
because (in each of these ways) some of the content of our minds is unrecoverable.
Extended self-knowledge (as is recognised within the anti-individualist paradigm),
by contrast, is susceptible to neither the von Restorff effect nor the reminiscence
bump, and in this respect, extended self-knowledge has a kind of stability which
self-knowledge retrieved from biological storage lacks.32

4.2  Expanded Self-Knowledge

The situationist challenge canvassed in §3 suggested that we have much less self-­
knowledge than we ordinarily take ourselves to have. This challenge, under closer
consideration, might be overstated. After all, the hindsight, egocentric and confir-
mation biases surveyed were representative of the kind of cognitive biases which
can influence our belief formation about our own mental states. If knowledge of our
own mental states demands (as, for example, is implied by robust virtue epistemol-
ogy) that our believing correctly about our mental states should be primarily
explained by our exercising cognitive abilities or virtues, then the empirical data
regarding the prevalence of our susceptibility to these biases would indeed suggest
that knowledge is often lacking in these cases.
This line of reasoning goes through only against the background of supposing
that if cognitive biases play some role in our belief formation, then we thereby fail
to know the target belief. But this claim, as we noted at the beginning of §4, is
highly suspect. As the arguments from epistemic dependence showed, propositional
knowledge does not demand that our cognitive abilities primarily explain our cogni-
tive successes. Rather, our cognitive abilities must at least play some weaker but
nonetheless significant role in explaining cognitive success in cases of knowledge.
But this weaker claim does not preclude that cognitive biases (including the

30
 See Karis et al. (1984).
31
 See, for example, Jansari and Parkin (1996).
32
 See Carter and Kallestrup (2014) for more detailed discussion on the comparative stability of
extended over biological storage.
46 J. Adam Carter and D. Pritchard

e­ gocentric, hindsight and confirmation biases) and other situational factors can play
some role in our acquisition of knowledge about the contents of our minds.
Against an epistemic anti-individualistic background—i.e., one which counte-
nances epistemic dependence—there is thus room to vindicate some of our beliefs
about our own mental states as bona fide knowledge even if they are to some degree
afflicted by cognitive bias or other situational factors, so long as these extra-agential
factors are not the overarching explanation for the belief’s correctness. The crux of
the matter is that so long as the explanatory load borne by the cognitive bias is rela-
tively small and doesn’t significantly undermine one’s cognitive performance, then
it can be entirely compatible with knowledge. Furthermore, even when it does play
a significant explanatory role in one’s cognitive performance, so long as the bias
promotes the cognitive goal—not all cognitive biases lead to error, after all—and so
long as the subject’s cognitive agency is also playing a significant explanatory role,
then this can also be compatible with knowledge. In such case, the cognitive bias
would rather be a kind of positive epistemic dependence. In these ways, we can
vindicate a significant part of our self-knowledge against what would otherwise, on
an epistemic individualistic picture, appear to be a more worrying empirically
driven sceptical challenge to the scope of our self-knowledge. The situationist chal-
lenge to our self-knowledge may be real, but it is not as extensive as proponents of
this challenge suppose.33

5  Concluding Remarks

Traditionally, the primary challenge for the prospect of vindicating self-knowledge


of a paradigmatic sort in contemporary epistemology and philosophy of mind has
been the thesis of content externalism, according to which one’s mental contents are
individuated by features of her physical or social environment. This paper has con-
sidered two other broadly externalist challenges to self-knowledge, both of which
advert to ways in which our thought content, cognitive processes and cognitive suc-
cesses depend crucially on our external environments. In particular, we’ve shown
how extended cognition and epistemic situationism stand, in various ways, in prima
facie tension with received thinking about the nature and scope of self-knowledge.
After outlining these challenges, we’ve then shown how, to the extent that these
purported challenges should be worrying, this is symptomatic of a tacit but we think
mistaken background commitment to epistemic individualism. We’ve concluded by
showing how, once we leave epistemic individualism behind, the purported chal-
lenges to our extended cognition and epistemic situationism can be attenuated, such

33
 For more on this point about the limits of the epistemic situationist challenge, see Pritchard
(2014). See also Carter and Pritchard (2017) which argues for this point in a specifically educa-
tional context.
Extended Self-Knowledge 47

that we are left in a better overall epistemic position with regard to the contents of
our own minds.34

References

Adams, F., and K. Aizawa. 2001. The bounds of cognition. Philosophical Psychology 14: 43–64.
———. 2008. The bounds of cognition. Oxford: Blackwell.
Alfano, M. 2012. Expanding the situationist challenge to responsibilist virtue epistemology. The
Philosophical Quarterly 62: 223–249.
———. 2014. Expanding the situationist challenge to reliabilism about inference. In Virtue epis-
temology naturalized, ed. A. Fairweather, 103–122. Dordrecht: Springer.
Boghossian, P. 1989. Content and self-knowledge. Philosophical Topics 17: 5–26.
Bonjour, L. 1992. Externalism/internalism. In A Companion to Epistemology, ed. J.  Dancy,
E. Sosa, and M. Steup, 132–136. Oxford: Blackwell.
Brueckner, A. 2002. The consistency of content-externalism and justification-internalism.
Australasian Journal of Philosophy 80: 512–515.
Burge, T. 1986. Individualism and psychology. Philosophical Review 95: 3–45.
———. 1988. Individualism and self-knowledge. Journal of Philosophy 85: 649–663.
Carter, J.A., and J. Kallestrup. 2014. Extended cognition and propositional memory. Philosophy
and Phenomenological Research. https://doi.org/10.1111/phpr.12157.
Carter, J.A., and S.O. Palermos. 2015. Active externalism and epistemic internalism. Erkenntnis
80 (4): 753–772.
———. 2016. Epistemic internalism, content externalism and the subjective/objective justification
distinction. American Philosophical Quarterly 53: 231–244.
Carter, J.A., and D.H. Pritchard. 2016. Cognitive bias, scepticism and understanding. In Explaining
Understanding: New Perspectives from Epistemology and the Philosophy of Science, ed.
S. Grimm and C. Baumberger. London: Routledge.
———. 2017. Epistemic situationism, epistemic dependence, and the epistemology of educa-
tion. In Epistemic Situationism, ed. M. Alfano and A. Fairweather, 168–191. Oxford: Oxford
University Press.
Carter, J.A., J.  Kallestrup, D.H.  Pritchard, and S.O.  Palermos. 2014. Varieties of externalism.
Philosophical Issues 24: 63–109.
Carter, J.  Adam, Andy Clark, Jesper Kallestrup, Orestis Palermos, and Duncan Pritchard, eds.
2018. Extended epistemology. Oxford: Oxford University Press.
Chase, J.  2001. Is externalism about content inconsistent with internalism about justification?
Australasian Journal of Philosophy 79: 227–246.
Christensen-Szalanski, J., and C.  Willham. 1991. The hindsight bias: A meta-analysis.
Organizational Behavior and Human Decision Processes 48: 147–168.
Clark, A. 2008. Supersizing the mind: Embodiment, action, and cognitive extension. Oxford:
Oxford University Press.
Clark, A., and D. Chalmers. 1998. The extended mind. Analysis 58: 7–19.
Doris, J.M. 1998. Persons, situations, and virtue ethics. Noûs 32: 504–530.
———. 2002. Lack of character: Personality and moral behavior. Cambridge: Cambridge
University Press.

 This paper was written as part of the AHRC-funded ‘Extended Knowledge’ (#AH/J011908/1)
34

project which is hosted by the University of Edinburgh’s Eidyn Philosophical Research Centre, and
we are grateful to the AHRC for their support of this research. Special thanks to Jesper Kallestrup,
Julie Kirsch, Patrizia Pedrini, and Orestis Palermos.
48 J. Adam Carter and D. Pritchard

Dretske, F. 1994. Introspection. Proceedings of the Aristotelian Society. (new series. Vol. 94,
263–278.
Forer, B.R. 1949. The fallacy of personal validation: A classroom demonstration of gullibility.
Journal of Abnormal and Social Psychology 44: 118–123.
Gerken, M. 2008. Is internalism about knowledge consistent with content externalism? Philosophia
36: 87–96.
Gertler, B. 2012. Renewed acquaintance. In Introspection and consciousness, ed. D. Smithies and
D. Stoljar. Oxford: Oxford University Press.
———. 2015. Self-knowledge. In Stanford Encyclopedia of Philosophy, ed. E.N.  Zalta http://
plato.stanford.edu/archives/sum2015/entries/self-knowledge/.
Goldberg, S. 1999. The psychology and epistemology of self-knowledge. Synthese 118: 165–199.
Goldman, A. 1979. What is justified belief? In Justification and Knowledge, ed. G. Pappas, 1–23.
Dordrecht: Springer.
Greco, J. 2010. Achieving knowledge. Cambridge: Cambridge University Press.
Harman, G. 1999. Moral philosophy meets social psychology: Virtue ethics and the fundamental
attribution error. Proceedings of the Aristotelian Society. 99: 315–331.
———. 2000. The nonexistence of character traits. Proceedings of the Aristotelian Society 100:
223–226.
———. 2003. No character or personality. Business Ethics Quarterly 13: 87–94.
Horgan, T. 2012. Introspection about phenomenal consciousness: Running the gamut from infal-
libility to impotence. In Introspection and consciousness, ed. D. Smithies and D. Stoljar, 405–
422. Oxford: Oxford University Press.
Jansari, A., and A. Parkin. 1996. Things that go bump in your life: Explaining the reminiscence
bump in autobiographical memory. Psychology and Aging 11: 85–91.
Kallestrup, J., and D.H. Pritchard. 2004. An argument for the inconsistency of content externalism
and epistemic internalism. Philosophia 31: 345–354.
———. 2012. Robust virtue epistemology and epistemic anti-individualism. Pacific Philosophical
Quarterly 93: 84–103.
———. 2013. Robust virtue epistemology and epistemic dependence. In Knowledge, virtue,
and action: Essays on putting epistemic virtues to work, ed. T. Henning and D. Schweikard.
London: Routledge.
———. 2014. Virtue epistemology and epistemic twin earth. European Journal of Philosophy 22:
335–357.
Karis, D., M. Fabiani, and E. Donchin. 1984. P300 and memory: Individual differences in the von
restorff effect. Cognitive Psychology 16: 177–216.
Ludlow, P. 1995. Externalism, self-knowledge, and the prevalence of slow switching. Analysis 55:
45–49.
———. 1997. On the relevance of slow switching. Analysis 57: 285–286.
Madison, B.J.C. 2009. On the compatibility of epistemic internalism and content externalism. Acta
Analytica 24: 173–183.
Nickerson, R. 1998. Confirmation bias: A ubiquitous phenomenon in many guises. Review of
General Psychology 2: 175–220.
Palermos, S.O. 2011. Belief-forming processes, extended. Review of Philosophy and Psychology
2: 741–765.
———. 2014. Knowledge and cognitive integration. Synthese 191: 1931–1951.
Parent, T. 2013. Externalism and self-knowledge. In Stanford Encyclopedia of Philosophy, ed.
E.N. Zalta http://plato.stanford.edu/archives/sum2013/entries/self-knowledge-externalism/.
Pritchard, D.H. 2009a. Apt performance and epistemic value. Philosophical Studies 143: 407–416.
———. 2009b. Knowledge, understanding and epistemic value. In Epistemology (Royal Institute
of Philosophy Lectures), ed. A. O’Hear, 19–43. Cambridge: Cambridge University Press.
———. 2010. Cognitive ability and the extended cognition thesis. Synthese 175: 133–151.
———. 2012. Anti-luck virtue epistemology. Journal of Philosophy 109: 247–279.
Extended Self-Knowledge 49

———. 2014. Re-evaluating the epistemic situationist challenge to virtue epistemology. In


Naturalizing Epistemic Virtue, ed. A.  Fairweather and O.  Flanagan, 143–154. Cambridge:
Cambridge University Press.
———. 2016a. Epistemic dependence. Philosophical Perspectives 30: 1–20.
———. 2016b. Epistemic risk. The Journal of Philosophy 113 (11): 550–571.
Pritchard, D.H., A. Millar, and A. Haddock. 2010. The nature and value of knowledge: Three inves-
tigations. Oxford: Oxford University Press.
Putnam, H. 1975. The meaning of “meaning”. Minnesota Studies in the. Philosophy of Science 7:
131–193.
Rupert, R. 2004. Challenges to the hypothesis of extended cognition. Journal of Philosophy 101:
389–428.
Schwitzgebel, E. 2014. Introspection. In Stanford Encyclopedia of Philosophy, ed. E.N.  Zalta
http://plato.stanford.edu/archives/sum2014/entries/introspection/.
Sosa, E. 1991. Knowledge in perspective: Selected essays in epistemology. Cambridge: Cambridge
University Press.
———. 2007. A virtue epistemology: Apt belief and reflective knowledge. Oxford: Oxford
University Press.
———. 2009. Reflective knowledge: Apt belief and reflective knowledge. Oxford: Oxford
University Press.
———. 2015. Judgment and agency. Oxford: Oxford University Press.
Taylor, S., and J.D. Brown. 1988. Illusion and Well-being: A social psychological perspective on
mental health. Psychological Bulletin 103: 193–210.
Vahid, H. 2003. Content externalism and the internalism/externalism debate in justification theory.
European Journal of Philosophy 11: 89–107.
Williamson, T. 2007. On being justified in one’s head. In Rationality and the good: critical essays
on the ethics and epistemology of robert audi, ed. M. Timmons, J. Greco, and A. Mele, 106–
122. Oxford: Oxford University Press.
Wright, C. 1989. Wittgenstein’s later philosophy of mind: Sensation, privacy, and intention.
Journal of Philosophy 86: 622–634.
Zagzebski, L. 1996. Virtues of the mind: An inquiry into the nature of virtue and the ethical foun-
dations of knowledge. Cambridge: Cambridge University Press.
———. 1999. What is knowledge? In The Blackwell guide to epistemology, ed. J.  Greco and
E. Sosa. Oxford: Blackwell.
The ‘Crux’ of Internal Promptings

Patrizia Pedrini

Abstract In Self-Knowledge for Humans (2014), Cassam defends a quite broad


inferentialist theory of substantial third-person self-knowledge, which he promises
to extend to virtually all mental states, including the so-called “internal promptings”
(Lawlor 2009). Internal promptings are spontaneous, self-intimated experiential
episodes that may not always be phenomenologically salient, or conceptually
clearly subsumed, to the extent that the subject may not always be able to identify
them. According to Cassam, however, their spontaneous surfacing does not pre-
clude our access to them actually being inferential.
I question the claim that internal promptings can really be covered by an inferen-
tialist theory of self-knowledge. While I agree with Coliva (2016) that an inferen-
tialist theory of self-knowledge does not in fact apply to self-knowledge of internal
promptings, I show that this failure does not depend on lacking a story about how
inferentialism can be extended to first-person self-knowledge, as Coliva diagnoses.
Rather, Cassam’s theory is flawed by an independent, and precedent, amphiboly
fallacy affecting the concept of self-knowledge he makes use of. That is why
Coliva’s objection may not apply immediately, even if her verdict on the non-­
extensibility of an inferentialist theory of self-knowledge to internal promptings is
unaffected.
I also raise and discuss the issue of under-determination of inner experience with
respect to conceptual schemes. Finally, by taking stock of the intrinsically elusive
nature of a vast portion of our own mental states, I express sympathy for a wider
geography of the mental.

Keywords  Third-person self-knowledge · First-person self-knowledge · Self-­


interpretation · Internal promptings · Internal evidence · Unconceptualized mental
states

P. Pedrini (*)
Department of Letters and Philosophy, University of Florence, Florence, Italy

© Springer Nature Switzerland AG 2018 51


P. Pedrini, J. Kirsch (eds.), Third-Person Self-Knowledge, Self-Interpretation,
and Narrative, Contributions To Phenomenology 96,
https://doi.org/10.1007/978-3-319-98646-3_4
52 P. Pedrini

1  Cassam’s Broad Theory of Third-Person Self-Knowledge

In his recent, seminal contribution Self-Knowledge for Humans (2014), Quassim


Cassam has offered a defense of the constant interest third-person self-knowledge
invites, especially from non-philosophers. Non-philosophers typically expect that
philosophers who study and talk about self-knowledge have in mind precisely
what they do roughly have in mind, i.e., the fascinating, Socratic enterprise of get-
ting to know our mental states, values, abilities, emotions, affective states, and
character traits that crucially matter to our own lives, conduct, choices, and self-­
understanding.1 This is the kind of knowledge about oneself to which the “self-
help” industry is in the business of contributing. It is also the fundamental goal of
most, if not all, “talking therapies”, which try to help people precisely to get their
psychological inner experience right, especially in the face of various forms of
suffering, which, more often than not, are unfortunately of psychological (if not
psychiatric) import. People have an interest in this kind of self-knowledge, and
place a value on it,2 precisely because very often self-ignorance does not lead to a
serene life.3 Indeed, self-­ignorance can induce us to make wrong choices, creates
fertile ground for cultivating troublesome emotions, and in general does not allow
us to comprehend ourselves, and thus act on the fairly firm basis of sufficiently
reliable self-­understanding.4 On the contrary, getting to know ourselves seems, in
principle, to enable us to be in the position to manage, or circumvent, all the psy-
chological characteristics that are detrimental to us, and in whose hands our lives
are passively doomed to remain until we get to know them. In this sense, the self-
knowledge in which both ordinary folks and philosophers in their non-academic
lives are greatly interested is this variety of self-knowledge, which Cassam then
calls substantial (2014, p. vii).
Now, there is a quite disappointingly trivial self-knowledge, as Cassam defines it
(p. vii) — the opposite, in a way, of the substantial self-knowledge that matters to
us  — which has, however, been the variety of self-knowledge that philosophers

1
 It is common to think that the kind of mental states one can get to know third-personally are dis-
positional states only. Part of the story I will try to clarify here is that there is an important variety
of occurrent mental states that can, and in fact are, known third-personally, in ways to be qualified,
and under a certain conception of self-knowledge.
2
 The value they put on it may be practical, or even moral. I am not interested here in adjudicating
this question, although I think there may be room to defend both kinds of values, depending on the
general self-knowing project a person embarks on. Cassam 2014 discusses the topic in Chap.15.
Cf. also Renz 2017.
3
 Some people are convinced that “ignorance is bliss”. A wide debate tries to adjudicate the ques-
tion, and I have addressed it in Pedrini 2013. My point here is just that, even if there may be cases
in which self-ignorance is beneficial to the individual, either in the short or in the long run, there
are certainly cases in which it nonetheless can be highly detrimental to the fulfillment of one’s life
projects.
4
 As above, it might well be the case that self-understanding is not always necessary to act in our
best interest. Again, I am just focusing on cases in which it is.
The ‘Crux’ of Internal Promptings 53

mostly study.5 An example of such trivial self-knowledge is my own knowledge that


I believe that I’m wearing socks. All those acquainted with the reasons why such
alleged triviality has been central to analytic philosophy of mind and epistemology
in the last few decades would not easily accept the idea that this variety of self-­
knowledge is indeed trivial. For, disappointing to a non-expert as this may be, an
important debate, both epistemological and metaphysical, revolves around the char-
acterization of such allegedly “trivial” self-knowledge.
To begin with, self-knowledge of important sets of mental states that is charac-
terized by Cassam as trivial is reached first-personally. It is precisely its first-­
personal character that has made it so central and worthy of investigation in
contemporary philosophy of mind and epistemology. For first-person self-­knowledge
of (often trivial) mental states has the interesting feature of being immediately avail-
able to us, without the mediation of evidential reasons, as happens instead for many
other kinds of knowledge. In order for me to know that I believe that I’m wearing
socks, I do not have to draw inferences from any sort of evidence allowing me to
reach the conclusion that I do have the mental state that I believe that I’m wearing
socks. Such a route to self-knowledge would be indirect, in so far as it is mediated
by the inferential calculus I have to perform from evidence to conclusion. Such
calculus amounts to a cognitive achievement on the part of its performer, and can
obviously suffer from all sorts of possibilities of failure. And, as a matter of fact, this
is the route an observer has to use to get to know my mental states — from the evi-
dence I offer the observer with my speech and behavior, to the conclusion that I
must have certain mental states. Obviously the observer can fail to get to know me,
her enterprise being a cognitive achievement aimed at the discovery of something
she can access indirectly only.
Now, immediacy is a crucial feature of first-person self-knowledge, which strik-
ingly singles it out within the general domain of knowledge, and which needs expla-
nation. Epistemologists have devoted much attention to it precisely because of that
peculiarity, and have also considered it a privileged form of knowledge that a sub-
ject enjoys, for each subject is “the first person of herself”, as it were, that is, she is
in the epistemic position to access her own mental states in a way that anyone else
is, in fact, impeded from doing. If this is so, then immediacy seems to present
another crucial feature of it, namely, its being authoritative: in so far as a subject
experiences a self-intimating mental state, that is, a state that is immediately avail-
able to her, she can hardly be wrong about having it.6 In addition, the capacity we
have for this kind of knowledge also seems to have a say in the kind of creatures we
are, from the metaphysical point of view.7
Now, being unacquainted with this epistemological and metaphysical debate,
non-philosophers can find it hard to grasp the sense of this major focus, compared
with substantial self-knowledge, considering the usefulness  that the latter

5
 I had drawn this distinction in Pedrini 2009.
6
 An excellent presentation of all these features in the literature and a theory of how they are inter-
locked is Coliva 2016.
7
 Cf. Rudder Baker 2013.
54 P. Pedrini

u­ ndoubtedly has in life. Obviously, defenders of first-person self-knowledge would


say that first-person self-knowledge is indeed as “useful” to our everyday life as
third-­person self-knowledge is, even if this is not immediately manifest to non-
philosophers, and very fundamentally so: it may not perhaps be instrumentally as
useful as third-­person self-knowledge has been described to be, but it contributes
the very constitutive basis of our mental well-being and basic rationality.8
It is worth noting, at this juncture, that it may not be the self-knowledge of certain
mental states that is trivial, as Cassam claims. Rather, it seems to be the content of
the mental states presented by Cassam and others when exemplifying instances of
first-person self-knowledge that sounds quite trivial. For in principle, and as a mat-
ter of fact, the content of mental states that one can first-personally know one has
could be far more interesting than the belief that I’m wearing socks. For instance, I
can know well, in a pretty first-personal way, that I love someone, or that I really
want to pursue a career as a pianist. So the charge of triviality that Cassam makes
against first-person self-knowledge is perhaps surreptitiously directed at the wrong
target, that is, at the way of knowing as opposed to known contents, which may be
trivial at times, but very interesting and constitutive of the our own psychological
identity on other occasions. By selecting an example of trivial first-personally self-­
known content, he makes it appear that the triviality regards the method of knowing
as such.
On the basis of such rhetoric, Cassam thus argues that, in so far as third-personal
self-knowledge contributes substantial self-knowledge  — which matters to us
because it discloses a vast portion of our mental life that could not be discovered
first-personally —, it deserves an investigation making it more central than it has
been in the last few decades in philosophy.9 Also, Cassam thinks that, most appear-
ances notwithstanding, possibly all forms of self-knowledge are reached third-­
personally, via an inferential route that is often psychologically hidden from the
subject. Other scholars are  unconvinced by the genuine character of first-person
self-knowledge. Some claim that there is no such thing as first-person self-­
knowledge, and even among those who think there is, many pose limits to it, by
claiming that it is restricted in scope.10 Virtually all agree that substantial self-­
knowledge in Cassam’s sense is reached third-personally instead, that is, by using
inferences, possibly to the best explanation.11 Those inferences start from the

8
 Cf. Coliva 2016 and elsewhere.
9
 There might be reasons to express discontent with such rhetoric. One might perceive Cassam’s
discourse as ultimately offering another among the numberless pictures of academic philosophy as
too little in touch with the understanding, and self-understanding, needs of non-academic folks.
Occupied as it is with subtle investigations, philosophy all too often would prove to be too distant
from the kind of engagement society would expect from the discipline and its representatives
(Coliva 2015).
10
 Cf. Coliva 2016, pp. 1–18.
11
 Coliva 2016 explains that inferential self-knowledge does not always use inference to the best
explanation, but also induction and abduction. In addition, in the context of presenting the wide
variety of methods of getting to know different kinds of mental states, she also discusses how simu-
lation and linguistic substitution of one’s behavior, along the lines described by expressivists, can
proficiently be used. Cf. Coliva 2016, pp. 217–240.
The ‘Crux’ of Internal Promptings 55

o­ bservation of one’s own verbal and non-verbal behavior, other people’s reactions,
and, as Cassam and others have recently pointed out,12 from taking into account the
psychological evidence of other mental states one has.
My target in this paper is the theory of substantial, third-person self-knowledge
that Cassam defends in his book, particularly his advocacy in favor of a quite broad
inferentialist theory of substantial third-person self-knowledge, which he promises
to extend virtually all mental states. I will specifically focus on his claim that it
extends not only to mental states, values, abilities, emotions, affective states, and
character traits that could not be known first-personally, but also, interestingly, to
the so called “internal promptings. It is the prospect of extending Cassam’s inferen-
tialist theory of third-person self-knowledge to internal promptings that I will
address next. Let me then begin by attempting a clarification of what “internal
promptings” are.

2  C
 assam’s Inferentialist Theory of Self-Knowledge
and the Problem of “Internal Promptings”

“Internal prompting” is a felicitous technical phrase Cassam borrows from Lawlor


(2009) to indicate experiential episodes whose phenomenology may not be entirely
grasped by the subject. Giving a clear definition of what internal promptings are is
not easy, and this seems in any case to depend directly on their very phenomeno-
logical and metaphysical nature. Cassam himself offers no strict definition of them.
Thus, one might try to approach a clarification of what they are by sketching out
how they tend to appear to the subject when they occur, and by trying to relate them
to the mental states they presumably stand for. For instance, people who have tried
to capture their nature say that they may often appear to be “some kind of inarticu-
lable hunch, intuition” (Kirsch 2015, p.189), the sense that something emotional, or
affective, or even judgmental,13 is going on inside oneself, without its metaphysical
basis and phenomenological character being explicitly manifest. Or else, they may
feel like an “incipient desire” (Boyle 2015, p. 344), whose propositional basis and
object are not totally revealed. In all, it does not seem incorrect to say that they are
somehow elusive experiential episodes, of the type one may find it really hard to
discriminate or identify. However, one thing about them seems to be patent: they
present themselves spontaneously, and in a way that many would be inclined to
characterize as psychologically direct, immediate, and self-intimating. Their direct,
immediate, self-intimating manifestation fits them prima facie to be the perfect can-
didates for representing mental episodes one knows precisely first-personally, even

12
 Cf. Coliva 2016, Chap. 8.
13
 I think it is not entirely obsolete to imagine the possibility of a feeling that “something judgmen-
tal” is going on, although this is the topic for another paper.
56 P. Pedrini

if it is hard to characterize, conceptualize, or clearly subsume them into fully cata-


logued major mental state categories.14
It is precisely here that the “problem” of internal promptings arises. On Cassam’s
view, their spontaneous surfacing does not indicate that our access to them is non-­
inferential, and thus first-personal. Rather, Cassam thinks that, on closer inspection,
they are often known precisely via the third-personal inferential route that he tries
to show is more ubiquitous than defenders of first-person self-knowledge have thus
far conceded. There is a passage in the book where he admits that there may be
promptings whose access might be first-personal (p.  154), but immediately after
(pp. 161–165) he develops the argument in such a way as to lay down an explicitly
all-encompassing inferentialist theory of self-knowledge, which is claimed to cover
internal promptings fully as well.
Some scholars have reacted immediately and expressed reasonable doubts about
the feasibility of such a project. Among critics, Coliva (2016, pp. 87–88) has clearly
defended the idea that Cassam really has no inferentialist story about how we access
internal promptings and objected that, until he does not offer such story, there will
be no credible inferentialist account of it. In a word, Cassam cannot justifiably claim
that his inferentialist theory can really cover self-knowledge of internal promptings
as well. Perhaps, that does not amount to a charge of petitio principiis, as Cassam
does not want to explain first-person self-knowledge by presupposing it. He actually
wants to say that there is no such thing as first-person self-knowledge, but rather an
ubiquitous form of third-person, inferential self-knowledge, which applies even to
what we think is genuine non-inferential, first-person self-knowledge.
Now, while I agree with Coliva that there is a clear sense in which Cassam in fact
lacks a story about how inferentialism can apply to internal promptings, I am also
convinced that there is a conception of knowledge under which Cassam could justi-
fiably claim that he does have such a story, although he neither makes use of it
explicitly nor acknowledges its existence. By failing to do so, Cassam unnecessarily
complicates the debate on the extent of third-person self-knowledge, and lays his
theory open to a far more serious objection. My general strategy will then be to
show that Cassam is right in claiming that it is by inference (possibly to the best
explanation) that we can adjudicate the epistemological question regarding what
internal promptings are. However, their initial manifestation in our psychological
experience as raw, unconceptualized promptings tells us that the capacity we have
to notice them, without identifying them, nonetheless amounts to some kind of

14
 Famously, in the context of emphasizing how bad human beings are at introspecting certain
mental states, especially emotions, Schwitzgebel (2008) claimed that a vast arrays of states we
have are unclear to us. His discussion was aimed to make a strong case against the reliability of
introspection, and brilliantly focused on the very opaque phenomenology of those states that we
are unable to introspect correctly. Contrary to Schwitzgebel, as it will be clear to the reader, my
project is not to invite any sceptical verdict on self-knowledge. Rather, I invite closer analysis on
cases of elusive phenomenology of mental states, with a view to adjudicating metaphysical and
epistemological questions that may turn out to be useful to chart more accurately both the mind
and the methods of self-knowledge. Also, I do not use those cases at the expenses of the existence
of fully conceptually clear mental states that we do have most of the time.
The ‘Crux’ of Internal Promptings 57

s­ elf-­knowledge, which does not seem to be third-personal. In order to accomplish


the task, in § 3, I will distinguish two epistemological conceptions of first-personal
self-­knowledge: (1) “self-knowledge by discrimination, or by identification”, and
(2) “self-knowledge without discrimination, or identification”. I will argue that
while the former may not apply to internal promptings, the latter does. This is the
reason why I agree with Coliva (2016) that an inferentialist theory of self-­knowledge
does not in fact cover self-knowledge of internal promptings. However, this failure
does not depend on lacking a story on the extensibility of inferentialism to internal
promptings as such. Rather, Cassam’s theory is faulted by an independent, and prec-
edent, amphiboly fallacy affecting the concept of self-knowledge he makes use of.
Cassam treats the concept of self-knowledge without bringing to light the distinc-
tion I make, while nonetheless trading on it, thus creating the conditions for a charge
of amphiboly. So, although Coliva’s objection may not immediately apply, her ver-
dict on the non-extensibility of an inferentialist theory of self-knowledge to internal
promptings remains unaffected. Bringing the distinction between the two concep-
tions of self-knowledge right to the fore may help to disentangle the hidden reasons
for the controversy on internal promptings.
If unidentified, raw internal promptings are first-personally, although non-­
discriminatively, known, i.e., if we can’t “give a name to them”, or subsume them
under concepts, they both promise to signal a serious crux lying at the very core of
Cassam’s allegedly inclusive inferentialism about third-personal self-knowledge,
and also provide the basis for a more general epistemological issue regarding our
very conception of knowledge that is worth exploring. For it may be true that self-­
knowledge without discrimination, or identification, may be perceived as quite
modest, as a kind of self-knowledge, but it has two advantages:
a) respecting the elusive, opaque, phenomenologically inexplicit nature of an

important class of internal promptings;
b) explicitly enriching our epistemological inventory of self-knowledge, within
which it is important to include varieties of self-knowledge that respect the elu-
sive, opaque, phenomenologically inexplicit nature of an important class of our
mental states, which are, nonetheless, known precisely for what they are: opaque,
vague, unclear, phenomenologically sub-indicative, and intrinsically elusive.
Accordingly, on the basis of such distinction, I will the defend a view of self-­
knowledge that explicitly allows both conceptions.
Having explored how raw promptings are most likely known, and under what
conception of self-knowledge, in § 4, I will discuss to what extent internal prompt-
ings can function as a basis for inferences to (discriminative, or identificatory) self-­
knowledge, as Cassam claims. Particular urgency seems to attach to the case in
which the vagueness of internal promptings reaches a considerable level and their
phenomenology gives no clear conceptual indication to the self-interpreting
subject.
In § 5, I will discuss further the topic of vague promptings, with a special focus
on a much neglected issue: the possible under-determination of inner experience
58 P. Pedrini

with respect to the conceptual schemes used to categorize it. I will argue that even
if conceptual and empirical investigation can help a subject to subsume at least
some raw experiential episodes under concepts, a portion of them may run the risk
of under-determination with respect to the conceptual schemes used, which in turn
indicates one important reason for failure of self-interpretation. This may lead us to
a threat of relativism that cannot easily be defused. I will make use of the notion of
“residual psychic tension after concept application” to intrinsically elusive or
opaque promptings to explore how and when such tension can prove to be a useful
clue to the soundness of concept application.
At the end of the paper (§ 6), I will take stock of the intrinsically elusive nature
of a vast portion of our own mental states, which cohabits with an ample domain of
mental states that are conceptually well (or better) defined instead. In light of the
empirically manifest diversity of internal promptings, I will express sympathy for a
wider view of the geography of the mental, which I recommend for methodological
reasons, guided by a principle of inclusiveness of evidence. Giving citizenship to
richness and diversity within the mind can greatly widen our descriptive picture of
the mind, which can then lend itself to more complete and refined explanations.
Also, it gives us the chance of elaborating a more refined epistemology of how, why,
and under what conditions we can know the most vague, opaque, and elusive parts
of our own mind.

3  S
 elf-Knowledge of Internal Promptings: Discrimination,
Identification, and Opacity

According to Cassam,“inferentialism hasn’t exactly been a popular approach to


self-knowledge”, and “(t)he assumption that intentional self-knowledge is normally
immediate (…) is rarely defended”, because “it’s just seen as obviously correct”
(2014, p. 137). Cassam disagrees: he thinks that this “is not obviously correct, and
that inferentialism is a live option” (ibidem). He specifies that he is mainly commit-
ted to applying inferentialism to “standing” attitudes rather than to “occurrent”
ones (p. 138), that is, to those kinds of attitudes (beliefs, desires, hopes, etc.) that are
often also characterized as “dispositional”. Standing attitudes are not conscious
mental events such as believing or judging. Rather, they are attitudes that “remain
in existence when you are asleep” (ibidem). He importantly thinks that the evidence
used to draw inferences to the standing attitudes we have need not be behavioral
only. There are other much neglected possibilities, such as our “judgments, inner
speech, dreams, passing thoughts and feelings” (p. 138). Also, the inferences he has
in mind when dealing with inferentialism need not be conscious, but rather supplied
by unconscious cognitive mechanisms that do not surface in consciousness and thus
operate unbeknownst to us (pp. 138–139). He claims that deploying unconscious
inferences is crucial to his inferentialism because it neatly defuses the risk of over-­
intellectualizing third-person self-knowledge, which would otherwise be
The ‘Crux’ of Internal Promptings 59

cognitively overloaded by what would, for the subject, be the quite demanding task
of performing all of the required inferences consciously (pp. 139–140).
He then goes on to analyze an interesting set of inferences, i.e., inferences from
“internal promptings” (p. 143). He borrows the phrase from Lawlor (2009, p. 60),
who has described inference from internal promptings in the context of explaining
how we know what we want. Lawlor focuses on desires that are not easy to know,
such as the desire for another child. She takes the example of Katherine, “who
feels,” as Cassam puts it, “there is a fact of the matter about her desire for another
child but struggles to know the answer to the question ‘Do I want another child?’”
(2014, p. 142). In the nice description of Katherine offered by Lawlor, the woman
is represented as “noticing her experiences and thoughts. She catches herself imag-
ining, remembering, and feeling a range of things” (2009, p. 57), including how a
newborn feels in one’s arms, or feelings perhaps pointing to an envy of another’s
pregnancy. This may be enough to make the self-attribution of wanting a child, and
once the attribution is made, Katherine “may feel a sense of ease or settledness”
(2009, p. 57).15 Clearly, such self-attribution is a cognitive achievement that does
full justice to the inferentialist character of third-person self-knowledge. The knowl-
edge Katherine acquires could not be reached first-personally. In fact, she did not
know much about her standing desire for another child before the reflection and the
inferential work on the evidence offered by most, if not all, of her internal
promptings.
Now, it would seem that, even if her third-person self-knowledge about her
standing desire for another child is reached inferentially, at the very least the knowl-
edge she has of her own initially unclear internal promptings must have been non-­
inferential. After all, they constituted the self-intimating psychological evidence,
immediately available to her, which formed the basis on which she performed the
necessary inferences. So, Cassam actively asks, “what’s the story about access to
internal promptings?” (p.  154). Giving such a story, Cassam specifies, would be
crucial for all those who are interested in understanding whether internal prompt-
ings could be considered as the basic self-justifying mental items that one usually
invokes to block the threat of a vicious regress to which an all-encompassing infer-
entialism could easily fall prey. As he says later in the book, this is why “the ‘infer-
ence all the way’ option looks unattractive” (p. 161). From the point of view of all
those fearing such a threat, it may seem “(b)etter to accept that self-knowledge of
internal prompting is non-inferential” (p.161), and thus subscribe to a hybrid view
of self-knowledge, which includes both inferential self-knowledge and first-person
self-knowledge. He admits that even if his “inferentialism is specifically a view
about knowledge of our own standing attitudes, (…) just because knowledge of
standing attitudes is inferential that doesn’t mean that all other self-knowledge is
also inferential” (p. 154). One alternative would then be “to say that we have non-­
inferential access to our own inner speech, fantasies, judgments, etc., and that, given
the appropriate theory of mind, we are then able to infer our own standing attitudes

 I will get back to the sense of ease or settledness in § 4, where I will discuss how we can use it
15

as a track towards the soundness of concepts application.


60 P. Pedrini

on this basis” (ibidem). Getting back to the example of Katherine, she is aware of a
range of feelings, emotions, and mental images from which she correctly infers that
she wants another child. As to the question “How, then, does Kathrine know her
evidence?” (…) (n)oticing that you have a particular feeling F is a way of knowing
that you have that feeling, and is different from inferring the you have F″, so that
“Katherine would discover her standing desire for another child by inference from
internal promptings which she knows about by means other than inference” (p. 163).
But “(s)aying that Katherine ‘notices’ her feelings doesn’t cast much light on the
nature of her knowledge of her feelings” (ibidem).
But Cassam does not think that embracing such a hybrid view is a good idea: “
(…) (E)ven if it is open to inferentialism to adopt such a hybrid approach to self-­
knowledge, it’s a further question whether it should adopt it” (ibidem, Cassam’s
emphases). There are two main reasons why Cassam thinks we should not “go
hybrid”: the first one is that no one will be terribly impressed, after all, by an account
of self-knowledge that only talks about how we inferentially know our standing
attitudes, while it “has nothing to say about any other self-knowledge” (pp. 159–
169). Perhaps, he thinks that, on a hybrid view, there will be nothing to say about
other forms of self-knowledge just because he believes that these other varieties are
both trivial and depicted along the lines of the immediacy thesis of self-knowledge,
which he finds overrated to say the least. The second reason why we should not
adopt a hybrid view is that the “‘inference all the way’ option doesn’t need to risk a
regress” (p. 162). In all, “(t)here are (…) excellent reasons for thinking that self-­
knowledge of internal promptings is inferential” (ibidem).16
I am not so much concerned here with the kind of defense Cassam offers about
how the threat of a vicious regress is not in order, although I will say something on
foundationalism later in the paper. Instead, I wish to focus on why he thinks that
self-knowledge of internal promptings is inferential. By appealing to Peter
Carruthers’ The Opacity of the Mind (2011) and Carruthers’ 2009 paper, Cassam
explains that when one infers a mental state from a set of internal promptings, one
is just embarking on an interpretative process (p. 162). Such a process is not simply
interpretative, but also clearly inferential. For it is “implausible that there is such a
thing as a ‘raw feel’ of a yearning for another child. The feelings we classify as such
are subtle and complex. (…) it takes cognitive effort to identify them as amounting
to a yearning for another child (…). When you identify your feelings as a yearning
for another child what you are doing is interpreting them, and your cognitive effort
is the effort of interpretation. Crucially, when you interpret your feeling you don’t
just go on ‘how it feels’. You also take account of contextual factors, (…) which
means that you are to some extent inferring what you feel from your background

16
 It is obviously important for Cassam to show that there needn’t be a regress, but I don’t think that
this is more than an additional reason to adopt the inferentialist model of self-knowledge of inter-
nal promptings. Even if he wouldn’t be able to block the regress, and even if some are convinced
he does not in fact block it (cf. Coliva 2016, pp. 86–88, and Coliva 2015), his idea that internal
promptings are inferentially known is, as I will try to show, based on a discriminative, or identifica-
tory conception of self-knowledge that allows him to defend inferentialism, however things stand
about the threat of a regress.
The ‘Crux’ of Internal Promptings 61

knowledge. Your inference is inference to the best explanation rather than inductive


or deductive” (ibidem, Cassam’s emphases).
With these cards on the table, we are in a favorable position to begin to disen-
tangle what Cassam is really talking about. While he may be right to claim that it is
by inference (possibly to the best explanation) that we can adjudicate the epistemo-
logical question regarding what internal promptings are, or what kind of conceptu-
alized mental states they induce us to form, the fact remains that their initial presence
in our psychological experience as raw, unconceptualized internal promptings tells
us that the mere ability we have to notice them as raw and unspecified does nonethe-
less amount to some kind of self-knowledge. I’m inclined to agree with Coliva
(2016) that such self-knowledge could hardly be inferential, and thus that it is first-­
personal. However, what matters for my present purposes is that it is a kind of first-­
person self-knowledge that does not involve the discrimination, or identification of
internal promptings.
To see why this kind of self-knowledge nonetheless qualifies as knowledge, I
think it useful to distinguish between two epistemological conceptions of self-­
knowledge, and of knowledge more generally. Let me call them, respectively, “self-­
knowledge by discrimination, or identification” (or self-knowledge*, for brevity),
and “self-knowledge without discrimination, or identification” (or self-­knowledge**,
for brevity):
–– Self-knowledge*:
This is a seemingly quite robust epistemological conception of self-knowledge
of our own mental states. It assumes that no one can be credited with having genuine
knowledge unless one can discriminate between kinds of states, or in any case iden-
tify them. On this conception, one can be required to tell a belief from a desire, say,
but the “contrastive” aspect of such knowledge need not be as central. For it may
well be enough for a subject to be able to identify a certain mental state, without
also having any contrastive opposition in view. Of course, given a reasonably
acceptable amount of mental holism, possessing the concept of belief, say, may be
relationally connected with the ability to distinguish it from other kinds of attitudes.
But I think that identification need not necessarily be connected with discrimina-
tion. One might imagine a subject who has beliefs only, say, and no other attitudes
at all. She could have a raw prompting about there being a certain belief pressing
under the surface of her consciousness to become manifest. If she can get to identify
the belief for which the prompting stands, or to which it points, then she can get to
identify it without also discriminating it in contrast with anything else. Be that as it
may, it is interesting that this not the only available conception of self-knowledge.
–– Self-knowledge**:
In a definitely less robust epistemological fashion, one can get to know one’s
own mental states without discriminating, or even identifying them at all, or not
sufficiently. On such a conception, the subject can still be credited with enjoying
some kind of knowledge in so far as she knows that she has certain experiences,
even if they are not conceptually or phenomenologically clear. She is unable to tell
62 P. Pedrini

what they are, nor can she say what attitude a prompting, or set of promptings,
they stand for. As a matter of fact, she can simply limit herself to “noticing” them,
to put it in Cassam’s terms, as they manifest — no more, no less. Such an account
has an obviously modest air: although the subject feels “something”, she does not
grasp what is really going on inside herself. From the point of view of a defender of
self-knowledge*, it seems fair to say that a subject having self-knowledge** cannot
really be said to know her psychological situation.
Now, despite the modesty of self-knowledge**, I think it is important to keep it
in the picture because it has the advantage, as we will see, of respecting the intrinsi-
cally elusive, opaque, phenomenologically inexplicit nature of an important class of
internal promptings. The intrinsic elusiveness, opacity, and phenomenologically
inexplicit nature of many internal promptings is an aspect that Cassam does not
consider, nor do his critics. In the passages quoted, he seems convinced that virtu-
ally all raw promptings stand for, or point to, attitudes that can in principle be speci-
fied —  barring of course all the usual possibilities of fallibility involved in
interpretation. Even if one grants that fallibility looms in any cognitive achieve-
ment, such as interpretative processes, a view that implies an underlying match
between raw promptings and their corresponding standing attitudes is a pretty opti-
mistic view indeed of the mental. Let us call it Cassam’s “matching assumption”.17
Not all promptings, however, are the “phenomenologically diminished” format
of the corresponding third-personally discoverable, full-blown standing attitude, or
set of attitudes. I will discuss this more extensively in § 4 and § 5. For the time
being, I think it is important to emphasize that, most likely, Cassam makes use of
the discriminative, or identificatory, conception of self-knowledge of internal
promptings (or self-knowledge*) when he claims that we acquire self-knowledge of
them through interpretation. If this is correct, then he really does have an inferen-
tialist story about self-knowledge of internal promptings. He has it because what he
is really after is an explanation of the self-knowledge* of internal promptings. If
things stand as I think they do, then when Cassam explains inferential self-­
knowledge* of internal promptings he is not also explaining non-inferential self-­
knowledge** of internal promptings. Thus, technically, he is not presupposing what
he is trying to explain. He certainly relies on non-inferential self-knowledge** of
internal promptings when he claims that they provide psychological evidence, but
his theory is not also in the business of clarifying inferential self-knowledge** of
the evidence as such. So, rather than being charged with a circularity fallacy, he may
be charged with a fallacy of equivocation between two different conceptions of self-­
knowledge that he does not explicitly bring to the fore when he maintains that his
inferentialism is all-encompassing.18

17
 In passim, it’s interesting that a defender of the less than ideal nature of human beings such as
Cassam has such a faith in there being a positive match between raw promptings and their corre-
sponding standing attitudes. Cf. Cassam’s discussion of how any of us can hardly be thought to be
the “homo philosophicus” that has too often been disputably used in philosophy to make theories
about self-knowledge right (2014, Chaps. 1 and 2).
18
 There is a locus where he gets close to gesturing more clearly towards this distinction, without,
however, thematizing it fully. Cf. Cassam 2014, p.168.
The ‘Crux’ of Internal Promptings 63

Disentangling the equivocation between self-knowledge* and self-knowledge**


also helps us to see why non-discriminative, or non-identificatory, self-knowledge
of internal promptings is rightly thought by Coliva to be non-inferential. What
Coliva might be willing to claim is that access to our promptings, prior to interpreta-
tion, can but be a case of self-knowledge**. The distinction in question thus disam-
biguates the epistemological disagreement between them created by internal
promptings. If raw internal promptings, before the interpretative work on them lead-
ing to self-knowledge*, tolerate an initial non-inferential access via self-­
knowledge**, internal promptings constitute a serious crux for Cassam’s allegedly
inclusive inferentialism about third-personal self-knowledge, even if they can be
used later on as a basis for inferences aimed at getting to know* them fully.
Acknowledging internal promptings accessed via self-knowledge** is important
both for a descriptively accurate epistemology of self-knowledge and for the meta-
physics of the mind. If mental items to which we can’t “give a name ”, and that we
can’t subsume under concepts, exist, they are still “something” that happens in our
psychology and that we notice. In addition, by accounting for how we get to know
that this “something” is there, we allow for the possibility of a variety of knowledge
that makes us know* that there are items that we don’t, or can’t, know**.19
One might ask whether the acknowledgment of the possibility of initially raw
psychological data in any way compels us to accept that such data should also be
thought to work as the foundation for the edifice of knowledge. I don’t think that
self-knowledge** need commit per se to accepting any form of foundationalism.
Foundationalism is a project. As such, it may be in order if one is in search of basic
epistemic items to be used as the basement of the edifice of knowledge. But such a
project about internal promptings is in no way necessitated by the very nature of
internal promptings; rather, it is an additional epistemological plan a thinker might
have if she is independently inclined to adopt a foundationalist view of knowledge
as a whole. It is not my aim here to single out non-inferentially accessed raw
promptings in order to block the threat of regress created by inferentialism by find-
ing the foundation of knowledge. I am developing an analysis aimed mainly at
defending the existence of raw, opaque, non-discriminated, or non-identified
promptings with a view to disentangling the issue of the conception of self-­
knowledge Cassam makes use of. But I am not in the business of defending raw
psychological data because I am looking for a theory of self-knowledge that shows
us that there is an indubitable basis for our knowledge of the mind, or the world, or

19
 Whenever one allows for self-knowledge**, there always seems to be a certain risk of conflation
between first-person awareness of one’s mental states and first-person self-knowledge of one’s
mental states. I have some ultimate sympathy for such view, but I think more should be done to
establish to what extent such theoretical move might still be deeply wanting. In particular, if we
deny the there is an epistemic relation between a subject and her object of knowledge, something
epistemically crucial seems to get lost. Coliva 2016 argues in favor of a view of first-person self-
knowledge of some kinds of mental states where such epistemic relation is lacking, and where the
term “knowledge” becomes “more the — ‘grammatical’, as Wittgenstein would have it — signal
of the absence of room for sensible doubts and ignorance” (p. 15).
64 P. Pedrini

both. On the contrary, as I will explain in the next paragraph, at least some raw
promptings can easily be shown to be entirely unfit to function as the basis for fur-
ther knowledge, for their opacity may on occasion be so deep that they can hardly
function as a reliable evidential ground for drawing inferences with any hope of
being conducive to self-knowledge*.
Outlining a more refined description of the mind, independently of any overarch-
ing epistemological project, analyzing the nature of the opacity of an important and
pervasive class of psychological items we all live with, and understanding the gen-
eral lesson we should reasonably draw from their existence is my primary interest.

4  Self-Interpretation and Internal Promptings

Independently of the story one is inclined to give about our initial access to internal
promptings, both Cassam and his critics (cf. Coliva 2016, Boyle 2015) grant that
internal promptings, along with the contextual features and theory of mind one has,
offer the crucial evidential basis upon which one can draw inferences to reach self-­
knowledge* of what internal promptings are, or what attitudes they stand for. In this
paragraph I wish to question to what extent internal promptings can function as a
basis for inferences aimed at leading the subject to acquire self-knowledge*. For
reaching self-knowledge* about what they are, or what attitudes they stand for, may
be hard to do in the event of their vagueness reaching a considerable degree and
their intrinsically inexplicable phenomenology giving no clear conceptual indica-
tion to the subject who self-interprets. Such features of some internal promptings
constitute a serious threat to the success of self-interpretation — a threat that has
been greatly underestimated in the literature. Authors recently dealing with third-­
person self-knowledge, including Cassam, are accurate at listing and analyzing
quite interesting risks of fallibility in a subject’s self-interpretative effort. Both
Cassam (2014), Coliva (2016, 2018), and Boyle (2015) carefully describe factors of
various nature affecting the reliability of self-interpretation, and cognitive achieve-
ments more generally. Cognitive enterprises such as the reconstruction and interpre-
tation of reality, be it mental or non-mental, can be negatively affected by purely
cognitive factors, such as the misapplication of concepts, sub-optimal inferential
operations, limitations of reasoning, the well-known limits to a sound cognitive
treatment of evidence due to the finitude of human psychology and resources, mere
ignorance or lack of concepts, and also motivational factors.20 Luck may also play a
role in our attempts to get to know reality as a whole.
Although I completely agree that all of these factors are relevant for the purposes
of defining and diagnosing how and why self-interpretation can go wrong, I am
interested in focusing here on a sub-class of internal promptings that provide

 The debate on motivated irrationality shows a wide variety of factors affecting cognition as a
20

whole, and hence any cognitive achievements. Cf. Carter & Pritchard 2018, Coliva 2018, Kirsch
2018, Sims 2018, Zawidzki 2018.
The ‘Crux’ of Internal Promptings 65

a ­troublesome basis for inference, not necessarily because of the influence of any of
the factors listed above, but rather because — for reasons to be qualified — they are
intrinsically opaque. In other words, I am rejecting Cassam’s “matching assump-
tion”, i.e., the idea that there is a match between virtually all internal promptings
and the corresponding standing attitudes they allegedly stand for. To my knowledge
there is no explicitly dedicated analysis of that possibility, which we should not
consider significantly far-fetched. After all, most of us can recall experiential cases
in which all of our efforts at self-interpreting did not really deliver any answer to our
question “What is it that I really feel?” I’m using the term “feeling” here not because
I wish to restrict the scope of opacity to feelings only, but because of the experiential
nature of internal promptings. Quite broadly, the term is meant to cover a vast array
of felt experiences, which might in principle signal all sorts of mental states, includ-
ing states that are not traditionally believed to have a phenomenal character or a
distinctive phenomenology, such as propositional attitudes (cf. Shoemaker 1994a,
1994b, 1994c). Personally, I don’t think that propositional attitudes really have no
phenomenology of their own, in the sense that having a propositional attitude, either
occurrent or dispositional, actually always make a more or less “felt” difference in
our overall experience, owing to its relational properties with other experiential
states, but defending such a view would obviously take me too far from my present
purposes, and it is, indeed, a project for another paper.21
Cassam’s assumption that internal promptings can at least offer indications to the
self-interpreting subject surfaces in other excellent analyses of Cassam’s work, such
as Boyle’s (2015) and Coliva’s (2016). In the context of discussing whether
Cassam’s analysis of the “Transparency Method,” as it is known, for self-knowledge
is ultimately tenable, Boyle seems to agree that internal promptings can give the
subject an input that can serve as a basis not for inferences, but for “making up one’s
mind” about what we really feel, think, desire, and so on (p. 344). Even if making
up one’s mind is ultimately a deliberative process, there is no reason to think that
such deliberation is arbitrarily voluntaristic; rather, it is often just necessitated by
what promptings reasonably suggest. So it can involve inferences as well. Now,
even if these are all lively psychological possibilities, there seems to be no princi-
pled reason why we should not think that at least some internal promptings simply
cannot offer any indication whatsoever, given their irresolvable, intrinsic opacity
and vagueness. An example may illustrate the point. Suppose that a subject, call him
Paul, is in a quite obscure psychological predicament. He is not completely settled,
but it is entirely unclear to him and to everybody else to whom he reports his state
of mind what, if anything, is wrong with him, and what is really going on inside
him. His life is not functionally compromised in any way; he eats, sleeps, and has
quite ordinary social relations. Yet, he has a vague sense of “something” going on,
something he does not identify. It seems to him that there are dissatisfying psycho-
logical items he somehow feels, to which, however, it is extremely difficult to give

21
 Coliva 2016 discusses Peacocke’s proposal (1999) that there is a phenomenology for occurrent
states (pp. 107–111). It goes beyond the scope of this paper to discuss the reasons of Peacocke’s
claim, and the objections Coliva moves to his account.
66 P. Pedrini

a name. He somehow feels a changeability, a moody state of mind that is quite elu-
sive. Perhaps he has an incipient, or transient, mood disorder, but of an absolutely
minor and almost imperceptible kind. Neither clinicians nor friends grasp what it is
that he is trying to voice and think there are no reasonable margins for either phar-
macological or psychotherapeutic treatment. What he feels is, and remains, really
obscure.
Another nice case is offered by the art of cinema. In an interesting scene in the
movie, Boyhood, the young main character is with his father in a bar and they are
somehow obscurely reflecting — more with silences than with words, as can be the
case only in a well-written and acted screenplay — about several things. Broken-off
sentences, uncertain facial expressions, and other almost imperceptible clues seem
to signal that what they are in some way doing is approaching questions about the
meaning of life. At a certain point, when the cinematographic time is ripe, the young
boy, with manifest uncertainty, as if he were not even sure that the very sentence he
is about to pronounce is appropriate, or makes sense, stunningly asks his father what
the point of “all that” is. Just that. Equally wonderfully, his father, with no further
prompting, seems to get the obscure point. In line with his own uncertainty, with a
sublime mix of surprise and human solidarity, he replies that if there is any point at
all, nobody knows what the point is. Just that. They (and we) are left with the sense
they were onto something, perhaps just the need for a sense about all they were
going through in life, but it remains totally undisclosed to them what they even
meant to ask, reply, and — ultimately — what exactly they were feeling at all. In a
Wittgensteinian spirit, we might say they were onto something we really can’t talk
about. Such a predicament is, and remains, intrinsically opaque.
It might be objected that in Paul’s case and in the Boyhood scene, the promptings’
opacity and vagueness might- just, in fact, be due, after all, to some of the factors of
fallibility listed above. If this were so, then there might be the risk of their opaque-
ness or vagueness not being intrinsic, but conferred, because it is traceable back to
something else, perhaps, in principle, also removable. I don’t deny that for some
promptings this might be precisely the case. However, what I am after is not the case
in which we can bring promptings back to clarity by virtue of removing influencing
factors. Rather, I am after cases in which, whether or not there are interfering factors,
promptings are in themselves irreversibly obscure. In this sense, when I refer to an
intrinsic character, I’m talking about a character of promptings which is either spon-
taneous or has a causal provenance that, however, irreversibly made it what it cur-
rently is — in the same sense in which, say, a piece of wood, once burned, is just ash.
Once we admit of the possibility of irredeemably obscure promptings, some con-
siderations regarding an old issue long thrown around in both philosophy of mind
and epistemology return to the fore, namely, the relationship between psychological
experience and concepts. Although it may be quite easy to apply most concepts
proficiently to a vast array of phenomenologically recognizable internal promptings
(the cases, referred to by Cassam, of inferential self-knowledge of “‘simple’ sensa-
tions”, 2014, p. 164), there may equally be promptings for which we simply lack
concepts. I cannot embark here on an investigation of why this may be so, nor do I
think any such investigation would necessarily turn out to be explanatory beyond a
The ‘Crux’ of Internal Promptings 67

certain point, as this might just be a brute fact of human psychology, the
Wittgensteinian point at which all explanations just come to an end. I am claiming,
descriptively, that it seems empirically evident that some felt experiences just do not
seem as if they can possibly be cut at their joints by any available concepts. It is
important to note that the lack of concepts I am referring to is not a local lack of
concepts that a subject suffers from, and which could be solved by giving her the
already existing concepts. There may obviously be cases in which an accurate con-
ceptual and empirical investigation can help a subject to subsume at least some
unclear promptings under concepts, and successfully so. For instance, in a clinical
setting, an analyst can progressively offer a patient a conceptual and evidential story
that can greatly help her to understand herself and what is really going on in her
psychological experience. Before being offered such a conceptual theory, and being
directed toward the relevant evidence and its connections with the concepts used,
the patient was utterly unable to interpret herself. After all, she went to the analyst
precisely because she wanted to reach a self-understanding that her lack of concepts
denied. Now that she has relevant and scientifically reliable concepts, which excel-
lently capture her inner life, she can identify what happens to her. Cases like these
show that once we come up with the appropriate concepts, self-interpretation can
eventually come about successfully.22 What I am trying to picture instead is the pos-
sibility of a virtually permanent lack of concepts capable of grasping some of our
raw experience, and which is due to impenetrable lacunas that are present in the
current human conceptual inventory.
Of course, we can still hope that, with the general progress made by human
inquiry into our psychological lives, we might be able to forge concepts for han-
dling those cases successfully. The worth and usefulness of such an ideal should, of
course, be assessed against the background of a wider philosophical project  —
about, perhaps, the very mission of philosophy — which I have no hope of begin-
ning even to outline. Thus, for present purposes, I suggest neither that we should
pursue such a line of thinking nor that we should not. I simply suggest that, at pres-
ent, we should record the evidence that we just don’t have all the concepts we would
need to discriminate and identify important pieces of our own experience.
This may sound like the reaffirmation of the popular refrain that the mind is
largely mysterious. But the brute inexplicability of interesting portions of our inner
life should not be considered as somehow epistemologically unique. There is com-
parable inexplicability regarding vast portions of outer reality as well. Reality as a
whole does not seem to bear complete coverage by available concepts. This must be
the sense of the famous, all in all optimistic line attributed to Bertrand Russell,
according to which “the world is full of magical things patiently waiting for our wits
to grow sharper”. Whether we should also be confident in the capacities of our
future intellect to develop far enough to grasp them is another issue.
Be that as it may, the very attempts, often unsuccessful, to grasp our internal
promptings and our inner life more generally clearly tell us that not all experience

 This is why Coliva correctly argues that we can improve and affinate our conceptual capacities
22

and concepts application proficiency. Cf. Coliva 2016, 2015, 2018).


68 P. Pedrini

is offered through concepts, nor can it all be harnessed by the available ones. This
fact invites us to continue to examine the relationship between concepts and experi-
ence, and between mind and reality, where reality, here, clearly includes the meta-
physical characteristics of the mind itself. If it turns out to be the case that there are
pieces of mental reality that are neither conceptualized nor offer themselves to post
hoc, interpretative conceptualization, not only are we faced with brute facts of nec-
essary ignorance, but we are also called upon to address scenarios of skeptical chal-
lenge and possibilities of relativism. In the next paragraph I will examine one of
them: the under-determination of internal promptings with respect to conceptual
schemes.

5  U
 nder-Determination of Internal Promptings with Respect
to Concepts

Thus far we have seen that some of our internal promptings may just be recalcitrant
to conceptualization. We just cannot grasp what they are. However, this is not the
only problematic aspect generated by opaque promptings. A quite thorny problem
created by unclear promptings has to do with the possibility of their under-­
determination with respect to available conceptual schemes. That is to say, to a
certain extent, some unclear promptings might, indeed, just be harnessed under
various competing theories, which in turn deliver quite different interpretative
reconstructions of what they really are, or point to, with no significant experiential
loss. In a word, different conceptual theories would just be experientially equiva-
lent. Suppose there is a subject, call him John, who obscurely lives with an inexpli-
cable raw sense of being threatened. He does not know by what or whom he might
be threatened, nor is he even totally sure it is a sense of threat that he feels. He finds
himself somehow feeling fear about unknown sources of menace, or possible events
he cannot even figure out completely. All he knows is that a sort of hidden anxiety
without a name and an object often takes over and leaves him with the sense that he
should find protection, or otherwise seek an ease that he thinks there might be a way
to acquire. He is also progressively in search of an explanation for his general pre-
dicament, perhaps a metaphysical one. Suppose now that John talks to two different
people voicing and proposing two quite different metaphysics of the world as a
whole. He first talks to a spiritual master who explains to him that the world is domi-
nated by opposing energies that deeply affect the individual. One set of energies is
benign and guides the individual to peace and universal love, while the other set
leads a person to indulge in states of anxiety and uneasiness aimed at spreading
disorder, suffering, and lack of faith in life — present, future, or even eternal. The
master offers an excellent explanation of what John is going through. As far as John
can see, there is nothing in the master’s description of the world and the mind that
fails to account for the totality of his experience. The master also explains a method
The ‘Crux’ of Internal Promptings 69

for reinforcing the positive action of the benign energies, and promises that after
some training John can really begin to feel better.
However, John also talks to another person, a clinician. The clinician offers a
general view of what is going on inside him by appealing to a sophisticated theory
of obsession and the related anxiety. He has a story accounting for both the cultural
roots of John’s obsessional thoughts and the neurological basis of such a deteriora-
tion of his psychological life. As above, there is nothing in the clinician’s descrip-
tion of the world and the mind that fails to account for the totality of John’s
experience. The clinician also explains a method for reinforcing the positive action
of non-obsessional thoughts, and promises that after some training John can really
begin to feel better.
Undecided about the validity of these views and methods, and not having a prin-
cipled preference for one over the other, John tries both. As a matter of fact, he real-
izes that he obtains benefits from both methodologies. Not only do both of them
capture the totality of his experience through the concepts used to identify, make
sense of, and causally explain his predicament, but they also offer real relief.
Now, if the case above is reasonably described and a credible live option, as I
think it is, it shows that the only reason why one could be entitled to suggest that
John should follow either ways is meta-theoretical, not experiential. It also high-
lights the possibility of a quite deep relativism elicited by the vary nature of some of
our experiences as such, which cannot easily be defused, at least not by appealing
to their experiential features themselves. Pace McDowell (1994), this kind of expe-
rience really does not seem to constitute a reliable tribunal. When raw promptings
are such as those described in John’s case, they are just under-determined with
respect to at least some conceptual schemes.23
Fortunately, not all sets of raw promptings are equally under-determined. There
may be psychological predicaments in which applying one conceptual scheme as
opposed to another does make an interesting difference, which experience itself
clearly tracks. For after the wrong concepts have been applied, there can remain a
feeling of residual psychic tension that, more often than not, signals that something
must have gone wrong in applying the concepts. Let us call it “residual psychic ten-
sion after concept application” to elusive promptings. If we vary some details of
John’s case, we can appreciate how this residual psychic tension presents itself.
Suppose that, despite the appearances of no experiential loss in adopting the con-
ceptual scheme offered by the spiritual master, John’s training with the latter’s
method leaves him with a residual sense of uneasiness that does not cease torment-
ing him. For instance, even if he works to neutralize the malign energies, he still
finds himself very worried about their very existence. Such worry fuels his anxiety,
and his obsession about threats does not quit. So, after a while, John decides to go

 It would take me too far to express a more defined view on the post-McDowellean debate about
23

conceptual vs. non-conceptual content. I just refer to it here as a general issue I think we should
explore further in light of the possibilities offered by a more fine-grained geography of the mind.
70 P. Pedrini

back to the clinician and, by dint of following the treatment he suggests, John is now
really better. All of his residual uneasiness has disappeared.24
Now, it seems reasonable to think that application without tension could be proof
that the proposed concept application is correct, in so far as it cuts the mental phe-
nomena in question at their joints. If this were so, then measuring the after-­
application residual tension could prove to be a useful clue to the soundness of the
concept application itself. This must be why Lawlor herself talks about ease or set-
tledness, when self-interpretation ends successfully.
Of course, we cannot rule out the possibility that a subject has, or has not, resid-
ual tension, owing  to factors independent of the metaphysical soundness of the
application. A third slight modification to John’s case can illuminate the point.
Suppose that after being unsuccessful with the master’s method, and resorting back
to the clinician, he feels better with the scientific training proposed, but at same time
he is quite afraid of having been diagnosed with an obsessional disorder. Here John
experiences residual tension because of a meta-cognitional fear about his health
overall. In this case, the residual tension would not be a reliable criterion for track-
ing the soundness of concept application, unless one clearly traces it back to the
side-effect of the diagnosis as such. In any case, even if the provenance of the ten-
sion is meta-cognitional, and not connected with the lower-order experience to
which the conceptual scheme was first applied, it becomes more problematic to
make a judicious use of psychic tension as an indicator of soundness of concept
application.
At the same time, at the other end of the spectrum of possibilities offered by the
great complexities, and complexity of levels, of the human mind, there are well-­
known cases of people finding ease in clearly wrong conceptual schemes just
because they are less costly for them than the correct ones. Self-deception is a case
in point.25 However, although it is said that self-deception has psychic tension as one
of its hallmarks, well-accomplished self-deception can offer significant relief to the
subject.26 Once again, the criterion of ease or settledness vs. psychic tension should
always be assessed as objectively as we can to be sure that it can really function as
a reliable guide to the evaluation of the soundness of concept application.

24
 In no way do I wish to assume the superiority of medicine in treating cases like these, as the well-
known nocebo effect sometimes renders medical treatments less beneficial than spiritual ones. Of
course, when it comes to the efficacy, or non-efficacy of a non-medical, or alternative, medical
treatment much depends on the neurological seriousness of the state.
25
 Cf. Mele 2001.
26
 I have argued in Pedrini 2013 that both tension and epistemic labor are two hallmarks of self-
deception. Cf. Noordhof 2009 and Funkhouser 2005.
The ‘Crux’ of Internal Promptings 71

6  C
 onclusion: Toward a Wider View of the Mental
and of Self-Knowledge

Let us now take stock of the intrinsically elusive nature of a vast portion of internal
promptings. It seems that their existence cohabits with an ample domain of mental
states that are, instead, conceptually well (or better) defined. Now, in the light of the
empirically manifest diversity of the nature of internal promptings from the concep-
tual point of view, we should try to adopt a wider geography of the mental. Guided
by a principle of inclusiveness of evidence, we should give citizenship to richness
and diversity within our own minds. This can greatly widen our descriptive picture
of the mind, so that it can then offer itself to more complete and refined explana-
tions. Instead of arguing in favor of a fully conceptualized conception of experience
as opposed to the unconceptualized one, it seems more promising to make room for
as many varieties of experience as our psychology manifests. Both the mind and
reality as a whole show a variety of events and states whose presence and operation
should not be downplayed or under-described, even if more accurate descriptions of
them may well create further metaphysical and epistemological issues, or resurrect
old ones. It is, after all, part of the fascinating mission of philosophy to try to develop
conceptual instruments to diagnose empirically when we are faced with one variety
of experience as opposed to another of a quite different nature. I thus recommend
further descriptive investigations into both the plurality and diverse richness of our
own mental experience and the epistemology that best applies to the unique, diverse,
often even irredeemably obscure mental items we all live with.

Acknowledgments  I thank Julie Kirsch for comments and revisions on an early draft of this
chapter.

References

Boyle, M. 2015. Critical study: Cassam on self-knowledge for humans. European Journal of
Philosophy 23 (2): 337–348.
Carruthers, P. 2009. How we know our own minds: The relationship between mindreading and
metacognition. Behavioral and Brain Sciences 32: 1–18.
———. 2011. The opacity of the mind: An integrative theory of self-knowledge. Oxford: Oxford
University Press.
Carter, J.A., and D. Prichard. 2018. Extended self-knowledge. In third-person self-knowledge, self-­
interpretation, and narrative, ed. Patrizia Pedrini and Julie Kirsch. Cham: Springer.
Cassam, Q. 2014. Self-knowledge for humans. Oxford: Oxford University Press.
Coliva, A. 2015. Review of Quassim Cassam self-knowledge for humans. Analysis. https://doi.
org/10.1093/analys/anv078.
———. 2016. The varieties of self-knowledge. Palgrave Macmillan.
———. 2018. In Self-knowing interpreters. In third-person self-knowledge, self-interpretation,
and narrative, ed. Patrizia Pedrini and Julie Kirsch. Cham: Springer.
72 P. Pedrini

Funkhouser, R. 2005. Do the self-deceived get what they want? Pacific Philosophical Quarterly
86: 295–312.
Kirsch, J. 2015. Review of Quassim Cassam Self-Knowledge for Humans. Philosophy in Review
4: 188–190.
———. 2018. In Interpreting things past. In third-person self-knowledge, self-interpretation, and
narrative, ed. Patrizia Pedrini and Julie Kirsch. Cham: Springer.
Lawlor, K. 2009. Knowing what one wants. Philosophy and Phenomenological Research 79:
47–75.
McDowell, J. 1994. Mind and world. Harvard: Harvard University Press.
Mele, A. 2001. Self-deception unmasked. Princeton: Princeton University Press.
Noordhof, P. 2009. The essential instability of self-deception. Social Theory and Practice 35-1:
45–71.
Peacocke, C. 1999. Being known. Oxford: Clarendon Press.
Pedrini, P. 2009. Epistemologia dell’autoconoscenza. Pisa: ETS.
———. 2013. L’autoinganno. Che cos’è e come funziona. Roma-Bari: Laterza.
Renz, U. 2017. Self-knowledge as personal achievement. Proceedings of the Aristotelian society
cxvii., part 3. https://doi.org/10.1093/arisoc/aox012.
Rudder Baker, L. 2013. Naturalism and the first-person perspectives. Oxford University Press.
Schwitzgebel, E. 2008. The unreliability of naive introspection. Philosophical Review 117 (2):
245–273.
Shoemaker, S. 1994a. Self-knowledge and ‘inner sense’: Lecture I: The object-perception model.
Philosophical and Phenomenological Research 54 (2): 249–269.
———. 1994b. Self-knowledge and ‘inner sense’: Lecture II: The broad perceptual model.
Philosophical and Phenomenological Research 54 (2): 271–290.
———. 1994c. Self-knowledge and ‘inner sense’: Lecture III: The phenomenal character of expe-
rience. Philosophical and Phenomenological Research 54 (2): 291–314.
Sims, A. 2018. In Causal inference in the clinical setting: Why the cognitive science of folk psy-
chology matters. In third-person self-knowledge, selfinterpretation, and narrative, ed. Patrizia
Pedrini and Julie Kirsch. Cham: Springer.
Zawidzki, T.W. 2018. In Self-interpretation as software: Toward a new understanding of why false
self-conceptions persist. In third-person selfknowledge, self-interpretation, and narrative, ed.
Patrizia Pedrini and Julie Kirsch. Cham: Springer.
Interpreting Intuitions

Marcus McGahhey and Neil Van Leeuwen

Abstract  We argue that many intuitions do not have conscious propositional con-
tents. In particular, many of the intuitions had in response to philosophical thought
experiments, like Gettier cases, do not have such contents. They are more like
hunches, urgings, murky feelings, and twinges. Our view thus goes against the
received view of intuitions in philosophy, which we call Mainstream Propositionalism.
Our positive view is that many thought-experimental intuitions are conscious, spon-
taneous, non-theoretical, non-propositional psychological states that often motivate
belief revision, but they require interpretation, in light of background beliefs, before
a subject can form a propositional judgment as a consequence of them. We call our
view Interpretationalism. We argue (i) that Interpretationalism avoids the problems
that beset Mainstream Propositionalism and (ii) that our view meshes well with
contemporary cognitive science.

Keywords  Intuition · Thought experiment · Proposition · Mental state ·


Philosophical method · Interpretation

1  I ntroduction: Interpretationalism versus Mainstream


Propositionalism

Philosophers tend to talk about intuitions as though they have propositional con-
tents, as in, “I have an intuition that p!” But we should be wary of such talk.
Sentences used to describe mental states often mislead in one way or another. One

Authorship is divided equally.


M. McGahhey (*)
Department of Philosophy, University of California, San Diego, La Jolla, CA, USA
N. Van Leeuwen (*)
Department of Philosophy and Neuroscience Institute, Georgia State University,
Atlanta, GA, USA
e-mail: nvan@gsu.edu

© Springer Nature Switzerland AG 2018 73


P. Pedrini, J. Kirsch (eds.), Third-Person Self-Knowledge, Self-Interpretation,
and Narrative, Contributions To Phenomenology 96,
https://doi.org/10.1007/978-3-319-98646-3_5
74 M. McGahhey and N. Van Leeuwen

tells a waiter, “I want the roasted Brussels sprouts,” when one’s desire is really just
for hot vegetables. Or one says, “I knew that she would go back on her promise!”
when one only really had a murky suspicion. It’s a mistake to confuse the nature of
a psychological state with the content of a sentence we happen to use in attempting
to describe it. This is a familiar point in philosophy of mind.1
We think talking about intuitions as though they have propositional contents is
also misleading. When reporting intuitions in response to thought experiments, phi-
losophers often use that clauses. “My intuition is that Smith’s belief is justified and
true but not knowledge.” “My intuition is that it is wrong to push the man off the
bridge.” Or, “My intuition is that there is no understanding of Chinese in the room.”
But such talk creates a misleading impression. The misleading impression is that
intuitions always have determinate, consciously-accessible propositional contents.
We call this standard, widely-shared view about intuitions Mainstream
Propositionalism and summarize it as follows.
Mainstream Propositionalism: the view that intuitions had spontaneously in response to
philosophical thought experiments always have determinate, consciously-­accessible propo-
sitional contents.

Much of this paper (our negative project) argues that Mainstream Propositionalism
is false. Many intuitions prompted by thought experiments (henceforth, thought-­
experimental intuitions) are not propositional. But if those intuitions are not propo-
sitional, what are they like? And how are they incorporated into philosophical
theory construction? We argue (our positive project) they are more like hunches,
twinges, internal promptings, urgings, or vague suspicions that need to be inter-
preted before their significance for thought experiments and related philosophical
theses is apparent. Our view is Interpretationalism.
Interpretationalism: the view that many intuitions had in response to philosophical
thought experiments must be interpreted in light of background beliefs before the intuiter
can judge what proposition they support.

We thus hold that there are two stages on the route from intuition to propositional
judgment. First, the person reading or hearing a thought experiment has an unarticu-
lated, conscious prompting that suggests something is or is not quite right about a
philosophical position in relation to a thought-experimental case; this prompting is
an intuition. Second, the person having that intuition, by thinking in light of
­background beliefs, comes to a propositionally articulated judgment2 about what
exactly is right or wrong about the philosophical position or vignette under consid-
eration. Mainstream Propositionalism—this is its great demerit—conflates the intu-
ition of the first stage with the judgment of the second.

1
 Lewis (1983) and Dennett (1987) both make this point.
2
 We’re using “judgment” here as a catchall term for what could really be a variety of conscious
cognitive attitudes, not just conscious beliefs. For example, the output of the interpretive process
might be a conscious hypothesis or working assumption. Those differences won’t make a differ-
ence, however, to the main arguments of this paper.
Interpreting Intuitions 75

Here’s a preview of the paper. Sect. 2 shows how Mainstream Propositionalism


breaks down into four distinct, widely-endorsed claims, where three of them (with
some modification) turn out to be still correct on our theory. This overlap gives us
confidence that we have many of the same mental states in mind under the term
“intuition” as people who hold Mainstream Propositionalism; so our dispute is not
merely terminological. Sect. 3 executes our negative project, giving three arguments
against Mainstream Propositionalism: one phenomenological; a second from diver-
gences over how intuitions are reported and described; and a third from unification
with the psychology of linguistic intuitions about fictional sentences. Sect. 4 pres-
ents Interpretationalism, a theory about what non-propositional intuitions are, where
they come from, and how philosophers in fact (often contrary to their metaphilo-
sophical musings) deploy them. Sect. 5 concludes with reflections on the signifi-
cance of our claims for philosophical method.
What is the scope of our project? That is, of all the mental states that could be
called “intuitions,” which ones are we talking about? As Nado (2014) argues, the
class of mental states called “intuition” is heterogeneous. So it’s folly to generalize
too broadly. We deal with this issue as follows. First, we restrict the scope of our
theory to mental states had spontaneously in response to philosophical thought
experiments, such as Gettier cases, which we treat here as paradigmatic of thought
experiments. We use intuition types from other parts of psychology mainly for com-
parison and illustration, and we set many intuitions aside for now, such as mathe-
matical intuitions. Second, even with this restriction, we acknowledge that some
thought-experimental mental states plausibly called “intuitions” might not fit our
theory. Accordingly, we present Interpretationalism as a theory of many (perhaps
the majority) of the intuitions people have in response to thought experiments. It is
an enlarged existential claim, not a universal one. This gives us a dialectical advan-
tage over someone who would defend Mainstream Propositionalism, who must
claim all thought-experimental intuitions are propositional; to win the dispute we
only have to show some paradigmatic intuitions are not. Third, we acknowledge that
how broadly our theory extends is in the end an empirical question; intuitions are
psychological states and psychology is a matter for empirical investigation. Our
theory applies at least to the examples we discuss here. But empirical evidence will
have to determine how often and when the cognitive structures we detail in Sect. 4
occur. In any case, since there are many philosophers to whom it has not even
occurred that intuitions about philosophical thought experiments might not be prop-
ositional, our main contribution is to put Interpretationalism on the map.

2  Mainstream Propositionalism

Four popular claims conjoin to make Mainstream Propositionalism. Each comes in


different forms, but we abstract from that variety in order to present evidence that
the four claims are popular enough that their conjunction amounts to being a stan-
dard view.
76 M. McGahhey and N. Van Leeuwen

(1) Intuitions are conscious.


Koksvik (2011: 10) writes, “Having an intuition makes a particular contribution
to the character of the overall experience of the person who is having it.” Put simply,
there is something it’s like to intuit. This what-it’s-likeness guides our interactions
with philosophical claims and theories. When we reflect on some philosophical
claim, our thinking may be punctuated by a peculiar phenomenology; suddenly, a
claim appears in a certain light—though we might not possess explicit, articulable
reasons for why this should be so. Or we experience an alarm-like surge in phenom-
enology, a tug, push, or feeling of attraction or repulsion. Bealer (1998: 207-208)
describes intuition as “a genuine kind of conscious episode” and calls intuitions
“intellectual seemings.”3
(2) Intuitions are spontaneous.
Intuitions “flow spontaneously from the situations that engender them, rather
than from any process of explicit reasoning” (Gopnik and Schwitzgebel 1998: 77).
Though consciously entertaining a proposition or thought experiment may result in
an intuition, intuitions themselves are not consciously formed. They are the con-
scious product of a non-conscious process. They cannot be formed via conscious
inference from other conscious mental states. Nor can intuitions be directly sum-
moned at will, though they can be indirectly facilitated by attention to appropriate
stimuli (like Gettier cases). For this reason, intuitions are “phenomenologically
basic; their inferential heritage is not introspectively available” (Kornblith 2002:
20). As far as first-person experience is concerned, intuitions spontaneously happen
in the presence of certain inputs.4
(3) Intuitions are non-theoretical.
Imagine that you are seeing the Müller-Lyer figure for the first time.

3
 The view that intuitions are phenomenally conscious is defended by Pollock (1974), Plantinga
(1993), Bealer (1998, 1999), Pust (2000), Huemer (2001, 2007, 2013), Koksvik (2011), Chudnoff
(2011, 2013), and Bengson (2015). Goldman and Pust (1998), Goldman (2007), and Ludwig
(2007) also intimate adherence to something like this claim. Even alleged deniers of intuitive phe-
nomenology, Williamson (2007) and Sosa (2007), maintain that there is something it is like to
intuit; however, they differ from many of the above authors in that they also maintain that this
phenomenology is exhausted by what it’s like to be consciously inclined to accept a proposition
and what it’s like to entertain a proposition, respectively.
4
 For more views that also suggest spontaneity, see Lycan (1988), Bealer (1998), Gopnik and
Schwitzgebel (1998), Goldman and Pust (1998), Kornblith (2002), Nichols et al. (2003), Goldman
(2007), Bengson (2015), and Taylor (2015).
Interpreting Intuitions 77

You are taken by the appearance of a longer line and a shorter one; you even form
the belief that one line is longer. Shortly thereafter, you discover your experience
was an illusion; neither line is longer than the other. A quick measurement assures
you that your eyes have deceived you. But looking again, you find the appearance
of inequality persists! Despite believing otherwise, one line just seems longer; “the
visual attraction to think them incongruent remains” (Sosa 2014: 41).
Thought-experimental intuitions seem to behave similarly. Like perceptual
seemings, intuitions often put pressure on a subject’s beliefs. And like perceptual
seemings, intuitions are not subject to revision simply as the conclusion of a rational
argument.5 Otherwise put, intuitions are non-theoretical.
Bealer (1998: 208) illustrates this point, writing, “Just try to diminish readily. ..
your intuition that your favorite Gettier example could occur.” Intuitions, Bealer
holds, just aren’t rationally sensitive to our beliefs and desires. This claim has
proved hugely influential. Pust (2000: 33), for instance, adds “[t]he appearance [in
the Müller-Lyer case]. .. is impenetrable by belief. .. in an analogous manner, our
intellectual seemings or intuitions can diverge from our beliefs.”
Bengson (2015: 721–22) unpacks the widely-endorsed analogy between percep-
tion and intuition in a way that highlights the intimate connection between non-­
theoreticality and spontaneity.
. . . one is not free to manage or get rid of [intuitions] in the way that one is, or at least
sometimes is, free to manage or get rid of one’s beliefs (e.g. by revisiting or resorting old
evidence or by seeking new evidence). Similarly, one is not free to pick whether, what, and
how to experience or intuit in the way that one is, or at least sometimes is, free to pick
whether, what, and how to imagine, guess, hypothesize, or judge. When looking around at
one’s immediate environment, whether one has a visual experience, what the content of the
experience that one has is, or whether the experience is clear and vivid, is not within one’s
conscious control. Likewise, when reflecting on a putative counterexample, thought experi-
ment, or elementary logical or mathematical proposition, whether one has an intuition,
what the content of the intuition that one has is, or whether the intuition is clear and vivid,
is not within one’s conscious control.6,7

(4) Intuitions have propositional content that is consciously accessible. [This is the
one we reject.]
According to consensus, intuitions are propositional—and the intuiter is aware
of her intuitions’ propositional content. Just as beliefs are beliefs that something is
the case, intuiting essentially has propositional content that is readily available for
use in downstream reasoning.

5
 It may be that they are revisable through a sort of practice, in which one attends to certain things
and trains oneself in a certain way.
6
 Note that Bealer, Pust, and Bengson extend the claim of non-theoreticality to mathematical intu-
itions, about which we reserve judgment. For our purposes, it is enough that those scholars endorse
the claim when it comes to thought-experimental intuitions.
7
 For more on non-theoreticality, see also Bealer (1998), Sosa (1998, 2007, 2014), Pust (2000),
Goldman (2007), Ludwig (2007), Williamson (2007), Koksvik (2011), Chudnoff (2011, 2013),
and Bengson (2015).
78 M. McGahhey and N. Van Leeuwen

The view that intuitions are propositional in this way is taken for granted. The
principal debate among theorists of intuition concerns what exactly intuiting that p
is, without much reflection on whether the “that p” even belongs there. There is the
doxastic view that intuiting involves believing, judging, or being inclined to believe
or judge that p. Van Inwagen (1997: 309), for example, claims, “[o]ur ‘intuitions’
are simply our beliefs . . . ”8 Then there is the sui generis view according to which
intuition is a distinct propositional attitude unto itself. Intuition, writes Bealer, is “a
sui generis, irreducible . . . propositional attitude” (1998: 207). For Cullison (2013:
33), the attitude of intuiting is “not analyzable in terms of the other mental phenom-
ena.” According to sui generis theorists, intuition essentially involves a conscious
seeming, presentation, or appearance that p, in which a subject is made consciously
aware of a proposition in a way that is epistemically favorable. Tucker (2010: 53)
writes, “[t]he phenomenology of a seeming makes it feel as though the seeming is
‘recommending’ its propositional content as true or ‘assuring’ us of the content’s
truth.”9
In this debate, advocates of the doxastic view and advocates of the sui generis
view, as noted, mostly take for granted that intuitions are always propositional and
that their propositional contents are consciously accessible (intuitions “recom-
mend” their propositional contents, as Tucker would put it).
Again, (4) is the claim we reject. Let’s get a little clearer about what we’re deny-
ing (and what we’re not denying) and why it matters.
We are not going to deny that it’s in principle possible to assign the intuitions we
investigate propositional content in some weak sense; rather, we deny that the intu-
itions we investigate have propositional content that is clear, determinate, and
immediately consciously available to the agent who has the intuition. A comparison
will help. Suppose you are in a room that has a buzzer in it that (it turns out) goes
off when the temperature of the room drops below freezing; you do not know (yet)
that this is what the buzzer indicates, though you might eventually figure it out.
Does the noise the buzzer makes carry propositional content? You might say it does
in some sense; maybe it carries the content that the temperature of the room is below
freezing. But importantly, the loud noise alone doesn’t make that propositional con-
tent available to you the agent hearing it; you have to do some work to figure out that
that’s what the buzzer noise means. Note also that you might assign the buzzing
noise any number of related contents, such as that the temperature of the thermom-
eter in the device is below zero Celsius, without there being cause to think that that
is a better or worse account of the noise’s content than some other account. But now

8
 He adds, interestingly, “or perhaps, in some cases, the tendencies that make certain beliefs attrac-
tive to us.” This latter portion could be interpreted along our lines.
9
 For the view that intuition is a propositional attitude, see also van Inwagen (1997), Goldman and
Pust (1998), Gopnik and Schwitzgbel (1998), Bealer (1998, 1999), BonJour (1998, 2001), Sosa
(1998, 2007, 2014), Pust (2000), Huemer (2001, 2007, 2013), Kornblith (2002), Ludwig (2007),
Tucker (2010), Koksvik (2011), Chudnoff (2011, 2013), Cullison (2013), and Bengson (2015). For
the view that intuiting involves being aware of a proposition, see Tolhurst (1998), Bealer (1998,
1999), BonJour (1998, 2001), Hales (2000), Pust (2000), Huemer (2001, 2007, 2013), Tucker
(2010), Chudnoff (2011, 2013), Lycan (2013), and Bengson (2015).
Interpreting Intuitions 79

suppose the buzzer is replaced with another device (set up to go off under the same
conditions) that has an electronic voice that says, “The temperature of the room is
now below freezing!” That sounded sentence carries a propositional content that
you grasp, if you understand English, without having to do any work to figure it out
(aside from the sub-conscious linguistic processing). Here’s the point of the anal-
ogy: we hold that some mental states are propositional only in a way that’s analo-
gous to the buzzer, if at all: a murky feeling like something is wrong might be
assigned a propositional content by some determined theorist, but that content isn’t
clearly and consciously accessible to the agent. Let’s call mental states that are
propositional in this buzzer-like sense minimally propositional. Other mental states,
such as articulated conscious thoughts, have propositional contents that are avail-
able for immediate conscious use in reasoning, analogous to the sentence. Let’s call
such states robustly propositional. Our arguments in what follows serve to deny that
many intuitions are robustly propositional (we don’t bother denying that they are
minimally propositional, so for brevity, “propositional” going forward will mean
robustly propositional, unless otherwise noted).
Importantly, the distinction between minimal and robust propositionality is phil-
osophically significant, even if it’s often overlooked. Many philosophers conceive
of intuitions largely in terms of the putative roles they play in philosophical theoriz-
ing. And according to one standard metaphilosophical story, intuitions recommend
certain claims as starting points for theory-development; they offer immediate jus-
tification for certain propositional commitments; their contents can be readily folded
into downstream inference; and so on. Though we do not focus primarily on the
epistemology here, we take this story to be an important motivator for the Mainstream
Propositionalist. Essentially, in claiming that intuition is a propositional attitude, the
Mainstream Propositionalist positions herself to offer a neat and straightforward
account of how it is that intuition fulfills these various methodological roles. Bearing
propositional content, one might think, is a prerequisite for entering directly into
inferential relations, as well as other relations of rational support. The wrinkle is
that the plausibility of this simple story turns on a particularly robust conception of
the propositionality of intuitions—one on which the mooted propositional contents
of the intuitions themselves are consciously-accessible. Thus, denying robust prop-
ositionality is sufficient to call into question a large fragment of philosophical meth-
odology—in addition to the psychological contribution of having a more accurate
picture of a certain mental state.
* * *
In sum, the purported properties of intuition are consciousness, spontaneity, non-­
theoreticality, and robust propositionality. Mainstream Propositionalism is the
­conjunction of the four claims attributing these properties. Importantly, Mainstream
Propositionalists hold that intuiters are conscious of a specific proposition, as the
above quotations confirm. This point is crucial to our refutation in the next section:
intuiters, we hold, are conscious of something, but it’s often not a specific proposi-
tion. Be that as it may, if we subtract claim (4) [that intuitions are robustly proposi-
tional] from this cluster, we get a triad with which we agree: the intuitions we
discuss are spontaneous, non-theoretical, and conscious (though their conscious
80 M. McGahhey and N. Van Leeuwen

aspect does not include consciousness of a proposition). Our positive view in Sect.
4 accounts for each of these properties.

3  O
 ur Negative Project: Intuitions Are Not Robustly
Propositional

We give three arguments against Mainstream Propositionalism. All three point in


the same direction: many intuitions do not have conscious propositional contents.
This section is not merely negative, however, since its observations set the stage for
our Interpretationalist theory.

3.1  A Non-Propositional Phenomenology

Let’s reflect on our conscious mental states as we approach the critical moment of
Gettier’s classic piece. Imagine it’s 1963, the year of publication. You begin reading
with a firm conviction that justified true belief is knowledge. True, there are ques-
tions about each of the three focal terms in this standard analysis, but whatever
waves rock the philosophical seas, knowledge and justified true belief ride them
together. Or so you think.
But concern sneaks in as you read Gettier’s set-up. “First…it is possible to be
justified in believing a proposition that is in fact false” (121). “Secondly…if S is
justified in believing P, and P entails Q, and S deduces Q from P…then S is justified
in believing Q” (121). These points both feel familiar. But seeing them together, so
simply put, is strangely unsettling.
Now comes the hammer.
. . . suppose that Smith has strong evidence for the following conjunctive proposition.
(d) Jones is the man who will get the job, and Jones has ten coins in his pocket.
. . . Proposition (d) entails:
(e) The man who will get the job has ten coins in his pocket.
. . . In this case, Smith is clearly justified in believing that (e) is true.
But imagine further that, unknown to Smith, he himself, not Jones, will get the job. And
also unknown to Smith, he himself has ten coins in his pocket. Proposition (e) is then true,
though proposition (d), from which Smith inferred (e), is false. (122)

Something is amiss. You frantically read back over the case, just to make sure
you understood everything. Yes. .. (d) is false, though justified. .. yes. .. (d) entails (e)
(but does anyone really have such a silly belief about 10 coins? does it matter? is
(e) really justified?). . .but yes. .. (e) is true! It appears that the conditions of justified
true belief are satisfied, though you have a lingering suspicion that maybe they
aren’t really. And importantly, when you think of using “knowledge” to refer to
Interpreting Intuitions 81

Smith’s belief (e), it feels off. That mental state—the something-feels-very-wrong


internal feeling—is, on our view, the intuition.
Let’s call this state “SFVW,” for “something feels very wrong.” In identifying
SFVW with the Gettier intuition, we mean something specific. SFVW comes spon-
taneously in response to learning details of the Gettier case; it also appears indepen-
dently of your background theory (it even feels contrary to it), so it is non-theoretical;
SFVW also appears consciously. Finally, and importantly, SFVW gets you started
on theory revision: it’s the unpleasant feeling that motivates you to revise your ear-
lier views in a way that, you hope, will make the unpleasant feeling go away. So
SFVW has three of the four properties standardly thought characteristic of intu-
itions, and it is a starting point for theorizing; that’s what we mean when we call it
“intuition.”
But, we think, SFVW is not propositional in a robust way. Uneasy, something-is-­
wrong feelings are not specific enough to have consciously graspable propositional
content. What would its proposition be? Is it that knowledge is not justified true
belief? Is it that (e) wasn’t really justified? Is it that Smith just doesn’t know? Is it
that a fourth condition is needed in the analysis? All of these possibilities (and
maybe more) lurk in your mind in the moment after you have SFVW and before you
revise your theory in a way you decide is right, which is an open-endedness you
would not expect if the intuition had a clear, consciously-accessible propositional
content.
SFVW has motivational force, pushing for you to look for a theoretical move that
will alleviate the sense of wrongness. After reflecting on your options, you judge,
reluctantly, that Smith’s belief (e) is not knowledge. But that judgment is not the
spontaneous intuition, it is a product, in part, of the background beliefs you have
that delineate the space of theoretical options before you. That is, your judgment is
the product of interpreting the significance of SFVW, in light of your background
beliefs. We might talk as if the intuition has the content that Smith’s belief is not
knowledge, but this loose talk conflates the intuition with its related follow-on judg-
ment. Follow-on judgments often happen very quickly after intuitions, but they are
still distinct.
Similar observations emerge from reflection on any number of thought experi-
ments. Searle’s (1980) “Chinese Room” elicits a sense of something’s being miss-
ing, but it’s not clear whether something is missing from the Chinese room or
something is missing from our understanding of what occurs in it (Dennett 2013).
Block’s (1978) “Nation of China,” Parfit’s (1984) splitting cases, and many others
will, on phenomenological reflection, yield similar results: the spontaneous, con-
scious intuition doesn’t determine a precise proposition. Even further reflection may
leave the intuition’s significance unstable, oscillating until our related beliefs stabi-
lize as well.
To summarize these points as sharply as possible, we assert the following. Many
thought-experimental intuitions are positively or negatively valenced spontaneous,
conscious, non-theoretical feelings that are directed at a cluster of ideas (broadly
construed) and that motivate revising or maintaining beliefs one has about those
ideas. A judgment, by way of contrast, is a conscious event in which a belief is
82 M. McGahhey and N. Van Leeuwen

formed or reaffirmed. The fact that a follow-on judgment has consciously accessible
propositional content does not imply that the earlier intuition that motivated it did.

3.2  Same Intuition, Varying Verbal Expression

Now we face a fascinating puzzle. Experimental evidence is converging to show


that there is a common intuition type among normal humans—across cultures,
among “experts” and non-experts—in response to Gettier cases. There is, however,
substantial variation in what people say about their Gettier intuitions. Philosophers,
thinking Gettier intuitions have propositional content, say divergent things about
that supposed content. And if you survey lay people about Gettier cases, whether or
not they say the Gettiered agent “knows” is sensitive to other response choices
offered in the survey (Cullen 2010). This is puzzling: if the Gettier intuition is com-
mon across people who read or hear Gettier cases, why are things people say in
reporting that intuition so various and divergent?
We offer a solution to this puzzle. We agree with Machery et al. (2015) that there
is, in human minds generally, a “universal core folk epistemology,” and we think this
core folk epistemology produces, in response to Gettier cases, tokens of a common
intuition type. But, we maintain, intuitions of that type—Gettier intuitions—do not
have consciously-accessible propositional content. The commonality of the intu-
ition type explains why experimental results converge to the extent they do; but the
fact that Gettier intuitions do not have consciously-accessible propositional content
explains why people’s expressions of those intuitions are so various: in grasping for
verbal options to report Gettier intuitions, people come up with various things, since
there is no conscious proposition affixed to the intuition to nail down what the ver-
bal choice should be.
Here’s the plan for this Sect. 3.2. Step 1: we review evidence that there is a uni-
versal core folk epistemology. Step 2: we show how varied verbal reports of Gettier
intuitions can be. Step 3: we say why our anti-propositionalist explanation of these
data is better than Mainstream Propositionalism.
Step 1. Two questions are important. Do people across cultures respond to Gettier
cases in similar ways? Do trained philosophers and lay people respond to Gettier
cases in similar ways? If the answer to both is “yes,” then it is fair to infer that nor-
mal humans generally have intuitions of a common type in response to Gettier
cases. Both questions elicit controversy. Weinberg et al. (2001) famously claim that
people from different cultures respond differently to Gettier cases, and Starmans
and Friedman (2012) argue empirically that lay people differ from philosophers in
their Gettier responses. However, the differences found by Weinberg et al. (i) aren’t
large and (ii) are contradicted by more recent, careful research. And the differences
found by Starmans and Friedman (2012), while interesting, betray deeper
similarities.
Interpreting Intuitions 83

The case for commonality is clear. Nagel et al. (2013a) tested lay subjects with a
range of Gettier cases and found they were significantly less likely to attribute
knowledge to Gettiered agents10 than to agents in matched True Belief cases, just as
philosophers do. And in their response to Starmans and Friedman (2013), who criti-
cize their work, Nagel et al. (2013b: 666) point out, “we all agree on a core point:
both philosophers and laypeople take some Gettier cases to exemplify justified true
belief without knowledge.” And this is true: Starmans and Friedman (2012) did find
that lay people denied knowledge to Gettier agents who had been duped by “appar-
ent evidence.” Thus, the controversy boils down to whether the overlap between lay
responses and philosopher responses to Gettier cases is partial or total. The overlap
is impressive in any case. A plausible construal of the evidence thus far, then, is that
lay people and philosophers start from a common core folk epistemology, and if
there are differences in responses, this is because philosophical training makes
adjustments to a common folk epistemological system.
Machery et  al. (2015), furthermore, survey individuals from several cultures
(Brazil, India, Japan, and the USA) and find that responses to Gettier and other
cases are similar across cultures. They present many subjects—245 in the final sam-
ple—with four cases: two Gettier-type cases, one case of clear knowledge, and one
false belief case. The main dependent variables were responses to two knowledge
probes. Their Knowledge 1 probe has the choices: “Yes, [s]he knows” and “No, [s]
he doesn’t know” (p.  5). Their Knowledge 2 probe has choices: “[Protagonist]
knows that [relevant proposition],” and “[Protagonist] feels like [s]he knows that
[relevant proposition] but [s]he doesn’t actually know [this]” (p. 5).11 Two striking
patterns emerge. First, across cultures, subjects deny knowledge in both probes sig-
nificantly more to Gettiered agents than to agents in the clear knowledge case (effect
sizes are substantial). Second, in response to the Knowledge 2 probe, rates of denial
of knowledge across cultures are more similar between the Gettier cases and the
False Belief case than between the Gettier cases and the Clear Knowledge case. The
question then is this: why would people in such different cultures respond to bizarre
and unfamiliar Gettier cases in more or less the same way? Machery et al. give a
compelling answer: there is a universal core folk epistemology among humans (see
Sect. 4 for how this idea might be expanded upon).
If Machery et al. are right—we think they are at least close—it is natural to think
that Gettier cases elicit intuitions of the same type12 across the majority of people

10
 By “Gettiered agent,” we mean a character whose belief seems to count as justified and true,
without seeming to count as knowledge. “Gettier cases” are situations described in vignettes, in
which Gettiered agents and their beliefs are prominently featured.
11
 The purpose of the Knowledge 2 probe is to address the possibility that some subjects might read
the Knowledge 1 probe as asking about felt knowledge from the point of view of the agent in the
vignette, as opposed to asking about knowledge itself.
12
 When we say there are intuitions of the same type, we assume that intuitions can be classified by
their etiological, phenomenological, and dispositional profiles and that to be of the same type is to
belong to the same class so characterized. Obviously, we think it is a mistake to type intuitions
according to their purported propositional content, as many philosophers would be inclined to do,
since we don’t think intuitions have such content.
84 M. McGahhey and N. Van Leeuwen

who encounter those cases; these are intuitions that emerge from core folk episte-
mology. In sum, the commonality of Gettier intuitions across lay people and phi-
losophers, as well as across different cultures, explains why the tendency to deny
knowledge to Gettiered agents is so widespread.
Step 2. Despite this commonality, there is still substantial variation in what peo-
ple say in response to Gettier cases, in surveys and otherwise. Three kinds of diver-
gence stand out.
First, Nagel et al. (2013a) observe that subjects who score lower on measures of
empathy are less likely to indicate that Gettiered agents are justified. But many phi-
losophers think that a verdict to the effect that the Gettiered agent is justified is part
of the Gettier intuition. So if we were to ask, “is it part of the Gettier intuition that
Gettiered agents are justified?” we would get divergent answers. Some would say
“yes;” some “no.” Furthermore, in informal discussion with other philosophers, we
have noticed that some characterize the intuition as being partly about justification,
while others only think it delivers a verdict on whether the Gettiered agent knows,
which would leave the question of justification open. So people’s reports (in surveys
and otherwise) diverge in terms of whether they suggest Gettier intuitions have jus-
tification as part of their subject matter or not.
Second, Cullen (2010), in another empirical study, finds that the proportion of
subjects who say the Gettiered agent “knows” is sensitive to whether the other
option on the survey is “only believes” or is simply “doesn’t know.” Subjects pre-
sented with [knows vs. doesn’t know] are more likely to pick knows for Gettiered
agents than subjects presented with [knows vs. only believes]. Somehow, having
only believes as an option attracts people away from knows. This, in our view, is
probably because subjects are reluctant to deny epistemic credentials to Gettiered
agents altogether, so they are more likely to pick knows when the only other option
seems to deny those agents epistemic credentials entirely; having only believes as an
option alleviates that pressure. Be that as it may, it is striking that context can cause
people’s responses to diverge even on the focal issue of knowledge.
Third, published metaphilosophy, which assumes Gettier intuitions have propo-
sitional content, has divergent accounts of what that content is. Malmgren (2011:
281), responding to Williamson (2005), argues that the content of Gettier intuitions
is a modal proposition of possibility, which could be expressed, “It is possible that
someone stands to p as in the Gettier case (as described) and that she has a justified
true belief that p but does not know that p.” Malmgren’s view contrasts with
Williamson’s, which portrays cognitive responses to the Gettier case as having
­primarily counterfactual content. And Ichikawa and Jarvis (2009) issue still another
view. They claim that the content of the Gettier intuition is a proposition of neces-
sity, roughly (see their paper for details): necessarily, if the Gettier vignette is com-
pletely true, then someone has a justified true belief but not knowledge. What can
we conclude from all this? All three accounts are similar enough in spirit that they
seem to be addressing the same intuition type (even though Williamson takes issue
with the notion of intuition as many philosophers deploy it). But they posit radically
different propositional structures as the intuition’s content. Let us note one further
Interpreting Intuitions 85

thing: all of these philosophers engage in a substantial amount of reasoning to fig-


ure out what the content of the Gettier intuition is supposed to be; it is not as though
the content is simply there to be read off.
Step 3. If Mainstream Propositionalism were right, all of this variation would be
strange. According to that view, a person with an intuition is thereby consciously
aware of some proposition (call it q): anyone with the same intuition is consciously
aware of the same proposition, q, and this awareness is immediate (not reasoned
toward), since intuitions are spontaneous. Thus, Mainstream Propositionalism sug-
gests that people shouldn’t have to think much about what the (supposed) contents
of their intuitions in fact are, and people with the same intuitions should be able to
agree on what the contents of those intuitions are supposed to be. But even philoso-
phers have to think a lot about what the supposed proposition of an intuition might
be, and often they disagree about it. Why should they have to work so hard to figure
out a proposition of which they are supposed to be immediately aware? It is strange,
from the perspective of Mainstream Propositionalism, that different philosophers
characterize the contents of an apparently common intuition type so differently. If
they all consciously intuit that q, why can’t they agree on what q is? Furthermore,
the contents posited by Williamson, Malmgren, and Ichikawa and Jarvis are all
rather baroque. Given that lay people have the same Gettier intuitions as philoso-
phers, are we to say that lay people are consciously aware of propositions as com-
plex as those described by the philosophers in question? In addition, the variation
that Nagel et al. observe in whether people think Gettier beliefs are justified puts
more pressure on Mainstream Propositionalism. Does q include something about
justification, or not? Finally, the sensitivity to alternate options on surveys, which
Cullen observes, is puzzling for Mainstream Propositionalism, since that sensitivity
appears to cause people to give different responses on the very focal issue of whether
the Gettiered agents “know.” Shouldn’t q, for the Mainstream Propositionalist, at
least include a clear verdict on knowledge, of which the person having the intuition
would be conscious?
Such puzzles evaporate once our view is adopted. For us, the intuition is the
vague but strong sense of something’s being wrong with the ascription of knowl-
edge to the Gettiered agents. That’s it; no conscious proposition. The commonality
of this sense/intuition across people who have core folk epistemology (most people)
accounts for the convergence in the experimental results that show people tend to
deny Gettiered agents have knowledge: if something feels wrong about ascribing
knowledge, it’s natural to deny knowledge. But since the Gettier intuition doesn’t
have propositional structure, it doesn’t include a clear verdict on questions like: is
the Gettiered agent’s belief justified? The lack of such verdicts explains the varia-
tion in observed Gettier reports, both among philosophers and lay people. The one
thing that seems difficult for our account to explain is that philosophers seem so
convinced that intuitions, including Gettier intuitions, are propositional. How could
such intelligent people be so misled? But this is in fact no surprise for us. Appeal to
propositional attitudes is so pervasive in philosophy—the approach often works—
that we expect it to be unreflectively overextended in certain domains. In fact, if you
86 M. McGahhey and N. Van Leeuwen

combine our view that intuitions are not propositional with the fact that philoso-
phers have the false impression that they are, it is easy to explain why philosophers
say intuitions have propositional content but then can’t agree what that content is. In
so saying, they just voice antecedent conviction; in not agreeing with one another,
they accidentally reveal that that conviction is false.
Let’s deal briefly with two apparent objections to the argument of this section.
First objection: one might say that thought-experimental intuitions have proposi-
tional content, but that the intuiters are not conscious of that content. We respond by
reminding the reader that we take no stand on this possible view (which would
accord intuitions minimal propositionality), since our focus is on refuting
Mainstream Propositionalism, which does imply consciousness of a proposition
(robust propositionality). And we add that even if intuitions did have non-conscious
propositional content, what we say in Sects. 3 and 4 about the need for intuitions to
be interpreted would still be true: since an intuition’s conscious component would
not be propositional, one would still need to engage in interpretation of that con-
scious component to figure out what the mooted non-conscious content is. Second
objection: one might maintain that one can explain the data reviewed in this section
by saying that the Gettier intuition’s conscious content is simply that x’s belief is not
knowledge (where x is the Gettiered agent).13 We respond to this suggestion in three
ways. (i) The suggestion will have a hard time with Cullen’s finding that people are
less likely to deny knowledge, given certain options as opposed to others in a forced-­
choice paradigm. If the intuition were just a conscious denial of knowledge, we
wouldn’t expect such variability due to pragmatic features of survey design. (ii) If
the Gettier intuition has that content and only that content, it is puzzling that phi-
losophers disagree so much about the exact content of the intuition. If one has prop-
osition q as a part of a conscious mental state, one should be able to agree with
others who have that same mental state about what the constituents of q in fact are.
Our view faces neither problem (i) nor (ii). Our response number (iii) comes at the
end of Sect. 3.3.

3.3  Comparison with Linguistic Intuitions About Fiction

Though our focus is on thought-experimental intuitions, it is worth asking whether


similar things can be said about other intuition types. The brain often uses the same
psychological structures for various purposes. So if intuitions of a certain form—
spontaneous, conscious, non-theoretical, and non-propositional—are found to occur
outside thought-experimental mental processes, that fact increases the likelihood
that intuitions of that form occur inside thought-experimental processes too.14

 Something like this would be the view of Goldman and Pust (1998).
13

 To make a similar point, it is scientifically desirable to unify theories as much as possible
14

(Friedman 1974), so unifying thought-experimental intuitions with others—provided it seems


empirically reasonable—should count in favor of any theory of intuition.
Interpreting Intuitions 87

Here we argue that linguistic intuitions about fictional sentences are non-­
propositional. Consider this as one step in a larger exploration of the terrain of non-­
propositional intuitions.
Consider two sentences about Hermione Granger, the clever wizard-in-training
at Hogwarts:
(HG1) Hermione can cast spells.
(HG2) Hermione has 42 fingers.
If you know even a bit about Harry Potter, you probably just had intuitions. HG1
seems right (in some way), and HG2 seems wrong (in some way). Let’s call the
intuition types had in response to these sentences I1 and I2, respectively. It’s highly
likely that, as psychological kinds, I1 and I2 have roughly the same characteristics
across speakers of English who are familiar with Harry Potter: the language is com-
mon (English); the stories are familiar; the capacity to process fiction is a general
human trait; so intuitions in response to sentences describing familiar figures, like
Hermione, shouldn’t differ much from person to person. I1 is a sort of internal,
conscious uh-huh; I2 is an internal, conscious nuh-uh.
Now, toward reductio, let’s assume that Propositionalism-f (that is, Mainstream
Propositionalism suitably modified to be about intuitions about fictional discourse)
is true of I1 and I2: people who have I1, on this view, are conscious of a proposition,
as are people who have I2. If this is correct, people should report their tokenings of
I1 and I2 using equivalent or at least almost equivalent propositional complements;
at the very least, their reports should be consistent with one another. If one English
speaker reports I1 as that p, another should report a similar content or at least a
content that coheres with p. But this is not what we are likely to find.15 One person
will say, “HG1 is true. She can cast spells.” But another will say, “Yeah, I know
HG1 isn’t true because it’s fiction, but it gets things right.”
Both speakers are trying to report the same “uh-huh” intuition to HG1. But they
differ in what they say due to different background views about whether fictional
sentences can be true. The two speakers in fact contradict one another, despite
reporting the same “uh-huh” intuition. The first speaker says HG1 is true, while the
second says it’s not true (though still somehow right). So not only is the same intu-
ition type (“uh-huh”) reported in ways that are non-equivalent; it is even reported in
ways that contradict. This, we think, would not happen if tokenings of I1 came with
conscious propositional content. So Propositionalism-f is not true of I1 after all; a
parallel argument can be run for I2. But I1 and I2 are conscious, spontaneous, and
non-theoretical (check your phenomenology), so the thing to reject is the element of
Propositionalism-f that says I1 and I2 are propositional. This argument generalizes
to most intuitions about fictional sentences.
The argument just given puts defenders of Propositionalism-f on their heels.
They might say that I1 simply has the content that HG1 is true and I2 the content
that HG2 is false; they would then claim that people who say otherwise are just let-

 We grant, of course, the question is empirical, but find it highly plausible that the data would
15

work out as we suggest.


88 M. McGahhey and N. Van Leeuwen

ting their verbal reports of an otherwise clear intuition be obfuscated by a pet theory
of fictional semantics. But this response is straining. A speaker who says “I know
HG1 isn’t true. ..” is unlikely to have had a conscious intuition that HG1 is true:
when this speaker verbalizes that HG1 is not true, she does not take herself to be
saying something that contradicts a consciously intuited proposition. Rather, her
saying “Yeah, I know HG1 isn’t true because it’s fiction, but it gets things right” is
a fumbling attempt at expressing that something feels right about HG1.
This murkiness about what to say about fictional sentences infects the philo-
sophical literature on fiction as well. And it does so in a way that further suggests
that intuitions like I1 and I2 don’t come with conscious propositional contents.
There is, as one would expect, widespread agreement that sentences along the lines
of HG1 are in some way correct and that sentences along the lines of HG2 are in
some way not. But in what ways are they “correct” or “incorrect”? Here the consen-
sus stops. Lewis (1978) regards sentences like “Sherlock Holmes lived in Baker
Street” as true and having an implicit “In fiction f . . . ” operator (hence his title:
“Truth in Fiction”), and he regards sentences like “Sherlock Holmes has three nos-
trils” as false (operator or no). Accordingly, he gives a theory of the meaning of
fictional sentences according to which they can come out as true or false; for him,
the contents of fictional sentences are complex counterfactuals. But Walton (1990),
in a rather different theory, characterizes such sentences in terms of fictionality, a
term of art that does not mean straightforward truth. HG1 would, for him, count as
fictional, and HG2 would count as not fictional. He would cash this out, roughly, by
saying that the Harry Potter game of make-believe prescribes imagining the propo-
sition expressed by HG1, while it does not prescribe imagining the proposition HG2
expresses. Importantly, for Walton, HG1 and HG2 both express literally false prop-
ositions, though one is fictional and the other is not. Taylor (2000), on the other
hand, holds a view strikingly at odds with both Walton’s and Lewis’. Taylor, a dyed-­
in-­the-wool referentialist, holds that sentences with empty names, like “Hermione,”
simply do not assert propositions, since a syntactically well-formed sentence still
doesn’t give you a proposition if some of its constituents are empty. Rather, for
Taylor, HG1 would express a “proposition-in-waiting” or an unsaturated “proposi-
tional scheme.” For him, sentences with empty names can’t be true or false; they
have no semantic value whatsoever. Taylor then argues that, though such sentences
assert nothing, they manage to be pseudo-assertions and some of what is pseudo-­
asserted can be propositional and true.16
Lewis, Walton, and Taylor, as normal readers of fiction and normal speakers of
English, presumably have the same intuitions about correct or incorrect fictional
sentences as anyone else. These intuitions are their data as theorists and are the
reason they’re not talking past each other. But note that there is no exact proposition
that the three theorists even could share as an “intuitive” starting point. Notice, for
example, how cautious and unspecific Taylor is about what the intuitions even are:

 Taylor focuses on cases that mix fictional names with claims about reality, such as “Santa Claus
16

does not exist” or “Santa Claus isn’t coming tonight.” But much of what he says about such mixed
sentences can carry over to purely fictional sentences as well.
Interpreting Intuitions 89

“I acknowledge the deference due to the widely shared intuition that one who utters
any of (1)-(3) may speak truly, while one who utters any of (4)-(6) may speak
falsely” (18). We think this caution (“may speak truly”) is appropriate. If you were
to press Lewis, Walton, and Taylor with the question, “What is the content of I1?”,
you would get very different responses, along the following lines:
Lewis would say: “I1 has the content that HG1 is true.”
Walton would say: “I1 has the content that HG1 is fictional.”
Taylor would say: “I1 has the content that HG1 is pragmatically appropriate.”

What causes this divergence? We have a simple explanation. I1 has no conscious


propositional content, so if theorists were confronted with saying what its content
was, they would be forced to make something up, which they would do partly by
consulting background beliefs. I1 conveys the sense of something’s being right
about HG1, but what exactly that is must be worked out by interpretation (what’s
causing it to feel right?) in consultation with one’s evolving set of background
beliefs.
To return to the main thread of this essay, thought-experimental intuitions appear
to have relatives in fictional-discourse intuitions: both kinds of intuition are sponta-
neous, conscious, non-theoretical, and non-propositional. Is this coincidence a coin-
cidence? We think not. Thought experiments, in the end, are miniature fictions. So
thought-experimental intuitions should have the same characteristics as intuitions
about fiction generally. Gettier’s case about Smith is purely fictional. There is no
Smith; correspondingly, there is no man with such a belief about ten coins. Now
take the fictional sentence: “Smith’s belief about ten coins is true.” One intuits that
there is something right about this sentence, but it is hard to make out exactly what
the supposed content of this intuition should even be, given that there is no Smith
and no actual belief. Your intuition pushes you to try to make sense of the sentence,
but it cannot be more specific than that.
Let’s return now to the objection to our thinking that we discussed at the end of
Sect. 3.2. That objection attempted to make sense of the data that motivate us by
saying that the content of the Gettier intuition just is that x’s belief is not knowledge
(where x is the Gettiered agent). Our response (iii) to this is as follows: since, for
any Gettiered x, x is a non-existent fictional character, it is entirely unclear exactly
what the mooted proposition is even supposed to be.17 That lack of clarity is made
palpable by the divergences between Lewis, Walton, and Taylor on the semantics of
sentences with fictional names. And if it’s not clear what the proposition is even
supposed to be, it’s unlikely that any layperson or philosopher has spontaneous
conscious access to that mooted proposition as part of an intuition. But it is an
entailment of Mainstream Propositionalism that people should have spontaneous
conscious access to a proposition; so much the worse for Mainstream
Propositionalism.18

17
 This, by the way, is why the Ichikawa and Jarvis (2009) theory comes out as complicated as it
does.
18
 At the risk of belaboring the point, we can put this response into the form of a dilemma. Any
proposition complex enough to be identified with truth conditions for a fictional sentence would be
90 M. McGahhey and N. Van Leeuwen

4  Our Positive Proposal: Interpreting Intuitions

We asserted earlier: many thought-experimental intuitions are positively or nega-


tively valenced spontaneous, conscious, non-theoretical feelings that are directed at
a cluster of ideas (broadly construed) and that motivate revising or maintaining
beliefs one has about those ideas. Where might such mental states come from?
Much information in our brains is buried in stores of sub-doxastic mental sys-
tems. This includes information about the biological world, the physical world, con-
tagion, threats, people’s minds and social relations, language, and even—it
seems—the structure of knowledge.19 “Sub-doxastic” means that much of this infor-
mation, though poised to guide various behaviors (like freezing or avoiding in
response to threats), is not poised for conscious, verbal expression. People “know”
things about biological kinds that they can’t consciously express—for example, that
a single organism cannot belong to two different kinds at the same “level” (no elm-­
beeches).20 People often don’t even realize that they have such sub-doxastic infor-
mation, since much of it guides inference and behavior without ever being coded as
explicit conscious principles. We propose that the core folk epistemology identified
by Machery et  al. (2015) is also a largely sub-doxastic information store, as are
many of the information stores probed by philosophical thought experiments, which
are, as Dennett (2013) aptly puts it, “intuition pumps.”
It is useful that not all information in the brain is conscious; if it were otherwise,
consciousness would be swamped and unable to support decision-making. But the
downside of this mostly useful arrangement is that it’s difficult to make sub-­doxastic
information conscious and explicit. The information humans have about object stat-
ics, for example, enables us to stack things (Povinelli 2000; Baillargeon 2002); it’s
not there to help us verbalize axioms of object statics. Your sub-doxastic fear system
(another example) may alert your consciousness to an apparent threat—you feel a
fright at it—without revealing the information it uses to categorize something as a
threat. The path from the sub-doxastic storage up to conscious verbal formulation is
rocky.

too complex to be borne in mind spontaneously and consciously (just look at Lewis’ final analy-
sis!). So either you accord the intuitions in question genuine propositional content or you don’t. If
you do, that content is too complex to be conscious. If you don’t, you don’t. Either way, Mainstream
Propositionalism loses. As it happens, Ichikawa and Jarvis (2009: fn. 14) express some reserva-
tions as to whether the thought-experimental processes they discuss are all conscious, so it seems
they might go with the first horn of the dilemma.
19
 [References.] Folk biology: Medin and Atran (1999). Folk physics: McCloskey (1983);
Baillargeon (2002); Proffitt and Kaiser (2006). Contagion system: Rozin et al. (1986); Springer
and Belk (1994). Fear and threats: LeDoux (1996). Mindreading / folk psychology: Baron-Cohen
(1995); Flavell (1999); Nichols and Stich (2003); Goldman (2006). Language: Chomsky (1965,
1986, 1995); Pinker (1994). Core folk epistemology: Machery et al. (2015).
20
 This is a cross-culturally shared “axiom” of folk biology (Atran et al. 1997). We put “know” here
in scare quotes to refer to implicit awareness of information that helps guide behavior, even if that
awareness doesn’t rise to the level of knowledge in a strict sense.
Interpreting Intuitions 91

We have, however, a cognitive tool for digging information out. We can volun-
tarily imagine situations and then see whether our sub-doxastic systems produce an
error or no-error signal of some sort (almost like a buzzer sound, to appeal to our
earlier analogy)—or any other reaction. Sub-doxastic systems produce fast reac-
tions to events in the world. But we can also imagine events in order to get similar
kinds of reaction. Since the world to which sub-doxastic systems react largely
arrives in a sensory way, imaginings, which often have sensory constituents (mental
imagery), can help get those reactions. These reactions are imagination-induced
intuitions. What is imagined can be chosen voluntarily, but the intuition the imagin-
ing induces is not subject to  voluntary control and is often not even anticipated
(that’s often the point): it’s spontaneous. And the intuition is non-theoretical in that,
instead of conforming to antecedent conscious thought, it conforms to sub-doxastic
information. Finally, though the intuition is conscious, the sub-doxastic information
generating it is not.
Our goal is to apply this framework to thought-experimental intuitions. But first,
let’s again observe a similar process in the linguistic case, just to have a parallel
example.
Suppose we consciously endorsed a grammatical rule that a grade school teacher
might have taught us: “nouns are pluralized whenever they refer to numerically
more than one object, but not otherwise.” But now imagine the following (ungram-
matical) strings of words:
*He ate zero piece of cake.
*On her plate was 0.5 piece of cake.
*They shared 1.0 piece of cake.

All three strings comport with the conscious principle just mentioned, since
“piece” in no case refers to numerically more than one. But they all elicit an error
signal from the sub-doxastic language system, since the more fundamental sub-­
doxastic rules of language (Chomsky’s 1986 “I-Language”) differ from the con-
sciously endorsed ones. This error signal is an intuition: spontaneous, conscious,
and non-theoretical (in the sense that it isn’t governed by one’s consciously endorsed
grammar). Conscious thought sends the imagined examples down to the s­ ub-­doxastic
system, which processes them unconsciously and then belches up an intuition to
consciousness. Systematizing such intuitions as evidence for a theory of the “lan-
guage acquisition device” is the grand project Chomsky (1957, 1965) initiated.
When a sub-doxastic system produces an error signal (or other signal), one is
often at a loss as to which element of one’s imagined example caused that error
signal (or other signal).21 Sometimes it’s obvious, as when I imagine a stone falling
upward. But sometimes it’s not at all obvious, as in the Gettier cases. Either way,
one must consciously interpret the error or no-error signal in light of other beliefs
in order to understand its significance for the imagined example. Other beliefs about

21
 We focus here mainly on error signals mostly for reasons of space and because those seem to
include the intuitions prompted by Gettier cases, which are our running examples. We think, how-
ever, that the broad structure of our account could apply to many “positive” intuitions as well.
92 M. McGahhey and N. Van Leeuwen

the imagined situation often give guidance as to the source of the error. But different
belief sets will assign different meanings to error or no-error signals of the same
type, as we saw with differing interpretations of sentences with fictional names.
Turning now explicitly to philosophical thought experiments, we arrive at the
following view. The original Gettier intuition is elicited by imagining the Smith “10
Coins” example, along with imagining applying the term “knowledge” to Smith’s
belief, and then sending that combined imagined case down to one’s sub-doxastic
conception of knowledge/core folk epistemology. The intuition in response to the
Gettier case is the error signal that the sub-doxastic system spontaneously belches
up. (We think this same process, mutatis mutandis, will turn out to be characteristic
of much philosophical thought experimentation.)
It is up to one’s subsequent interpretation to make sense of this error signal (or
whatever the response is) and formulate a judgment. One might judge that justified
true belief isn’t sufficient for knowledge. Or one might judge that Smith’s belief
wasn’t really justified. Different background beliefs (or different levels of empathy)
yield different interpretations of the same error signal/intuition.
We sketch the mental processing of Gettier’s thought experiment below, where
the dashed line represents the muddy boundary between conscious and unconscious
processing and where arrows trace the order of processing:

Conscious thought, which first considers the case, begins in the dark about what
is or isn’t “intuitively” problematic about it; the detailed processing first happens in
sub-doxastic systems; consciousness then ponders significant features of the
thought-experimental case to which the intuition is a response. If all goes well, it
comes up with a reasonable interpretation of the case, in light of the intuitive
response.22

 This process may of course be iterative, with repeated cycles of slightly varied examples being
22

sent down to the sub-doxastic systems. Dennett (2013) calls this “turning the knobs.”
Interpreting Intuitions 93

How does the information in sub-doxastic systems get there? The answer is vari-
ous. Some information in sub-doxastic systems may be innate (this is Chomsky’s
view of the language organ); other information may be what McCauley (2011) calls
“maturationally natural” (it occurs as a reliable product of a maturing person’s inter-
action with normal environmental inputs); and some might be acquired from experi-
ences that manage to influence how low-level processes work. This last possibility
is easily overlooked: the fact that some information is acquired through experience/
learning does not entail that it is conscious or easily revisable. As Ernst Mach writes:
Everything which we observe imprints itself uncomprehended and unanalyzed on our per-
cepts and ideas, which then, in their turn, mimic the process of nature in their most general
and striking features. In these accumulated experiences we posses a treasure-store, which is
ever close at hand, and of which only the smallest portion is embodied in clear articulate
thought. The circumstance that it is easier to resort to these experiences than it is to nature
herself, and that they are, notwithstanding this, free, in the sense indicated, from subjectiv-
ity, invests them with high value. (1883/1960: 36)

Gendler (1998: 415) interprets this passage as saying that much of the valuable
and reliable information we have is in our minds “not organized under any theoreti-
cal framework.” We endorse Gendler’s interpretation and note that it fits well with
the aims of this paper. Intuitions issue from the non-theoretical “treasure-store” of
sub-doxastic information, when that store is prompted by imaginings, including
thought experiments.
In sum, intuitions are spontaneous, conscious products of sub-doxastic, non-­
theoretical systems, which can be elicited by way of imagining. This appears to be
so for intuitions in response to many sorts of imagined situation, including espe-
cially those portrayed in philosophical thought experiments.
How does the conscious interpretation of intuitions work?
There is no single answer to this question, since interpretation is holistic. But
let’s sketch how it goes sometimes. When one has an intuition, one forms a belief
that one had an experience with a certain phenomenology. Call this the experience
belief. One then forms linking beliefs that link the experience belief, which is about
the intuition, to the body of theoretical beliefs under consideration. A linking belief
could be: “If theoretical propositions p1, . . . , pn were true, then the thought experi-
mental case wouldn’t have prompted the negative intuition it did.” This linking
belief, in conjunction with the experience belief, prompts one to reject at least one
of the theoretical propositions p1, . . ., pn. But which one? Here one must use other
beliefs to help figure out which of p1, . . ., pn is most worth rejecting. One might then
judge that p3 is false. If one does, one might carelessly say, “I have an intuition that
p3 is false.” But this talk, though tempting, conflates the intuition with the follow-on
judgment. We should not attribute p3 to the intuition because (i) the denial of p3 is a
theoretical belief that has its significance partly in virtue of other surrounding theo-
retical beliefs and (ii) one may just as well have come to reject p2 or p4 instead of p3
in light of the intuition and experience belief, if one had had somewhat different
94 M. McGahhey and N. Van Leeuwen

background beliefs.23 So linking the intuition specifically to denial that p3 is true,


even if that’s the judgment one comes to have, is a mistake.
This concludes our positive theory of interpreting intuitions, which, once
grasped, is fairly simple. For convenience, we restate its basic components here:
1 . Thought experiments are represented in conscious imaginative acts.
2. Those imaginative acts trigger further processing in sub-doxastic systems.
3. As a result of that processing, which is unconscious, the sub-doxastic systems
bring about an intuition in consciousness.
4. That intuition is non-propositional (at least not robustly propositional), though it
often has a positive or negative valence.
5. Conscious thought interprets the intuition in light of other beliefs in a way that
yields a propositional judgment.24
We think this account covers a wide range of thought-experimental intuitions,
though perhaps not all. It is an empirical question in the end how far the account
extends. Interpretationalism has the virtue of comporting well with both phenome-
nology and psychology, and it preserves most of what we want to say about the
psychology of intuitions: they are spontaneous, conscious, and non-theoretical. Our
only major revision to the psychological picture of intuitions is in giving up robust
propositionality.

5  Conclusion: The Use of Intuition in Philosophy

We have given reasons for thinking that many thought-experimental intuitions,


including especially the ever-fascinating Gettier intuition, do not have consciously-­
accessible propositional contents. Mainstream Propositionalism is false. The phe-
nomenology of having intuitions in response to important thought experiments does
not comport with that view. The view also stumbles at addressing the puzzle of why
people with the same intuition type say strikingly different things in expressing their
intuition tokens. And the psychology of related intuitions about fictional sentences
appears not to be propositional either.
A better view is on offer: Interpretationalism. This view avoids the problems of
Mainstream Propositionalism and sits well with contemporary cognitive science.
On this view, thought-experimental intuitions are conscious states that occur in rela-
tion to imagined examples and philosophical positions that pertain to those exam-
ples; such intuitions are non-propositional in the relevant sense, so their philosophical
significance in propositional terms needs to be worked out through a process of
interpretation in conjunction with background beliefs. One cannot simply “read off”

23
 Here the spectre of the Quine-Duhem problem hovers over the interpretation of thought-experi-
mental intuitions.
24
 Points 4. and 5. here correspond to the two “stages” of intuition and judgment mentioned in the
introduction.
Interpreting Intuitions 95

propositional content from a thought-experimental intuition. The appearance to the


contrary is largely the product (i) of confusing the intuition with follow-on proposi-
tional judgments (which may occur in quick succession)25 and (ii) of wishful
thinking.
What wishful thinking do we mean? Doing philosophy would be a neater, more
straightforward process if our thought-experimental intuitions had clear, conscious
propositional contents. Propositions would be convenient places to begin theoriz-
ing, even if we treated them as defeasible. Propositions, on many construals, are
structured in a way that can serve as the basis for logical derivation, a process at
which most philosophers have some skill. We thus suspect, though few would
openly endorse it, that many philosophers are still quietly tempted by the lull of an
image of philosophy in which intuited propositions plus logical inference forms
straightforwardly justify philosophical theories. But that’s just not how it works.
Rather, intuitions cast shade or sunlight over certain regions of intellectual space,
inviting us to look further. The resulting philosophical work we must do is far more
difficult than the Mainstream Propositionalist picture would have us think. But it is
also, in the end, much more creative and rewarding.

Acknowledgements  We would like to thank the audience at the Belief and Intuition workshop
held at the University of Antwerp in May 2016, at which Neil Van Leeuwen presented this joint
work. We would also like to thank, in particular, James Andow, John Bengson, and Stephen Stich
(who was entirely unconvinced by our view) for stimulating exchanges. This work was supported
in part by the Marie Skłodowska-Curie Fellowship that Neil Van Leeuwen received from the
European Commission [call identifier: H2020-MSCA-IF-2014; contract number: 659912].

References

Atran, S., P. Estin, J. Coley, and D. Medin. 1997. Generic species and basic levels: Essence and
appearance in folk biology. Journal of Ethnobiology 17 (1): 17–43.
Baillargeon, R. 2002. The acquisition of physical knowledge in infancy: A summary in eight les-
sons. In Blackwell handbook of childhood cognitive development, ed. U. Goswami, 47–83.
Baron-Cohen, S. 1995. Mindblindness: An essay on autism and theory of mind. Cambridge, MA:
MIT Press.
Bealer, G. 1998. Intuition and the autonomy of philosophy. In Rethinking intuitions: The psychol-
ogy of intuition and its role in philosophical inquiry, ed. M. DePaul and W. Ramsey, 201–239.
Maryland: Rowman & Littlefield.
———. 1999. A theory of the a priori. In Philosophical perspectives, vol. 13, 29–55.
Bengson, J. 2015. The intellectual given. Mind 124 (495): 707–760.
Block, N. 1978. Troubles with functionalism. Minnesota Studies in the Philosophy of Science 9:
261–325.

25
 One more point on terminology: even if Mainstream Propositionalists wanted to stipulate that
they use the word “intuition” to refer to judgments, they would still need a word for the things we
call intuition, which are non-propositional and no doubt exist; also, they would have to give up the
non-theoreticality of intuition, which is supposed to be its chief epistemic virtue, since the judg-
ments that issue from thought experiments are theoretically loaded.
96 M. McGahhey and N. Van Leeuwen

BonJour, L. 1998. Defense of pure reason. Cambridge: Cambridge University Press.


———. 2001. Precis to in defense of pure reason. Philosophy and Phenomenological Research
63 (3): 625–631.
Chomsky, N. 1957. Syntactic structures. Berlin: Walter de Gruyter.
———. 1965. Aspects of a theory of syntax. Cambridge, MA: MIT Press.
———. 1986. Knowledge of language. Westport: Praeger Publishers.
———. 1995. Language and nature. Mind 104 (413): 1–61.
Chudnoff, E. 2011. What intuitions are like. Philosophy and Phenomenological Research 82 (3):
625–654.
———. 2013. Intuitions. Oxford: Oxford University Press.
Cullen, S. 2010. Survey-driven romanticism. Review of Philosophy and Psychology 1 (2): 275–296.
Cullison, A. 2013. Seemings and semantics. In Seemings and justification: New essays on dogma-
tism and phenomenal conservatism, ed. C. Tucker, 33–51. Oxford: Oxford University Press.
Dennett, D.C. 1987. The intentional stance. Cambridge, MA: MIT Press.
———. 2013. Intuition pumps and other tools for thinking. W. W. Norton.
Flavell, J.H. 1999. Cognitive development: Children’s knowledge about the mind. Annual Review
of Psychology 50: 21–45.
Friedman, M. 1974. Explanation and scientific understanding. Journal of Philosophy 71 (1): 5–19.
Gendler, T. 1998. Galileo and the indispensability of scientific thought experiment. British Journal
for the Philosophy of Science 49 (3): 397–424.
Gettier, E. 1963. Is justified true belief knowledge? Analysis 23: 121–123.
Goldman, A. 2006. Simulating minds: The philosophy. In Psychology, and neuroscience of mind-
reading. Oxford: Oxford University Press.
———. 2007. Philosophical intuitions: Their target, their source, and their epistemic status.
Grazer Philosophische Studien 74 (1): 1–26.
Goldman, A., and J. Pust. 1998. Philosophical theory and intuitional evidence. In Rethinking intu-
itions: The psychology of intuition and its role in philosophical inquiry, ed. M. DePaul and
W. Ramsey, 179–198. Lanham: Rowman & Littlefield.
Gopnik, A., and E. Schwitzgebel. 1998. Whose concepts are they, anyway? The role of philosophi-
cal intuition in empirical psychology. In Rethinking intuitions: The psychology of intuition and
its role in philosophical inquiry, ed. M. DePaul and W. Ramsey, 76–91. Lanham: Rowman &
Littlefield.
Hales, S. 2000. The problem of intuition. American Philosophical Quarterly 37: 125–147.
Huemer, M. 2001. Skepticism and the Veil of perception. Lanham: Rowman and Littlefield.
———. 2007. Compassionate phenomenal conservatism. Philosophy and Phenomenological
Research 74 (1): 30–55.
———. 2013. Phenomenal conservatism Uber Alles. In Seemings and justification: New essays on
dogmatism and phenomenal conservatism, ed. C. Tucker, 328–350. Oxford: Oxford University
Press.
Ichikawa, J., and B. Jarvis. 2009. Thought-experiment intuitions and truth in fiction. Philosophical
Studies 142: 221–246.
Koksvik, O. 2011. Intuition, Diss. Australian National University.
Kornblith, H. 2002. Knowledge and its place in nature. Oxford: Oxford University Press.
LeDoux, J.E. 1996. The emotional brain: The mysterious underpinnings of emotional life.
New York: Touchstone/Simon & Schuster.
Lewis, D.K. 1978. Truth in fiction. American Philosophical Quarterly 15 (1): 37–46.
———. 1983. Philosophical papers: Volume I. Oxford: Oxford University Press.
Ludwig, K. 2007. The epistemology of thought experiments: First person versus third person
approaches. In Midwest studies in philosophy, vol. 31, 128–159.
Lycan, W. 1988. Judgment and justification. Cambridge: Cambridge University Press.
———. 2013. Phenomenal conservatism and the principle of credulity. In Seemings and justi-
fication: New essays on dogmatism and phenomenal conservatism, ed. C. Tucker, 293–305.
Oxford: Oxford University Press.
Interpreting Intuitions 97

Mach, E. 1883/1960. On thought experiments, In Knowledge and error, 6 edn, Trans. T. McCormack
and P. Foulkes, Dortrecht: Reidel, pp. 134–147.
Machery, E., S. Stich, D. Rose, A. Chatterjee, K. Karasawa, N. Struchiner, S. Sirker, N. Usui, and
T. Hashimoto. 2015. Gettier across cultures. Nous (online). https://doi.org/10.1111/nous.12110.
Malmgren, A. 2011. Rationalism and the content of intuitive judgments. Mind 120 (478): 263–327.
McCauley, R.N. 2011. Why religion is natural and science is not. Oxford: Oxford University Press.
McCloskey, M. 1983. Intuitive physics. Scientific American 248 (4): 122–130.
Medin, D.L., and S. Atran. 1999. Folkbiology. Cambridge, MA: MIT Press.
Nado, J. 2014. Why intuition? Philosophy and Phenomenological Research 86 (1): 15–41.
Nagel, J., V.  San Juan, and R.  Mar. 2013a. Lay denial of knowledge for justified true beliefs.
Cognition 129 (3): 652–661.
Nagel, J., R.  Mar, and V.  San Juan. 2013b. Authentic Gettier cases: A reply to Starmans and
Friedman. Cognition 129 (3): 666–669.
Nichols, S., and S. Stich. 2003. Mindreading. Oxford: Oxford University Press.
Nichols, S., S. Stich, and W. Weinberg. 2003. Metaskepticism: Meditations in ethno-epistemology.
In The Skeptics, ed. S. Luper, 227–247. Burlington: Ashgate.
Parfit, D. 1984. Reasons and persons. Oxford: Oxford University Press.
Pinker, S. 1994. The language instinct: How the mind creates language. New  York: Harper
Perennial Modern Classics.
Plantinga, A. 1993. Warrant and proper function. Oxford: Oxford University Press.
Pollock, J. 1974. Knowledge and justification. Princeton: Princeton University Press.
Povinelli, D.J. 2000. Folk physics for Apes: The Chimpanzee’s theory of how the world works.
Oxford: Oxford University Press.
Proffitt, D. R. and Kaiser, M. K. 2006. Intuitive physics, Encyclopedia of Cognitive Science.
Pust, J. 2000. Intuitions as evidence. Routledge.
Rozin, P., L. Hammer, H. Oster, T. Horowitz, and V. Marmora. 1986. The child’s conception of
food: Differentiation of categories of rejected substances in the 16 months to 5 year age ranges.
Appetite 7 (2): 141–151.
Searle, J.R. 1980. Minds, brains, and programs. Behavioral and Brain Sciences 3 (3): 417–457.
Sosa, E. 1998. Minimal intuition. In Rethinking intuitions: The psychology of intuition and its
role in philosophical inquiry, ed. M. DePaul and W. Ramsey, 257–269. Rowman & Littlefield.
———. 2007. Intuitions: Their nature and epistemic efficacy. Grazer Philosophische Studien 74
(1): 51–67.
———. 2014. Intuitions: Their nature and probative value. In Intuitions, ed. A.  Booth and
D. Rowbottom, 36–49. Oxford: Oxford University Press.
Springer, K., and A. Belk. 1994. The role of physical contact in early contamination sensitivity.
Developmental Psychology 30 (6): 864–868.
Starmans, C., and O.  Friedman. 2012. The folk conception of knowledge. Cognition 124 (3):
272–283.
———. 2013. Taking ‘know’ for an answer: A reply to Nagel, San Juan, and Mar. Cognition 129
(3): 662–665.
Taylor, K. 2000. Emptiness without compromise: A referentialist semantics for empty names. In
Empty names, fiction, and the puzzles of non-existence, ed. A. Everett and T. Hofweber, 17–36.
Stanford: CSLI Publications.
Taylor, S.A. 2015. What seemings seem to be. Episteme 12 (3): 363–384.
Tolhurst, W.E. 1998. Seemings. American Philosophical Quarterly 35 (3): 293–302.
Tucker, C. 2010. Why open-minded people should endorse dogmatism. Philosophical Perspectives
24 (1): 529–545.
van Inwagen, P. 1997. Materialism and the psychological continuity account of personal identity.
In Philosophical Perspectives, vol. 11, 305–319.
Walton, K. 1990. Mimesis as make-believe: On the foundations of the representational arts.
Cambridge: Harvard University Press.
98 M. McGahhey and N. Van Leeuwen

Weinberg, J.M., S. Nichols, and S. Stich. 2001. Normativity and epistemic intuitions. Philosophical
Topics 29 (1–2): 429–460.
Williamson, T. 2005. Armchair philosophy, metaphysical modality and counterfacteual thinking.
Proceedings of the Aristotelian Society 105: 1–23.
———. 2007. The philosophy of philosophy. Malden: Blackwell.
Interpreting Things Past

Julie Kirsch

Abstract  Our memories of past experiences influence the way that we think about
ourselves and our relationships with others. But, as a growing body of empirical
research has demonstrated, these memories are often biased and distorted. We
sometimes play an active role in this process though our attempts at interpreting the
past and ‘uncovering’ its true meaning; friends, lovers, and therapists often encour-
age us to do just this. In what follows, I argue that our memories are not always free
from moral evaluation; for how we remember the past can enhance or impair our
ability to relate to others in morally appropriate ways. In showing that an ethics of
memory is possible, I look at the various forms of indirect control that we can exer-
cise over our memories. While we cannot remember at will, we can influence our
memories by using external aids, guarding against biases, and revising the interpre-
tations that we develop of our past experiences.

Keywords  Memory · Narrative · Interpretation · Self-knowledge · Self-deception ·


Ethics

1  Introduction

Our understanding of who we are is largely shaped by memories of things past. But
these memories do not always provide us with an accurate representation of what
we have experienced. Remembering the past is often a matter of interpreting the
past and searching for meaning. In this chapter, I look at the effect that interpretive
and sometimes distorted memories can have upon our moral identities and relations
with others. I propose an ‘ethics of memory’ and examine the extent to which we
may be able to control the formation of our memories and interpretation of the past.
Among other things, I show that we can exercise some control over the extent to
which our interpretations are deluded or self-serving. Interpretation can play a role

J. Kirsch (*)
Department of Liberal Arts, D’Youville College, Buffalo, New York, USA
e-mail: kirschj@dyc.edu

© Springer Nature Switzerland AG 2018 99


P. Pedrini, J. Kirsch (eds.), Third-Person Self-Knowledge, Self-Interpretation,
and Narrative, Contributions To Phenomenology 96,
https://doi.org/10.1007/978-3-319-98646-3_6
100 J. Kirsch

in the formation of our memories without rendering them wholly fictitious and
unreliable. Any interpretation that we accept will have to contend with the world of
facts and the sometimes incompatible interpretations of others.
I will proceed as follows: In §2, I will look at various distortions and biases that
can influence the way that we form and recall memories of past events. In §3, I will
explore a variety of moral considerations surrounding memory. I will argue that we
can have moral reasons for both remembering and forgetting; these moral reasons
extend to the interpretations that we develop of past events, interpretations that often
give our memories shape and direction. Finally, in §4, I will show that we can exer-
cise some control over the way that we interpret and remember the past, thus defend-
ing the ethics of memory developed here against the charge that it is psychologically
impossible.

2  Memory Distortion and Interpretation

We often take memory to be a kind of time machine through which we can revisit
the past. When we want to get in touch with our earlier selves, we return to the past
through our memories of how things used to be. Unfortunately, we have reason to
doubt that our memories always provide us with a faithful representation of past
events. A growing body of empirical research suggests that the memories of ordi-
nary people are subject to various forms of bias and distortion.
Take, for instance, what psychologists refer to as the ‘misinformation effect’.
The misinformation effect occurs when a person who has witnessed an event is
misinformed about it and on this basis forms a distorted or inaccurate account of
what happened (Loftus and Hoffman 1989, p. 100). The typical case goes some-
thing like this: Aaron has witnessed an armed robbery in a jewelry store. The robber
rummages through the store and throws a gold necklace into his bag. Immediately
following the robbery, Aaron speaks to Sara, who also witnessed the robbery. Sara
tells Aaron that she saw the robber throw a gold bracelet into his bag. When the
police arrive, they question Aaron about the robbery. Speaking with remarkable
confidence, he tells them that he saw the robber throw a gold bracelet into his bag.
This case involves the misinformation effect because the misinformation here, i.e.,
Sara’s testimony, causes Aaron to hold a false or distorted memory about the
robbery.
While there is some disagreement about whether or not the later memory replaces
an earlier memory, it is generally agreed that the misinformation has some effect
upon the later memory. The implications of this study are not limited to highly con-
trived experimental contexts. As Loftus and Hoffman see it, they have “tapped a
phenomenon that occurs quite often in real life whenever people who experience the
same event talk to one another, overhear each other talk, or gain access to new infor-
mation from the media, interrogators, or other sources” (p. 103). Ordinary experi-
ence confirms this observation: at some point in our lives, most of us have suspected
that our own memories have been influenced by the conflicting accounts of others.
Interpreting Things Past 101

There is also evidence that mood can have a distorting influence upon our memo-
ries of past events. Numerous studies have established that it is easier for clinically
depressed individuals to recall negative, rather than positive, experiences. Typically,
these studies involve making comparisons between individuals with high levels of
depression and non-depressed controls. Collectively, the studies have found that
“clinically depressed subjects show a strong bias to recall negative, especially self-­
referential, information” (Mineka and Nugent 1997, p. 178). This finding contrasts
sharply with non-depressed subjects who show a tendency to remember positive
information (p. 178). Reinforcing the research on positive illusions, this suggests
that mentally healthy individuals perceive—in this case, recall—the best in the
world and themselves.1
This tendency of mentally healthy people to see the best in the world and them-
selves can play out in other ways as well. By the time we have reached adulthood,
most of us have recognized that our memories of past events are not always mutu-
ally compatible. Two people who have experienced the same event can walk away
from it with profoundly different memories. In many cases, we suspect that this has
something to do with our tendency to form memories that cast ourselves in a posi-
tive light. While investigating this tendency, Michael Ross and Roger Buehler con-
sider one couple’s conflicting memories of how they first met:
The husband’s account was approximately as follows: “I was sitting in ----- airport, waiting
for my flight to be announced, when a woman came into the room. She chose to sit beside
me even though many other seats were available. We talked, hit it off, and the rest is his-
tory.” His wife’s version of this episode was identical except for one detail: She recalled the
waiting room as being entirely full, except for the empty seat beside her future husband.
The spouses had discussed their discrepant memories on a number of occasions. Both
reported that they firmly believed their own version of the story. (1994, p. 228)

Ross and Buehler (1994) point out that almost all people who are presented with
this case believe the wife’s version of the story. They offer two reasons for believing
the wife’s account: “First, a woman would be unlikely to sit immediately next to an
unfamiliar man when other seats were available…. Second, the man’s story seems
to flatter him more than his spouse’s account seems to flatter her. In the face of con-
flicting accounts, audiences distrust memories that appear self-serving” (p.  228).
This case not only illustrates that our own memories are often biased, but also that
we are quite good at detecting similar biases in the memories of others.
Some theorists have argued that our memories are at times biased for the purpose
of self-enhancement. We selectively remember events, or details of events, that sug-
gest that we have any number of positive or desirable character traits. Certain per-
sonality types appear to be more susceptible to memory bias and distortion than
others. It may come as no surprise that narcissists appear to be particularly prone to
memory distortion in response to threatening ego-relevant feedback. Frederick
Rhodewalt and Stacy K. Eddings conducted a study that examined the relationship
between negative feedback and memory distortion in narcissistic men. Participants
in the study were undergraduate males who scored in the upper third or lower third

 See, for example, Positive Illusions by Shelly Taylor (1991).


1
102 J. Kirsch

of the distribution of scores on the Narcissistic Personality Inventory (or NPI)


(2002, p. 101). In the first phase of the study, participants were asked a series of
questions via telephone by an attractive young woman whom they had seen in a
photograph. They were told that the woman believed that she was responding to a
computer dating service advertisement in the school newspaper. The true purpose of
the study, they were told, was to explore “factors that affect romantic attraction
independently of physical appearance” (2002, p. 100). They were also told that the
woman would interview four other men and believed that she would go out on a date
with her top choice (p. 102). After the interview, the participants answered a series
of questions about the interview and their personal dating histories (p. 100).
During the second phase of the study, one week later, each participant listened to
a recording of the woman in which she revealed that “she either liked him the best
or liked him the least of the five men she interviewed” (p. 100). After listening to the
recording, participants again answered a series of questions about their experience
and personal dating histories (p. 100). The results of the study were striking. Among
other things, high and low NPI men exhibited completely different patterns of mem-
ory distortion in response to ego-relevant feedback. Among low-narcissistic
participants,
[e]go-threatening feedback produced recall of personal romantic histories that were less
successful than what were reported at the initial interview. Low-narcissistic participants
also displayed feedback-congruent changes in affect and self-esteem.. .. Narcissists who
received ego-threatening feedback recalled romantic pasts that were more successful than
they had recalled initially [my emphasis]. Moreover, this positive distortion appeared to
serve a self-esteem protective function in that greater self-enhancing distortion was associ-
ated with increases in self-esteem following threatening feedback compared to what they
recalled at the initial interview. (111–112)

What this shows is that the narcissistic men who participated in the study
responded to the news that they were the woman’s last choice by enhancing their
personal dating histories. Rhodewalt and Eddings attribute this to the fact that they
needed to protect their fragile egos. Distorted memories occur when narcissists con-
front ambiguous but threatening feedback that has the potential to wound their
“grandiose but fragile” self-concepts (114). If Rhodewalt and Eddings are right,
memory distortion is a self-defense mechanism that serves to preserve or enhances
a narcissist’s self-esteem.
None of the theorists discussed thus far argue that memory distortion results
from the intentional implementation of a bias. But this should come as no surprise
given that there tends to be disagreement about the extent to which any form of
biased thinking is intentional. Still, we are at times active participants in the recon-
struction, if not distortion, of our past experiences. We are often encouraged by
others to reflect upon our lives and search for the meaning of past events. The pro-
cess of doing so inevitably involves focusing our attention upon certain events, or
details of events, and dismissing others as less important. While the less important
ones tend to fade away, the important ones gain new meaning and develop a life of
their own.
Interpreting Things Past 103

Some therapists, particularly those who practice narrative therapy, consciously


encourage this of their patients; their goal is to help patients – quite deliberately –
replace narratives that interfere with psychological well-being with narratives that
contribute to mental health and happiness. Take, for example, Stephen Madigan’s
story of Tom. Tom, a 66-year-old married middle class White man, was a psychiat-
ric in-patient who had been diagnosed with chronic depressive personality follow-
ing his retirement (Madigan, 1999, p.  151). He had received various forms of
treatment for his condition, including electroconvulsive therapy, group and indi-
vidual psychodynamic therapy, and six different medications (p. 151). Still, there
was little improvement. The staff of the in-patient unit, and those around him,
emphasized that his condition was severe and hopeless by referring to it as chronic
(p. 152). As a last resort, Madigan, a narrative therapist, was brought in to work with
Tom.
During their first session, Tom told Madigan that he had been ‘depressed.’ When
asked whether or not the word was his own, Tom told Madigan that it was a ‘hospi-
tal word’ – that what he really felt was ‘bored and unaccomplished’ (p. 155). During
later sessions, Madigan looked at the way that Tom’s personal narrative was shaped
by cultural themes involving “men’s culture and training, expectations, economics,
and love” – especially as these pertained to men of retirement age (p. 155). He also
asked Tom a series of discursive questions that encouraged him to free himself from
the hospital’s certainties and make space for other ways of understanding his situa-
tion (pp. 155–156). He called upon friends and family members to assist Tom in this
process. Madigan, Tom, and his wife, Jane, sent out a written request for alternative
descriptions of the story that Tom had been telling himself about his life—descrip-
tions that challenged Tom’s view of his life as boring and unaccomplished. Madigan,
Tom, and Jane received 41 letters in response that countered the story that Tom had
been telling himself (p. 161). The purpose of the letter writing campaign was to free
Tom from “the dominant domain of psychological theory and practice” and re-­
remember aspects of his past that he may have forgotten or misinterpreted (pp. 160–
161). Ultimately, Madigan’s efforts proved successful (or so he claims) and he was
able to provide some relief for Tom’s depression by helping him revise the story of
his life.
As this case illustrates, Madigan and other narrative therapists encourage their
patients to consider various readings of the same set of events, especially those that
might be conducive to mental health and well-being. According to Wendy Drewery
and John Winslade, the basic idea behind narrative therapy is that “people make
meaning, meaning is not made for us” (p. 3). Practitioners of narrative theory take
the perspective that
it is more useful for people to think of themselves as drawing on the stories or discourses
that are available to them for making sense of the world than to think of themselves as in
direct contact with reality. This kind of philosophy raises lots of questions about the nature
of reality. We do not argue that reality does not exist—only that we cannot know it directly.
Further, we certainly accept that different people describe reality differently, and that some-
times their descriptions are at odds with those of their therapists and others in their social
world. (p. 35)
104 J. Kirsch

Narrative therapy is committed to the view that there is more than one way of
telling the same story. Moreover, it holds that certain stories can be more damaging
to storytellers than others. While an interpretation cannot completely dismantle the
past, it can make a significant difference to the way that we think about it. And, as
the story of Tom illustrates, the importance of the way that we interpret the past
should not be underestimated; it can make an enormous difference to our lives and
sense of self. By focusing upon certain events, and ignoring others, a person may
come to see herself as an unlovable, unaccomplished loser. But a shift in perspec-
tive, and an openness to others, may encourage her to accept a more favorable inter-
pretation of her life.
Our memories play an important role in shaping our identity or the way that we
think about ourselves. Even memories of negative events—a breakup, death, or vio-
lent attack—can matter to us because they are woven into our identity and play a
role in making us who we are. Indeed, a study conducted by Eryn Newman et al.
found that, if given the opportunity, most people would not want to eliminate mem-
ories of traumatic events, even if doing so would dramatically decrease their risk of
post-traumatic stress disorder (2011, pp.  678–680). People appear to find some
value in their memories even when they are the source of considerable suffering. In
part, this may be due to the fact that a single memory is not an island, but is bound
up with many others in a complex and interconnected archipelago of memories. Or
it may just be the case that we value this information for its own sake, or for the role
that it plays in making us who we are.
Our memories are not just a matter of personal significance; they also affect our
interactions with others, and the way that we think and feel about them. And thus,
how we interpret the past can play an important role in determining our moral rela-
tions with others. An inaccurate interpretation of the past can both cause us to over-
look moral obligations that we have, and to perceive moral obligations that we do
not have. We might then say that our personal narratives shape our moral identities.
For our purposes, a person’s moral identity involves the beliefs and attitudes that she
has about her general moral standing and relations with others. My moral identity
may include the belief that I am compassionate and, as such, am compassionately
engaged with the world around me. How I interpret the past will affect whether or
not I view myself as living up to this moral identity. If I wildly misinterpret the past,
I may cling to a moral identity that is at odds with the facts of my behavior and way
of life. If I am mistaken about what I have done, or why we have done it, then I may
fail to respond to my actions in morally appropriate ways. For these and other rea-
sons, it makes sense to consider whether or not our interpretations of the past are
subject to moral evaluation.
Interpreting Things Past 105

3  The Ethics of Memory and Interpretation

For those who hold that there can be moral reasons to believe, it will come as no
surprise that there can also be moral reasons to remember. After all, our memories
are bound up with our beliefs such that we believe whatever it is that we remember.
But many of the concerns that have been raised about an ethics of belief can also be
raised about an ethics of memory. For example, it is sometimes argued that mental
states, like beliefs and memories, are not subject to moral evaluation because they
are not under our direct control. This is a legitimate concern, and I will address it
later in this chapter. However, before I do, I want to provide a brief overview of the
moral considerations surrounding memory. I will begin by considering some of the
moral reasons that we have for remembering.

3.1  Reasons for Remembering

Memories Are Sometimes Educational We sometimes think that a person ought to


remember what she did in the past so that she does not make similar mistakes in the
future. Just as I may remember that putting too much salt in my pasta sauce ruined
it and ought not to be repeated in the future, so also I may remember that deceiving
my friend poisoned our friendship and ought not to be repeated in the future.
Memory, when put to good use, helps us learn from our past moral failings and
respond to them in morally appropriate ways. To be sure, memory alone cannot do
this, but memory along with reflection, feeling, and reason can go a long way
towards getting the job done.
For the purposes of this discussion, let us say that reflection involves rehearsing
a memory in one’s mind with a critical eye; thus, we can remember something with-
out reflecting upon it. When we engage in reflection, we do not just relive an experi-
ence, we rethink it and often notice what passed us by the first time around. We gain
new insight from reflection. For example, while reflecting upon a conversation with
a friend who experienced a sudden loss, I may come to see my words as heartless,
distant, and cold. The memory, and my reflection upon it, may also prompt me to
feel more deeply about my friend. I may empathize with her and experience feelings
of regret in response to my words. Finally, reason may help me decide how best to
handle the situation and make amends for my earlier response. Without memory,
this entire process of moral discovery and growth would be lost. Thus, our memo-
ries can be educational and play an important role in enabling us to learn from our
past moral failings.
Memories Alert Us to Particular Obligations to Others We can have both general
and particular moral obligations to others. I have general obligations to others in
virtue of their being human or satisfying certain morally relevant criteria, whatever
such criteria may be. But, in addition to these, I can have particular moral obliga-
tions to others because of my relationship with them and the way that this ­relationship
106 J. Kirsch

has unfolded in the past. For example, I may be obligated to help you find a new job
since—in some morally objectionable way—I caused you to lose the old one. Or I
may have an obligation to someone in virtue of a morally innocent act that I have
performed. I may be obligated to build you a website if I promised to do so in
exchange for your landscaping services. In this case, the promise, and the comple-
tion of your part of the deal, is the source of my obligation to you. But with no
memory of this, I would be left unaware of my obligation.
Memories Can Serve as a Sign of Respect for Others It may be unkind of a per-
son to erase or distort the memory of some immoral action that she performed in the
past. Forgetting in this situation may suggest a level of insensitivity towards the
person wronged. Remembering, in contrast, suggests that our concern for the other
person has not faded away with the passage of time. Perhaps this is why we desig-
nate certain days as ‘remembrance’ days. On such days, we reflect upon a harm
done, or sacrifice made, and show our concern and gratitude for others. We may
have similar reasons for remembering non-moral events. Failing to remember cer-
tain significant non-moral events often reflects poorly upon a person’s character. In
the right kind of case, we take those who forget to be callous or coldhearted. Again,
the details of the case make all of the difference: If my spouse cares even less about
the date of our anniversary than I do, then I may do him no harm by forgetting it.
But if I am so wrapped up in my professional affairs that I forget the highly antici-
pated birthday of my ten-year-old son, then the harm done may be comparatively
great. A virtue-based theory of ethics is perhaps best able to account for our concern
with forgetting in this context. Forgetting, in such cases, reveals a degree of self-­
centeredness, or narcissism, that many of us would find objectionable.

3.2  Reasons for Forgetting

None of this implies that we ought to retain our past in its entirety, even if we could.
Gordon Bell, co-author of Total Recall: How the E-Memory Revolution Will Change
Everything (2009), has set about video recording every moment of his life. If there
is an ethics of memory, it is unlikely that it would require anything this extreme.
Indeed, as I intend to show, our reasons for remembering must at times be weighed
against our reasons for forgetting, reasons that can be impressively, if unexpectedly,
strong. Let us now consider when, if ever, we may have moral reasons to forget.
Forgetting Can Reduce Personal Suffering Rememberers are sometimes better
able to move on and heal themselves when they forget certain unpleasant moments
of their lives. And even if they do not forget these events altogether, we think that
they may benefit from giving their memories a rest, so to speak. Surely there is a
difference between remembering that something has occurred, and rehearsing that
occurrence over and over again in one’s own mind. When we say that a person has
good reasons to forget the past, it is generally the latter, what we might call, ‘mem-
ory rehearsal’, that we have in mind. This is because the control that we have over
our memories tends to be indirect. Ordinarily, we cannot will that we forget
Interpreting Things Past 107

s­ omething with any degree of success. But, at least some of the time, we can directly
control whether or not we rehearse a particular memory. And one consequence of
not rehearsing a memory is that it may fade into the background, as it were, or dis-
appear altogether. This is an issue that I will explore more fully later in the chapter
when I consider the degree of control that we have over our memories.
Forgetting Can Reduce the Suffering of Others Reliving the past in the form of
one’s memories can also harm individuals who appear in these memories. If I have
forgiven another person for something that he has done, and am in the process of
repairing a broken relationship with him, it can be unproductively damaging to reg-
ularly revisit the past. Moving forward in a relationship often demands that we quiet
memories of the past, even if we do not rid ourselves of them altogether. To be sure,
there are times when we are unwilling or unable to do this. But this is generally
because, at such times, we are also unwilling or unable to preserve the relationship.
Clearly, there are cases when we cannot move forward in a relationship because the
damage done is too great. In still other cases, we take ourselves to be justified in
causing eternal torment in the form of memories to those who have harmed us.
When we do hold fast to these memories, we – for better or worse – fail to let the
wrongdoers off the moral hook.
It follows from all of this that forgetting, like remembering, can have a place in
a morally decent life. Forgetting can make life easier to bear. Friedrich Nietzsche
once wrote that “it is possible to live almost without memory.. .. but it is altogether
impossible to live at all without forgetting” (1983, p. 62). Nietzsche was making a
somewhat different point here, but his words are instructive nonetheless. Living, or
moving forward in life, often demands that we distance ourselves from the past. If
we continuously relive every heartbreaking, gut-wrenching moment of our lives, we
will find ourselves at an emotional standstill or worse.

3.3  Moral Considerations Surrounding Interpretation

Now how we interpret the past is not straightforwardly a matter of remembering or


forgetting. The way that we interpret the past is more a matter of giving meaning or
shape to our memories; it is a way of making sense of them and putting them
together. Still, our interpretations of the past appear to have some influence over
what we remember and what we forget. Events that play a crucial role in the way
that we interpret our lives tend to stay with us while others fade away with the pas-
sage of time. A personal anecdote may be of some use here. When I was an adoles-
cent about the age of 12, I attended an outdoor summertime birthday party at a
friend’s house. We were all having a great time until the boys down the street
decided to crash the party armed with various water containers and devices. A cer-
tain boy was particularly well prepared and chased me for several blocks with a
bucket full of water. Exasperated and unable to run any farther, I finally accepted my
fate and allowed him to pour it over my head. Now I cannot be sure what I thought
or felt at that time, but as I interpret things now, this is the moment when I fell in
108 J. Kirsch

love with him (in the way that a 12-year-old can)—as I stood there exhausted,
humiliated, and dripping with water. Arguably, having a bucket of water poured
over one’s head is a pretty momentous event no matter how you look at it (although
it is unclear why it would prompt feelings of love). But it seems to me that the role
that this event played in the story of my life gave it a special meaning—a meaning
that made it resilient to the passage of time.
This example does not obviously raise any moral concerns, except for the fact
that the young man in question still owes me a serious apology, but we can imagine
other cases that do. When a relationship, romantic or otherwise, falls apart, there is
often more than one way to interpret what happened. The interpretation that a per-
son eventually settles upon will shape the way that she sees things and guide her
future actions. For example, a cold and distant mother may consider various inter-
pretations of why her son, and only child, has severed all ties with her. According to
one interpretation, she may be fully responsible for the failure of the relationship
because she made a series of selfish and insensitive decisions. But, according to
another, she may be free from blame because her actions resulted from a lifelong
struggle with a debilitating form of depression. The interpretation that she accepts
will influence what she does next: Will she embrace her son’s decision and view him
as an ungrateful write-off? Or will she reach out to him with the hope of repairing
the broken relationship? If so, where will she begin?
As time passes, we tend to interpret and reinterpret the earlier parts of our lives.
In the course of doing this, we sometimes take on the third-person point-of-view
and theorize about what our true motives may have been. Indeed, our tendency to do
this seems to increase as we move farther away in time from our earlier selves; it is
as though great lapses of time make it easier for our present selves to relate to our
past selves as different persons. Whatever the case may be, the process of interpret-
ing and reinterpreting the past matters because, as we have seen, it opens and closes
doors of morality for us.

4  Taking Control of Our Memories

None of the moral considerations explored thus far will have any force if it turns out
that memory is entirely beyond our control. Just as we may doubt that there can be
an ethics of belief because we cannot believe at will, so also we may doubt that there
can be an ethics of memory because we cannot remember at will. In response to this
worry, I will now show that we can exercise at least some indirect control over our
memories by exercising direct control over the processes that contribute to them. If
we do have at least some control over these processes, then we may be able to
express legitimate concern about the way that people remember, and as we shall see,
interpret, the past.
Popular science writer, Joshua Foer, spent years investigating the techniques that
memory champions use to accomplish their astonishing feats of memory. Foer him-
self employed these techniques while training to become the memory champion of
Interpreting Things Past 109

the United States. In Moonwalking with Einstein (2011), he argues that a person’s
ability to remember is not entirely determined by genetic or innate factors. Most
memory champions are not prodigies that are born with an unusual ability to remem-
ber; instead, they spend years developing and perfecting techniques and training
their memories. What this shows, argues Foer, is that we can develop and implement
strategies for remembering information that might otherwise escape us. Foer is not
particularly interested in autobiographical memories. But if we can use these strate-
gies to remember poems or telephone numbers, then it seems possible that we can
use them to remember important personal events from the past. By putting our
minds to it, as it were, we can exercise at least some control over our memories.
Memory is not always, or typically, under our deliberate control. But we can, in
some indirect sense, ‘remember at will’ by implementing any number of memory
strategies and techniques.
Foer focuses upon competitive memorization and the use of techniques that are
internal to the mind, such as memory palaces. But conveniently for those of us who
are not competing in tournaments, we can assist or enhance our memories by ‘cheat-
ing’, or using various external aids. We can help ourselves remember past events
through the use of journal entries, photographs, mementoes, and even social media
sites like Facebook and Twitter. More than ever before, we are able to document,
and thus remember, much of our lives. While I have a collection of fading photo-
graphs to assist me in remembering my early childhood, my children will have
thousands of digital photographs and videos to assist them. For better or worse, our
lives are being documented and immortalized in an unprecedented level of detail.
We may also be able to exercise control over how accurate or representative our
memories are by guarding ourselves against various documented memory biases.
In, “The Least Likely of Times; How Remembering the Past Biases Forecasts of the
Future,” Carey K. Morewedge, Daniel T. Gilbert, and Timothy Wilson (2005) inves-
tigate one kind of memory bias and explain how we may be able to avoid it in our
personal lives. Their research looks at our tendency to recall unrepresentative mem-
ories of the past and use them to make biased predictions about the future. For
example, even though the United States of America has experienced a variety of
terrorist attacks at home and abroad, Americans immediately think of the destruc-
tion of the World Trade Center on September 11, 2001 when they hear the phrase
‘terrorist attack’. Moreover, when asked to predict how they will respond to future
terrorist attacks, they “tend to overestimate the intensity and duration of that reac-
tion” (p. 626). Morewedge, Gilbert, and Wilson show that we can curb this form of
biased thinking by either (A) explicitly asking people to recall an atypical case, or
(B) asking them to consider several cases within the same class (p. 627). Now peo-
ple will tend to recall an atypical case whether or not they are explicitly asked to do
so. However, by (A) explicitly asking them to recall an atypical case, we alert them
to the fact that their memory is atypical and nonrepresentative of the class in ques-
tion. We can also curb biased thinking by asking people to consider several cases
within the same class. By prompting people to consider more than one case, we can
ensure that they do not just hit upon a single atypical incident (p. 627).
110 J. Kirsch

While these strategies were implemented in an experimental context, there is


reason to think that they can also be effectively employed in everyday life. Indeed,
Morewedge, Gilbert, and Wilson advise that we give them a try when we witness
people engaging in the kind of biased thinking described in their study:
[W]hen a person thinks about “a trip to the dentist,” a lone root canal may come to mind
more readily than numerous checkups and cleanings, and this may leave the patient feeling
so apprehensive about an upcoming appointment that he or she furtively plans to miss it.
Family members who are concerned about the person’s health may be naturally inclined to
do what they can to keep the person from recalling and dwelling on the single most dreadful
day he or she ever spent in a dentist’s chair, but the present research suggests that they might
be wiser to encourage the person to recall and recount that awful day in painful detail. When
biased recollection is unavoidable, it may make sense to explicitly promote it, thereby alert-
ing people to the unrepresentativeness of the events they are remembering. (p. 630)

Surely, if other people can prompt us to recognize our biases, then we can prompt
ourselves to do the same. In the case at hand, we can learn to recognize the nonrep-
resentative events that we recall in our own lives and, when appropriate, make an
effort to recall a more representative class of events in the future. Studying biased
thinking is often consciousness raising because it alerts us to various biases that can
infect our own thinking. Learning about a bias is often the first step to avoiding the
bias in our own lives. Just as we may be able to prevent ourselves from engaging in
negative thinking by recognizing this thought pattern and prompting ourselves to
change it, so also we may be able to prevent ourselves from developing a distorted
sense of the past by recognizing this bias and prompting ourselves to overcome it.
Finally, and of particular importance for the purposes of this chapter, we can
exercise some control over how we interpret what we remember. If, as narrative
therapists believe, more than one interpretation can be offered for the very same
series of events, then we may have some control over which interpretation we ulti-
mately settle upon. We may be able to borrow the narrative therapist’s strategy of
reconsidering, and when appropriate, revising the narratives of our lives. To be sure,
there is no guarantee that our narratives will become more accurate through this
process. But if we are generally and sincerely motivated by a concern for truth, then
the process of reconsidering a narrative may help us settle upon one that is, as it
were, more fact than fiction.
We can also try to keep our interpretations in check by comparing them with the
interpretations of others. The development of a personal narrative is often an inter-
personal or collaborative process. If our collaborators are reliable, then they can
help us settle upon a narrative that is largely truthful or compatible with the facts of
our lives. Therapists may offer contrasting interpretations of our lives, but so also
may friends, family members, and—as most of us have figured out on our own—
lovers. As research involving long-term couples reveals, the process of remember-
ing and composing a personal narrative is often an interpersonal affair, one that we
share with those closest to us.2 This same group of people may be in a privileged
position to correct our interpretations and call attention to what we have forgotten.

2
 See, for example, “We Remember, We Forget: Collaborative Remembering in Older Couples,” by
Celia B. Harris, Paul G. Keil, John Sutton, Amanda J. Barnier, and Doris J. F. McIlwain.
Interpreting Things Past 111

Even with the assistance of others, we may have a difficult time settling upon the
best or most truthful interpretation of our lives. In part, this is because what we
make of our lives and experiences is to some extent up to us. While there are
undoubtedly certain facts of the matter that cannot be contested, we can exercise
some creative control over the way that we thread these bits and pieces of our lives
together. Suppose, for example, that a recent philosophy Ph.D. decides to abandon
the field after failing to secure a tenure-track job. There are bound to be certain facts
of the matter that would appear in any narrative involving this series of life events.
Let us assume that our philosopher, whom we shall Sarah, sent out job applications
5 years in a row, worked as an adjunct at three different universities, and had two
failed on-campus interviews. Now any narrative that she constructs will include
these facts of her life. But there may be more than one way of weaving these events
together and making sense of them. She may, for example, think of her failure to
secure a tenure-track position as a life-shattering disappointment, or she may—
more optimistically— think of it as a gift that led her to her true passion, a career in
non-profit media.
There is, then, a certain amount of ‘space’ for creative intervention when we act
as interpreters or narrators of our lives. Thus, it is in some sense true that “people
make meaning, meaning is not made for us,” as Drewery and Winslate would have
it (1997, p. 3). The challenge, though, is to find some method for sorting through
these various narratives and discarding those that are blatantly self-deceptive or
fictitious. Like the narcissists in the Rhodewalt and Eddings study, we are often
quick to settle upon a view of ourselves that is flattering but false. In the case under
consideration, Sarah may attempt to comfort herself by devising a self-deceptive
interpretation of her two failed on-campus interviews according to which both hir-
ing committees were biased and unfair in their treatment of her. She may do this for
the purpose of minimizing the harm done to her wounded and delicate ego.
When it comes to memory, interpretation walks a precarious line between search-
ing for meaning and spiraling into self-deception. As interpreters, we do not pas-
sively take in the facts of our lives; instead, we develop narratives around these facts
that give them meaning and tie them together. What we make of an event, or signifi-
cant life occurrence, is to some extent up to us. The particular interpretation that we
accept helps to position this occurrence along a kind of time line that projects into
the future, giving it shape and meaning. There is, then, a sense in which our inter-
pretations can be self-fulfilling. Viewing a significant life occurrence as, say, a
blessing in disguise and not as a tragedy can set us along a positive path of life,
prompting us to be mindful and self-aware. In this way, the interpretations that we
accept can guide our behavior and make themselves true. There is a kind of boot-
strapping effect at play whereby our interpretations feed into the future, become
self-sustaining, and ultimately validate the past.
There are, of course, limitations here; only a certain set of interpretations can be
self-fulfilling in this way. Other things being equal, believing that I was justified in
murdering my spouse will not bring it about that I really was justified in murdering
112 J. Kirsch

my spouse. Any interpretation that we consider will have to contend with a set of
incontestable facts. But it is often the case that more than one interpretation will
work, and that the interpretation that we settle upon will be self-fulfilling.

5  Conclusion

We do not always think of ourselves as deliberately or explicitly interpreting the


past. Ordinarily, this is something that we do automatically and with little or no
forethought. But, to some extent, we can intervene in the process and exercise cre-
ative control over how it happens. As we have seen, we can be vigilant and learn
how to guard against any number of well-documented biases and distortions. And
we can also make ourselves aware of the interpretations that we accept and any oth-
ers that might be offered in their place. As interpreters of our lives, we may have the
last word, but others can sometimes steer our interpretations towards the truth and
help us see things more clearly. We should take advantage of the control that we
have over our interpretations because they not only shape our understanding of our-
selves, but of the social and moral world around us.

References

Bell, Gordon, and Jim Gemmell. 2009. Total recall: How the e-memory revolution will change
everything. New York: Penguin Group.
Drewery, Wendy, and John Winslate. 1997. The theoretical story of narrative therapy. In Narrative
therapy in practice: The archaeology of hope, ed. Gerald D.  Monk, John Winslade, Kathie
Crocket, and David Epston, 32–52. San Francisco: Jossey-Bass.
Foer, Joshua. 2011. Moonwalking with Einstein. New York: Penguin Books.
Loftus, Elizabeth F., and Hunter G. Hoffman. 1989. Misinformation and memory: The creation of
new memories. Journal of Experimental Psychology 1 (118): 100–104.
Madigan, Stephen. 1999. Inscription, description and deciphering chronic identities. In Perspectives
on psychotherapy: Deconstructing psychotherapy, ed. Ian Parker, 150–163. Sage Publications.
Mineka, Susan, and Kathleen Nugent. 1997. Mood-congruent memory biases in anxiety and
depression. In Memory distortion: How minds, brains, and societies reconstruct the past, ed.
Daniel L. Schacter, 173–193. Cambridge: Harvard University Press.
Morewedge, Carry K., Daniel T. Gilbert, and Timothy Wilson. 2005. The least likely of times; How
remembering the past biases forecasts of the future. Psychological Science 16 (8): 626–630.
Newman, Eryn J., Shari R. Berkowitz, Kally J. Nelson, Maryanne Garry, and Elizabeth Loftus.
2011. Attitudes about memory dampening drugs depend on context and country. Applied
Cognitive Psychology 25: 675–681.
Nietzsche, Friedrich. 1983. Untimely meditations. Trans: R.  J. Hollingdal. London: Cambridge
University Press.
Rhodewalt, Frederick, and Stacy K.  Eddings. 2002. Narcissus reflects: Memory distortion in
response to ego-relevant feedback among high- and low- narcissistic men. Journal of Research
in Personality 36: 97–116.
Interpreting Things Past 113

Ross, Michael, and Roger Buehler. 1994. Creative remembering. In The remembering self;
Construction and accuracy in self-narrative, ed. Ulric Neisser and Robyn Fivush, 205–235.
Cambridge: Cambridge University Press.
Taylor, Shelly. 1991. Positive illusions: Creative self-deception and the healthy mind, 1991.
New York: Basic Books.
Self-Interpretation as Software: Toward
a New Understanding of Why False
Self-­Conceptions Persist

Tadeusz Wiesław Zawidzki

Abstract  Typically, action-relevant, false beliefs do not persist. Human beings


with false beliefs about how to secure food, social partners, and physical security do
not succeed at these tasks as well as human beings with true beliefs about them. For
this reason, mechanisms operating at phylogenetic, ontogenetic, and cultural scales
weed out such false beliefs. However, there is voluminous empirical evidence, gath-
ered over many decades, that many highly persistent self-conceptions are false. In
this paper, I propose a solution to this apparent puzzle. Rather than evaluating self-­
directed beliefs for truth, I argue that we should evaluate them in the way we evalu-
ate computer software. Software is not evaluated in terms of the degree to which it
truly describes the workings of the computer that runs it. Instead, software is evalu-
ated in terms of the costs and benefits of what it gets the computer that runs it to do.
Similarly, I argue that we can understand the persistence of false self-conceptions
only if we evaluate them in terms of the costs and benefits of what they get human
beings to do, instead of how well they describe the workings of human cognition.
This perspective is inspired by Dennett’s notion of “cognitive tools”, but it departs
from Dennett in defending a socialized version of this notion. Rather than tools for
improving individual cognition, self-conceptions are tools for interacting with
others.

Keywords  Self-interpretation · False self-conceptions · Cognitive functions ·


Freedom of the will · Conscious self · Propositional attitudes

T. W. Zawidzki (*)
Department of Philosophy and Mind-Brain Institute, George Washington University,
Washington, DC, USA
e-mail: zawidzki@gwu.edu

© Springer Nature Switzerland AG 2018 115


P. Pedrini, J. Kirsch (eds.), Third-Person Self-Knowledge, Self-Interpretation,
and Narrative, Contributions To Phenomenology 96,
https://doi.org/10.1007/978-3-319-98646-3_7
116 T. W. Zawidzki

1  Introduction

One over-arching lesson of the last century of social and cognitive science is that
many of our most cherished self-conceptions are deeply misleading. Common sense
has it that we make free, rational decisions, based on conscious reasons, i.e., mental
states like beliefs and desires, to which we have transparent access. These conscious
mental states, and the reasoning we apply to them, are taken to constitute a unified,
coherent whole that persists through time. Many of them are thought to be accurate
records of past mental states. Many of them are thought to bear contents that are
precisely expressible in simple sentences of natural language, and to be governed by
similar inferences, e.g., believing that P precludes believing that not P, in the same
way that asserting the sentence P precludes asserting the sentence not P. The last
century of social and cognitive science has provided strong evidence that all of these
components of our self-conception constitute false and deeply misleading assump-
tions about how the human mind actually functions.
This raises a number of puzzles. One is why we seem so convinced of erroneous
assumptions about our own minds. This is easily explained however. Most people
are not aware of the evidence against our commonsense self-conceptions; so, it is
unsurprising that they remain convinced of their truth. But a deeper puzzle is not so
easily dismissed. In domains other than self-interpretation, false and deeply mis-
leading, action-relevant beliefs tend not to persist, independently of whether or not
those who hold them are aware of evidence against them. The reason is that such
beliefs lead to unsuccessful actions, and this leads to selection against such beliefs
and the mechanisms that generate them, at phylogenetic, ontogenetic, and cultural
scales. For example, false beliefs about food location, social partner viability, and
environmental dangers do not typically persist in populations of natural cognizers.
But it seems that our commonsense self-conceptions consist in many false, action-­
relevant beliefs. For example, surely assumptions one makes about one’s own
capacities are action relevant. So, even if most people are not explicitly aware of
evidence against the truth of many of their self-conceptions, it is still puzzling that
false and deeply misleading ones persist. Surely, such self-conceptions should lead
to unsuccessful actions, just like false beliefs about food location, and hence be
selected against, either in phylogeny, ontogeny, or cultural evolution.
This chapter is about this latter puzzle. The aim is to explain why false and
deeply misleading self-conceptions persist, unlike false and deeply misleading
beliefs in other domains. Key to this explanation is a reconceptualization of the
cognitive role of self-conceptions. I argue that, unlike beliefs about the environ-
ment, many beliefs about the self do not aim to correctly represent independently
constituted facts. Instead, self-directed beliefs, particularly those over which per-
sons are taken to be authoritative, aim to create facts. In much the same way that
computer software aims not to describe facts about computers that run it, but, rather,
to transform them into machines that respect constraints it specifies, on the view
explored here, many self-directed beliefs aim not to describe facts about human
beings, but, rather, to transform human beings into the kinds of persons who respect
Self-Interpretation as Software: Toward a New Understanding of Why False… 117

the constraints these self-directed beliefs specify. If this hypothesis is correct, then
it can explain why many deeply misleading self-conceptions persist: their raison
d’être is not to correctly represent facts about the brain and other features of human
beings; rather, it is to constrain human behavior in certain ways. In what ways? I
argue that self-conceptions function to turn human beings into good coordination
partners relative to their social environments. If this is true, then their persistence
depends not on their truth, but on their fit with other self-conceptions that prevail in
the ambient social environments of the persons who harbor them. Self-conceptions,
on this view, are frequency-dependent phenomena, akin to languages: they persist
as long as they help coordination in virtue of complementing the self-conceptions of
those with whom we need to coordinate.
The chapter proceeds as follows. In Sect. 1, I review a number of central compo-
nents of self-conceptions that appear, at least according to many western philoso-
phers, to prevail in many cultures. I also discuss the empirical evidence against their
truth. That is, I argue that, if we treat these self-directed beliefs as descriptions of
facts about human beings, then there is strong evidence that they are false. In Sect.
2, I review common assumptions about the function of self-conceptions, on which
the persistence of these empirically disconfirmed self-directed beliefs is problem-
atic. I then discuss a number of theorists who urge the kind of reconceptualization I
defend in this chapter: that many self-directed beliefs be treated more as software,
i.e., as aiming to transform human beings into the kinds of persons that respect the
constraints they specify. In Sect. 3, I sketch an account of how the persistence of
self-conceptions can be explained when they are reconceptualized as software. Sect.
4 explores some questions, implications, and speculations arising from this
proposal.

2  Empirical Evidence against Prevalent Self-Conceptions

There is strong empirical evidence that the following six prevalent self-conceptions
constitute false descriptions of human beings, particularly of how the human mind/
brain functions: (1) the mind is transparent to itself (self-transparency), (2) human
reasoning and decision-making occur in a unitary, enduring, causally efficacious,
conscious self (the Cartesian ego), (3) human decision-making is free from con-
straints, in the sense that the Cartesian ego is exclusively causally responsible for a
significant proportion of human decisions, (4) a person’s memory accurately records
what the person’s mind has previously experienced, (5) a significant proportion of
the mental states that issue in human decisions and actions have contents that are
fully expressible in simple sentences of natural language, and are governed by the
same inferential rules as these sentences, (6) human inference, on the whole,
respects standard norms of rationality. I now turn to a brief overview of empirical
evidence against and conceptual problems with these self-conceptions, considered
as descriptions of how the human mind/brain functions.
118 T. W. Zawidzki

2.1  Self-Transparency

This is perhaps the most important of the six self-conceptions on which I focus here.
The reason is that it is a higher-order self-conception that grounds the other, lower-­
order ones. That is, self-transparency is a conception of our other self-conceptions,
namely, that they are by and large accurate. If human minds are transparent to them-
selves, then, if human minds conceptualize themselves as consisting in Cartesian
egos that come to free and rational decisions, based on the conscious consideration
of linguistically expressible mental states, many of which are accurate records of
past mental states, then these self-conceptualizations are justified and likely true.
Unfortunately, there is very strong empirical evidence against the assumption of
self-transparency.
The assumption of self-transparency has been known to be problematic for some
time (McGeer 1996). Many of the most infamous and long-standing results from
social psychology appear to confute it. Carruthers (2011, Chap. 11) is a very thor-
ough survey. For example, Nisbett and Wilson (1977) famously show that consum-
ers are systematically misinformed about the reasons for their product preferences:
even when product placement (typically on the right side) best explains choice,
consumers persist in confabulating more reasonable, yet spurious justifications, cit-
ing differences between products that are not actual. Wegner (2002) shows that
subjects hypnotized to move an object tend to confabulate reasons for the action,
e.g., tidying the room. Brehm (1956) identifies a “choice effect”: subjects rate two
products equally along a number of dimensions, and then must choose one. When
rerating the products after the choice, subjects rate the one they chose more highly
than they did initially. Carruthers (2011, 356–365) also provides a detailed overview
of the empirical literature on “counter-attitudinal advocacy”, arguing that it is best
explained on the hypothesis that human beings systematically misconstrue their
own beliefs in order to reduce cognitive dissonance. For example, after having to
write an explicit defense of a position with which they have previously strongly
disagreed, subjects express lower levels of disagreement when asked again, but only
if they have not been substantially compensated for writing the defense. The idea is
that if they have not been paid sufficiently, subjects feel like their counter-attitudinal
advocacy is pursued freely, which leads to cognitive dissonance that can be reduced
only by revising initial self-attributions of attitudes such that they are more consis-
tent with the advocacy.
Thus, there appears to be strong empirical evidence that one of the central com-
ponents of prevalent self-conceptions is false: human minds are not self-transparent.
To appreciate how jarring such results are, consider how one might react to the
claim that one does not actually think what one sincerely claims to think. It seems
obvious that anyone accused of such self-opacity would react with incredulity and
probably offense. So, there is little doubt that self-transparency is an extremely
important component of prevalent self-conceptions. Yet, there is strong evidence
that it is false. And if we are wrong about this higher order self-conception, then the
truth of lower order self-conceptions to which it applies is also at risk.
Self-Interpretation as Software: Toward a New Understanding of Why False… 119

2.2  The Cartesian Ego

I use the term “Cartesian ego” very loosely here. I do not intend to refer to the
details of Descartes’ actual model of the mind. Rather, I use the term to pick out a
vaguer notion: roughly, the model of the mind that Descartes seems to have inspired
in philosophy and psychology. It is, in any case, the view of the mind that many
anti-Cartesians take themselves to be critiquing (Ryle 1949; Dennett 1991). The
assumption that the human mind is a Cartesian ego amounts to the view that human
reasoning and decision-making take place in a unitary, enduring, causally effica-
cious, conscious self. As Descartes puts it, “it is so obvious that it is I who doubt, I
who understand, I who will … indeed I am also the same one who imagines; for …
the very power of imagining … constitutes part of my thought … I am the same one
who senses or who takes note of bodily things as if through the senses. For example,
I now see a light, I hear a noise, I feel heat” (1649/1979, 20). This unified mental
substance is taken to be the most important cause of intelligent action. For example,
many intelligent actions are assumed to be products of long-term plans formulated
via processes of multimodal, conscious reasoning, taking place in the mind. In order
for long-term plans to have such efficacy, the substance in which they are formu-
lated and enacted must endure through time: the same substance must first formu-
late plans and then, when the time comes, execute them. Although it is unclear how
long-standing and cross-culturally robust this concept of the mind is, the fact that it
continues to exercise such intuitive pull on philosophers and cognitive scientists,
decades and perhaps centuries after serious empirical and conceptual objections
were first raised against it, suggests that it is a strongly rooted denizen of our com-
monsense self-conception. And, indeed, recent empirical studies suggest that human
adults have default conceptions of the mind that look something like the Cartesian
ego (Forstmann and Burgmer 2015).
Despite its intuitive pull, there are strong empirical and conceptual reasons to
think that the Cartesian ego is a highly misleading model of the mind. It has been
known for some time that the mind is “compartmentalized”, i.e., that, far from a
multimodal field, unified in consciousness, much of the mind’s work is the result of
unconscious activity by quasi-independent modules, with little access to each other
(Carruthers 2006). The causal efficacy of conscious thought has also been subject to
persuasive, empirical and conceptual critique, for centuries (Wegner 2002). Despite
centuries of repeated attempts, philosophers in both Asian and European traditions
have failed to make sense of the notion that the mind is an enduring substance
(Hume 1738/2000; Garfield 2015). In fact, as a number of philosophers have
recently argued, disorders of the self suggest that the unitary, enduring, causally
efficacious, conscious self is a fragile, socially-scaffolded construct, rather than the
natural, causal basis for intelligent action (Flanagan 1991; Neisser and Fivush 1994;
Tekin 2011).
120 T. W. Zawidzki

2.3  Freedom of the Will

The assumption that human decisions can be entirely free from causal influence is
closely related to the assumption that the human mind constitutes a Cartesian ego.
As philosophers have long noted (Locke 1689/1997), one of the principal functions
of the enduring self is forensic: we need an enduring, conscious locus of decision-
making in order to hold persons responsible for their actions. For example, if the
perpetrator of a criminal act is not, in some sense, the same as the agent who planned
the act, then how can she be held responsible? But it makes little sense to hold per-
sons responsible for the decisions they enact if those decisions are products of fac-
tors beyond their control. Hence, the decisions reached in the Cartesian ego must, in
some sense, be free from causal constraint. There is now a lot of empirical evidence
that this link between freedom and responsibility is central to, at least western, com-
monsense self-conceptions. For example, it appears that most naïve subjects attri-
bute responsibility for actions only to the extent that the actions are products of
decisions free from all causal constraints. This would explain why, e.g., encourag-
ing doubt about freedom of the will in naïve subjects leads to increased cheating
(Vohs and Schooler 2008). Since it is clearly a widespread assumption that people
are responsible for their actions, this suggests that belief in freedom of the will is
equally widespread.
As with the Cartesian ego, however, there have long been conceptual and empiri-
cal reasons for doubting that human decision-making can be free from all causal
constraint. First of all, as many philosophers have long noted, it is hard to see how
to even make sense of this concept of freedom of the will. Presumably, something
causally explains why persons decide one way rather than another. Whatever this
explanatory factor is, it causally determines one’s decisions. But then how can
human decisions be free from all causal constraints? This conceptual objection to
the assumption that human decisions are free has, over the last few decades, been
joined by a growing variety of empirical data revealing the causal determinants of
human decisions. Perhaps most dramatically, Libet (1985), and work this paradigm
has inspired, appear to show that the neural signatures of behavioral decisions pre-
cede the moment at which we consciously think we freely make them, suggesting
that they are not really products of our free will, which appears to be nothing more
than a post-hoc interpretation of activity in which our brains engage due to causal
factors beyond our ken. Wegner (2002) is a compendium of empirical results sug-
gesting that much behavior that subjects think is freely chosen is demonstrably a
product of factors beyond their consciousness and control. Harris (2012) has
recently summarized the empirical case against freedom of the will, updating philo-
sophical arguments dating to the Stoics in the ancient era, and Spinoza in the mod-
ern era. Even sophisticated contemporary defenses of freedom of the will against
such evidence and arguments rely on reconceptualizations of it that are at odds with
the commonsense concept. For example, Dennett (1984, 2003) argues that freedom
Self-Interpretation as Software: Toward a New Understanding of Why False… 121

of the will is compatible with determinism, a view that appears at odds with ­common
sense, at least in some contexts (Nichols and Knobe 2007).1 And Kane (1996),
though he defends incompatibilism, grounds free decisions in truly random pro-
cesses, also arguably at odds with the commonsense notion.

2.4  Memory as Accurate Copy

Common sense clearly recognizes that memory is not perfectly reliable. After all,
forgetting and forgetfulness are commonsense concepts. However, forgetting is
only one of two types of memory failure. To forget is to have experienced something
in the past that one cannot remember in the present. But what of the converse kind
of memory failure: consciously misremembering something in the present that one
has not experienced in the past? Common sense clearly does not, typically, recog-
nize this possibility. On our commonsense self-conception, if one clearly and dis-
tinctly remembers having experienced X in the past, one must actually have
experienced X in the past. That this is a central assumption of common sense is
revealed by how jarring we tend to find fictional illustrations of its denial. Consider
the science fiction of Philip K.  Dick, which routinely explores the possibility of
manufactured memories in works such as “We Can Remember It for You
Wholesale” (Dick 1997), and Do Androids Dream of Electric Sheep? (Dick 1996)
The overwhelming lesson of the last forty years of research into human memory
is that, not only are the memory illusions explored in Dick’s fiction possible, to an
alarmingly large degree, they are actual. Human beings routinely remember having
experiences that they demonstrably never underwent. To illustrate just two exam-
ples of this, consider Neisser’s work on “flashbulb memory” (Neisser and Harsh
2000), and Loftus’s work on eyewitness testimony (Loftus 1996). Flashbulb mem-
ory is the phenomenon of clearly remembering, in detail, where one was when one
first learned of a publicly prominent event, like the assassination of John F. Kennedy,
or the 9/11/2001 terrorist attacks on New York City. Neisser tested this on students
in a class he was teaching when the space shuttle Challenger exploded in 1986. The
day after the event, he asked his students to write down in detail where they were,
and what they were doing, when they first learned of the event. A number of years
later, he contacted the students, asking the same question. Those who replied gave
completely different answers than before, with equal confidence. Neisser speculates
that prominent events are often relived, and the contexts in which they are relived
come to be confused with the contexts in which they are first experienced, leading
to false memories. Loftus’s work shows how distressingly easy it is to induce false
memories in witnesses through leading questions and other means. For example, in

1
 Though see Nahmias (2014) for a nice discussion of the empirical debate regarding whether or
not the folk are incompatibilists.
122 T. W. Zawidzki

one paradigm, she showed subjects a film of a traffic accident. Later, she asked
subjects whether, for example, the green car was turning at the time of the accident,
even though there had been no green car in the original film. Subjects overwhelm-
ingly reported having seen a green car in response to this question.
There is now strong consensus among memory researchers that human memory
is reconstructive, rather than an accurate copy of prior experiences (Roediger 1996;
Boyer 2009). But this is clearly not the commonsense conception of memory.
People take their clear, vivid memories to be accurate records of what they actually
experienced in the past. This is a particularly dramatic illustration of a false yet
persistent self-conception. In fact, it is doubly false. At one level, it involves the
false interpretation of particular memories, i.e., as accurately reflecting prior experi-
ence. At another level, it involves a false assumption about how the mind works, i.e.,
as including an accurate record of prior experiences.

2.5  Language as Window on Thought

We often attribute thoughts to ourselves and others using public language. If we


assume that such attributions are accurate, this means that we also assume that the
contents of our thoughts can be captured accurately using words of public language.
Since it is also widely assumed that such reported thoughts are independent of their
attribution in public language, common sense seems to assume that the contents of
many of our language-independent thoughts can be accurately represented using
public language vocabulary. In addition, we often justify our behaviors by self-
attributing trains of reasoning involving inference over mental states specified using
public language vocabulary. We take such practical inferences to be accurately rep-
resented by the inferential rules linking the public language sentences that represent
our thoughts. This has persuaded many influential philosophers that cognitive sci-
ence should respect the assumption that sentences of public languages, and inferen-
tial rules governing them, constitute accurate representations of, respectively, the
contents of language-independent thoughts, and the processes that operate on them
(Fodor 1975, 1978).
However, there are good reasons to doubt this assumption. First, some philoso-
phers of cognitive science have argued that it is highly unlikely that the language-
independent cognitive states of our brains have contents that can be captured
precisely in words of natural language. For example, Dennett suggests that, unlike
the opinions we formulate and endorse in public language, language-independent
cognitive states bear contents that are specifiable only in the language of the cogni-
tive sciences, referring to the biological functions of the neural systems of which
they are states (Dennett 1978, 306). This general point is elaborated in great detail
by Kathleen Akins (1996), who argues that the contents of most natural, sensory
system states are inherently narcissistic in a way that specification in terms of public
language words, referring to an objective reality, cannot capture. This makes sense,
given the inherently intersubjective nature of public language: language must speak
Self-Interpretation as Software: Toward a New Understanding of Why False… 123

of a perspective-independent world, since it aims to facilitate communication and


coordination among different perspectives. But there is no such constraint on
language-independent neuro-cognitive representations, which aim only to support
adaptive behavior relative to one idiosyncratic perspective.2
Besides concerns about the use of public language to specify the contents of
language-independent thought, many philosophers of cognitive science have argued
that the structure and processing of language-independent thought cannot be any-
thing like the syntax and inferential rules governing public language. One reason for
this appeals to what is known of the computationally relevant structural properties
of the brain. These seem to be nothing like those of a language (Churchland 1981,
1990). For example, Ramsey et  al. (1990) argue that, in some neurally plausible
cognitive models of propositional memory, mental states are not functionally dis-
crete in the way sentences of public language are: every state is involved in every
computation, unlike in inferences in public language. Most dramatically, Frankish
(2004, 2009) has recently argued that language-independent belief differs from lan-
guage-dependent belief in that the former and not the latter can be partially active.
For example, in language, one cannot endorse both P and not-P: this is what Frankish
calls “flat-out” belief. But there is strong evidence that language-independent belief
is not categorical in this way: it is possible to accord 75% credence to P and 25%
credence to not-P simultaneously. This leads Frankish to posit a distinction between
“mind”, i.e., non-categorical, language-independent cognitive processing imple-
mented in the brain, and “supermind”, i.e., “flat-out”, categorical beliefs imple-
mented via “premising policies” applied to sentences of public language.
There is little doubt that when we attribute mental states, like beliefs and desires,
to ourselves, and others, using public language sentences, we take ourselves to be
accurately capturing the contents of these mental states, yet also understand these
mental states as independent of the sentences we use to attribute them. And when
we justify our behavior by appeal to inferences applied to public language sentences
expressing our mental states, we take ourselves to be accurately depicting the cogni-
tive processes that give rise to them. However, there are persuasive reasons to view
these components of our self-conception as a false and misleading picture of how
our minds actually work.

2.6  Rationality

Philosophers in the European tradition have, with some notable exceptions, long
assumed that human beings are essentially rational. This long-standing assumption
seems to be endorsed by common sense. It is widely assumed that, with the excep-
tion of individuals suffering from psychiatric disorders, human decision-making
typically respects norms of rationality. This is so central to our practice of interpret-
ing ourselves and others that it leads many philosophers to argue that rationality is

 For a roughly analogous point, see Clark (1997, Epilogue).


2
124 T. W. Zawidzki

constitutive of interpretation: we cannot treat an individual as minded unless we


also treat the individual as rational (Davidson 1984; Dennett 1987).
However, perhaps the most robust result of the past four decades of empirical
work in social psychology is the degree to which everyday decision-making fails to
respect norms of rationality. Much of the research cited above bearing on the non-
transparency of the mind is relevant here. It is because much of our decision-making
is based on irrelevant factors that we tend to confabulate rationally kosher justifica-
tions for decisions arrived at by different means. This is the point of Nisbett  and
Wilson’s work on product preferences: the placement of products in the visual field
of the consumer is not relevant to their value, yet it predicts consumer choice better
than relevant variables; a fact that consumers hide from themselves and others
through confabulation. Tversky and Kahnemann have documented in detail the irra-
tional biases to which human beings are prone (Tversky and Kahnemann 1974). This
sort of research leads many theorists to posit a distinction between fast, efficient,
automatic, innate, unconscious, and biased reasoning processes that fail to respect
norms of rationality (System 1), and slow, laborious, culturally acquired, conscious
reasoning process that respect norms of rationality (System 2) (Stanovich 1999;
Frankish 2004; Kahneman 2011). System 2 does not come naturally to human beings:
it is a slow, highly fallible system dependent on pedagogical instruction. This seems
to contradict the commonsense assumption that human beings are naturally rational.

2.7  The General Lesson

It is possible to question these reasons for skepticism regarding central components


of our self-conception. It is even possible to question whether all of these assump-
tions actually constitute important components of our self-conception. Nevertheless,
even if the details reviewed above can be questioned, the broader puzzle remains:
we have good reasons for thinking that important components of our self-concep-
tion are false and radically misleading, yet they are also remarkably persistent. In
other words, one can treat the puzzle as hypothetical: if the strong reasons we have
for doubting central components of our self-conception turn out to be dispositive,
then how do we account for their remarkable persistence? Even if the foregoing is
wrong in detail, the very existence of any philosophy of mind or cognitive science
appears to presuppose that common sense can be very wrong in its assumptions
about how the mind works.3 Thus, any interesting view in philosophy of mind or
cognitive science must somehow account for the persistence of mistaken compo-
nents of our self-conception. This is no different from the task facing philosophical
or scientific amendments to other components of our world-view, e.g., explaining
why our assumptions about the physical world can be radically false yet remarkably

3
 Even as ardent a defender of folk psychology as Jerry Fodor argues that some components of our
commonsense self-conception are radically misleading, i.e., the assumption that we acquire con-
cepts from experience (Fodor 1975).
Self-Interpretation as Software: Toward a New Understanding of Why False… 125

robust. In what follows, I propose and defend a theory of why the components of
our self-conception reviewed above might persist even if radically false.
According to one strategy for addressing why misleading conceptions persist,
often employed to explain the persistence of misleading assumptions about the
physical world, although false in detail, commonsense conceptions approximate the
truth at spatiotemporal scales relevant to everyday life. Some have proposed a simi-
lar strategy for explaining the persistence of misleading assumptions about the mind
(Smolensky 1988). This will not be my strategy. While it makes sense to talk of
illusions of scale in physical domains, it is not clear how this transfers to the psycho-
logical domain. We can understand why the human visual system is insensitive to
features at extremely microscopic scales, or in non-visible parts of the electromag-
netic spectrum, or of objects moving at extreme velocities. However, it is difficult to
explain why brains can know themselves only as conscious, transparent, unified,
persisting, rational decision-makers, free from causal constraint, with access to
accurate records of past experiences. Even if the details of neural processing are
inevitably invisible to us, why conceive of the mind in these ways, rather than oth-
ers? What makes these abstract assumptions about how the mind works so much
more obvious than those supported by cognitive science and rational reflection? The
mind does not have a visible portion, like physical objects do, so we cannot trace its
misleading, apparent properties to scale constraints on the human visual system.
Rather than pursuing such a strategy, I argue below that the persistence of false self-
conceptions can be explained only if we appreciate that their point is not to describe
the mind at any scale or in any sense; rather, their point is to shape or transform
minds in ways that make their bearers better at coordinating with each other.

3  Against the Spectatorial Model of Self-Interpretation

Persistent yet false self-conceptions pose a problem only if we accept what might be
called the “spectatorial” (Hutto 2008) model of self-interpretation. On this model,
self-knowledge is entirely analogous to knowledge of the environment. Self-
conceptions are descriptions of facts about our minds that exist independently of
such self-conceptions, in much the way that perceptions and beliefs about the physi-
cal environment are descriptions of facts about the physical environment that exist
independently of such perceptions and beliefs.
This understanding of self-knowledge derives from Descartes, yet it continues to
be very influential even in the wake of centuries of critiques of Cartesianism
(Nietzsche 1886/1989, 23–24; Ryle 1949; Dennett 1991). For example, David
Armstrong conceives of consciousness as the mind’s capacity for self-scanning,
entirely analogous to sense perception (Armstrong 1980). On the “Higher Order
Thought” theory of consciousness, first-order thoughts become conscious when
they are represented by second-order thoughts (Rosenthal 2005). Even theories of
self-knowledge that critique the perceptual model of introspection, still conceive of
self-knowledge as analogous to knowledge of the environment (Carruthers 2011): it
126 T. W. Zawidzki

involves the mind’s use of theory-like inference to ascertain independently consti-


tuted facts about itself. All such views have a hard time explaining the persistence
of false self-conceptions. If the mind’s assumptions about how it works are based on
some sort of internal perception, then how can the mechanisms supporting such
perception be so unreliable? If the mind’s assumptions about how it works are based
on theory-like inferences from sensory evidence, then how can the mechanisms sup-
porting these inferences be so unreliable? Surely actions based on such unreliable
information could not be successful, and this would lead to selection against such
unreliable introspective mechanisms at phylogenetic, cultural, or ontogenetic scales.
One of the earliest explanations of the unreliability of self-conceptions in west-
ern philosophy is due to Spinoza.4 He argues that our self-conceptualization as
capable of decision-making that is entirely free from causal constraints is simply a
reflection of our ignorance of the mechanisms of decision-making: “experience
itself, no less clearly than reason, teaches that men believe themselves free because
they are conscious of their own actions, and ignorant of the causes by which they are
determined” (Spinoza 1677/1955, IIIP2S).5 Although this is a step in the right direc-
tion, it fails to explain why our ignorance takes the form that it does. Granted that
we are ignorant of the causes of our actions, but why should we think of them as
free, rather than in some other way, or in no way at all? Nietzsche takes a further
step in the right direction. He offers a sociological explanation of the origins of
some false components of our self-conception, specifically, morally relevant ones,
like conscience and freedom of the will. According to Nietzsche, these prevalent
self-conceptions are, roughly, means by which religious representatives of “slave
morality” wield influence (Nietzsche 1887/1989, 140). I think the insight that false
self-conceptualizations are maintained for sociological reasons is fundamentally
correct, and elaborate upon it substantially below.
More recently, a number of philosophers have proposed alternatives to the spec-
tatorial model of self-interpretation that have the potential to explain the persistence
of false self-conceptualizations. Victoria McGeer (1996, 2007) has been perhaps
most explicit about this strategy:
…we must give up a fiction to which many philosophers and psychologists have clung …
that self-knowledge consists in coming to know (perhaps via theoretical mediation) a col-
lection of facts: facts about one’s own thoughts, feelings, intentions, and so on … the
reporter-predictor model of authoritative self-knowledge… In place of this, I propose we
move toward an agency model of authoritative self-knowledge. (McGeer 1996, 506)

On McGeer’s model, we do not “interact with one another as scientist to object, as


observer to observed, since the ‘objects’ themselves – viz. … [our]selves and other
agents  – are changing under pressure of the ‘explanatory-predictive’ attributions
that are made to them” (McGeer 2007, 146). The idea is that the “direction of fit”
(Searle 1969) between self-conceptualizations and the domain to which they apply

4
 Skepticism about self-knowledge has much older roots in Asian philosophy; it has always been a
central preoccupation of Buddhist philosophers (Garfield 2015).
5
 I thank Bryce Huebner for the reference.
Self-Interpretation as Software: Toward a New Understanding of Why False… 127

is the reverse of the direction of fit between our representations of the physical envi-
ronment and the domain to which they apply. Self-conceptualizations specify tar-
gets for our minds and behavior to approximate, rather than representing
independently constituted facts about our minds and behavior. As McGeer puts it,
… we are able to ensure a fit between the psychological profile we create of ourselves in
first-person utterances and the acts our self-attributed intentional states are meant to predict
and explain simply by adjusting our actions in appropriate ways. Thus, because we do not
just wait to see if our actions make sense in light of intentional self-attributions, but rather
make them make sense, the tale we tell of ourselves from the intentional stance is impor-
tantly unlike the tale we tell of other people (or even of other things). I cannot make it the
case that you behave in ways coherent with what I say you hope, desire, or fear any more
than I can make it the case that the world is a certain way by announcing how (I think) it is;
but I can and do govern my own actions in ways that fit with the claims I make about
myself. If so-called "knowledge” of our own minds thus consists largely of claims we have
both made and acted in light of, it is no surprise that such "knowledge” is peculiarly authori-
tative. (McGeer 1996, 507)

This is not yet an explanation of why particular false self-conceptualizations per-


sist. It provides no explanation of why the psychological profiles we attempt to
maintain have the character that they do; why we attempt to maintain the profile of
a self-transparent, unified, enduring, Cartesian ego, making free, rational decisions
based on conscious, linguistically expressible mental states, in accordance with
norms of rationality that govern public language use. But it opens the door to such
an account. For it removes the deepest problem with the spectatorial model of self-
knowledge: if self-interpretation aims not to correctly represent independently con-
stituted facts about the mind, but rather to specify constraints that the mind seeks to
approximate, then the falsity of many of our self-conceptions is not necessarily
relevant to their persistence. An analogy to computer software can make this clear.
The point of computer software is not to provide a true description of the machine
that runs it. Rather, its point is to specify constraints for the machine to approximate
through running the software. In other words, the direction of fit between software
and computer is the reverse of the direction of fit between our descriptions of the
physical environment and the physical environment. Software does not aim to
describe an independently constituted domain of facts pertaining to the machine
that runs it; it aims to create facts pertaining to the machine that runs it. For exam-
ple, word processing software does not describe the operation of a computer, until
it first transforms the operation of the computer such that it approximates a word
processor. For this reason, if we want to explain why certain kinds of software per-
sist, we do not refer to the accuracy with which they represent independently con-
stituted facts about computers that run them. Rather, we refer to the utility of the
operations that they get computers to perform. Analogously, if we pursue McGeer’s
reconceptualization of self-knowledge, the question of why components of our self-
conceptualizations persist should be addressed by attending to the utility of what
they get us to do, not to their accuracy as descriptions of facts about our minds.
Since any physical computer will only approximate the capacities specified by the
software it runs, the software will always be an inaccurate description of the
­computer’s operation. But it can nonetheless persist in virtue of working to make the
128 T. W. Zawidzki

computer operate in useful ways. Similarly, perhaps our self-conceptualizations


persist, despite their inaccuracies, because they work to make us operate in useful
ways.
But this raises the question of what these useful ways are. Suppose it is true that
conceiving of ourselves as self-transparent, unified, enduring, Cartesian egos, mak-
ing free, rational decisions, based on conscious, linguistically expressible mental
states, in accordance with norms of rationality that govern public language use,
helps turn us into agents that, in some sense, approximate these idealizations. Why
should this be useful enough for these particular self-conceptualizations to persist,
even though, as Sect. 2 reviewed, we often fall short of them? Recently, Daniel
Dennett and Keith Frankish have proposed similar answers to this question.
Daniel Dennett has explicitly invoked the software metaphor in numerous dis-
cussions of our self-conceptualizations. In general, Dennett thinks the secret to the
relative biological success of our species is our facility with what he calls “cognitive
tools” (Dennett 2014). We are able to invent and culturally transmit techniques of
problem solving that vastly expand our inborn cognitive capacities. For example,
Arabic numerals made possible vastly greater mathematical capacities than were
available prior to their invention and dissemination. Dennett views this as a perva-
sive phenomenon that explains the distinctively human cognitive profile. Everything
from natural languages, to symbol systems, to rituals, to mnemonic devices, to
physical technologies count as cognitive tools that simplify and otherwise make
tractable problem domains that would be intractable otherwise. Dennett explicitly
likens these cognitive tools to software. What makes the human brain special is that
it behaves like a general purpose computer, in the following sense: it has the capac-
ity to learn an indefinite variety of techniques of problem solving via socio-cultural
means.
One particularly important example of such naturally and culturally evolved
software for Dennett is natural language; for, according to Dennett, the human
brain’s installation of this bit of software makes consciousness possible. On his
view, consciousness is the controlled, linear, reactivation of cognitive states made
possible through habits of self-directed speech acquired via cultural learning
(Dennett 1991, 2005). The brain is a machine consisting of multiple streams of
processing taking place simultaneously, i.e., in parallel, and unconsciously.
Consciousness arises when it implements a serial virtual machine on this massively
parallel hardware via the linear stream of self-directed speech. Whether or not this
is an adequate model of consciousness, there is little doubt that something like this
happens in the brain. Furthermore, Dennett’s theory is an explicit answer to the
central puzzle addressed in this chapter: why a factually false self-conception might
persist. The idea that the brain is actually a serial processor, consisting of a con-
scious self that generates a linear stream of thoughts, though technically false, is, for
Dennett, an extremely useful piece of software that the brain runs. It enables an
unparalleled degree of cognitive control. Unlike other animals, human attention is
not at the mercy of environmental triggers that enhance now this processing stream,
now another. We have endogenous control of attention, and hence, of the sequence
in which various processing streams get enhanced, in virtue of talking to ourselves
Self-Interpretation as Software: Toward a New Understanding of Why False… 129

in a linear medium, i.e., public language. Running this software is thus very useful
for tasks like long-term planning, in which our species out-performs all others.
Despite its factual inaccuracy, the utility of our self-conceptualization as conscious
selves, generating linear streams of thoughts, explains its persistence.
Keith Frankish also appeals to the utility of cognitive self-control in order to
explain why the brain implements a language-based self-conceptualization, or, in
his words, why our “mind” implements a “supermind” (2004). On his view, the
natural mind traffics in partially active beliefs: propositions are endorsed with vary-
ing degrees of credence, which allows for simultaneously endorsing contradictory
propositions, as long as each is endorsed with a different degree of credence, all of
which add up to no more than 100%. The “supermind” is implemented by the natu-
ral mind via “premising policies” applied to declarative sentences of natural lan-
guage. Accepting a sentence as a premise precludes accepting any sentence that
contradicts it as a premise. Hence, the supermind traffics in the kinds of beliefs we
tend to attribute to persons in everyday life: beliefs constrained by inferential norms
governing sentence use, such as the Law of Non-Contradiction. What advantages
might such premising policies provide? Frankish (2009) speculates:
The short answer, I suggest, is that doing so affords us a measure of personal control over
our thought processes. A lot of our mental life goes on below the surface. Much of our
behaviour is the product of automatic, non-conscious mental processes of which we are
unaware and over which we have no direct control. It is these processes that succumb to
description in Bayesian terms (which is not to say that they involve actual Bayesian calcula-
tions, of course). Premising policies, on the other hand, are consciously formed and exe-
cuted and are subject to personal supervision and control. By adopting premises and goals,
and engaging in conscious reasoning, we can take manual control of our thought processes.
We can decide what to think about, and when, and can direct our minds to theoretical prob-
lems with no immediate behavioural relevance. We can evaluate propositions and goals in
the light of appropriate norms and choose which to endorse. And we can reflect on the
inferential procedures we use, assess them, and enhance them. In effect, by forming and
executing premising policies we create a highly flexible general-purpose reasoning system,
whose procedures are open to continual refinement and extension.

Thus, both Dennett and Frankish sketch a kind of “software” model of two of the
misleading components of our self-conception discussed above, in order to explain
their persistence. For Dennett, the enduring Cartesian, conscious self is imple-
mented on the brain via self-directed speech, and persists due to the higher-level
cognitive control it enables. For Frankish, premising policies involving categorical
acceptance and rejection of sentences of natural language likewise make possible
various feats of higher-level cognitive control. Although both of these accounts
illustrate the general strategy for explaining the persistence of misleading self-
conceptions that I explore in this chapter, they also suffer from a flaw: an excessive
focus on individualized functionality. According to both Dennett and Frankish, mis-
leading self-conceptions persist because of the advantages they provide to individ-
ual brains, i.e., they make possible better cognitive self-control. However this
overlooks the even more important social functions of these and other misleading
self-conceptions. I think these social functions are essential to understanding the
130 T. W. Zawidzki

persistence of the specific forms that our misleading self-conceptualizations take. I


turn now to a detailed explanation and exploration of this idea.

4  Self-Conceptions as Frequency-Dependent Phenomena

One problem with Dennett and Frankish’s exclusive focus on individualized func-
tionality is that both accounts presuppose the availability of a structurally complex,
linear, public language with which to implement self-conceptions that enable cogni-
tive self-control. But how did such a language evolve? Clearly, as Dennett recog-
nizes (Dennett 1991, 194–195) social functionality, i.e., communication, must have
played a role. In Dennett’s view, the brain begins to implement an illusory, Cartesian,
conscious self when dispositions to inform others using a linear, public language,
accidentally succeed in revealing to speakers information tucked away in inacces-
sible compartments of their own brains. But this suggests that there may have
already been social functions served by the fiction that human minds consist in
beliefs the contents of which can be specified using simple sentences of public lan-
guage, prior to individualized functions like cognitive self-control.
A second problem with excessive focus on individualized functionality concerns
the other misleading self-conceptions discussed above. Many of these appear to
have a primarily social functionality. Consider the forensic functions of the fiction
of a lifelong, enduring self, with unconstrained control over decisions. It is not clear
why such concepts are necessary for higher-level cognitive control; yet their social
roles are obvious. Persons can be held responsible for their behavior only if that
which is responsible for their decisions persists in time indefinitely, and is not fully
determined by factors that it cannot control. In his discussions of freedom of the
will, Dennett is explicit about this fundamentally social function: taking ourselves
to be responsible turns us into more responsible, and hence reliable, social partners
(Dennett 2003, 297–305).
In general, an exclusive focus on individualized functionality has a hard time
explaining the origins and essentially social nature of the misleading self-concep-
tions discussed above. Furthermore, it is not clear why solving the problem of indi-
vidual, higher-level, cognitive control requires our self-conceptualizations to take
the specific forms that they do. Why does this function require the brain to think of
itself as self-transparent, free from casual constraint, enduring (not just for the length
that it takes to implement a particular plan, but across multiple, interacting and
orthogonal plans), containing accurate records of past experiences, trafficking in
cognitive states the contents of which can be exhaustively specified using short sen-
tences of public language, and respecting public norms of rationality? Or, to address
Frankish’s proposal specifically, why must our premising policies take the particular
forms that they do, e.g., as implementing flat-out, categorical belief, in order to make
possible higher-order cognitive control? Why is partial belief, governed by Bayesian
principles, incompatible with such higher-order cognitive control? I think
that such features of prevalent self-conceptions can be explained only on the
Self-Interpretation as Software: Toward a New Understanding of Why False… 131

assumption that they serve social functions. In particular, they work to transform
individuals into reliable coordination partners on cooperative projects.
Individualistic accounts of the functions that sustain misleading self-concep-
tions, like Dennett’s and Frankish’s, miss an important feature of the distinctively
human cognitive niche. The tasks at which we most dramatically outperform other
species, and which are most central to our overall biological advantages over other
species, are overwhelmingly social. What set us apart and explain our competitive
advantages are: (1) an unparalleled ability to coordinate on large-scale cooperative
projects with enormous numbers of conspecifics with whom we are personally
unacquainted, and (2) an unparalleled ability to learn from each other, quickly, effi-
ciently, and accurately, complex techniques of social interaction and resource
extraction, upon which we can then innovate, yielding cumulative cultural evolution
(Sterelny 2012). These two unique, pervasive features of the human cognitive niche
interact in rich ways: we could not learn from each other so efficiently if we weren’t
cooperative, and we could not cooperate so effectively if we did not culturally
evolve communicative and normative frameworks governing our interactions.
Furthermore, these two, towering socio-cognitive accomplishments rely on a diverse
palette of socio-cognitive talents at which members of our species excel, including
acquiring systems of norms and communication, and tracking and anticipating the
behavior, goals, and knowledge of our fellows.
Underestimating this social dimension of the human socio-cognitive niche, as
individualistic theories like Dennett’s and Frankish’s do, risks overlooking a key
feature of social cognition: in this domain, cognizer and cognized are one and the
same. This has important implications for the view that our self-conceptualizations
function as cognitive tools, or software. On an individualistic view, the point of our
cognitive tools is to transform us into better cognitive agents. For example, facility
with Arabic numerals makes us better at computing certain mathematical functions.
Facility with self-directed speech makes us better at formulating long-term plans.
Facility with conscious premising policies makes us better at higher-level cognitive
control. But the objects of these forms of cognition are treated as static problem
domains: these cognitive tools do not alter the nature of mathematical functions, or
long-term plans, or higher-level cognitive control. However, if the goal is more
effective social cognition, besides improving our capacities as cognitive agents,
cognitive tools can also work to transform us into better cognitive objects. That is,
our socio-cognitive needs are served not just by cognitive tools that make us better
at solving socio-cognitive problems, but also by cognitive tools that make the
domain of social cognition, i.e., our behavior, easier to compute by others. This lat-
ter function is particularly important given the specific kinds of socio-cognitive
tasks at which humans excel: cooperation and learning from each other. If our socio-
cognitive tools can make us more cooperative, and more reliable sources of infor-
mation, i.e., if they can transform us qua socio-cognitive objects, then the tasks
facing us qua socio-cognitive agents can be simplified.
Although Dennett shows some appreciation of this special feature of some socio-
cognitive tools in his discussions of freedom of the will (Dennett 2003, 297–305), it
has much broader application. In fact, all six misleading components of our
132 T. W. Zawidzki

s­ elf-conception discussed above arguably function to make us better socio-cogni-


tive objects, i.e., more regular, reliable, and hence more easily computable. As
Nietzsche puts it: “the task of breeding an animal that is permitted to promise
includes, as condition and preparation, the more specific task of first making man to
a certain degree necessary, uniform, like among like, regular, and accordingly pre-
dictable” (Nietzsche 1887/1989, 58).6
Consider the myth of self-transparency. If, as McGeer suggests, this is grounded
in our unique capacity to make ourselves match our publicly declared psychological
profiles, then the myth of self-transparency can function to make us more easily
computable by our fellows. For example, declarations of belief or desire lead to
expectations of behavior compatible with such beliefs and desires, expectations that
we feel pressure not to disappoint, based on a concern with reputation for reliability.
Such social pressure helps simplify the computational tasks facing our interpreters:
they can rely on straightforward expectations derived from our explicit avowals to
predict our future behavior.
Similar considerations apply to the myth of the Cartesian ego. Conceiving of
oneself and others as enduring, unified, Cartesian egos, coming to conscious, multi-
modally constrained decisions directly responsible for behavior, might exert pres-
sure to take great care in trying to live up to this profile, however imperfectly.
Dennett is probably right that this makes possible long-term planning and other
kinds of cognitive self-control. But it also makes one a much more tractable object
of social cognition, especially among fellows who share this (self-) conceptualiza-
tion. Interpreters can count on the fact that the behavior of their targets will derive
from decisions that take into account information from multiple modalities, gath-
ered and integrated over indefinite periods of time, over which they will take some
form of ownership.
The myth of freedom of the will likewise has clear social functions, as has long
been recognized. As Dennett argues, whether or not we are truly causally responsi-
ble for our behavior, taking responsibility for it yields multiple social and individual
benefits (Dennett 2003). For one thing, one is able to learn from one’s mistakes in
virtue of “owning” them. But more importantly, from a social perspective, one
becomes more reliable if one views unreliability as one’s responsibility, opening
oneself up to sanction.
Likewise, the myth of memory as accurate record can serve important social
functions, particularly in the domain of social learning. If Sterelny (2012) is right,
our species’ distinctive niche consists in reliable cooperation and reliable informa-
tion sharing. But information sharing can be reliable only if individuals work to turn
themselves into reliable reporters of facts they have witnessed. If individuals expect
themselves and others to have memories that consist in accurate records of past
events, and if conforming to such expectations is incentivized through both endog-
enous and social means, e.g., shame and reputation, respectively, then the myth of
memory as accurate record can work to make humans more reliable information
sharers.

 I thank Bryce Huebner for the reference.


6
Self-Interpretation as Software: Toward a New Understanding of Why False… 133

Social functions make better sense of the myth of language-as-window-on-the-


mind than individual functions. It is unclear why an individual function, like higher-­
level cognitive control, requires us to conceive of our mental states as perfectly
expressible in simple, public language sentences, or as governed by the same infer-
ence rules as such sentences. But this self-conception makes perfect sense on the
hypothesis that it serves important social functions. Consider how much easier it is
to track a person’s behavior if she expresses her beliefs, desires, and intentions in
simple sentences, and then works to behave in ways which conform to simple infer-
ence rules that govern such sentences, the limited number of attitudes one can bear
to them, and their connections to behavior. Think how much more difficult it is to
determine the levels of credence that some person’s brain distributes over a set of
potentially contradictory propositions, and how these feed into her decision-­making.
If we simply assume that when a person says that she believes P, she will act only in
ways publicly deemed compatible with P, and never in ways publicly deemed com-
patible with not-P, and the person is heavily incentivized by endogenous (e.g., guilt
or shame), and social (e.g., reputation or sanction) means to conform to this profile,
a computationally very difficult socio-cognitive task is considerably simplified.
Similar considerations can explain the persistence of the myth of rationality: if per-
sons assume that human behavior naturally respects public norms of rationality,
they set up expectations that all are endogenously and socially incentivized to meet,
thereby making behavioral anticipation more tractable.
It is relatively straightforward, therefore, to identify social functions that can
explain the persistence of misleading components of prevalent self-­
conceptualizations. It is hard to dislodge myths about the self, despite growing evi-
dence, because these myths function to make us more reliable at cooperation and
information sharing, the two main sources of human biological success. It is impor-
tant to note, however, that there is nothing inevitable about these myths. The reason
is that their social functions depend on their frequency. It is only in a population in
which the assumption prevails that persons are self-transparent, conscious, unified,
enduring, Cartesian egos, that come to free decisions respecting public norms of
rationality, based on mental states, including accurate memories, fully expressible
in simple sentences of public language, that this self-conceptualization can support
social functions, like reliability in coordination on cooperative projects, and infor-
mation sharing. In this regard, mythical self-conceptions are rather like languages.
We do not think that English persists because it constitutes a more accurate descrip-
tion of independently constituted facts than other possible languages; rather, it per-
sists because, once its use reached a certain prevalence, especially among politically
and economically influential players, its further dissemination and persistence were
incentivized. Similarly, according to the view advanced here, for mythical self-­
conceptions: they persist not because they constitute accurate descriptions of inde-
pendently constituted facts about human beings; rather, they persist because, once
they reach a certain prevalence in specific populations, their further dissemination
and persistence are incentivized, in virtue of the coordinative and communicative
134 T. W. Zawidzki

feats they facilitate as socio-cognitive tools, or software, that transform individuals


into more tractable socio-cognitive objects.7

5  Questions, Implications, Speculations

In conclusion, let me address a number of questions, implications, and speculations


that arise from this account of why false self-conceptions persist. I focus on 7: (1)
The importance of distinguishing between questions of how self-conceptions per-
sist, how brains implement them, and how they ought to be evaluated; (2) how facts
about the brain might still constrain the nature of our self-conceptions; (3) how
brains might implement self-conceptions; (4) the differences and relations between
self-interpretation and the interpretation of others; (5) the relations between false
self-conceptions and other useful fictions; (6) how the software model of self-­
interpretation might address conflicting self-conceptions; (7) how the social distri-
bution of self-conceptions might change historically, and the role of the arts and
humanities in this. My goal here is not anything approaching an exhaustive explora-
tion of these issues; I intend only to briefly characterize some interesting issues and
perspectives that arise for the software model of self-interpretation.

5.1  Three Distinct Questions

There are at least three distinct questions that one can raise about our self-­
conceptions: (1) why they persist, even if false; (2) how the brain implements them;
(3) how they ought to be evaluated. In this chapter, I have addressed only the first. It
is important not to conflate these questions. If the view I defend is correct, the rea-
sons why many false self-conceptions persist are social. This suggests that the per-
sistence of self-conceptions has little to do with the operation of the brain. This may
be true. On the other hand, there may be strong constraints that neural architecture

7
 This frequency-dependence of self-conceptualizations is crucial to forestalling the following
objection, suggested by an anonymous referee. Suppose these self-conceptions are culturally rela-
tive, e.g., prevalent only among populations derived from European cultures. If they persist because
they make social coordination feasible, would we not expect much poorer social coordination in
populations lacking such self-conceptions, e.g., those derived from East Asian cultures? Not nec-
essarily. Their utility to social coordination depends on their prevalence: conceiving of myself as a
Cartesian Ego helps because I interact with others animated by similar or complementary self-
conceptions. In populations where different self-conceptions are prevalent, coordination relies on
those different self-conceptions. This is what I mean by the analogy to language: English helps
coordination in populations of English speakers, but very different languages can have similar
effects in different populations. Thus, if, for example, it is true that Asian cultures are marked by
more “collectivist” self-conceptions, this does not imply that they coordinate less smoothly.
Relative to those cultures, such self-conceptions are better coordination tools, since the efficacy of
such tools depends, to a large extent, on their prevalence.
Self-Interpretation as Software: Toward a New Understanding of Why False… 135

places on the nature of our self-conceptions, even if these self-conceptions are false
descriptions of neural architecture. I discuss this possibility below. Furthermore,
even if it is true that many self-conceptions persist for social reasons, it does not
follow that these reasons legitimize these self-conceptions. Perhaps self-­
interpretation should aim at truth, as many philosophers have maintained, and false
self-conceptualizations should be replaced by true ones, no matter how socially
useful the former are. Since the question of how to evaluate our self-conceptions is
entirely independent of the question of what makes them persist, the software meta-
phor explored here need not have implications for the former question.

5.2  Neural Constraints on Viable Self-Conceptions

On the spectatorial model of self-interpretation, there is a straightforward answer to


what are the neural constraints on viable self-conceptions: since self-conceptions
aim to correctly describe the brain, they should persist only to the extent that they
correctly describe facts about the brain. The software model of self-interpretation
suggests a more complex understanding of the neural constraints on viable self-­
conceptions. It is inspired by how computer hardware constrains software. Reflection
on this metaphor suggests that neural constraints on the viability of self-conceptions
admit of a spectrum of cases. That is, just as it makes sense to say that some specific
computer hardware is more or less effective at implementing a particular bit of soft-
ware, it makes sense to say of a particular brain that it is more or less effective at
implementing a particular self-conception. Note the striking difference from the
spectatorial model, on which a self-conception is either accurate or not. On the
software model, such a binary understanding of the relation between brains and
their self-conceptions is completely inadequate. A particular self-conception can be
more or less implementable on a particular brain, and whether that brain imple-
ments it will depend on cost-benefit calculations, just as with computer software
and hardware.
In particular, if the view explored above is on the right track, particular brains
will have to balance the processing costs of thinking of themselves in certain ways,
against the social benefits of being the kinds of persons into which such self-­
conceptions transform their owners. Thus, relative to some social environments, it
might be worth implementing self-conceptions that are very difficult for human
brains to implement. For example, relative to communities where some kinds of
extreme fundamentalism prevail, the social benefits of conceiving of oneself as mor-
ally responsible for all of one’s thoughts, emotions, and behaviors might outweigh
the extreme costs of such a self-conception, both in terms of processing by brains
like ours, and long-term psychological harm. This would be roughly analogous to
running very complex, advanced software on out of date hardware, because one has
no other option. On the other hand, some self-conceptions might track neural pro-
cessing so closely as to be virtually cost-free in implementation. This would be
roughly analogous to running software for evaluating truth functions on a s­ tandard
136 T. W. Zawidzki

Von Neumann computer architecture, which includes hardware dedicated to calcu-


lating truth functions.
The latter two examples suggest a nice feature of the software model of self-­
interpretation. Unlike the spectatorial model, it can explain the viability of both
accurate and inaccurate self-conceptions using the same conceptual resources.
Accurate self-conceptions are just at one extreme of the spectrum of implementabil-
ity: those that are effortlessly implementable, because they correctly represent rel-
evant facts about the brain, as truth function evaluating software correctly represents
relevant facts about computer hardware. So, the software model does not preclude
the possibility that some components of human self-conceptions are accurate: these
are the ones that it takes no effort to implement. However, at the same time, and
unlike the spectatorial model, the software model accommodates persisting yet false
self-conceptualizations, using the same conceptual framework: these are the ones
that it takes a lot of effort for brains like ours to implement, yet which are worth
implementing nonetheless due to their social utility.

5.3  How Brains Implement Mythical Self-Conceptions

Although I have explored some reasons why human brains implement false self-­
conceptions, I have said little about how human brains do this. This is largely an
empirical question, ripe for exploration by cognitive neuroscience. One possibility
alluded to above is that brains employ a kind of internal surrogate of social sanc-
tions to motivate conformity to public constraints governing publicly declared
self-conceptions.
Research on the evolution and neural realization of moral emotions, like guilt
and shame, might be relevant here (Frank 1988). Other relevant work includes evi-
dence that human neural activation in response to social rejection overlaps signifi-
cantly with neural activation correlated with physical pain (Kross et al. 2011), and
evidence that failure to conform to majority opinion in one’s judgments triggers an
error signal in regions also active during all forms of conditional learning, in
response to mistaken behavior (Klucharev et  al. 2009). Such mechanisms would
allow for easy internalization of potential social sanctions, like ostracism and other
forms of punishment. All of this is compatible with Carruthers’s (2011) suggestion
about the implementation of counter-attitudinal advocacy; the tendency, discussed
above, to endorse views one has previously disavowed in order to dampen cognitive
dissonance. Carruthers appeals to Damasio’s “somatic marker system” (Damasio
1994) to hypothesize that simulation of the affective consequences of different self-­
interpretations selects for those that minimize negative feelings associated with cog-
nitive dissonance.
There are a number of other possibilities as well. For example, self-conceptions
might be encoded in grammatically obligatory structures of public languages, and
implemented via neural constraints on linguistic processing. Dan Slobin has pro-
posed that a mild variety of Whorfian linguistic relativity might consist in “thinking
Self-Interpretation as Software: Toward a New Understanding of Why False… 137

for speaking” (Slobin 1996). Much of our thought involves preparing to make reports
and other linguistic acts in our primary public language, to interlocutors. This sug-
gests that we code experience in ways that respect the grammars of our primary
means of communication, public languages. This might make possible the neural
implementation of some components of our self-conceptions. For example, accord-
ing to Slobin, “…in Turkish you are obliged to choose between two past-­tense
inflections, one for witnessed and one for non-witnessed events” (Slobin 1996, 74).
This grammatical constraint on thinking for speaking may contribute to a Turkish’
speaker’s self-conceptualization as an accurate tracker of epistemic authority.
Frankish also proposes a model of how the mind, trafficking in partial, Bayesian
belief, might implement the supermind’s “flat-out” belief: “…when we perform an
action in response to our flat-out beliefs and desires we do so because we are highly
confident that we have premising policies that dictate the action and attach a high
desirability to sticking to them” (Frankish 2009). The basic idea is that the Bayesian
mind includes unconscious, partial, metacognitive beliefs and desires, which, if they
reach a high enough degree of confidence, implement the supermind’s conscious,
flat-out beliefs, in virtue of motivating behavior consistent with premising policies.

5.4  Interpreting Oneself Versus Interpreting Others

In this chapter, I have focused on self-interpretation. But, of course, we use the same
concepts in self-interpretation as in the interpretation of others. This raises ques-
tions about how these two interpretive tasks are related. Of course, these are among
the oldest questions in philosophy, pertaining specifically to the problem of other
minds. How is it that our access to our own minds in self-interpretation seems so
different from our access to the minds of others when we interpret them? The soft-
ware model of self-interpretation has the potential to address this question in a
highly effective way: because self-interpretations can make their targets, i.e., inter-
preters, satisfy the constraints they specify, while interpretations of others cannot,
the asymmetry between self-interpretation and the interpretation of others is not
surprising. Indeed, explaining this asymmetry is precisely what motivates McGeer’s
(1996) pioneering discussion of this model of self-interpretation.
Still, it remains puzzling why we use the same concepts to interpret ourselves as
we use to interpret others, given the very different functions of these different modes
of interpretation. At least, this is puzzling if we focus on individualistic accounts of
the functions of self-interpretation, as Dennett and Frankish do. Why would we
expect concepts useful to transforming oneself into a better long-term planner, or
cognitive self-controller, also to work in the interpretation of others? However, one
of the advantages of the socio-functionalist account explored above is that this mys-
tery disappears. On this view, the point of self-interpretation is to transform oneself
into a good coordinator on cooperative projects, including social learning. Self-­
conceptions are frequency-dependent phenomena: they work only if they are
­widespread. It is only then that partners in coordination can have appropriate, mutual
138 T. W. Zawidzki

expectations, based on interpretations of self and other. Hence, concepts we use to


transform ourselves into good coordinators can, in virtue of the fact that our poten-
tial partners must be doing the same for these concepts to have these positive effects
on coordination, also be used to describe and predict the behavior of others. It is only
if we appreciate the role of potentially false self-conceptions in making us good
socio-cognitive objects, for the expectations of our likely coordination partners, that
we can explain why the same concepts are useful both in regulative self-­interpretation,
and descriptive/predictive interpretation of others.
Most regulative uses of interpretation are self-directed, as McGeer (1996) notes.
However, this does not preclude some regulative uses of interpretation of others. For
example, it is possible that interpreting infants as believing, desiring, and intending
things, including intending communicative acts, helps bootstrap these forms of cog-
nition in human ontogeny (Bruner 1983; McGeer 2001; Mameli 2001). More wor-
ryingly, the interpretation of infant behavior in terms of implicit social categories
might trigger conformity to, often unjust, social expectations. For example, there is
evidence that adults interpret the same behavior (crying) by the same infant in very
different ways, depending on the gendered clothing the infant wears: when the
infant is dressed in stereotypically female colors, crying is interpreted as being
upset, and requiring comfort; when the infant is dressed in stereotypically male
colors, crying is interpreted as being angry, and requiring a more aggressive response
(Mameli 2001). If such early experiences are taken as models for appropriate behav-
ior, i.e., as regulative, by infants, then this is a mechanism whereby interpretation of
others can come to play a regulative, software-like role. Perhaps such interpreta-
tions are analogous to the installation of (cultural) software on a new “machine”.
This has potentially deep political implications, depending on the prevalence of
such regulative uses of other-interpretations. On some views (e.g., Haslanger 2012),
deeply immoral and unjust conceptions of human beings persist in virtue of such
deep-seated, tacit, interpretive/regulative practices.

5.5  Mythical Self-Conceptions and Other Myths

I have focused on components of our self-conception that we have good reason to


believe are false and misleading. However, it is not only such myths that seem
extraordinarily hard to dislodge. Myths about the external, non-human world also
show remarkable stability. Above, I suggested that some of this might be a function
of spatiotemporal scale: certain myths about space, time, mechanics, and dynamics
approximate truth at scales for which our visual system is adapted. But not all myths
about the nonhuman world can be explained in this way. Myths about biology, such
as the eternalism and essentialism that characterize commonsense concepts of spe-
cies, and about cosmology, such as the origin and destiny of the universe consisting
in the plan of an intelligent designer, appeal to highly abstract concepts that have
little connection to the spatiotemporal scales for which our sensory systems are
adapted.
Self-Interpretation as Software: Toward a New Understanding of Why False… 139

It is possible that the software account of mythical self-conceptions can be


extended to account for the persistence of such abstract myths about the non-human
world. The reason is that our self-conceptions may derive a lot of their authority
from how they fit into a broader, over-arching, mythical framework that includes
mistaken ideas about the non-human world. For example, conceiving of oneself as
embarked on a mission derived from the plans of the creator of the universe can
plausibly help motivate behavior consistent with this self-conceptualization, help in
its widespread dissemination and long-term stability, and hence help sustain its role
in coordination.

5.6  Self-Interpretive Conflict and Mental Health

The account explored above also suggests an interesting perspective on why self-­
conceptions often conflict. If our brains implement our self-conceptions primarily
to serve certain social functions, like smoother coordination on cooperative proj-
ects, then they will be as variable as the social contexts in which we require such
functionality. Individuals who must routinely coordinate on cooperative projects in
diverse social contexts, defined by very different “ecologies” of self-interpretation,
will have to implement very different, and sometimes conflicting self-­interpretations.
Since Durkheim first proposed it (Durkheim 1897/2006), this has been a very influ-
ential characterization of modernity. It is arguable that, unlike in traditional societ-
ies, modern humans are confronted with a complex array of coordinative contexts,
each calling for very different self-conceptualizations. As Erving Goffman argues in
his classic, The Presentation of Self in Everyday Life, there seems to be little more
to the self than the capacity to inhabit different social roles relative to different pub-
lic contexts (Goffman 1959), and, in the modern world, the diversity and mutual
incompatibility of such contexts seems to be extreme.
For this reason, the software metaphor of self-interpretation has the potential to
shed some light on certain psychiatric disorders. To the extent that many such dis-
orders can be characterized in terms of the disintegration of the self (Tekin 2011), it
is possible that some are triggered by the excessive demands on self-­conceptualization
that characterize modernity. In particular, on the software metaphor, different self-­
concepts are just like different software running on the same hardware. Just as the
demands that different kinds of software make can be mutually incompatible, in the
sense that the hardware can’t run them simultaneously, the demands of different
self-conceptualizations can likewise be incompatible. For example, it might be very
difficult to play the role of upstanding parent or community member, and effective
corporate executive at the same time. Such conflicts can lead to cognitive disso-
nance, and, perhaps, in extreme cases, trigger psychiatric conditions.
140 T. W. Zawidzki

5.7  D
 isrupting Interpretative Socio-Ecologies and the Role
of the Arts and Humanities

As suggested above, one way to conceive of the social roles of self-­conceptualizations


explored here is in terms of “ecologies” of self-interpretation. The idea is that dif-
ferent self-interpretations can be adapted to each other, and come to constitute a
tight-knit, self-reinforcing whole, much as real ecologies of biological species do.
Gendered self-conceptualizations are obvious examples. Consider gender stereo-
types taken by many to have characterized 1950s United States. Female self-­
conceptualizations, as homemakers and primary childcare providers, complemented
male self-conceptualizations, as primary breadwinners and sociopolitical/legal rep-
resentatives of the family. If the primary function of self-conceptualization is mak-
ing possible coordination on cooperative projects with similarly socialized others,
then we should expect the emergence of such socio-ecologies of complementary
self-conceptualizations.
But this raises a problem. How can such socio-ecologies be disrupted, as they
have been, historically? Recall that on the view defended here, self-conceptions are
frequency-dependent phenomena: their stability depends on their frequency, just as
with languages. But then, given an extremely stable socio-ecology of self-­
interpretation, how can a new one emerge? Any divergent self-conceptualization
will, at first, pay no dividends, since its functionality depends on its prevalence, i.e.,
the availability of other similar, or complementary self-conceptualizations. Any
individual trying to coordinate based on a deviant self-conceptualization will find
no suitable partners. For example, presumably, in 1950s United States, females who
did not conceive of themselves as primary homemakers and childcare providers had
a hard time securing coordinative partners for the project of cooperative breeding.
There just weren’t enough males with complementary self-conceptions available
with whom to partner. Due to the seemingly inevitable lack of success of atypical
self-conceptualizations, they appear extremely unlikely to spread, and the typical
ones ought to remain indefinitely stable. Yet, as history attests, this has not been the
case, especially in the modern era.
One response to this problem is to note that socio-ecologies of self-interpretation
are not typically, in the modern era, isolated enough from exogenous factors to
remain stable. For example, it is arguable that the socio-ecology of 1950s United
States was inherently unstable due to the experience of women in World War Two.
During this period, due to the new conditions on military success, i.e., mass-­
produced, industrial-scale weaponry, women had to take on many traditional male
jobs in factories, while men fought in the armed forces. This collective experience
of economic self-reliance and efficacy possibly set the stage for the extreme instabil-
ity of the attempt to return to pre-Twentieth Century socio-ecologies of gender iden-
tity in the 1950s. Women weren’t buying the restriction on their roles to homemaking
and childcare, because of the collective experience of alternative roles in World War
Two. All that was needed were some exogenous factors to actualize a potential self-
Self-Interpretation as Software: Toward a New Understanding of Why False… 141

reconceptualization on a mass-scale: perhaps the availability of cheap and safe birth


control, and the pop-cultural revolution of the 1960s.
Thus, all sorts of contingent, exogenous factors, having to do with independent
economic or political or cultural developments might trigger disruptions in socio-­
ecologies of self-interpretation. It is plausible that communication technologies that
allow for widespread dissemination of new self-conceptualizations can be particu-
larly disruptive of entrenched socio-ecologies. Consider for example the disruptions
to socio-ecologies based on certain male self-conceptualizations triggered by the
mass marketing of popular music acts, like the Beatles, in the 1960s. Suddenly men
could wear their hair long, pay special attention to pattern and adornment in clothing,
affect a more “sensitive” demeanor, etc. This suggests an important reconceptualiza-
tion of public art. If the software model of self-interpretation explored here is correct,
and, in particular, if it is on the right track regarding the role of self-­interpretation in
solving coordination problems, and how this promotes socio-­ecologies of self-inter-
pretation, then public art can be seen as a very powerful tool for disrupting otherwise
stable, mythical self-conceptions. The reason is that public art, broadcast through
modern media, enables the widespread dissemination of alternative self-conceptual-
izations in very short periods of time. This allows for new, complementary self-
conceptualizations to be implemented on lots of brains at roughly the same time,
making possible their efficacy as coordination tools. As a consequence, traditional
self-conceptions no longer have the advantage of prevalence over divergent ones;
due to mass, virtually instantaneous, public dissemination, brains implementing
divergent self-conceptualizations will almost always find partners.
This reconceptualization of the role of public art has important implications for
what biologist E.  O Wilson (1998) calls the problem of “consilience”: how the
world as portrayed by science can be reconciled with the world as portrayed by the
arts and humanities. On the software model of self-interpretation, our interpretive
frameworks aim not only to describe our selves and our role in the world, but also
to regulate and transform our selves and our role in the world. If the role of science
is exclusively descriptive, the role of art is arguably regulative and transformative:
it produces and disseminates new self-conceptions that disrupt the socio-ecologies
based on established self-conceptions. Humanities, like literature, philosophy, and
history, can be conceived of as providing systematic, critical scrutiny of the vari-
ous regulative and transformative socio-ecologies of self-interpretation that pre-
vail today, or have prevailed historically. Such scholarly endeavors are not in
conflict with science, because the goals of the self-conceptions they study are not
descriptive: they aim to transform human beings into better socio-cognitive
objects, relative to their ambient social environments, rather than to describe their
natural constitutions. It is for this reason that there is no conflict, for example,
between science’s description of human beings as biological products of natural
selection, controlled by nervous tissue constrained by the laws of physics and
chemistry, and our self-conceptualization as self-transparent, unified, persisting,
Cartesian egos, making rational and free decisions based on conscious contempla-
tion of accurate memory records, and other mental states expressible in simple
sentences of public language. Where the former, scientific conceptualization aims
142 T. W. Zawidzki

to describe, the latter, humanistic conceptualization, like software, aims to regu-


late and transform human nature.

6  Conclusion

There are good empirical and conceptual reasons to think that a number of beliefs
about the human mind that appear central to prevalent self-conceptions are false and
deeply misleading. There are good reasons to think that human minds are not, con-
trary to commonsense self-conceptions, self-transparent, unified, persisting,
Cartesian egos, coming to free, rational decisions, based on conscious contempla-
tion of mental states, including accurate memory records, and propositional atti-
tudes accurately expressed by simple public language sentences. The persistence of
such self-conceptions is puzzling because it is unclear how such false and mislead-
ing, action-relevant beliefs could support successful actions, or how the mecha-
nisms generating them could persist otherwise. In the foregoing, I have proposed an
explanation of the persistence of false and misleading-self conceptions: they func-
tion not to describe the human mind, but to transform human beings into better
socio-cognitive objects, relative to the cooperative and communicative tasks at
which our species excels. In other words, self-interpretations are akin to computer
software the human brain runs in order to turn us into better coordinative and com-
municative partners, relative to our ambient social environments. They persist as
long as they are effective in this task, and this is largely a function of how prevalent
they are: coordination succeeds only if most potential coordinators are animated by
the same self-conceptions.
This is but one of a number of possible responses to the stubborn persistence of
false and misleading self-conceptions. It is also possible to: (1) question the reasons
for thinking them false; (2) question their centrality to common sense, and (3) argue
that relative to ordinary spatiotemporal scales and contexts they approximate truth.
Due to limitations of space, I have not adequately addressed these alternative
responses. Hence, it is best to think of the view defended here as hypothetical: if it
turns out that many assumptions central to our self-conception are false and mis-
leading, even relative to ordinary spatiotemporal scales and contexts, one can still
explain their persistence in terms of the socially-functional software metaphor. This
metaphor also raises a rich set of questions and ideas for studying human self-­
conceptions, including questions about how facts about the brain constrain the via-
bility of its self-conceptions, the notion that self-conceptions are frequency-dependent
phenomena, like languages, and the different roles that artistic/humanistic and sci-
entific depictions of human beings might play.
Self-Interpretation as Software: Toward a New Understanding of Why False… 143

References

Akins, K. 1996. Of sensory systems and the “aboutness” of mental states. Journal of Philosophy
93 (7): 337–372.
Armstrong, D. 1980. The nature of mind and other essays. Queensland: University of Queensland
Press.
Boyer, P. 2009. What are memories for? Functions of recall in cognition and culture. In Memory in
mind and culture, ed. J. Wertsch and P. Boyer. New York: Cambridge University Press.
Brehm, J.W. 1956. Postdecision changes in the desirability of alternatives. Journal of Abnormal
and Social Psychology 52: 384–389.
Bruner, J. 1983. Child’s talk: Learning to use language. New York: W. W. Norton.
Carruthers, P. 2006. The architecture of the mind. New York: Oxford University Press.
———. 2011. The opacity of mind: An integrative theory of self-knowledge. Oxford: Oxford
University Press.
Churchland, P.M. 1981. Eliminative materialism and the propositional attitudes. Journal of
Philosophy 78: 67–90.
———. 1990. A neurocomputational perspective. London: MIT Press.
Clark, A. 1997. Being there. London: MIT Press.
Damasio, A.R. 1994. Descartes’ error: Emotion, reason, and the human brain. New.York: Grosset
and Putnam.
Davidson, D. 1984. Belief and the basis of meaning. In Inquiries into truth and interpretation,
141–154. Oxford: Oxford University Press.
Dennett, D. 1978. Brainstorms. London: MIT Press.
———. 1984. Elbow room. London: MIT Press.
———. 1987. The intentional stance. London: MIT Press.
———. 1991. Consciousness explained. New York: Little, Brown, and Company.
———. 2003. Freedom evolves. New York: Viking.
———. 2005. Sweet dreams. London: MIT.
———. 2014. Intuition pumps and other tools for thinking. New York: Penguin.
Descartes, R. 1641/1979. Meditations on first philosophy. Indiana: Hackett.
Dick, P. 1996. Do androids dream of electric sheep? New York: Del Rey/Ballantine Books.
———. 1997. We can remember it for you wholesale. In The Philip K. Dick Reader. New York:
Carol.
Durkheim, E. 1897/2006. On suicide. New York: Penguin.
Flanagan, O. 1991. Varieties of moral personality. Cambridge: Harvard University Press.
Fodor, J.A. 1975. The language of thought. Cambridge: Harvard University Press.
———. 1978. Propositional attitudes. The Monist 61: 501–523.
Forstmann, M., and P.  Burgmer. 2015. Adults are intuitive mind-body dualists. Journal of
Experimental Psychology – General 144 (1): 222–235.
Frank, R.H. 1988. Passions within reason. New York: W. W. Norton.
Frankish, K. 2004. Mind and supermind. Cambridge: Cambridge University Press.
———. 2009. Partial belief and flat-out belief. In Degrees of belief, ed. F. Huber and C. Schmidt-­
Petri. Dordrecht: Springer.
Garfield, J. 2015. Engaging Buddhism. Oxford: Oxford University Press.
Goffman, E. 1959. The presentation of self in everyday life. New York: Anchor Books.
Harris, S. 2012. Free will. New York: Free Press.
Haslanger, S. 2012. Resisting reality: Social construction and social critique. Oxford: Oxford
University Press.
Hume, D. 1738/2000. A treatise of human nature. Oxford: Oxford University Press.
Hutto, D.D. 2008. Folk psychological narratives: The sociocultural basis of understanding rea-
sons. Cambridge: MIT Press.
Kahneman, D. 2011. Thinking fast and slow. New York: Farrar, Straus and Giroux.
Kane, R. 1996. The significance of free will. Oxford: Oxford University Press.
144 T. W. Zawidzki

Klucharev, V., K.  Hytönen, M.  Rijpkema, A.  Smidts, and G.  Fernández. 2009. Reinforcement
learning signal predicts social conformity. Neuron 61: 140–151.
Kross, E., M.G. Berman, W. Mischel, E.E. Smith, and T.D. Wager. 2011. Social rejection shares
somatosensory representations with physical pain. Proceedings of the National Academy of
Sciences of the United States of America 108 (15): 6270–6275.
Libet, B. 1985. Unconscious cerebral initiative and the role of conscious will in voluntary action.
Behavioral and Brain Sciences 8: 529–566.
Locke, J. 1689/1997. An essay concerning human understanding. New York: Penguin.
Loftus, E.F. 1996. Eyewitness testimony. Cambridge: Harvard University Press.
Mameli, M. 2001. Mindreading, mindshaping, and evolution. Biology and Philosophy 16: 597–628.
McGeer, V. 1996. Is “self-knowledge” an empirical problem? Renegotiating the space of philo-
sophical explanation. Journal of Philosophy 93 (10): 483–515.
———. 2001. Psycho-practice, psycho-theory and the contrastive case of autism. Journal of
Consciousness Studies 8 (5–7): 109–132.
———. 2007. The regulative dimension of folk psychology. In Folk psychology re-assessed, ed.
D.D. Hutto and M. Ratcliffe, 137–156. Dordrecht: Springer.
Nahmias, E. 2014. Is free will an illusion? Confronting challenges from the modern mind sci-
ences. In Moral Psychology, Volume 4: Free will and moral responsibility, ed. Walter Sinnott-­
Armstrong. London: MIT Press.
Neisser, U., and R. Fivush, eds. 1994. The remembering self: Construction and accuracy in self-­
narrative. New York: Cambridge University Press.
Neisser, U., and N.  Harsh. 2000. Phantom flashbulbs: False recollections of hearing the news
about challenger. In Memory observed, ed. U. Neisser and I.E. Hyman Jr. New York: Worth
Publishers.
Nichols, S., and J. Knobe. 2007. Moral responsibility and determinism: The cognitive science of
folk intuitions. Nous 41 (4): 663–685.
Nietzsche, F. 1886/1989. Beyond good and evil. New York: Vintage.
———. 1887/1989. On the genealogy of morals. New York: Vintage.
Nisbett, and Wilson. 1977. Telling more than we can know: Verbal reports on mental processes.
Psychological Review 84 (3): 231–259.
Ramsey, W., S. Stich, and J. Garon. 1990. Connectionism, eliminativism, and the future of folk
psychology. Philosophical Perspectives 4: 499–533.
Roediger, H. 1996. Memory illusions. Journal of Memory and Language 35: 76–100.
Rosenthal, D. 2005. Consciousness and mind. Oxford: Clarendon Press.
Ryle, G. 1949. The concept of mind. Chicago: University of Chicago Press.
Searle, J. 1969. Speech acts. Cambridge: Cambridge University Press.
Slobin, D.I. 1996. From “thought and language” to “thinking for speaking”. In Rethinking linguis-
tic relativity, ed. J.J. Gumperz and S.C. Levinson. Cambridge: Cambridge University Press.
Smolensky, P. 1988. On the proper treatment of connectionism. Behavioral and Brain Sciences
11 (1): 1–23.
Spinoza, B. 1677/1955. The ethics. Illiois: Dover.
Stanovich, K. 1999. Who is rational? Mahwah: Lawrence Erlbaum.
Sterelny, K. 2012. The evolved apprentice. Cambridge: MIT Press.
Tekin, Ş. 2011. Self-concept through the diagnostic looking glass: Narratives and mental disorder.
Philosophical Psychology 24 (3): 357–380.
Tversky, A., and D.  Kahnemann. 1974. Judgments under uncertainty: Heuristics and biases.
Science 185 (4157): 1124–1131.
Vohs, K.D., and J.W. Schooler. 2008. The value of believing in freewill: Encouraging a belief in
determinism increases cheating. Psychological Science 19 (1): 49–54.
Wegner, D.M. 2002. The illusion of conscious will. Cambridge: MIT Press.
Wilson, E.O. 1998. Consilience. New York: Vintage.
Self-Interpretation and Social Cognition

Shaun Gallagher

Abstract I contrast narrative and theory-of-mind (ToM) approaches to self-­


understanding and social cognition. A narrative approach is a clear alternative to
strict ToM views on self-understanding, some of which deny that we have first-­person
access to our own mental states. On a narrative approach, self-understanding is more
than just familiarity with one’s mental states. It involves a rich knowledge of one’s
own embodied comportments and skills, one’s affective life, one’s autobiography,
and a knowledge that derives from one’s relations with others. I consider evidence
from developmental studies which shows that through our narrative understanding of
others we begin to shape our own self-narrative, registering not only their actions and
attitudes but also our own experiences in a way that differentiates self and other.

Keywords  Narrative · Self-understanding · Action · Interaction · Social cognition

In contemporary discussions of intersubjectivity, narrative plays opposite to theory.


That is, in contrast to mainstream theory of mind (ToM) approaches to social cogni-
tion, which emphasize mindreading by inference or simulation, alternative
approaches, inspired by phenomenology and hermeneutics, give narrative practices
a central role in our ability to understand others (e.g., Gallagher and Hutto 2008). In
a similar way, a narrative approach is a clear alternative to strict ToM views on self-­
understanding, some of which deny that we have first-person access to our own
mental states, and suggest that we require theoretical inferences to make sense of our
own experience (e.g., Gopnik 1993). In contrast, on a narrative approach, self-­
understanding is more than just familiarity with one’s mental states. It involves a rich
knowledge of one’s own embodied comportments and skills, one’s affective life,
one’s autobiography, and a knowledge that derives from one’s relations with others.

S. Gallagher (*)
Department of Philosophy, University of Memphis, Memphis, TN, USA
Faculty of Law, Humanities and the Arts, University of Wollongong (AU),
Wollongong, NSW, Australia
e-mail: s.gallagher@memphis.edu

© Springer Nature Switzerland AG 2018 145


P. Pedrini, J. Kirsch (eds.), Third-Person Self-Knowledge, Self-Interpretation,
and Narrative, Contributions To Phenomenology 96,
https://doi.org/10.1007/978-3-319-98646-3_8
146 S. Gallagher

In this chapter I explain why a narrative approach to self-understanding offers a


better alternative to the standard ToM approaches, precisely because it offers a more
parsimonious approach to other-understanding than those theories that focus on
mindreading by theoretical inference (the so-called “theory theory” [TT] approach)
or by simulation (simulation theory [ST]). Although the relation between narrative
and ST is more complicated, I’ll argue that narrative competency is not reducible to
competencies in theoretical inference or simulation, whether in regard to under-
standing others or understanding oneself. I’ll also argue that self-understanding is
already shaped by embodied intersubjective relations. In this respect, it is never a
solipsistic affair, but is already structured by social and cultural norms and institu-
tions that are often communicated in narrative form.

1  Intersubjectivity First

Narratives have a pervasive presence in our daily lives and are found in all cultures
(Flanagan 1992). From childhood, by being exposed to stories, we develop specific
kinds of narrative competencies that allow us to develop nuanced understandings
(and sometimes mis-understandings) of others. Competency with different kinds of
narratives, including what Jerome Bruner (1990) calls ‘folk psychological narra-
tives’, enables us to understand others in a variety of ways. Daniel Hutto terms this
‘the narrative practice hypothesis’: “children normally achieve understanding by
engaging in story-telling practices, with the support of others. The stories about
those who act for reasons – i.e. folk psychological narratives – are the foci of this
practice. Stories of this special kind provide the crucial training set needed for
understanding reasons” (Hutto 2007, 53).
The stories that we tell children allow them to understand that others do various
things, in various situations, and have reasons for acting in such ways. Children are
educated in narratives by being exposed to and by engaging with stories presented
by caregivers, and through a variety of media. These include what Menary (2008)
calls embodied narratives. “This little piggy …” is a nice and very basic example
where the caregiver enacts the story on the child’s body (e.g., the toes) and draws
the child into the play of the rhyme. In some instances of storytelling children are
actively engaged, prompted to answer questions about characters and their actions;
their attention is directed to (and by) particular events in the story. This will nor-
mally involve jointly attending (with caregivers) to portrayed actions, but also, in
folk psychological narratives, to thoughts and reasons when discussing what the
story characters know, feel and want. During this process children learn how others
behave in relation to each other and they start to learn a vocabulary of intentional
attitudes and a psychological understanding of intentions and emotions. Importantly,
narrative allows the child to see these attitudes in a wider context so they learn how
and why these attitudes matter not only to the story protagonists, but also, since
these early narrative encounters are themselves interactive engagements with
Self-Interpretation and Social Cognition 147

c­ aregivers, to their caregivers and to themselves. Reasons for acting, in their great
variety and complexity, are put on show in this way (Gallagher and Hutto 2008).
Through such narrative practices children develop an implicit practical under-
standing of others and their situations, and of how to make sense of their actions.
This is quite different from the account given by ToM approaches where we require
theoretical inference or simulation to mindread the mental states of others. One
problem with such approaches is a version of the frame problem that I’ve called the
‘starting problem’ (Gallagher 2011).
For example, the theory theorist will claim that we formulate our theoretical
inference about the other’s mental states by appealing to some specific folk psycho-
logical rule that will explain the other person’s behavior. That assumes, however,
that we already know that this specific situation calls for this specific rule. In many
circumstances, however, a certain action or behavior situation remains ambiguous. A
wave of the hand may mean ‘hello’ or ‘goodbye’ or ‘stop’ or ‘enough’ or ‘get out of
here’, and so on – and it may mean one thing or the other in different circumstances
and if performed by different people with different authority statuses. The rules of
folk psychology, however, are abstract – they supposedly apply to human behavior
in general, and, in part, that’s what makes them theoretical. The problem is, faced
with a particular and ambiguous situation, how do we know which rule to apply?
The simulationist faces the same problem. For example, Alvin Goldman describes
the steps involved in running a simulation routine as follows.
First, the attributor creates in herself pretend states intended to match those of the target. In
other words, the attributor attempts to put herself in the target’s ‘mental shoes’. The second
step is to feed these initial pretend states [e.g., beliefs] into some mechanism of the attribu-
tor's own psychology ... and allow that mechanism to operate on the pretend states so as to
generate one or more new states [e.g., decisions]. Third, the attributor assigns the output
state to the target ...” [e.g., we infer or project the decision to the other’s mind].
(Goldman 2005, 80–81)

The first step seems problematic. How do I know which pretend state (belief or
desire) matches what the other person has in mind? Indeed, this is what the simula-
tion is meant to deliver. If I already know what state matches the target, then the
problem, as defined by the simulationist approach, would already be solved.
Starting the process, then, seems to be a problem for both TT and ST. A hybrid
version of ToM, fares no better. If folk psychology is supposed to get my simulation
off the ground, I am still caught attempting to identify what folk psychological rule
is appropriate for the particular situation; and if to know what rule of folk psychol-
ogy to apply I need to begin by simulating pretend states, I still run into the problem
of already having to know what the other person is thinking. Hybrid approaches
simply push the problem a step back  and lead into a questionable (and non-­
hermeneutical) circle that turns from abstract rules to unsure suppositions and then
returns to abstract rules. ToM provides no account of how one obtains the right kind
of particularistic or contextual knowledge that would be the ground for getting min-
dreading off the ground.
ToM fails to account for two important aspects of social cognition that would
help us avoid the starting problem. First, by narrowly focusing on mindreading or
148 S. Gallagher

attributing mental states, ToM fails to account for our dynamic embodied abilities
for perceiving and interacting with others. These “primary intersubjective” abilities
develop early in infancy (Trevarthen 1979), and involve face-to-face interaction and
the perception of the other’s movement, posture, gesture, facial expression, vocal
intonation, etc. “Secondary intersubjective” abilities start around 9 months of age.
They involve joint attention and getting a grasp on the other’s contextualized situa-
tions through joint action in pragmatic and social contexts. In this regard we begin
to understand others in terms of how they engage with the surrounding environ-
ment. Interaction theory (Gallagher 2001, 2005, 2008; Hobson 2004; Ratcliffe
2007; Reddy 2008) argues that in our dynamic interactions with others we perceive
their intentions and emotions in their bodily and contextualized actions and often
engage with them in a co-constitution of meaning.
Second, ToM fails to account for what I’ve called, following terminology sug-
gested by Bruner and Kalmar (1998), the massive hermeneutical background
(MHB) of particularistic knowledge that enters into our understanding others
(Gallagher 2011). For Bruner and others, we gain the MHB, not only through our
embodied experience of others, but importantly by exposure to narratives. Together,
primary and secondary intersubjectivity and the narrative extension of the MHB
provide the ability for understanding others that is sufficient in most of our everyday
interactions without the need for mindreading in the sense defined by ToM.
In contrast, on a nativist view of TT, which contends that an innate ToM module
for social cognition simply comes online early in development, and allows us to
reason our way into an understanding of others (e.g., Carruthers 1996; Scholl and
Leslie 1999), the starting problem is simply ignored rather than solved. On this
view, we simply have the capability and start using it when our brain is sufficiently
developed. One would still need to explain how we are able to apply folk psycho-
logical knowledge to particular situations.
On a more empiricist view, however, one might argue that there is a natural con-
nection between the MHB and folk psychology. On this view one could conceive of
folk psychology as a set of generalizations based on the MHB. We gain the MHB in
lots and lots of observations of others, and from such experiences we simply
abstract, through an inductive process, the general rules and theory of human behav-
ior that constitute folk psychology. Gopnik and Meltzoff (1997), for example, argue
that young children are like scientists, constantly doing experiments (having experi-
ences, playing, observing others) and generalizing across those experiments. Folk
psychology, then, might be considered an abstract set of principles generated from
the particularities of the MHB. Accordingly I can simply draw on that background –
the kind of very particular knowledge which comes from our experience of how
people behave – to set the stage for the application of folk psychological rules.
One might make a similar argument about simulation skills. On this view we
could consider the MHB to consist in a learned set of skills or practical knowledge
of how to deal with people. Thus, for either TT or ST, I know what rule or principle
to apply, or what simulation to run, for any particular situation, because I draw on
the particular knowledge or set of skills I have in the MHB.
Two conceptual problems follow from this way of thinking. First, if these empiri-
cist accounts are accounts of how we acquire the MHB to support folk psychology
Self-Interpretation and Social Cognition 149

or, mutatis mutandis, simulation skills, they presuppose that the MHB already
involves an implicit understanding of others. That understanding would be based on
embodied interaction and narrative rather than theory or simulation (since the latter
two would run us back into the starting problem). On the one hand, if there is not an
intersubjective understanding already implicit in the MHB, then it’s not clear how
we could rely on it to specify a relevant folk psychological rule or how it could be
the basis for activating a simulation. On the other hand, if there is an intersubjective
understanding already implicit in the MHB, then it undermines the typical universal
claim that normally goes along with ToM – namely that folk psychology or simula-
tion, or some hybrid form of mindreading is the primary and pervasive way in
which we understand others. Rather, the embodied and contextualized processes of
primary and secondary intersubjectivity, together with an intersubjective narrative
understanding that is already implicit in the MHB, would have a necessary primacy.
Moreover, if we rely on the MHB to get folk psychological inferences or simula-
tions off the ground, then MHB would be at least as pervasive as folk psychology or
simulation ability. In that case, rather than primary and pervasive, folk psychology
or simulation would be derivative and perhaps specialized, put into use only in situ-
ations where our interactions with others break down and the resources of the MHB
are not sufficient to deliver a good understanding of others.
Second, if there is already an intersubjective understanding implicit in the MHB,
it is not clear whether mindreading is simply a continuation of this kind of under-
standing, or constitutes something different; and if the latter, what that difference is.
Is the application of folk psychology a way of breaking away from this primary
understanding, or a continuation of it? (see Astington 1990; Guajardo and Watson
2002; Lewis 1994; Nelson 2007). Are simulation skills of a different nature than the
intersubjective understandings implicit in our embodied interactions and in the
MHB, or if not, does that mean that the MHB is already a matter of simulation – is
it simulation all the way down? It has been argued by simulation theorists, for exam-
ple, that to understand a narrative, one needs to empathize with, or put oneself into
the place of the protagonist or antagonist, etc., and this requires simulation (e.g.,
Stueber 2006), or more generally some kind of ToM (Astington 1990). To develop
an answer to this argument, and to see its implications for the concept of self-­
understanding, we need to get clear about some distinctions between first-person,
second-person, and third-person perspectives.

2  Missing Persons

Barresi and Moore (1996) provide a straightforward and influential characterization


of the distinction between first-person and third-person in terms of different kinds
of information about intentional relations. The latter concept is close to the phenom-
enological conception of intentionality  – a directedness of attention or activities
towards some object or state of affairs. Third-person information is the information
an observer has of another person’s intentional relations, strictly in terms of that
150 S. Gallagher

person’s observable behavior – her movements, gestures, facial expressions, and so


on. First-person information is the information an agent has of her own intentional
relations; specifically, it is information about what the agent perceives, and the
agent’s own activities. These characterizations of third- and first-person information
are consistent with ToM approaches to understanding others and ourselves. Baressi
and Moore carefully steer away from notions of direct experience of self or direct
perception of other’s intentions and emotions. For them, we are, in all cases, for
purposes of self- or other-understanding, observers – even when we are agents. We
have no direct access to our own or the other person’s mind.1 First-person involves
information about our own actions; third-person involves information about another
person’s actions. This is the way Barresi and Moore’s distinction has been taken up
in much of the experimental literature (see e.g., Carpenter et al. 1998; Decety and
Jackson 2004; Tomasello et al. 2005; Uithol and Paulus 2013).
Barresi and Moore, however, go further and suggest that there are various com-
binations of first- and third-person information in what they term ‘intentional sche-
mas’ (also Tomasello et  al. 2005). For example, cases of joint attention and
interaction involve an integration of first- and third-person information. As these
experiences emerge during the first year of life, the “resulting representation of the
[shared] intentional relation is one that spans the self and other without the inten-
tional relation being identified as existing in either individual alone” (Barresi and
Moore 1996, 116). The principle of this integration depends on a matching or an
imitative contingency existing between first- and third-person information, and the
integration is such that there is a lack of differentiation between self and other. This
easily becomes the basis for a simulationist interpretation.2 Accordingly, Barresi
and Moore point out that as children develop they can imagine the other’s first-­
person information (as in cases of empathy, which appears in the second year of
life); and likewise can imagine their own first-person information in terms of third-­
person information (as in cases of mirror self-recognition, and later in self-­conscious
emotions, such as coyness and embarrassment, where self is represented in
third-­
­ person information). Mindreading (at around 4  years of age or earlier3)
emerges when this ability to imagine or simulate goes meta-representational and the
child is able to imagine both first-person and third-person information about the
other at the same time.

1
 This has been called the ‘unobservability principle’ (Krueger 2012), and has been subject to
some debate (see Bohl and Gangopadhyay 2013; Lavelle 2012; Gallagher 2015a).
2
 As simulation theorists explain it: “There is something shared between our first- and third-person
experience of these phenomena: the observer and the observed are both individuals endowed with
a similar brain-body system. A crucial element of social cognition is the brain’s capacity to directly
link the first- and third-person experiences of these phenomena” (Gallese et al. 2004).
3
 At the time of Barresi and Moore’s (1996) paper, it was thought that children pass false belief tests
only around the age of 4 years. Since that time, experiments show that much younger children,
around age 1 year, pass implicit false belief tests. These new results present a problem for their
theory, and for any theory theory or simulation theory of this sort. See Gallagher (2015b) for fur-
ther discussion.
Self-Interpretation and Social Cognition 151

What’s missing in this kind of account is any conception of a second-person


experience. Something that resembles second-person experience seemingly has to
be constructed from an integration of first- and third-person information – a process
that, when mature, seems more distinctively additive than integrative. Indeed, on
ToM models, it’s not clear what something like a second-person experience would
be, since our experience of the other person is equated with having third-person,
observational information about that person. What’s missed in such accounts is the
importance of intersubjective interaction.
To get a sense of what second person means, consider a recent paper by Eleonore
Stump (2013). She begins with a quotation from Peter Hobson: “a system [of child-­
in-­relation-to-others] not only exists, but also takes charge of the intellectual growth
of the infant. Central to mental development is a psychological system that is greater
and more powerful than the sum of its parts. The parts are the caregiver and her
infant; the system is what happens when they act and feel in concert” (Hobson 2004;
cited in Stump 2013, p.  153). For Hobson, as for other interaction theorists, the
child-in-relation-to-others involves a dynamical system that adds up to more than
the sum of its parts. These dynamical interactions are what constitute second-person
experience. Here is Stump’s characterization: “a second-person experience is differ-
ent in character from a first-person or a third-person experience because it is neces-
sary for a second-person experience, as it is not for a first- or third-person experience,
that you interact consciously and directly with another person who is conscious and
present to you as a person, in one way or another” (2013, 163). If we retain the
characterization of first-person experience (or information) as involving only aware-
ness (or information about) my own intentional attitude, and if third-person experi-
ence (or information) consists only of observational (i.e., non-interactive) awareness
(or information) about the other person’s behavior, then it’s clear that second-person
experience offers something different.
If the second-person is missing in ToM accounts, so is the first-person plural.
Much has recently been written about the “we” (e.g., Rakoczy 2007). But let me go
back to an early essay (1986) by David Carr where this notion was already expli-
cated in a way that could inform our current attempts to understand self-­understanding
and other-understanding. Carr nicely captures the irreducibility of the “we” experi-
ence in his example of seeing the Eiffel Tower. If you and I saw the Eiffel tower at
different times, or at the same time but we were unaware of each other, we could
say, “we saw the Eiffel tower.” But, as Carr points out, this would not mean anything
more than “you saw the tower” and “I saw the tower.” Even if I knew that you saw
the tower the day before I did, that would still amount to an addition of third-person
information and first-person information about seeing a common object, and it
would still not add up to the experience of you and I together seeing the Eiffel tower
in a joint way, since “in this case each of us saw the tower and was aware that the
other was seeing it too” (Carr 1986, 525). Joint attention is itself already a basic
joint action (Fiebich and Gallagher 2013), and setting out on a walk together to see
the tower and accomplishing our goal is a complex joint action that requires a “we”
that transcends anything we could accomplish on our own.
152 S. Gallagher

I indicate this insight about the “we” to try to capture what I mean by the idea
that second-person experience in interaction or in joint action (with a joint inten-
tion) transcends (or is irreducible to) first-person or third-person, or some addition
of the latter two. This notion of transcendence, even if it is difficult to explicate, is
nothing mysterious. It’s an everyday occurrence, on the occurrence of interaction.
To reduce it to an addition of first- and third-person information is to miss the phe-
nomenon and in fact to take a third-person perspective on it.4
Carr, who goes on to consider the notion of collective intentionality, makes it
clear that one important ingredient of the “we” experience is the sense of participat-
ing (together) in the action or interaction, and making sense of the action or interac-
tion in terms of that participating.
Stump takes a different route by suggesting that one way to differentiate second-­
person experience is to say it is not a kind of knowledge-that. This may differentiate
it from third-person, observational awareness, but it is not clear that first-person
experience, on some conception, is necessarily a form of knowledge-that. I agree,
however, that “second-person experiences cannot be reduced to first-person or third-
person experiences without remainder” (2013, 164). Despite what Stump takes to be
an irreducibility to a form of knowledge-that, she suggests that second-­person expe-
rience can be communicated by means of narrative, or more precisely, a story. What
she has in mind is the kind of encounter one might have with a protagonist in read-
ing a good piece of literature. Take as an obvious example the kind of emotion-
eliciting experience one has when reading a tragedy or comedy, or cheap romance.
We can put the point the other way around by noticing what we lose if we try to reduce a
narrative to expository (that is, non-narrative) prose. If we boil a story down to non-­narrative
propositions, so that all the knowledge it conveys is knowledge that, then we lose the
knowledge that the story distinctively provides just because we cannot convey by means of
expository prose alone even a simulacrum of a second-person experience. (2013, 165).

The simulationist might claim victory here, however. Clearly, the simulation
theorist would say, second-person experience just is a form of simulation, whether
that is in the mode of face-to-face encounters with others, or face-to-page ­encounters
with fictional characters in, for example, a novel. And ST may contend that empathy
is clearly a kind of experience that is not reducible to knowledge-that, although we
might also note that narratives and stories do communicate some amount of
knowledge-that.
This is one way to get narrative into the story about social cognition, but it is not
the most important role for narrative in this regard. That is, the question is not
whether our encounters with characters in fictional narrative or stories are second-­
person encounters. Rather, as we indicated above, the question is: What do narrative
practices contribute to establishing a second-person understanding? Narrative com-
petency is not simply the ability to comprehend and produce stories. Narrative in a

4
 It’s an interesting question whether second-person experience is something more or less, or in
some sense different from the first-person plural. I set this question aside for my purposes here, but
I would argue that second-person experience is more basic in the same way that intersubjective
interaction is more basic than joint action (see Fiebich and Gallagher 2013).
Self-Interpretation and Social Cognition 153

wider sense includes certain oral and written communications and self-reports on
experience, structured monologues as well as conversationally based interactions.
In this sense narrative is more basic than story, and is not necessarily characterized
by a formal plot structure (Abbott 1998). Furthermore, narrative practices include
not just abilities for understanding narratives, but also capacities for narrative under-
standings, which allow us to frame our understanding of self and others in a narra-
tive way, and thence to form/produce self-narratives and narratives about things,
events and other people. We should distinguish between narrative framing, as an
implicit process or practice of seeing/understanding events in a narrative frame-
work, and narrative production, which is the explicit construction of stories or nar-
rative reports. When Bruner (1990) describes how we come to know our world and
construct our representation of reality through the use of narrative, and when he
suggests that “we organize our experience and our memory of human happenings
mainly in the form of narrative – stories, excuses, myths, reasons for doing or not
doing, and so on” (p. 4), he is describing narrative framing.
We make sense out of our own actions and out of the actions of others by placing
them in a narrative framework. This is not a new theory of understanding. It has its
roots in hermeneutical traditions. Thus, for instance, Dilthey recognized that it is
not sufficient to focus on grasping the mental states of others in order to understand
their actions.
It is necessary to distinguish the state of mind which produced the action by which it is
expressed from the circumstances of life by which it is conditioned. … [In some cases]
action separates itself from the background of the context of life and, unless accompanied
by an explanation of how circumstances, purposes, means and context of life are linked
together in it, allows no comprehensive account of the inner life from which it arose.
(Dilthey 1988, 153).

Dilthey expresses this in an overly mentalistic way, however; it is not the “inner
life” or the mental life that we attempt to access – if this is understood as a series of
causally efficacious mental states — but rather the other’s life in its worldly/situa-
tional contexts, and that is best captured in a narrative form (see Gallagher in press).
What we are attempting to understand is much richer than a set of mental states; it
is the other’s actions, and their reasons as they figure against the larger history and
set of projects. Ricoeur (1992) is closer to the mark: I encounter the other person,
not abstracted from their circumstances, but in the middle of something that has a
beginning and that is going somewhere. I see them in the framework of a story in
which either I have a part to play or I don’t (see Gallagher 2003). The narrative is
not just about what is going on inside their heads; it’s about what is going on in our
shared world. To understand the story of what this person is doing does not require
a mentalizing inference or simulation. Our understanding of others is not based on
attempts to get into their heads, to access a “landscape of consciousness” since we
already have access to a “landscape of action” (Bruner 1986), which is constituted
by their embodied actions and the rich worldly contexts within which they act –
contexts that operate as scaffolds for the meaning and significance of actions and
expressive movements. Crucially, coming to appreciate the other’s story  – to see
why they are doing what they are doing – does not require a capacity for mentaliz-
154 S. Gallagher

ing inferences or simulations. It rather requires what I have called elsewhere, an


enactive hermeneutics (Gallagher 2016).
Accordingly, I would return to Hobson’s idea and say that second-person experi-
ence is the experience that is had “when [two people] act and feel in concert,” but,
as we develop, this process is in part informed by the practice of narrative framing.
Second-person experience is one that emerges in a way that transcends the individu-
als (and the first-person experiences) involved. In this respect, no simulation is nec-
essary since the other person is encountered directly in fully embodied interaction,
and we make sense of their actions in context through a practice of narrative
framing.5

3  Self-Understanding

So far my discussion has focused on intersubjectivity. I’ve argued that social cogni-
tion is not reducible to the kinds of processes described by ToM approaches. Rather,
social cognition is a second-person phenomenon that gets off the ground with the
help of intersubjective embodied interactions and the influence of the MHB, par-
tially constituted by narrative practices. How does this relate to self-understanding?
I want to suggest that second-person experience transforms first-person experience;
our experiences of others transform our self-experiences. Accordingly, our under-
standing of others transforms our self-understanding, and narrative has a role to play
in this. It is not just, as McIntyre (1981, 212) pointed out, “because we live out nar-
ratives in our lives and because we understand our own lives in terms of narratives
that we live out, that the form of narrative is appropriate for understanding the
actions of others”; it begins by going the other way: because we frame our under-
standing of the actions of others in narratives, the form of narrative is appropriate
for understanding ourselves.
As already indicated, we learn to make sense of persons (others as well as our-
selves) in part through narrative practices, starting as young children. Since we
develop in social contexts and normally acquire the capacity for narrative in those
contexts, then the development of self-narrative obviously involves others. Katherine
Nelson (2003) points out that “with respect to the child’s own experience, which is
forecast and rehearsed with him or her by parents,” competency for self-narrative
starts to emerge in 2-year olds. Self-narrative requires building on our experiences
of others and their narratives, so “children of 2–4 years often ‘appropriate’ someone
else’s story as their own” (Nelson 2003). Furthermore, to carve out a self-position

5
 It is a different question whether simulation is necessary to re-enact this kind of experience in
reading or hearing a story (as suggested by simulationists as well as by Stump). In this regard,
activation of the same neural system in each case is not a sufficient reason to conclude that under-
standing a literary character is equivalent to understanding a real and present person. The latter can
gaze back and engage interactively; the former cannot. To claim that in each case these are simula-
tion processes would be like claiming that reading about or simulating eating is much the same as
eating since the same neural processes are activated in each case.
Self-Interpretation and Social Cognition 155

within a set of narratives requires a self that is aware of itself as having a point of
view that is different from others. By the time infants are two years of age and well
practiced in understanding immediate environments and events as other people
understand them, the acquisition of language, plus the capacity to recognize their
own image in the mirror, feed a developing conceptual understanding of themselves
that is essential to the onset of autobiographical memory.
By 18–24 months of age infants have a concept of themselves that is sufficiently viable to
serve as a referent around which personally experienced events can be organized in mem-
ory…. The self at 18–24 months of age achieves whatever ‘critical mass’ is necessary to
serve as an organizer and regulator of experience…. This achievement in self-awareness
(recognition) is followed shortly by the onset of autobiographical memory …” (Howe
2000, 91–92).

Autobiographical memory is one aspect that shapes narrative competency – an


ability to see things in a narrative framework. Along with a growing linguistic com-
petency, a developing conceptual sense of self, and the interactions associated with
secondary intersubjectivity, autobiographical memory helps to kick-start narrative
abilities during the second year of life. Two-year olds may start this process by
working more from a set of short behavioral scripts than from full-fledged narra-
tives; and their autobiographical memories have to be elicited by questions and
prompts (Howe 2000; Nelson 2003). From 2–4 years, children fine-tune their narra-
tive abilities via further development of language ability, autobiographical memory,
and a more stable objective sense of self.
When children listen to stories, or play-act (and the same continues in adulthood
when we are exposed to parables, plays, myths, novels, films and other media) they
become familiar with characters in a range of ordinary or extraordinary situations,
and the sorts of actions appropriate to such characters, all of which helps to shape
their expectations about others and about themselves (Richner and Nicolopoulou
2001; Nelson 2003). Children are usually well supported in their narrative practices.
Typically, caregivers provide them with running commentaries on stories that teach
them not only which actions are suited to particular situations but also which rea-
sons for acting are acceptable and which are not (Guajardo and Watson 2002).
Narratives thus contribute to our normative understanding of what others can expect
from us, and what we can expect from others in certain situations. They provide an
important source of guidance for staking out the boundaries of what is acceptable
and what is not. In that regard, narratives may be part of the structures that shape
social and cultural institutions (Gallagher and Tollefsen 2017), that in turn inform
our intersubjective interactions. Through them we learn the norms associated with
social roles that pervade our everyday world and our social practices, in homes,
schools, playgrounds, shops, restaurants, etc. As children discover why characters
act as they do in particular cases, they become accustomed to standard scripts, sce-
narios, characters, plots, etc., and these begin to structure the child’s own actions
and her understanding of those actions.
Developmentally, in our narrative understanding of others, we begin to shape our
own self-narrative, registering not only their actions and attitudes but also our own
experiences and, at the same time, in a way that differentiates self and other. Other
156 S. Gallagher

larger narratives shape our self-understandings and our understanding of others;


narratives that help to constitute the shared normative practices that inform our cul-
tural and commonsense understandings. Self-narratives are initiated and shaped by
others and by those kinds of narratives that are common and possible in the culture
surrounding the child.

References

Abbott, P. 1998. Narrative and the evolution of intelligence. Paper presented at Department of
English, University of California Santa Barbara. April 17, 1998. http://www.anth.ucsb.edu/
projects/esm/PorterAbbott.html
Astington, J.W. 1990. Narrative and the child’s theory of mind. In Narrative thought and narrative
language, ed. B.K. Britton and A.D. Pellegrini, 151–171. Hillsdale: LEA.
Barresi, J., and C. Moore. 1996. Intentional relations and social understanding. Behavioral and
Brain Sciences 19 (1): 107–121.
Bohl, V., and N. Gangopadhyay. 2013. Theory of mind and the unobservability of other minds.
Philosophical Explorations., (ahead-of-print: 1–20.
Bruner, J.S. 1986. Actual minds, possible worlds. Cambridge, MA: Harvard University Press.
———. 1990. Acts of meaning. Cambridge, MA: Harvard University Press.
Bruner, J.S., and D.A. Kalmar. 1998. Narrative and metanarrative in the construction of self. In
Self-awareness: Its nature and development, ed. M.D.  Ferrari and R.J.  Sternberg, 308–331.
New York: Guilford Press.
Carpenter, M., K. Nagell, M. Tomasello, G. Butterworth, and C. Moore. 1998. Social cognition,
joint attention, and communicative competence from 9 to 15 months of age. Monographs of the
Society for Research in Child Development 63 (4): 1–174.
Carr, D. 1986. Cogitamus ergo sumus: The intentionality of the first-person plural. The Monist 69
(4): 521–533.
Carruthers, P. 1996. Simulation and self-knowledge: A defence on theory-theory. In Theories of
theories of mind, ed. P. Carruthers and P.K. Smith, 22–38. Cambridge: Cambridge University
Press.
Decety, J., and P.L. Jackson. 2004. The functional architecture of human empathy. Behavioral and
Cognitive Neuroscience Reviews 3 (2): 71–100.
Dilthey, W. 1988. The understanding of other persons and their life-expressions. Trans. K. Mueller-­
Vollmer. In The hermeneutics reader, ed. K. Mueller-Vollmer, 152–164. New York: Continuum.
Fiebich, A., and S. Gallagher. 2013. Joint attention in joint action. Philosophical Psychology 26
(4): 571–587.
Flanagan, O.J. 1992. Consciousness reconsidered. Cambridge, MA: MIT Press.
Gallagher, S. 2001. The practice of mind: Theory, simulation or primary interaction. Journal of
Consciousness Studies 8 (5–7): 83–108.
———. 2003. Self-narrative, embodied action, and social context. In Between suspicion and sym-
pathy: Paul Ricoeur’s unstable equilibrium (Festschrift for Paul Ricoeur), ed. A. Wiercinski,
409–423. Toronto: The Hermeneutic Press.
———. 2005. How the body shapes the mind. Oxford: Oxford University Press.
———. 2008. Intersubjectivity in perception. Continental Philosophy Review 41 (2): 163–178.
———. 2011. Narrative competency and the massive hermeneutical background. In Hermeneutics
in education, ed. Paul Fairfield, 21–38. New York: Continuum.
———. 2015a. The new hybrids: Continuing debates on social cognition. Consciousness and
Cognition 36: 452–465.
———. 2015b. The problem with 3-year olds. Journal of Consciousness Studies 22 (1–2):
160–182.
Self-Interpretation and Social Cognition 157

———. 2016. Enactive hermeneutics and natural pedagogy. In Neuroscience and education, ed.
Clarence Joldersma, 176–193. London: Routledge.
Gallagher, S. (In press). Dilthey and empathy. In E.  S. Nelson (ed.), Interpreting Dilthey.
Cambridge: Cambridge University Press.
Gallagher, S., and D. Hutto. 2008. Understanding others through primary interaction and narra-
tive practice. In The shared mind: Perspectives on Intersubjectivity, ed. J. Zlatev, T. Racine,
C. Sinha, and E. Itkonen, 17–38. Amsterdam: John Benjamins.
Gallagher, S. and Tollefsen, D. 2017. Advancing the ‘we’ through narrative. Topoi. doi: https://doi.
org/10.1007/s11245-017-9452-1. [http://rdcu.be/p8rd].
Gallese, V., C. Keysers, and G. Rizzolatti. 2004. A unifying view of the basis of social cognition.
Trends in Cognitive Science 8: 396–403.
Goldman, A. 2005. Imitation, mind reading, and simulation. In Perspectives on Imitation II, ed.
Hurley and Chater, 79–93. Cambridge, MA: MIT Press.
Gopnik, A. 1993. How we know our minds: The illusion of first-person knowledge of intentional-
ity. Behavioral and Brain Sciences 16: 1–1.
Gopnik, A., and A.N. Meltzoff. 1997. Words, thoughts, and theories. Cambridge: The MIT Press.
Guajardo, N.R., and A. Watson. 2002. Narrative discourse and theory of mind development. The
Journal of Genetic Psychology 163: 305–325.
Hobson, P. 2004. The cradle of thought. Exploring the origins of thinking. Oxford: Oxford
University Press.
Howe, M.L. 2000. The fate of early memories: Developmental science and the retention of child-
hood experiences. Cambridge, MA: MIT Press.
Hutto, D.D. 2007. The narrative practice hypothesis: Origins and applications of folk psychology.
Royal Institute of Philosophy Supplement 60: 43.
Krueger, J.  2012. Seeing mind in action. Phenomenology and the Cognitive Sciences 11 (2):
149–173.
Lavelle, J.S. 2012. Theory-theory and the direct perception of mental states. Review of Philosophy
and Psychology 3 (2): 213–230.
Lewis, C. 1994. Episodes, events and narratives in the child’s understanding of mind. In Children’s
early understanding of the mind, ed. C. Lewis and P. Mitchell, 457–480. Hillsdale: Erlbaum.
McIntyre, A. 1981. After virtue: A study in moral theory. Notre Dame: University of Notre Dame
Press.
Menary, R. 2008. Embodied narratives. Journal of Consciousness Studies 15 (6): 63.
Nelson, K. 2003. Narrative and the emergence of a consciousness of self. In Narrative and
consciousness, ed. G.D.  Fireman, T.E.J.  McVay, and O.  Flanagan, 17–36. Oxford: Oxford
University Press.
———. 2007. Young minds in social worlds. Cambridge, MA: Harvard University Press.
Rakoczy, H. 2007. Play, games, and the development of collective intentionality. New Directions
for Child and Adolescent Development 2007 (115): 53–67.
Ratcliffe, M. 2007. Rethinking commonsense psychology. Palgrave Macmillan.
Reddy, V. 2008. How infants know minds. Cambridge, MA: Harvard University Press.
Richner, E.S., and A. Nicolopoulou. 2001. The narrative construction of differing conceptions of
the person in the development of young children’s social understanding. Early Education and
Development 12: 393–432.
Ricoeur, P. 1992. Oneself as Another, Trans. K. Blamey. Chicago: University of Chicago.
Scholl, B.J., and A.M.  Leslie. 1999. Modularity, development and ‘theory of mind’. Mind &
Language 14 (1): 131–153.
Stueber, K.R. 2006. Rediscovering empathy: Agency, folk-psychology and the human sciences.
Cambridge, MA: MIT Press.
Stump, E. 2013. Narrative and the knowledge of persons. Euresis 5: 153–169.
Tomasello, M., M. Carpenter, J. Call, T. Behne, and H. Moll. 2005. Understanding and sharing
intentions: The origins of cultural cognition. Behavioral and Brain Sciences 28 (5): 675–690.
158 S. Gallagher

Trevarthen, C. 1979. Communication and cooperation in early infancy: A description of pri-


mary intersubjectivity. In Before speech: The beginning of interpersonal communication, ed.
M. Bullowa, 321–347. Cambridge: Cambridge University Press.
Uithol, S., and M. Paulus. 2013. What do infants understand of others’ action? A theoretical account
of early social cognition. Psychological Research Psychological Research 78 (5): 609–622.
Hermeneutics, Self-Knowledge
and Self-Interpretation

Bruce B. Janz

Abstract  The purpose of this chapter is to explore the ways in which hermeneutics
might contribute to understanding third-person self-interpretation. This chapter pro-
ceeds as follows: First, we show that the “persons” form of analysis is used differ-
ently in linguistic, literary, and philosophical contexts. Second, we make the whole
structure more complex by recognizing that there are different applications we can
have for this “persons” form, and in particular, the goal of providing reliable episte-
mology for self-construction is different from providing self-understanding, in the
hermeneutic sense. Third, we look at the Truman Show Delusion, which is a case in
which there seems to be a systematically mistaken third-person self-understanding.
We argue here that the Golds are looking to provide an actual mechanism to account
for this form of delusion, instead of just a symptomology as has been done in the
past. Fourth, we argue that the real issue is navigating between regimes of knowl-
edge, and that we require forensics to accomplish this. The result of these steps is to
maintain that incommensurabilities in the regimes of knowledge, for instance
between first-person and third-person accounts (or between multiple third-person
accounts) might be ways of checking the reliability of a narrative, but they might
also end up being part of the construction of the self. Holding accounts at different
regimes of knowledge is not necessarily the sign of a deficiency in one’s epistemol-
ogy, or even the sign of a delusion. We might simply be highlighting the inadequacy
of a forensics between two forms of reason, and that could be a creative moment for
the self, rather than a dissolution of the self. Finally, we introduce the idea that the
second-person can be an element of the forensics of self, in particular the openness
to the voice of the other, which makes the third-person self-understanding
possible.

Thanks to Dr. Lisa Roney for comments on this chapter


B. B. Janz (*)
Department of Philosophy, The Center for Humanities & Digital Research, and Core Faculty
in the Texts and Technology Ph.D. Program, University of Central Florida, Orlando, FL, USA
e-mail: bruce.janz@ucf.edu

© Springer Nature Switzerland AG 2018 159


P. Pedrini, J. Kirsch (eds.), Third-Person Self-Knowledge, Self-Interpretation,
and Narrative, Contributions To Phenomenology 96,
https://doi.org/10.1007/978-3-319-98646-3_9
160 B. B. Janz

Keywords  Hermeneutics · Truman show delusion · Reference · Persons ·


Forensics

The purpose of this chapter is to explore the ways in which hermeneutics might
contribute to understanding third-person self-interpretation. Hermeneutic thought is
often seen as a way of framing first-person self-reflection by defining the situated-
ness and commitments of the self’s horizon of meaning. To the extent that the
“third-person” designates something anonymous, generalizable, publically avail-
able, scientific, testable, or objective, it would seem that hermeneutics would have
little to say about third-person self-understanding. The goal here will be to examine
the concept of the third-person (and in relation to that, the other persons as well)
with a view to clarifying just what we might mean when we use it. This will help to
clarify the ways in which philosophy uses the term, and will furthermore open the
door to thinking about what interpretive value third-person experiences might have.
The place to start will be to think about the conceptual travel which happens
when we use a frame of first-, second- and third-person as applied to an understand-
ing of the self. This frame has its roots in linguistics and passes through literature
before arriving as a philosophical concept. The interaction of these areas affects
what we mean when we ask about a third-person understanding of the self.
Once this is clarified, we will look at an example which I think highlights the
potential of hermeneutics in third-person self-understanding. Hermeneutics has a
long history of placing the subject in an already meaningful world, which means
that the task of self-interpretation and self-understanding (these two may be differ-
ent, as we shall see) will always be a continuous one between various modes of
reference and perspective. If the task of self-knowledge is one of getting things right
about the self, this will have to be understood in the context of being in an already
meaningful world. In other words, knowledge about the self is not the primitive
building-block of understanding the self, and we do not first get the facts about the
self (or about our self) right and only then come to an understanding of the self.
Once we examine a range of senses of the third-person, we will unpack the ways
in which the third-person contributes to reasoning about the self and the construc-
tion of the self. The useful case here will be the Truman Show Delusion, identified
and discussed by Joel Gold and Ian Gold in Suspicious Minds: How Culture Shapes
Madness (Gold and Gold 2014). I will argue that they have provided a compelling
third-person account of the unique delusion they have identified, and that it has the
virtue of bringing explanatory power in the form of a mechanism to the phenomena.
A hermeneutic understanding of a third-person self-understanding does not intend
to unseat their account with a better mechanism, but rather to think about how we
use the rational capacity I call “forensics” to navigate diverse and incommensurable
forms of reason. Hermeneutics enables a second-person capability, that of being
open to the voice of the other, to aid in forensically sorting and interpreting between
third-person accounts. While the central questions for an epistemological theory of
the third-person are “how can we adjudicate between competing third-person
accounts?”, and “how can we account for less adequate ones?”, the central question
Hermeneutics, Self-Knowledge and Self-Interpretation 161

for the hermeneuticists is “how can we account for the construction of the self and
the world amidst the competing forms of reason which the self deploys?”, and “is
there a way of recognizing misinterpretation of the self and the world while not rul-
ing out the possibility that a minority third-person account or an unpopular account
might be correct after all?”

1  T
 he Function of Numbering Persons in Linguistics,
Literature, and Philosophy

If we are to think about third-person self-understanding, it will be important to


understand just what these terms mean and, as important, where they came from.
We use frame of the “person” – first-, second-, third-, and as we shall see, pos-
sibly even fourth-person, in linguistics to designate modes of address. The first-­
person stance is the speaker speaking about him/herself. The second-person stance
has a speaker directing language at another, and the third-person stance has a
speaker speaking to someone else about someone apart from either of them. These
are well-known ways of describing language, and not just language but interactive
and dialogical language. It is a way of designating referents, or perhaps put more
phenomenologically, objects of intention.
Is there a fourth-person stance? Linguists do sometimes use this category for an
obviative as opposed to proximate reference (Fleck 2008: 281). So, a third-person
referent would be to a specific person not being directly addressed, whereas a
fourth-person referent is to a non-specific person not being directly addressed. If I
say to you “Let’s talk about Chris,” we are talking about a specific third-person,
whereas if I say to you “Let’s talk about them,” we might be talking about an indefi-
nite fourth-person. I say “might” because English does not have specific pronouns
to designate a fourth-person, as some languages do (Fleck 2008: 281–284), and so
representing a true fourth-person stance in English is difficult.
Why is it important to be clear on these stances? At the most basic level, to ask
about the constitution of the self in terms of first-person self-understanding or some
other form is to use a form of conceptual organization from linguistics within philo-
sophical space. Every conceptual travel of this sort brings with it opportunities but
also assumptions which, if we are not aware of them, could structure conceptual
creation in a new area in unintended ways.
The second reason this is significant is that the linguistic organization of persons
is about referentiality and relation. It is not, in itself, about persons themselves, and
to some extent the use of the word “person” is a contingent fact of language. In
linguistics, first-, second-, third-, or even fourth-person referentiality tells us little or
nothing about persons as such, but rather about sets of relations, and in a limited
sense, forms of affect, that is, ways one faces a world. The “persons” in this concep-
tual structure already exist, and the forms of reference or engagement they have do
not in themselves affect who they are. To the extent that these modes of reference
match up to pronouns, Emile Benveniste argues that they are “empty signs” waiting
162 B. B. Janz

to be filled by a speaker (Benveniste 1971: 219). The pronouns lack material refer-
ence, but take on reference as they are used by someone, for some purpose. Their
referentiality, then, exists as potential in language itself, and is actualized (made
into a real reference) in use. “Tree” refers to an object in the world whether or not
someone speaks it, whereas “you” means nothing until it is spoken by someone.
Benveniste’s point leads to a related issue, which is that the mode of reference is
fundamentally about relation. This can already be seen in the observation that actual
persons are irrelevant to the structure of first-person and the rest. The linguistic
mechanics of these modalities of reference and relation are far more complex than
can be explored here, and of course they vary across different languages (for more
on this, see Esposito 2012: 104ff).
If we move to literature, we find that these forms of reference and relation take
on a different role. Characters are shown to be discovering (or not) their world and
themselves depending on how these forms of reference are used. Skilled writers use
these various forms of reference as perspectives, and can create (or destabilize) nar-
ratives based on these different perspectives. We draw implications about the kinds
of knowledge available which goes beyond differences in forms of referentiality.
And, we can extend this to particular genres of literature which foreground the use
of perspective, in particular, the first-person (e.g., autobiography and memoir;
Lenart-Cheng 2016), the second-person (e.g., epistolary writing), and the third-­
person (e.g., the majority of novels). There are variations on these, of course, and
one can find exceptions even in the genres mentioned here, but the point here is not
to analyze literary uses of these frames of reference but to note that they stand as
perspectives which allow for different levels and kinds of knowledge for the author,
readers, and the characters. They are, in other words, a different use of these frames
than the linguist’s interest in forms of reference.
It is worth noting that a writer will summarize the use of first-, second-, and
third-person as “voice.” In other words, the work that is done by these modalities is
different from the work they do for linguists. A voice suggests that there is a person
present, and that person matters to the relation between persons, the kind of knowl-
edge that is generated, and how all of that relates to the aesthetic created. While
linguists will sometimes refer to the fourth-person, it is rare to find this in literature.
The voice, in other words, matters, because the world that is created in the aesthetic
moment of literature uses those voices to come into being. If in linguistics what
matters is the relation, and the persons themselves are irrelevant, in literature what
matters is the voice, and the relations are inferred as we hear those voices. They are
reliable or unreliable narrators; they are omniscient or not (or perhaps systemati-
cally deluded); they have various levels of awareness of their own worlds, and in
their encounters they come to know or fail to do so. The form of voice used in a
work of literature matters to the world that is created, and moving between forms of
voice (from, for instance, first- to second-person, or first- to third-person), when
done well, can add layers and nuances to the world of the work.
This structure means that narrative will be much more likely to arise from the
literary version of these modalities than it will from the linguistic version. Or, put
more carefully, in order for the linguistic emphasis on reference and relation to be
Hermeneutics, Self-Knowledge and Self-Interpretation 163

useful as narrative, it will need to take on the literary emphasis on perspective and
voice. If we think of narrative as being more than just a sequence of components
(events, objects, propositions) but as a major way in which language structures and
makes available a world, then perspective and voice will be what shows the nuances
of that world. It makes available the lived meaning of the world.
In literature, then, we have a sense of perspectives, which operate within a lim-
ited epistemological space. So in other words, a first-person perspective in literature
might signal specific things about the reliability of knowledge, the status of the nar-
rator, and so forth. Those might, in the end, be questioned (a first-person narrator
does not necessarily give us a more reliable account of the self – after Freud we
understand more clearly that it is entirely possible that narrators might be systemati-
cally deluding themselves), but we start from an assumption about what that narra-
tion stands for. It is important to note that the linguistic uses of this frame of first-,
second-, third-, and fourth-person are not lost or forgotten in the literary uses. The
context of referentiality and relationality is put to specific uses, in this case aesthetic
uses. Likewise, as we move to the philosophical context for the use of these terms,
the literary is not lost, but there are different questions at stake.
Philosophy, I want to argue, has come to the concept of the modalities of forms
of address relatively recently. It is an example of “travelling concepts” (Bal 2001),
that is, concepts which are developed in one conceptual ecology travelling to
another. This happens often – “evolution” travels from biology to economics and a
host of other areas, “text” travels from literature to social analysis, “performance”
travels from theatre to anthropology, and so forth. There is, of course, nothing wrong
with any of this. It is part of the vitality of human inquiry. It is nevertheless impor-
tant to be clear on the nature of the travel, and the shifts, adjustments, and even
violence that occur when concepts travel. Most importantly, concepts have func-
tions as well as intensions and extensions. In other words, they do work in contexts,
and that work may change as they travel to a new place. A concept might superfi-
cially seem to be the same in different conceptual ecologies, but in fact it might be
activating completely different concepts as it is used.
Philosophical uses of these modalities, when they have not just been theorizing
their use in linguistics or literature, have tended to focus on the implications for
theories of consciousness and/or theories of the self. In doing so, philosophers have
drawn on both the linguist’s vocabulary of reference and relation and the writer’s
vocabulary of perspective and voice. Philosophy itself has, of course, reflected on
the nature of experience, subjectivity and the self, and so forth, since ancient times,
but these modalities bring a particular frame to those reflections. To speak of the
first-person, second-person, and third-person is to put these modes of epistemology,
narrative, and experience on an even plane, and ask how they relate to each other.
There are philosophies which would reject this even plane – solipsism would be an
obvious example. If there is nothing but the self, then the second- and third-person
are illusions. We could imagine a philosophy that privileged the third-person and
regarded the first- and second-person as illusory (although in order to do so, we
might have to deal with Descartes’ cogito argument, but versions of Buddhism
would likely qualify in this regard). Martin Buber arguably privileged the second-­
164 B. B. Janz

person (Buber 1937). The point is that something new comes to philosophy when
we decide to use these modalities, and that new frame comes either from the linguis-
tic preoccupation with reference and relation, or from  the literary preoccupation
with perspective and voice.
We move from thinking about the structure of narrative and the aesthetic poten-
tial made available by the use of different perspectives, to the question of how the
self integrates or fails to integrate fundamentally different epistemologies. Cartesian
versions of the self solve the problem of multiple narrative sources for the self by
giving one version regulatory power. The first-person version of the self is the “real”
version. I think, therefore I am, and anything else that happens after that must be
filtered through that fundamental reality.
But this pride of place for the first-person has been questioned since at least the
middle of the nineteenth century. Marx, Freud, Saussure, Lacan, Derrida, and a host
of others showed the limits of thinking about the first-person as reliable or regula-
tory over other frames of reference. My first-person understanding may be insuffi-
ciently historically advanced to understand my debts to something larger than
myself (Marx), or I may be systematically deluded by my attempts to rationalize the
non-rational in my unconscious (Freud), or I may be irrevocably tied to the stream
of utterances in my language and thus unable to fully understand the underlying
order from the perspective of those utterances (Saussure). The point here is not to
endorse any of these critiques, but to simply observe that the primacy of the first-­
person as a source of reliable knowledge about the self has been under sustained
attack, and at this point it is hard to see how any theory of the self which simply
starts from that point without recognizing the problems with it will fare well.
As we move to thinking about the philosophical implications of all this, we make
another move. It is more than an epistemic one, for thinking in terms of first-,
second-­person and so forth raises further issues about intentionality and care. To say
that we have a third-person context of referentiality means that the two or more
people in conversation about someone or something else are intentionally directed
at that person or thing. There is at least a minimal level of care, or perhaps a minimal
level of questioning. If I am talking to someone about someone else, that “someone
else” stands as an object of concern, in a Heideggerian sense. There is an intentional
object under consideration, which is neither so thoroughly part of a context of
meaning that the object becomes transparent, instrumentalized for a further purpose
rather than an object in its own right (think of a pencil, used for writing, which only
becomes  an object of intentional thought when it breaks, that is,  when it ceases
being useful for its function), nor so wholly other that it resists intentional thought
(think of Levinas’ other, the one which can never be an object of thought because
that would require it to be part of a meaningful world). The object between com-
plete transparency and complete opacity is the object of intentional thought.
What comes with intentionality is what Gadamer would call a “horizon”. In
other words, the act of focusing on something means that there is a context of mean-
ing which makes intentional consciousness possible. In other words, the first-person
mode of understanding is never sufficient for hermeneutics. Understanding is
always rooted in an already meaningful world, and that meaning is emplaced, that
Hermeneutics, Self-Knowledge and Self-Interpretation 165

is, it exists within and is made possible by a set of material and narrative boundaries.
Our understanding of ourselves is simultaneously our understanding of our place in
that meaningful world, with a particular horizon. Is the horizon a third-person con-
text, an objective and anonymous space? For hermeneutics, it is closer to a second-­
person orientation, a recognition that we are always in a world of encounter, and we
are always being created by those encounters, to the extent that we are or can be
open to them.
Within hermeneutics there is a breakdown of the first-, second-, third-, and
fourth-person framework inherited from linguistics and given aesthetic form in lit-
erature. To some extent, the Cartesian heritage can be seen even in the numbering of
these perspectives. “First” person suggests a priority, a starting point within the
individual which becomes more diffuse as one involves other people. Hermeneutics
(and phenomenology more generally) does not start with that assumption or the
prioritizing of the individual. All self-understanding occurs within a context where
all the forms of referentiality, perspective and relationality exist simultaneously.
This means that the problematic of these frames must change from the Cartesian
assumption about self-understanding. The central question is not, “How can third-­
person self-interpretation correct, amend, supplement, or fill in the gaps for first-­
person self-understanding?” Rather, from a hermeneutic point of view the question
is, “How do the various modes of relation and perspective that contribute to our
self-understanding relate to each other, and in particular, what happens when these
modes are in some manner incommensurable with each other?”
The overall point here is that there is a conceptual travel that happens when we
go from thinking about first-person or any other form of referentiality, to its use in
literary settings for aesthetic production, to the move to philosophical questions
about actual persons and their self-knowledge. It is not that this shift cannot happen,
but that as we make the move we have to be clear about what is being transported in
that move and what is not, and perhaps more importantly, what aspects of the con-
cept can take on a new light in the transition. The relationships between these con-
ceptual ecologies must be clearly understood in order to see how concepts adapt to
a new set of questions and new uses.

2  First, Second, Third, Fourth

The use of these modalities, then, have differences depending on whether the con-
text is linguistics, literature, or philosophy. In the first, the purpose of using the
modalities is to work out the nature of reference. What is highlighted are modes of
relation, and persons themselves are merely place-holders. In the second context,
literature, the purpose of using the modalities is to work out the nature of perspec-
tive. The context is aesthetic, for the most part, and these modalities serve the aes-
thetic end of the construction and exploration of worlds and those who inhabit them.
In the third context, philosophy, the use of these modalities is perhaps more com-
plex. We might think that it is solely epistemological, and if we think that, we would
166 B. B. Janz

be highlighting the ways in which knowledge is constructed by an actor, and how


that knowledge might be used by that actor. These modalities, if viewed epistemo-
logically, are sites for the production of knowledge by the self, and the central ques-
tion would be whether one site has priority over others. So for instance, in a Cartesian
version of the world, first-person knowledge must be the starting point and hence
the ultimate judge for the reliability of any knowledge. We can imagine other epis-
temic spaces in which first-person knowledge would not have priority, but some
other mode of knowledge production might.
We might not think about these modalities, though, in primarily epistemological
terms. Or rather, while epistemology is clearly at stake in thinking about these
modalities, it might not be the first question we ask. This would bring the inquiry
into a phenomenological and hermeneutical space.
So, for instance, what might second-person self-understanding look like? Martin
Buber, in I and Thou, presents an argument for something like this (Buber 1937). He
frames the I-Thou relation quite explicitly as a form of second-person knowledge,
that is, knowledge in which the two persons engaged in a relationship are formed by
that relationship, in other words, the “poles” of the relation do not precede the rela-
tion itself. The “I” of the “I-Thou” is not the same as the “I” of the “I-It” relation,
which might be seen as a set of relations of objects in the world (perhaps closer to a
third-person modality). Theorists of empathy (e.g., Evan Thompson 2001) argue
that intersubjectivity is a fundamental feature of self-knowledge, and along with
that, empathy is central to consciousness and therefore central to self-knowledge.
What is important to recognize here is that second-person knowledge is not primar-
ily about reference, as in linguistics, nor is it about perspective, as in literature. It is
about the formation of consciousness as a task. In other words, it is not that con-
sciousness exists first and then it analyses the various modalities of knowledge in
order to determine what it can and cannot know about itself. It is that consciousness
becomes what it is through a relation, and at least some knowledge comes along
with that. A consciousness sees itself as a consciousness because of a relationship
which allows it to emerge as a consciousness.
So, between Buber, theorists of empathy, and others such as Levinas (Bernasconi
1999), a great deal has been done on second-person self-consciousness, or rather,
the emergence of self-consciousness in a second-person setting. There is not the
space here to go into that literature, except to point out one significant thing for
third-person self-understanding. The third-person is not the generalization of the
second-person, as if we might abstract elements out of our intersubjective experi-
ence to arrive at the third-person. Just as it would be a mistake to suppose that a
first-person perspective could easily become a second-person perspective (linguisti-
cally, the reference is obviously different, but even in literary or philosophical terms,
we would have something like a projection or an imposition of the first-person that
would undermine any real intersubjective relation), it would also be a mistake to
suppose that an intersubjective second-person relation can easily be rendered as an
objective, scientific, or public third-person relation.
Having said that, it is still possible that one hermeneutic element of the second-­
person might appear in the third person. That element is openness to the voice of the
other.
Hermeneutics, Self-Knowledge and Self-Interpretation 167

What, then, does all this say about third-person self-understanding? Several
things. First, the fact that these modalities are used in differing contexts means that
those contexts will travel. In particular, when we are thinking about a third-person
context in philosophy, we cannot help but think about relation and reference, and
perspective. The philosophical understanding of third-person self-understanding is
not reducible to these other areas, but it is also not wholly different. The projects are
different, but they draw on some similar elements.
Second, these modalities do not in themselves solve the question of whether we
are doing epistemology or hermeneutics. The modalities can be used in both ways
in a philosophical setting. We might be asking about differing kinds of knowledge
produced in these different modalities, or we might be asking about the different
kinds of life that are made available in these different modalities. We might want to
say that the epistemological questions in philosophy draw more on the linguistic
than the literary, and the hermeneutic questions draw more on the literary than the
linguistic, but both are a matter of degree.
Third, given that we might be doing either epistemology or hermeneutics, we
will inevitably have different accounts of what is actually happening in the third-­
person. If we think we are doing epistemology, the third-person is about a particular
frame of knowledge construction, which comes with forms of justification of that
knowledge. If, on the other hand, we think we are doing hermeneutics, the issues of
justification are less important and issues around narratives (how they are judged,
whether they are trustworthy, what their rhetorical force is) will be more prominent.
A narrative account of the self will draw on elements of the literary, and in particular
introduces temporality to the question of how the third-person might relate to the
self in a manner other than simply being stories about the self told in a public space.
Narratives necessarily introduce temporal elements, and frame issues of the under-
standing of the self as including an unfolding of realization. It is not just a question
of whether or not a third-person narrative is true about the self, it is that the function
of the narrative may be one of the self coming to realization, or a self-consciousness
rooted in existing meanings in the world but simultaneously a free and unique cre-
ation in that world. This is Heidegger’s version of truth as alethia, or an unforgetting
and unfolding of a way of being. The issues about third-person accounts of the self
remain issues of truth, but not issues of the correspondence between an account and
an external reality.
Fourth, given that there are different ways of thinking about the relationship
between the first-person and the third-person, we might have to have a clear way
of addressing their relationship, whether we are focusing on epistemology or
hermeneutics. This is particularly important since the relationship between a ver-
sion of the first-person, lived experience, and a version of the third-person, scien-
tific data, has not been easy to reconcile. We have some candidates for this. Daniel
Dennett has argued for “heterophenomenology” (Dennett 1991: 66ff; van de Laar
2008), which states that phenomenology is committed to a Cartesian privileging
of the first-person, and so a method is needed which decenters that view and
places more weight on the third-person. Neurophenomenology starts from expe-
rience as it is actually lived, and so for at least some commentators is seen as
168 B. B. Janz

being more open to first-person experience (van de Laar 2008: 368ff), although
an argument can be made that it makes possible a second-person approach to
consciousness (Olivares et al. 2015). The problem that these methods try to solve
is the problem of studying human consciousness scientifically. To the extent that
science is only about the “objective” in the sense of something available in the
same manner to anyone, the gulf between lived experience and scientific study
will remain vast. Neurophenomenology, at least, offers the possibility that a
third-person account might be able to work with first-person perspectives without
reducing them to data points or subjective (and hence unreliable and ungeneraliz-
able) reports.
Fifth, there is a further difference between the contexts of epistemology and
hermeneutics, which is that the very nature of the third-person relationship itself
might have to be defined differently in each case. For epistemology, the specific
nature of that third person might not matter much, nor does our stance toward that
third person matter. It functions, in other words, more like what we previously saw as
the fourth-person stance, that is, the stance in which the person outside of the imme-
diate exchange between the self and an interlocutor is a generalized person rather
than a specific one. There is no necessity for third-person knowledge to be about any
specific third person. For hermeneutics, though, the situation is different. The third
person functions more like an actual third person, and the question of the nature of
that person matters. Levinas gives us the best picture here (Bernasconi 1999). An eth-
ics of the third person is woven into the third-person knowledge, including third-
person self-understanding. It matters what the status of that third person actually is,
and what our relationship is. In other words, we do not simply get straight the claims
to knowledge first, and then later ask about ethical or political considerations.
We can see different but related philosophical projects that would both lay claim
to third-person self-understanding. Focusing on epistemology leads us to questions
about the reliability and applicability of third-person narratives, and gives the means
to adjudicate between rival third-person accounts based on objective factors.
Focusing on hermeneutics leads us to questions about the production of the self in
third-person narratives. As important, as Otto Doerr points out in connection with
Jaspers’ work on psychopathology (Doerr 2014: 22ff), hermeneutics emphasizes
the importance of language, and the world that language creates. In other words, it
is not that language is a vehicle for truth-values, as in the epistemological approach.
It is that language itself is important in creating a space in which truth or falsity
makes sense at all. There is a distance between the third-person account and the
first-person account (and, for that matter, there is a distance between multiple third-­
person accounts). This distance is not a problem but an opportunity for hermeneu-
tics. It is the very basis for understanding, since it sets up the possibility for seeing
the limits of understanding. So unlike the Cartesian privileging of the first-person,
which renders the third-person as a problem to be solved, hermeneutics regards the
third-person as the basis for working out what the first-person might be. It is part of
Gallagher’s “pattern theory of self” (Gallagher 2015).
So, we have a range of disciplinary approaches for the first-, second-, and third-­
person structure, from the linguistic to the literary to the philosophical. We have the
Hermeneutics, Self-Knowledge and Self-Interpretation 169

possibility of a fourth-person, as we consider whether third-person applies only to


reference to a particular person as opposed to any non-present person. The fourth-­
person, if we are to use it, would be a generalized reference, or perhaps a universal
observer standpoint. And, we have a range of philosophical possibilities for the
third-person, from epistemological to hermeneutical.

3  Culture and Third-Person Self-Understanding

Hermeneutics suggests that context matters to third-person understanding of any


sort, including self-understanding. If context matters, then culture matters (Hutchins
2014). The horizons of meaning we inhabit and interpret become solidified and
replicated as cultural objects and cultural patterns. Our narratives do not spring
anew each time they are needed or apprehended. They are dependent on prior pat-
terns, meanings which have become available as pathways and borders and are told
and retold with variations, bringing something new into being by articulating some-
thing familiar. There is nothing new about this – Gadamer worked this out at length
in Truth and Method and in other places (Gadamer 1989, 2000).
What this means, though, is that the primary question of third-person self-­
understanding is not how we adjudicate between competing third-person narratives.
That approach holds to a Cartesian version of the world in which it is the introspec-
tive self which in the end determines what it will use and what it will not. But this
does not mean that every third-person narrative is as good as every other, and that is
it merely an act of the will or, alternatively, an act of social construction that deter-
mines the self. Hermeneutics is not voluntarism, nor is it social determinism.
Instead, we could follow Gallagher’s “midline self” theory (Gallagher 2013), a par-
ticular pattern theory of self which allows us to avoid the stark version of a third-­
person self-understanding as presented in an epistemological application of the
modality of the third-person, which would focus on the question of whether the
third-person account was true or not, based on a correspondence to reality, including
the reality of the self. To the extent that the self is an assembly (but not a complete
fictional construction), it is possible to have a version of third-person self-­
understanding which allows for the emergence of the self through third-person
accounts, and as such allows the third-person to capture the act of living as a person
rather than simply attributes or characteristics of the person. Part of allowing the
third-person to capture the act of living as a person is to see an element of the
second-­person in the third person – being open to the voice of the other.
To see the difference this makes, we can consider a recent example of a culturally
based delusion. Joel and Ian Gold posit a cultural basis for mental illness in their
recent book Suspicious Minds: How Culture Shapes Madness (Gold and Gold 2014;
see also Gold and Gold 2012). They speak of an affliction they call the “Truman
Show Delusion” in which a patient believes that his or her entire life is being lived as
the subject of a reality show. All the people he encounters are actors, his or her entire
life is scripted or at least directed, and there are cameras everywhere filming his
170 B. B. Janz

or her every action. In 1998 The Truman Show depicted just such a scenario, and
after the release of the movie Joel Gold, in his capacity as a psychiatrist at Bellevue
Hospital in New York, saw a steady stream of patients who exhibited this delusion.
Suspicious Minds is largely about the spotty history of the search for a medical
account of delusion. The Golds argue that the lack of success, and the resultant
continuation of a reliance on symptomology to diagnose delusion, suggest that we
have failed to fully appreciate the role of cultural content in delusions. The Truman
Show Delusion is, they argue, a function of the age of surveillance and control that
we live in, and can find correlates in earlier ages in which technology operates
differently.
Delusions, as the Golds describe them, raise the question of the tension between
self-knowledge and self-understanding, with self-interpretation in the middle. If
something like a Truman Show Delusion can be identified, that suggests that the
self-knowledge of a subject has significantly to do with his or her self-­understanding.
That in itself is not new – a common account of paranoid delusions, for instance,
assumes that a subject is a particular kind of actor, with a specific role that requires
the scaffolding of a narrative. The Golds’ argument is that an adequate account of
the Truman Show Delusion cannot just follow this pattern, since there is no mecha-
nism given. The Golds construct an account of the rise of delusions that draws on
the evolutionary theory formulated by others, to posit a “suspicion system” which
enables us to interpret complex social structures with a view to potential threats to
those systems. These threats are to potential non-immediate, high-level breakdowns
in the social fabric. The Truman Show delusion is the breakdown between a first-­
order suspicion system and a second-order reflective system, and has both an evolu-
tionary basis (as we have built-in mechanisms that allow us to make quick decisions
for the sake of threat reaction) and a cultural basis (as the threats we experience can
manifest themselves with cultural coding and context, and a person needs to be
competent in a culture for at least some of the threats to be understood). In other
words, the Golds’ account is less about the failure of a system of understanding (as,
for instance, in the case of paranoid delusions), than the failure of mechanism that
uses elements of understanding but which has a reason for existing, based on an
evolutionary account. Delusions are mal-adaptions within our current social con-
text, using culturally specific content, of otherwise useful features of our social
evolution.
Why is this example significant? For one thing, the fact that a delusion of this
sort could be so closely tied to a cultural event (the movie The Truman Show) sug-
gests that for those suffering from this affliction, they are searching for a useful
story which is complete and consistent. For another, we have here a third-person
self-understanding, that is, a confirmation of an account of the self in these patients
that depends not only on recognizing patterns in the world that confirm an account,
but also on a dismissal of the second-person moment in the third person. The
­openness to the voice of the other is absent, in that there is no possibility that any
other can disrupt the third-person account. It is this that makes the Truman Show
Delusion a delusion.
Hermeneutics, Self-Knowledge and Self-Interpretation 171

The Golds posit an explanation for this cultural delusion, rooted in evolutionary
theory. They are seeking to explain a clearly erroneous account of the self which
leads subjects to project a third-person narrative on the world, and see that narrative
as borne out in every action by other people. It is a systematic, self-justifying third-­
person narrative which confirms a first-person narrative on the part of the subjects.
Their account is certainly plausible in the sense that it is internally consistent to
them, and it gives an explanation for a third-person self-understanding which rein-
forces what, to the rest of the world, is clearly an error. In other words, their account
admirably fulfills the requirements of epistemology.
The account given of the Truman Show Delusion is plausible, and has the virtue
of explanatory power. It has the hallmarks of a third-person scientific account, that
is, an account which does not primarily depend on the subject’s version of the world,
but rather on an underlying mechanism, having an adaptive logic, which accounts
for the phenomena presented in a form of delusion which has culturally contingent
elements beyond what we would normally find in delusions. The role of hermeneu-
tics is not necessarily to offer a better or more adequate explanation, if by that we
mean overturning the existing account with a more adequate mechanism. It is to be
able to connect a third-person account to the understanding, and to recognize the
second-person moment in the third-person self-understanding. We have a third-­
person account, and that account might even be held (at least in part) by the indi-
vidual in question. In other words, someone with the Truman Show Delusion could
in fact believe that the mechanism described by the Golds is a plausible account of
the Truman Show Delusion in general, while at the same time either denying that he
or she has that delusion, or that he or she has it but that in itself does not invalidate
the truth of the belief at the center of the delusion.
The point is that having a third-person explanation in itself does not tell us any-
thing about the truth or falsity of the narrative in some specific case. It would be
analogous to having a naturalistic explanation for having a belief that God exists,
and a believer accepting that that explanation is plausible, but nevertheless continu-
ing to hold that God exists. Some might see this as cognitive dissonance, but the
relationship between our belief structures and our sense of self in the world is more
complex than that.
So, is this just someone being delusional or irrational? Perhaps. And yet, it is also
possible that accounts could exist at different levels of explanation, or at levels of
explanation and understanding. I have argued elsewhere (Janz 2008: 305ff) using
Emmanuel Eze’s book On Reason: Rationality in a World of Cultural Conflict (Eze
2008) that in the case of cultural reasoning, there can be simultaneous, complete and
consistent, but incommensurate versions of reason, differing narratives which are
nevertheless held and which form part of a coherent self. We might, on the other
hand, hold different accounts of the self, again simultaneous complete and consis-
tent, which are nevertheless commensurate with each other. We are inclined to
think, in the second case, that we have a well-ordered and rational self. I would like
to argue, though, that whether or not we have commensurate accounts, we are still
faced with a transitional point between those accounts. We are still faced with
engaging in what we might call “forensics”, that is, the move between regimes of
172 B. B. Janz

knowledge (in this case, first-, second-, and third-person forms of knowledge, along
with the complexities outlined earlier). The point of forensics in investigations is to
take two or more regimes of knowledge and draw inferences between them. One
can, for instance, examine the debris at a vehicle crash site and painstakingly detail
the position of that debris, thereby inferring things like the speed at which vehicles
were travelling prior to a crash, what kind of collision happened, what the vectors
were for the vehicles, and so forth. Those inferences all exist within a regime of
knowledge, that of the motion of physical bodies and the application of physical
laws to events in time and space. That regime is complete and consistent in itself –
no extra information is needed other than what is given by observation and laws of
motion and matter. There are other possibly relevant regimes of knowledge, also
complete and consistent in themselves  – the intentional regime (what were the
intentions of the actors and how does the relationship of cause and effect operate in
relation to those intentions?), the legal regime (what were the relevant laws and
what was violated, under what conditions?), and even the religious regime (what
cosmic forces preordained or influenced what happened?). These regimes may be
real or imaginary, but they have logics of their own, and evidence that fits their terms
of reference (for a more complete explication of this example, see Janz 2008).
It is possible to move between these regimes, but it takes another kind of logic to
do this. The forensic expert is the person who specializes in that logic. He or she is
asked to draw inferences on a variety of things outside of the evidence in that regime
of knowledge. The ultimate goal is to arrive at intent, culpability, and so forth. None
of that is directly implied by the data drawn from the physical evidence in the debris,
but we make the leap between regimes of knowledge, in this case between physical
evidence and intention, or physical evidence and the legal regime. Understood epis-
temically, forensics is the logic of moving between regimes of knowledge.
Ideally, of course, we might want all these to line up, that is, we want the implica-
tions to seamlessly lead us from one regime of knowledge to the next. Emmanuel
Eze argues that we have a range of types of reason available to us, and we are always
working with multiple versions, prioritizing different ones at different times. In this
he follows Hilary Putnam’s “internal realist” model of reasoning (Putnam 1988:
107ff), but he adds something to it. He adds rationality, that is, he adds the idea that
these various forms of reason have different applications at different times, and
rationality is the capacity to order these. That capacity shows itself differently in
different places, at different times, and for different people. In other words, people
from different cultures and places might rationally use the forms of reason differ-
ently from each other, ordering or giving them priority in different ways. And ratio-
nality then functions in a forensic capacity, moving between these forms of reason,
determining what is appropriate at different times and interpreting between these
forms. One of the ways it achieves its forensic function is through the second-­person
hermeneutic capacity of being open to the voice of the other.
In view of all this, it is possible to see the Truman Show Delusion in a different
manner. The Golds are rightly looking for a mechanism that goes beyond sympto-
mology and can deal with delusions which have a cultural base to them, that is,
which can adjust in more than just details to the narratives of their surrounding
Hermeneutics, Self-Knowledge and Self-Interpretation 173

culture. They leverage one form of reason, scientific reason, to accomplish this. The
strength of this account is that it allows for diagnosis and prognosis. The weakness
is that it does not actually shed any light on why it is a delusion. It is logically pos-
sible, after all, that someone is indeed under surveillance at all times (ironically, in
the movie The Truman Show, that is in fact the situation for Truman Burbank – there
was no delusion). So, there is no epistemological third-person guarantee. There is,
however, a hermeneutic response, which is the second-person standard of openness
to the voice of the other. If someone insists on a particular account, whether first-­
person or third-person, against any and all possible evidence, it is worth asking what
the creation of the person through dialogue actually looks like in this case. Is the
person capable of self-creation through an encounter with the other, or is the other
subsumed under a prior narrative? How would one tell the difference? And yet, as
with the fictional Truman Burbank, it is at least possible that someone who claims
to be observed and broadcast to the world at all times could, in fact, be correct about
this. It is an unpopular and minority position, an exercise of what Ricoeur called the
“hermeneutics of suspicion” (Ricoeur 1974: 147ff). If we only use the Golds’ mech-
anism, we must already be committed to the idea that there is in fact a delusion
which needs to be explained. And, there is little question that the cases that show up
in a psychiatric hospital are very likely suffering from delusions. It is not difficult to
imagine, though, that someone could have both a third-person and a first-person
account which seems highly improbable but which is nevertheless true. The herme-
neutical test, the forensics that would be applied, would be the question of the open-
ness to the voice of the other. There are no guarantees with this  – this is not an
epistemological method. It is an opening of the self, all the selves implicated in this
scenario, subject and researcher alike. In the case of a delusional subject, of course,
this may well not be possible, and it would point to the possibility that there is more
going on than the failure of an evolutionary mechanism designed to protect us from
threat. It is also the failure of an existential aspect of humanity, the constructive
second-person openness, and so any therapeutic approach would have to take that
into account along with the failure of the mechanism.

4  Conclusion

How do we now move to the question which opened this discussion, that is, how
third-person self-understanding relates to other kinds of self-understanding? The
path we have taken is the following. First, we showed that the “persons” form has
its roots in multiple disciplines, and it matters whether we are thinking in linguistic,
literary, or philosophical terms. Second, we made the whole structure more complex
still by recognizing that there are different applications we can have for this “per-
sons” form, and in particular, the goal of providing reliable epistemology for self-
construction is different from providing self-understanding, in the hermeneutic
sense. Third, we looked at the Truman Show Delusion, which is a case in which
there seems to be a systematically mistaken third-person self-understanding. I
174 B. B. Janz

argued there that the Golds are looking to provide an actual mechanism to account
for this form of delusion, instead of just a symptomology as has been done in the
past. Fourth, I argued that the real issue is navigating between regimes of knowl-
edge, and that we require a forensics to accomplish this. The result of these steps is
to maintain that incommensurabilities in the regimes of knowledge, for instance
between first-person and third-person accounts (or between multiple third-person
accounts) might be ways of checking the reliability of a narrative, but they might
also end up being part of the construction of the self. Holding accounts at different
regimes of knowledge is not necessarily the sign of a deficiency in one’s epistemol-
ogy, or even the sign of a delusion. We might simply be highlighting the inadequacy
of a forensics between two forms of reason, and that could be a creative moment for
the self, rather than a dissolution of the self. Finally, I introduce the idea that the
second-person can be an element of the forensics of self, in particular the openness
to the voice of the other, which makes the third-person self-understanding
possible.
My purpose in taking this line of thinking has been to interject a positive role for
hermeneutics into what is often seen as a space governed by scientific, objective
reason – the third-person. The point is not to undermine those scientific accounts,
but as with Husserl, Merleau-Ponty, and many contemporary neurophenomenolo-
gists, it is to find a way to make the self a subject of reflection without turning it into
data and thus losing what most people see as human. Moreover, it has also been my
goal here to introduce the idea that we are not simply contrasting a Cartesian privi-
leging of the first-person with a scientific privileging of the third-person, and then
deciding how the self comes to understanding in this tension. There is much more
happening – a range of disciplinary frames for the “persons”, the need for a strategy
to deal with different rational accounts which goes beyond simply subjecting the
accounts to one form of rationality (e.g., scientific), and finally the recognition that
self-understanding is a task rather than an accomplishment, one which happens in
the context of an element of the second-person orientation.

References

Bal, M. 2001. Travelling concepts in the humanities: A rough guide. Toronto: University of
Toronto Press.
Benveniste, E. 1971. The nature of pronouns. In Problems in general linguistics, 217–222. Miami:
University of Miami Press.
Bernasconi, R. 1999. The third party. Levinas on the intersection of the ethical and the political.
Journal of the British Society of Phenomenology 30 (1): 76–87.
Buber, M. 1937. I and Thou. Trans. Ronald Gregor Smith. T. & T. Clark, Edinburgh.
Dennett, D. 1991. Consciousness explained. New York: Back Bay Books.
Doerr, O. 2014. Hermeneutical and dialectical thinking in psychiatry and the contribution of Karl
Jaspers. In Karl Jaspers’ philosophy and psychopathology, ed. Thomas Fuchs, Thiemo Breyer,
and Christoph Mundt, 19–32. Heidelberg: Springer.
Esposito, R. 2012. Third person: Politics of life and the philosophy of the impersonal. Cambridge:
Polity Press.
Hermeneutics, Self-Knowledge and Self-Interpretation 175

Eze, E. 2008. On reason: Rationality in a world of cultural conflict. Durham: Duke University
Press.
Fleck, D. 2008. Coreferential fourth-person pronous in matses. International Journal of American
Linguistics 74 (3): 279–311.
Gadamer, H. 1989. Truth and method. 2nd ed. New York: Sheed and Ward.
———. 2000. Subjectivity and intersubjectivity, subject and person. Continental Philosophy
Review 33: 275–287.
Gallagher, S. 2013. A pattern theory of self. Frontiers in Human Neuroscience 7: 443.
———. 2015. Self and narrative. In The Routledge companion to hermeneutics, ed. J. Malpas and
H. Gander, 403–414. New York: Routledge.
Gold, J., and I. Gold. 2012. The ‘Truman Show’ delusion: Psychosis in the global village. Cognitive
Neuropsychiatry 17 (6): 455–472.
———. 2014. Suspicious minds: How culture shapes madness. New York: Free Press.
Hutchins, E. 2014. The cultural ecosystem of human cognition. Philosophical Psychology 27 (1):
34–49.
Janz, B. 2008. Reason and rationality in Eze’s on reason. South African Journal of Philosophy 27
(4): 296–309.
Lenart-Cheng, H. 2016. Paul Ricoeur and the “Particular” case of autobiography. a/b: Auto/
Biography Studies 31 (2): 355–372.
Olivares, F., E.  Vargas, C.  Fuentes, D.  Martínez-Pernía, and A.  Canales-Johnson. 2015.
Neurophenomenology revisited: Second-person methods for the study of human conscious-
ness. Frontiers in Psychology 6: 673.
Putnam, H. 1988. Representation and reality. Cambridge: MIT Press.
Ricoeur, P. 1974. The conflict of interpretations. Evanston: Northwestern University Press.
Thompson, E. 2001. Empathy and consciousness. Journal of Consciousness Studies 8 (5–7): 1–32.
van de Laar, T. 2008. Mind the methodology: Comparing heterophenomenology and neurophe-
nomenology as methodologies for the scientific study of consciousness. Theory and Psychology
18 (3): 365–379.
Identification and Self-Knowledge

Luca Malatesti and Filip Čeč

Abstract  Recently, Matt King and Peter Carruthers have argued that the Real Self
accounts of moral responsibility or autonomy are under pressure because they rely
on a questionable conception of self-knowledge of propositional attitudes, such as
beliefs and desires. In fact, they defend, as a plausible assumption, the claim that
transparent self-knowledge of propositional attitudes is incompatible with mount-
ing evidence in the cognitive sciences. In this chapter, we respond to this line of
argument. We describe the types of self-knowledge that might plausibly be involved,
as psychological prerequisites, in the processes of identification and integration that
lead to the constitution of the real self of an agent. We argue that these forms of self-­
knowledge do not require the type of transparent knowledge of propositional atti-
tudes that, according to King and Carruthers, is incompatible with the results of
contemporary cognitive science.

Keywords  King and Carruthers’s criticism to the “Real self” · “Real self”
accounts of responsibility · Identification with our mental states · Self-knowledge ·
Conscious attitudes · Self

1  Introduction

Some maintain that the notion of “Real Self” is needed to account for moral respon-
sibility or autonomy.1 This thesis can handle cases where agents act on the basis of
mental states, such as desires or beliefs that they, nonetheless, do not endorse and,

1
 One of the most important advocates of this view is Harry Frankfurt, who advanced a famous
version of this position in Frankfurt 1971. For further refinements of his view, see several of his
essays that are collected in Frankfurt 1988b. Another influential advocate of the position is Gary
Watson, see Watson 1982a and Watson 2004.
L. Malatesti (*) · F. Čeč
Department of Philosophy, Faculty of Humanities and Social Sciences, University of Rijeka,
Rijeka, Croatia
e-mail: lmalatesti@ffri.hr; fcec@ffri.hr

© Springer Nature Switzerland AG 2018 177


P. Pedrini, J. Kirsch (eds.), Third-Person Self-Knowledge, Self-Interpretation,
and Narrative, Contributions To Phenomenology 96,
https://doi.org/10.1007/978-3-319-98646-3_10
178 L. Malatesti and F. Čeč

even, would prefer not to have. Obsessive compulsive disorders might exemplify
these situations. The guiding idea of the Real Self accounts is that an agent is
accountable or autonomous when she acts or thinks on the basis of those mental
states that belong to her real self. Against these views, Matt King and Peter
Carruthers have argued that the best plausible account of our self-knowledge of
propositional attitudes, such as beliefs and desires, supports, as a hypothesis that
demands careful consideration, the conclusion that there are no conscious proposi-
tional attitudes and decisions (King and Carruthers 2012; Carruthers 2011, p. 383–
383). Given that they take the existence of conscious mental states and processes to
be necessary prerequisites for Real Self views, they conclude that these accounts of
moral responsibility and autonomy are untenable.
The aim of this chapter is, in responding to King and Carruthers’s objection, to
investigate the forms of self-knowledge that are necessary prerequisites for moral
responsibility or autonomy. We do not engage with their claim that it is a plausible
hypothesis to assume that there are no conscious propositional attitudes. Instead, we
focus on the capacities for self-knowledge that are required by Real Self accounts.
King and Carruthers argue that these doctrines necessarily include the implausible
and widespread assumption that an agent, from her own first-person perspective,
knows her propositional attitudes transparently. This notion of transparency
involves the assumption that the subject, having a direct access to her mental states
that is not mediated by interpretations or explicit inferences, has a privileged epis-
temic position on her mental states over the indirect access that others might have to
them.2 We argue, instead, that a Real Self account can be formulated by spelling out
the two main types of self-knowledge required by it without assuming their
transparency.
We maintain that two types of self-knowledge or stances towards our mental
states are involved in the formation and preservation of the real self. The first one is
a stance in virtue of which the agent knows that a certain mental state occurs in her,
without endorsing it. In this case, we say that the mental state has for the agent an
objective character. The other stance involves accessing the mental state as related
with other mental states of the agent in a way that enables the agent to endorse or
reject that mental state, thus we say that the mental state has for the agent a rela-
tional character. While Frankfurt argues that the identification that is fundamental
in the constitution of the real self does not involve decisions based on reasons, oth-
ers have seen as relevant to it the capacity to recognize mental states as satisfying,
under certain requirements of practical and theoretical rationality, commitments
that stem from other mental states, values, and plans of the agent.3 We argue that in
any case identification impacts the psychological life of the agent, and thus grounds
its role in the formation and persistence of the real self, by offering reasons that are
relevant in assessing whether or not to endorse mental states and to respond to
changing circumstances.

2
 On the transparent knowledge assumption, see Carruthers 2011. For a classical analysis of the
epistemic privileged access that a subject has to her mental states, see Alston 1971.
3
 For elaborations of the notion of identification along these lines, see Moran 2002.
Identification and Self-Knowledge 179

Specifically, we will examine how certain capacities for mental time travel are
relevant to sustain this psychological role of identification. In relation to the past,
mental time travel is the capacity of an agent to have memories of past episodes in
which she was personally involved. In relation to the future, mental time travel
involves what is called “prospection”. This is the capacity of an agent to imagine
future situations where she might be involved. Relying on a proposal by Jeannette
Kennett and Steve Matthews, we focus on certain capacities of mental time travel
that relate the present state of the agent to the “normatively rich” dimension of her
past self and imagined possible future selves (Kennett and Matthews 2009; Gerrans
and Kennett 2010). Within this conception, identifications with past intentions,
desires, beliefs, and plans are relevant in shaping the normative commitments that
constitute and constrain the present self.
In this chapter we proceed as follows. First, we delineate the main tenets of the
Real Self account in terms of the notion of identification. Then, we consider the
prerequisite epistemic capacities that might enable an agent to identify with mental
states and possess a real self. In the fourth section, we respond to King and
Carruthers, by arguing that although consciousness has a role in the process of iden-
tification, our account does not require a transparent access to our propositional
attitudes. Instead, we will maintain that the access to the sensory and perceptive
data, that according to King and Carruthers are the only objects of transparent
knowledge, is enough to support the existence of epistemic capacities necessary for
the types of self-knowledge and the stances that are required by a Real Self view.

2  The Real Self and Identification

The intuitive appeal of Real Self doctrines can be illustrated with the case of indi-
viduals who are afflicted by obsessions or compulsions. Let us consider, for exam-
ple, Mary, who is constantly late for work because every morning, before going out,
she rearranges several times her wardrobe due to an obsessive belief that this action
will prevent disastrous events in her life. Let us also assume, as it often happens in
these cases, that Mary would like to get rid of these obsessive thoughts, and the
consequent compulsive behavior, because she realized that they are pointless and
harmful. For example, she understands that they affect her chances to keep a job that
she does not want to lose. The intuition is that Mary should not be held responsible
for her lateness. One plausible way to explain this intuition is that she does not have
control over the obsession and the consequent compulsive behavior. This lack of
control appears to amount to the fact that, although in a sense the obsessive thought
and the consequent behavior are her own, they do not belong to her in the right kind
of way.4 The requirement of a mental state to belong to the real self of an agent has

4
 Some authors characterize this difference with a distinction concerning the sensed sources of the
actions that might be at the basis of minimal conception of a phenomenological self. The sense of
ownership is the sense that the agent is causing or generating the action or mental process. The
180 L. Malatesti and F. Čeč

thus been introduced to describe a relation an agent needs to have with her mental
states, and subsequent actions, to be responsible or autonomous.
Harry Frankfurt has offered several influential refinements of the notion of real
self and the conditions under which a certain mental state belongs to it.5 At the core
of his different proposals lies the idea of a process of identification of an agent with
motives, desires, beliefs and/or other mental states. This process is taken to lead the
agent to embrace and internalize these mental states in a way that they come to
reflect what she really is. The mental states processed by the exercise of identifica-
tion come to be constitutive of one’s selfhood, and consequently of agency, includ-
ing also moral agency when the relevant mental states are at stake.
In one of Frankfurt’s most recent formulations, identification with mental states
expresses the agent’s care for herself, for her current wellbeing, as well as her future
being. Not caring about the current desires, beliefs, and intentions that are present
in one’s self would render that person a helpless bystander that would passively
observe as she becomes overwhelmed by the surge of her current desires. A lack of
attention for one’s plans, beliefs, and desires would rip apart the care for the future
that an everyday, average person usually expresses and would turn her into a mere
vessel for thoughts, desires, and plans that would pass through her, and uncontrol-
lably direct her behavior. Harry Frankfurt coined a name for such a creature, such a
helpless bystander: a “wanton” (Frankfurt 1971, p. 16).
According to Frankfurt, the capacity to choose which desires one will identify
with, and ultimately act upon, distinguishes the person from a wanton. Frankfurt
uses the distinction between first-order desires, for example Giorgio’s desire to
mine for gold on the shores of river Klondike and second-order desires, Giorgio’s
desire not to want to mine for gold on the shores of river Klondike. A wanton is a
creature which cannot form second-order volitions or internalized second-order
desires that ultimately are the grounds of his behavior. The wanton will whimsically
act on the basis of whichever first-order desire occurs in him. For instance, he will
dig for gold until another desire occurs in him and, when this happens, he will fol-
low it. Giorgio, on the other hand, will stick to the motivational state to which he
identifies with a second-order volition. If he desires not to want to dig for gold on
the shores of river Klondike, he identifies with that desire, even if, within himself, a
first-order desire to dig might be still present.
Let us consider in the next section how the features of the process of identifica-
tion, and their role in determining the real self, constrain the capacities for self-­
knowledge that are necessarily required for such a process to take place.

sense of agency is the sense that the agent is the one who is causing or generating an action. On this
distinction, see Gallagher 2000. However, although the sense of agency might be a prerequisite for
the real self, we are going to argue that processes involving explicit reasoning must be involved as
well. Probably, these processes would fall under what Shaun Gallagher and others call the “narra-
tive self”.
5
 Several contributes by Frankfurt on this topic are collected in Frankfurt 1988b. For critical discus-
sions of his account of agency, and Frankfurt’s replies, see Buss and Overton 2002.
Identification and Self-Knowledge 181

3  Identification and Self-Knowledge

It is important from the outset to clarify that the specific control on mental states that
we derive by identifying with them does not occur because the identification causes
the mental state at issue. Frankfurt is explicit and correct on this (Frankfurt 2002,
p.  218). We could even voluntarily produce certain mental states in ourselves.
Consider the case of someone who, let us suppose for science’s sake, takes drugs
because he wants to have the desires of an addict, although without then identifying
with them. Instead, identification involves a certain stance towards certain already
existing mental states. Let us consider the forms of knowledge and self-knowledge
that are involved in this stance.
Minimally, to identify with a certain mental state an agent would need to know
that a certain mental state is occurring in her. This knowledge involves a stance
towards a, so to speak, “objectified” mental state. Within this stance, the agent rec-
ognizes that she has a certain thought or desire, but she does not express second-­
order desires about it that would involve an identification. So, further capacities are
involved in identification, that might require further types of self-knowledge.
One theoretical option is to introduce a sensitivity to reasons as a prerequisite for
the identification with mental states. Thus, identification might require a type of
self-knowledge that permits such a sensitivity. Richard Moran, for instance, has
argued that the identification of a subject with a mental state requires a way of being
active in relation to our mental states that is based on reason-sensitive endorsements
(Moran 2002, p. 194). Specifically, Moran argues that the process of identification
should be a norm-governed activity.
However, Frankfurt has replied to Moran that identification is not a process gov-
erned by reasons and norms; instead it is a volitional state that is constitutive of
having reasons:
Identification and wholeheartedness are volitional states that necessarily create reasons but
that do not otherwise depend upon them. We can identify with various psychic elements,
and we can be wholehearted in various thoughts and attitudes, without having any reasons
for doing so. On the other hand, it is in virtue of these states of our wills that certain things
count for us as reasons. (Frankfurt 2002, p. 218)

In particular, Frankfurt maintains that the process of identification cannot be based


on a decision that is reached by means of some practical forms of reasoning, given
that it would require that any premise that enters in that decision would have to be a
mental state the agent identifies with.
Different criticisms can be levelled against Frankfurt’s interpretation of identifi-
cation. For example, Gary Watson, responding to the original formulation of the
notion of identification by Frankfurt, asked: “Can’t one be a wanton, so to speak,
with respect to one’s second-order desires and volitions?” (Watson 1982b).
According to Frankfurt’s account, theoretically, for every desire of a certain order,
there should be a desire or volition of a superior order, that effectuates the identifica-
tion. However, Frankfurt also recognizes that:
182 L. Malatesti and F. Čeč

It is possible ... to terminate such a series of acts [i.e. the formation of ever higher order
volitions] without cutting it off arbitrarily. When a person identifies himself decisively with
one of his first-order desires, this commitment ‘resounds’ throughout the potentially end-
less array of higher orders . . . The fact that his second-order volition to be moved by this
desire is a decisive one means that there is no room for questions concerning the pertinence
of volitions of higher orders ... The decisiveness of the commitment he has made means that
he has decided that no further question about his second-order volition, at any higher order,
remains to be asked. (Frankfurt 1988a, p. 91)

However, it might be still asked what constitutes the “decisiveness” of the commit-
ment, and how the agent decides that there is no need for further questions about the
grounds of his second-order volition. The proposal that the identification derives
from a reasoning that is undertaken by the agent, especially if grounded in value
judgments, might better capture the idea of the autonomy of the agent (Velleman
1982). However, we do not need to settle the issue here. In fact, in any case, in Real
Self accounts there must be space for processes that are responsive to reasons.
The constitution and preservation of the real self by means of identification
implies a certain connectedness and integration between the mental states of the
agent. As stated by Frankfurt in the quote above, identification brings about reasons.
We read this consequence of identification with a mental state as determining cer-
tain commitments that, normatively, determine the agent’s stance towards new
desires or other mental states. In this way, the agent expresses in relation to new
mental states that she might acquire and in response to changing situations the care
for what she is, what she does, what she wants to be and what she will do. Thus, the
process of identification constitutes the real self by constraining whether a certain
mental state fits or competes with other states that are already part of the real self.
Thus, a new desire might easily become part of the real self of an agent, because it
is a means to an aim already endorsed by the agent. Giorgio, for example, will
smoothly endorse changing his digging technique when an experienced gold-digger
offers him some mining tips. However, the agent would not endorse a desire that is
conducive to the satisfaction of an obsessive thought with which the agent does not
identify. Mary surely does not want to endorse the obsessive desire to rearrange her
wardrobe that occurs in her. Thus the self-constitutive commitments stemming from
identification operate in situations where the agent has several competing desires
which enter in the “space” of mutual relations of conflict, and which she might want
to endorse, by comparing them, weighing them in order to be able to accept them as
compatible with her real self. The notions of competition or fitness between mental
states appear to emerge on the background of connections that are regulated by
certain principles of practical or epistemic rationality that are primitively compel-
ling for the agent.
The role of identification in structuring the real self of the agent requires, thus,
that the agent has a specific form of self-knowledge. We can call it a “relational”
form of self-knowledge, which involves capacities that fall under the rubric of
“mental time travel”. While in general the capacity of “mental time travel” indicates
the capacity to think about the past and the future, cognitive scientists and philoso-
phers in studying the phenomenon have focused their attention on a more specific
Identification and Self-Knowledge 183

class of capacities that enable agents to control their behavior on the basis of past
experiences and imagine future situations. Here, we suggest that the capacity to
relate to occurring motivational states, to the commitments we have incurred in the
past, to approve or reject them, falls in this class of capacities (for a similar view,
see Kennett and Matthews 2009).
Thus, the model of real self that we are offering requires that accessing our men-
tal states also enables us to be responsive to the commitments we incur in having
them. Given that these commitments guide how we deliberate, they also participate
in the formation of our self, both in everyday and morally relevant situations. Thus,
our capacity for self-knowledge is a prerequisite for certain forms of moral agency.
Possessing the capacities for knowing mental states in an “objectified” manner and
in terms of a “relational” knowledge that allows us to appreciate their normative
relations with other mental states is necessary to have contrastive decisions, a plan
or a life project.
In the next section, we consider whether these preconditions for identification,
and its role in our psychological life, render the resulting Real Self account
implausible.

4  The Real Self and Consciousness

Matt King and Peter Carruthers have argued that plausible theoretical interpreta-
tions on recent empirical data concerning our capacities for self-attribution of men-
tal states puts under considerable pressure the notion of real self and identification
that we have considered so far (King and Carruthers 2012; Carruthers 2011, p. 381–
383). They maintain that identification requires that conscious judgments, deci-
sions, or evaluations exist. Moreover, they argue that contemporary cognitive
sciences offer evidence for taking seriously the thesis that there are no conscious
propositional attitudes and decisional processes. Let us examine first this latter
thesis.
According to King and Carruthers, the theory of the self-attribution of proposi-
tional attitudes that best explains extensive empirical data, and that sits well with
several theoretical proposals that are widely accepted in the cognitive sciences,
implies that there are no conscious attitudes. Carruthers, in a book-length treatment,
has called this theory the Interpretive Sensory Access (ISA) theory (Carruthers
2011). According to this theory, our access to our current thoughts and thought pro-
cesses, that are exemplified, for instance, by judging, actively wanting, and decid-
ing, is accounted by three main theses:
(a) there is a single mental faculty underlying our attribution of propositional atti-
tudes, whether to ourselves or to others;
(b) this faculty has only sensory access to its domain;
(c) its access to our attitudes (or rather to most kinds of attitudes) is interpretive
rather than transparent. (Carruthers 2011, p. 1–2)
184 L. Malatesti and F. Čeč

Let us describe in some details what the theses involve.


According to Carruthers, the ISA theory coheres with the global broadcasting
model of our cognitive architecture.6 This model requires that in the mind there are
several perceptual systems that process sensory and perceptual information that is
broadcast and made available to cognitive, conceptual systems. These systems
include belief-forming systems amongst which there is the mindreading faculty.
This faculty can only receive information from the sensory systems (thesis b) and
produces judgements and beliefs about the occurrence of mental states in other
individuals, but also in the individual to whom it belongs (thesis a). For example,
John will come to believe that Mary believes that there is a door in front of her
thanks to the workings of the mindreading faculty that processes perceptual infor-
mation about Mary, that tracks what she says and Mary’s other forms of behavior,
as well as her environment: the presence of a door in front of her. However, also
when John self-ascribes a certain propositional attitude, this is due to the operation
of the mindreading faculty that exploits sensory and perceptual information about
himself, including his behavior, his body, his visual imagery, and inner and outer
speech. Given that the mindreading faculty cannot have as input our propositional
attitudes, it follows that they are not accessed directly, or transparently (thesis c);
instead their self-ascription is based on interpretation:
on this account the mindreading system has no access to the subject’s own propositional
attitude events of judging, deciding, endorsing, and so on. These can only be self-attributed
by turning the mindreading system’s interpretative powers onto the self. (King and
Carruthers 2012, p. 213)

This account of the self-knowledge of our propositional attitudes and the events of
judging, deciding and endorsing and so on has relevant implications for the possibil-
ity to be conscious of them.
According to King and Carruthers, it follows from the ISA theory that we cannot
be conscious of our propositional attitudes and of our deciding on the basis of them.
In fact, both most relevant accounts of when a certain mental state is conscious
require an access to propositional attitudes that contrast with the ISA theory. The
first-order accounts of consciousness state that a mental state is conscious when it is
available or poised to systems for reasoning, decision-making, and verbal expres-
sion and so on.7 However, as we have seen, while the ISA states that sensory and
perceptual states are so broadcast, propositional attitudes are not. According to the
higher-order account of introspection, a mental state is conscious when we are
aware of it by targeting it by means of a second-order mental state, a quasi-­perception
or a thought (Carruthers 2000; Lycan 1987). On this view, while the mental states of
other individuals are known by means of some interpretation of their behaviour and
situation, we have a direct non-interpretive access to our mental states that renders
them conscious (King and Carruthers 2012, p. 209–211). However, the ISA theory

 For a systematic treatment and defense of the global broadcast hypothesis, see Baars 1988.
6

 Significant version of this theory are offered in Dretske 1995 and Tye 1995; see also Tye 2000.
7
Identification and Self-Knowledge 185

asserts that we can only have an interpretive access to propositional attitudes (King
and Carruthers 2012, p. 211–215).
We can grant to King and Carruthers that they build their argument on an account
of self-attribution of mental states that is compelling and that is supported by a wide
range of empirical results and hypotheses in contemporary cognitive science.
Therefore, we can regard their thesis that there are no conscious propositional atti-
tudes as a plausible one that should force us to rethink the Real Self accounts of
responsibility and autonomy. However, if there are no conscious attitudes, does it
follow that any plausible account of the real self is undermined? Let us, thus, focus
on King and Carruthers’s claim that this is so.
The problem, according to King and Carruthers, is that the distinction between
mental states that are alien to the agent and the ones that she endorses should be a
product of processes that presuppose that the agent is conscious of her propositional
attitudes. The agent should consciously decide that these urges are alien to her real
self, and that she does not want them to guide her behavior. But if such an endorse-
ment of motivational states is not conscious then the distinction fails. If the klepto-
maniac’s identification with the desire to end his thieving would be an unconscious
one, an epiphenomenon of the mental attitudes that she happens to have, then it:
... would seem that desires receiving unconscious endorsement would be equally “outside”
the real self as pathological compulsions, and an appropriation system or evaluation system
that bypassed conscious thought would by all appearances bypass the real self. (...) It seems,
therefore, that no desire could count as belonging to one’s real self unless consciously
desired to be operative, or cohering with conscious value judgments. (King and Carruthers
2012. p. 220)

However, is it the case that any Real Self account needs such a commitment?
What we have argued in the previous section seems to be consistent with the
previous premise of King and Carruthers’s argument. The identification with mental
states and their integration in the real self involves a preliminary objective stance
towards a mental state and/or a decision either to identify with a mental state or to
endorse a mental state on the basis of the reasons that stem from previous identifica-
tions. Thus, it could be further maintained that both the identification with mental
states or their integration in the real self require being conscious of our proposi-
tional attitudes. However, we think that the argument by King and Carruthers is
based on the wrong assumption, i.e., the Real Self accounts require necessarily a
direct, non-interpretive transparent access to our propositional attitudes, that, in
accordance to first-order or higher-order accounts of consciousness, would render
them conscious.
If the process of identification would require a special first-personal access to our
mental states that would render them conscious, then we would not be able to iden-
tify or integrate with certain unconscious desires in our real self. However, the iden-
tification with certain mental states or their integration with other mental states do
not require that type of first-personal access to the relevant mental states. We can
illustrate this point by using an example that has been used by David Velleman in a
different context.
186 L. Malatesti and F. Čeč

Suppose that I have a long-anticipated meeting with an old friend for the purpose of resolv-
ing some minor difference; but that as we talk, his offhand comments provoke me to raise
my voice in progressively sharper replies, until we part in anger. Later reflection leads me
to realize that accumulated grievances had crystallized in my mind, during the weeks before
our meeting, into a resolution to sever our friendship over the matter at hand, and that this
resolution is what gave the hurtful edge to my remarks.” In short, I may conclude that
desires of mine caused a decision, which in turn caused the corresponding behaviour; and I
may acknowledge that these mental states were thereby exerting their normal motivational
force, unabetted by any strange perturbation or compulsion. (Velleman 1992, p. 464)

The later reflection of the protagonist, by exploiting the remembered perceptual


content of that situation, but also that of several weeks before, his behavior on that
occasion and on previous ones, enables him to ascribe to his past self certain desires,
although not conscious at that time. In addition, the agent identifies with or inte-
grates that intention with previous identifications. For instance, he might now self-­
ascribe the commitments of not tolerating invasive acquaintances. Therefore, the
decision to terminate the friendship is not alien to the agent, on the contrary it is part
of his real self because it can be integrated with his other mental states or because it
can be an object of a second-order volitional act. Specifically, he is morally respon-
sible for the act of terminating the friendship. In fact, by a present interpretation of
perceptual, sensory, and emotional data, concerning his past behavior, including his
outer and inner speech, he forms an evaluative stance of his past mental states that
were unconscious when he acted. Therefore, sometimes we are not conscious of a
decision we made but that does not mean that the decision is alien to us, and that
later we cannot identify with its motives or integrate them in our real self.
It could be objected that our example of identification and integration of mental
states in the real self simply shows that unconscious attitudes, that guided a certain
behavior, can become part of the real self in virtue of becoming conscious in a later
time. We would not, therefore, be out of the reach of King and Carruthers’s
criticism.
However, our example also highlights the kind of evidence that is used for a post
hoc identification with a mental state or its integration in our real self. In accordance
with the ISA theory, conscious perceptual and sensory states associated with our
behavior and its circumstance, imagistic states, including anticipations of what
would happen to us if we acted in certain ways, and verbal expressions of our atti-
tudes are given to the mindreading faculty.8 Thus the mindreading system offers
self-attribution of mental states, and so the necessary precondition for identification
and integration, by relying on these conscious data. Therefore, we can say that the
self-attribution of propositional attitudes does not require that we are directly aware
of them. Instead, it is enough that the sensory and perceptual inputs to which the

8
 Neil Levy, in Levy 2012, offers a quite similar line of argument to defend the consistency of Real
Self views with the hypothesis that there are no conscious propositional attitudes. However, he
maintains that self-ascription of propositional attitudes is enabled by our being conscious of their
propositional content. We think that this is too strong a requirement that, in any case, is not com-
patible with King and Carruthers’s position. We are instead exploiting the central thesis in the ISA
theory that the same mindreading faculty operates in self-attributions and in attributions of mental
states to others.
Identification and Self-Knowledge 187

mindreading faculty applies are conscious. Additionally, once we have self-­


attributed mental states, these perceptual and sensory data relative to this attribution
can be globally broadcast and thus made available for future self-attributions. We
maintain, thus, that identifications and integrations can be based on interpretative
self-attributions of propositional attitudes. These self-attributions are formulated on
the basis of sensory and perceptual evidence concerning our present and past behav-
ior and the surrounding conditions and imaginative anticipations of future events.
Although we cannot identify with or integrate in our real self the token mental
states of other people, on the basis of perceptual information about their behavior
we can judge about the mental states they might identify with and those that are
integrated with their real self. That evidence provides enough data for us to build a
narrative about them and discriminate what can be plausibly taken to be the voli-
tional states they identify with.
The same process might be directed at ourselves. However, in our case, the sen-
sory and perceptual data have a wider range, scope and depth. There is a qualitative
difference in these data that is worth mentioning briefly. Consider the case of the
agent who breaks a friendship. Her memories of her discussions with her friends are
amongst the perceptual and sensory data her mindreading faculty accesses to self-­
ascribe certain propositional attitudes, and thus, to be then capable to identify or
integrate them. Although, by hypothesis, she was not conscious of the mental states
that she was not ascribing to herself, these memories present the discussion as a
situation she was experientially involved in. We would like to suggest, then, that as
part of the inputs in the mindreading faculty, to produce self-ascription of mental
states relevant for identification and integration, we should include also self-­
specifying perceptual and sensory information. Without entering into details, this is
the kind of information that minimally supports a certain sense of ownership that is
taken to be explained in terms of ecological self-awareness built into movement and
perception (Gallagher 2000, p. 16). Besides, there can be also self-specifying infor-
mation that involves the position of our body and the relation of a certain perceived
or remembered situation to it. In this way, sensory and perceptual information is
available to the mindreading faculty and grounds the self-ascription of intentional
mental states that are not so available to others.
King and Carruthers’s line of argument, however, appears also to imply that there
cannot be real selves, if our self-attributions of propositional attitude might not
amount to knowledge. By conceding to them the central tenets of the ISA theory, we
are also committed to the view that we lack direct and infallible knowledge of the
mental states that we identify with or integrate in our real self. Rather our ascrip-
tions are the result of the interpretation of the mindreading faculty that might con-
fabulate certain of our self-ascriptions.9 Thus, we are conceding that we might be
wrong about the mental states that we have. It follows that we might build our
identifications or integrations on false self-ascriptions of our mental states. This
might attract the objection that we are admitting that there is no real self; there is

9
 For a discussion of current experimental literature on confabulated decisions, intentions, and
judgements and how it relates to the ISA theory, see Carruthers 2011, pp. 339–345.
188 L. Malatesti and F. Čeč

just a construct. However, this criticism fails when we realize that the discussion
regarding the real self has been introduced to distinguish the aspects of ourselves
that we consider to be alien to us, from the ones that we consider to be part of us.
In everyday life, people might have distorted pictures of themselves, some over-
estimate themselves, others underestimate, and subjects often confabulate attitudes
that they do not have but report as having. Often these distorted pictures are prod-
ucts of identification and therefore, as such, report what these agents consider to be
their real selves. However, that does not mean that they do not know their real
selves. They do; it is just that the way they see themselves is a product of a set of
mental states. The distortion is due to a difference between how we see them and
how they see themselves, and not to their lack of a real self. What is alien to us in
normal, non-pathological situations, depends on our decisions and thus on self-­
attributions that are constitutively interpretative. Therefore, for example, Dr.
Strangelove might identify with his “Nazi-arm” that, at random, every now and then
salutes the Führer and because of that he would be responsible for breaking a law
that prohibits such a salute. However, if he would have considered it as non-­
conductive of his behavior, then we should deem the random arm rising as alien to
him.

5  Conclusion

We have argued that a Real Self account of autonomy or responsibility does not
necessarily involve conscious propositional attitudes and a direct access to them.
Instead, we have argued that both the process of identification and its role in the
constitution of our real selves can plausibly rely on a process of self-interpretation
that is based on sensory and perceptual information. This information will concern
our present and past behavior, inner and outer speech, and visual imagery, including
the presentation of possible future outcomes. We have also argued that a key role in
this interpretation can also be played by some self-specifying sensorial
information.

Acknowledgments  Many thanks to Marko Jurjako for reading and commenting on previous ver-
sions of this chapter. We presented and discussed some of the ideas in this chapter at the conference
Contemporary Philosophical Issues: Society, Agency and Knowledge, Rijeka (Croatia),
24-25/05/2016, many thanks to the organizers and participants. The Croatian Science Foundation
(HRZZ) funds our research that is a part of the project: Classification and explanations of antiso-
cial personality disorder and moral and legal responsibility in the context of the Croatian mental
health and care law (CEASCRO), grant n. 8071.
Identification and Self-Knowledge 189

References

Alston, William. 1971. Varieties of privileged access. American Philosophical Quarterly 8:


223–241.
Baars, Bernard. 1988. A cognitive theory of consciousness. Cambridge: Cambridge University
Press.
Buss, Sarah, and Lee Overton, eds. 2002. Contours of agency: Essays on themes from Harry
Frankfurt. Cambridge, MA: The MIT Press.
Carruthers, Peter. 2000. Phenomenal consciousness: A naturalistic theory. Cambridge: Cambridge
University Press.
———. 2011. The opacity of mind: An integrative theory of self-knowledge. Oxford: Oxford
University Press.
Dretske, Fred. 1995. Naturalizing the mind. Cambridge, MA: MIT Press.
Frankfurt, Harry. 2002. Reply to Richard Moran. In Contours of agency essays on themes from
Harry Frankfurt, ed. Sarah Buss and Lee Overton, 218–225. Cambridge, MA: The MIT Press.
Frankfurt, Harry G. 1971. Freedom of the will and the concept of a person. Journal of Philosophy
68: 5–20. Reprinted in Gary Watson ed. Free Will. Oxford: Oxford University Press: 81–95.
———. 1988a. Identification and wholeheartedness. In The importance of what we care about:
Philosophical essays, ed. H.G. Frankfurt, 159–176. Cambridge: Cambridge University Press.
———. 1988b. The importance of what we care about. Cambridge: Cambridge University Press.
Gallagher, Shaun. 2000. Philosophical conceptions of the self: Implications for cognitive science.
Trends in Cognitive Sciences 4 (1): 14–21.
Gerrans, Phillip, and Jeanette Kennett. 2010. Neurosentimentalism and moral agency. Mind 119
(475): 585–613.
Kennett, Jeanette, and Steve Matthews. 2009. Mental time travel, agency, and responsibility. In
Psychiatry as cognitive neuroscience: Philosophical perspectives, ed. Matthew R. Broome and
Lisa Bortolotti, 327–349. Oxford: Oxford University Press.
King, Matt, and Peter Carruthers. 2012. Moral responsibility and consciousness. Journal of Moral
Philosophy 9: 200–228.
Levy, Neil. 2012. A role for consciousness after all. Journal of Moral Philosophy 9: 255–264.
Lycan, William. 1987. Consciousness. Cambridge, MA: MIT Press.
Moran, Richard. 2002. Frankfurt on identification: Ambiguities of activity in mental life. In
Contours of agency essays on themes from Harry Frankfurt, ed. Sarah Buss and Lee Overton,
189–217. Cambridge, MA: The MIT Press.
Tye, Michael. 1995. Ten problems of consciousness: A representational theory of the phenomenal
mind. Cambridge, MA: MIT Press.
———. 2000. Consciousness, color, and content. Cambridge, MA/London: MIT Press.
Velleman, David J. 1992. What happens when someone acts? Mind: 461–481.
Watson, Gary. 1982a. Free agency. In Free will, ed. Gary Watson, 205–220. Oxford: Oxford
University Press.
———. 1982b. Free will. Oxford: Oxford University Press.
———. 2004. Agency and answerability. Oxford: Oxford University Press.
Causal Inference in the Clinical Setting:
Why the Cognitive Science of Folk
Psychology Matters

Andrew Sims

Abstract  In psychodynamic practice, irrational thought and action is explained


within a framework that licenses a range of personal-level psychological attribu-
tions that are not ordinarily present in the quotidian interpretation of others. So, the
thought goes, psychodynamic explanation is best construed as a theoretical exten-
sion of ordinary personal-level psychology, or folk psychology. If this view is cor-
rect, I argue, then it commits the psychodynamic explainer to some fairly weighty
commitments with respect to the sub-personal explanation of folk psychologi-
cal — or mindreading — capacities. For instance, it commits the psychodynamic
explainer to the domain-generality and cognitive-extensibility of these capacities.
Does the present state of affairs in the relevant sciences bear these commitments
out? A discussion of dual-process theories of mindreading seems to show that
although strong conclusions cannot be drawn at present, a complete model at the
sub-personal level has the potential to add more epistemic detail to the standard
philosophical story and even to epistemically vindicate key clinical phenomena
(like counter-transference). On the other hand, it also has the potential to invalidate
the commitments of the psychodynamic explainer, assuming she wishes to appeal to
the “extension of folk psychology” argument that causal inference is warranted in
the clinical setting.

Keywords  Psychoanalysis · Folk psychology · Minimal mindreading · Dual-­


process · Causal inference

In general, analytic philosophy has not much been concerned with psychoanalysis,
especially in comparison to the vast body of work on conceptual issues in cognitive
science. The obvious exception is negative critique, for example in McIntyre (1958),
Popper’s (1962) rather perfunctory remarks, and in Grünbaum’s (1984) substantial

A. Sims (*)
Institut supérieur de philosophy, Université Catholique de Louvain,
Louvain-la-Neuve, Belgium
e-mail: andrew.sims@uclouvain.be

© Springer Nature Switzerland AG 2018 191


P. Pedrini, J. Kirsch (eds.), Third-Person Self-Knowledge, Self-Interpretation,
and Narrative, Contributions To Phenomenology 96,
https://doi.org/10.1007/978-3-319-98646-3_11
192 A. Sims

and important work. But there is nonetheless a positive research tradition in analytic
philosophy which seeks to make clear the structure, warrant, and import of psycho-
analytic explanation. Broadly speaking, that research tradition concerns itself with
three sorts of question.
The first of these is grounded in the philosophical analysis of explanation: what
is the conceptual structure of psychoanalytic explanation? Given that psychoanaly-
sis not only aims to cure but to explain irrational thought and action, how exactly
does it do so? What is the nature of its explananda and central concepts, and what
role do these play in the explanation? For instance, Gardner (1993) has argued that
the psychoanalytic explananda — like symptoms that putatively fulfil unconscious
wishes: a young man suffers obsessional fears of his father tortured because he
hates his father, cf. Freud (1909) — are distinct from ordinary irrational phenomena
like self-deception and wishful thinking, and that the unconscious processes used to
explain these are sub-intentional; they do not require that some symptom be brought
about intentionally to be wish-fulfilling (for example, by some Freudian “censor”).
Pataki (2014) has argued the contrary view; he believes that wish-fulfilments are
irreducibly intentional in nature and involve an unconscious and intentional self-­
gratification on the part of the agent.
The second is an epistemological question: what is the epistemic warrant of psy-
choanalytic explanation? Grünbaum has argued that causal inference in the clinical
setting lacks warrant without corroborating evidence from extra-clinical sources. In
response, Hopkins (1988) presents the by-now widespread view that it is an exten-
sion of the warrant of ordinary mental ascription, and this analysis is revisited in
Lacewing’s (2012a, b) treatment of the issue, where he distinguishes between tiers
of causal inference which have ordinally different levels of epistemic warrant. The
general theory that obsessional traits in adults are caused by events during toilet
training, for example, has a lesser warrant than some particular ascription of an
unconscious desire to an individual on the basis of a verbal slip on a particular occa-
sion. That is because the former is an inference to a large-scale causal aetiology,
while the latter is the more proximal ascription of a single causal fact (these exam-
ples are quite specifically Freudian, but the point generalises).
The third is ontological: what must the mind be like, if psychoanalytic explana-
tion is cogent? This question has to do with what Gardner (1993, p. 175), following
Freud, calls “meta-psychology”: “[…] the conditions of possibility of psychoana-
lytic explanation.” For instance, does psychoanalytic explanation entail that the
mind is divided into distinct systems, or sub-systems? Are these systems best con-
ceptualised in terms of persons, or in terms of the sub-personal “systems” or “mod-
els” discussed in cognitive science (Hopkins 2012)? Does it entail the existence of
unconscious propositional attitudes? What must be true of the mind, in other words,
for such explanations to be cogent?
I aim to establish two theses in what follows. The first is that there is an important
relationship between those epistemic and the ontological questions that is not
already apparent, or at least that has not yet been treated properly in the literature.
This is not the more obvious assumption that the answers to those two questions
would be related in the following way: that the cogency of the ontological question
Causal Inference in the Clinical Setting: Why the Cognitive Science of Folk Psychology… 193

is dependent on a satisfactory settling of the epistemological issues. That is to say,


on this view it only makes sense to wonder about the ontological commitments of
psychoanalytic explainers if that mode of explanation is warranted. Otherwise the
antecedent of the entailment (if psychoanalytic explanation is warranted, the mind
is like this) is false, and it makes no empirical difference. But I will be arguing that
in fact there is an important relationship that goes in the other direction. In particu-
lar, if philosophers of psychoanalysis want to be basing the epistemic warrant of
psychoanalytic explanation on folk psychology, then they are held hostage to
advances in our understanding of folk psychology that come from cognitive science
(more often called “mindreading” in that literature).1
My second thesis follows up on the consequences of the first; it is that contem-
porary work in the cognitive science of mindreading can fill out extant analyses of
psychoanalytic explanation in interesting ways. The work I have in mind is not a
theory but rather a controversy. I make this choice out of necessity: there is not a
settled theory of mindreading, though there are points of consensus. In particular,
theorists of mindreading agree that there are tensions in the empirical findings that
can need to be accounted for either by one-system or two-system accounts. I will be
arguing that these issue in different pictures for the psychoanalysis-as-extension
thesis, and that the epistemic warrant of psychoanalytic explanation is held hostage
to such accounts if that warrant is based on folk psychological warrant. But like-
wise, knowledge of the mechanisms of social cognition has the potential to make a
positive contribution to the epistemology of clinical inference, as well: if, for exam-
ple, there is some modular part of the overall mechanism that responds to particular
patterns of behaviour in fixed and reliable ways, then analysts should be more trust-
ing of their intuitions regarding those types of patterns.
Let me say a quick word at this point about the generality of my claims. I intend
this essay to be of interest beyond the philosophy of psychoanalysis—and it ought
to be, since the problem I identify is quite general to any talk therapy which licenses
novel forms of causal inference on the basis of ordinary social cognition. I think this
description would apply to a broad range of approaches, and possibly even cognitive-­
behavioural therapy (CBT), since CBT requires that the practitioner engage in
causal inference and introduces novel concepts like “cognitive distortions” (Beck
and Alford 2009) that are based upon clinical observation rather than experimental
testing. All such practitioners had better make sure that they are familiar with the
way in which social cognition works, if social cognition is at the basis of their
explanatory practices. Otherwise it might be that the therapeutic success of those
practices, to the extent that there is such success, has nothing to do with the correct-
ness of the theories that they are based on. It may, for instance, lend plausibility to
the much discussed “dodo effect” (Rosenzweig 1936).
All this is to say that the problems here are more general than the focus on psy-
choanalysis may suggest. Presumably, if the extension thesis is right, then the

1
 I will be using the terms “folk psychology,” “mindreading,” “commonsense psychology,” “social
cognition,” and so on interchangeably. These are just different names for the capacity to ascribe
mental states to others in order to explain and predict their behaviour.
194 A. Sims

­ ethods that are used in psychotherapy have continuity with the ordinary capacity
m
for self-knowledge, at least insofar as we endorse the weaker thesis that self-knowl-
edge contains a third-person component. Freud, for instance, is well known for the
self-­analyses that comprise much of his work.2 So far from being a question that is
limited to a specialised clinical practice, the epistemological reach of social cogni-
tion should concern anyone who believes that some of the knowledge we have about
ourselves results from the capacity that we have to understand others. Let me start
by discussing what it means for psychoanalytic explanation to be an extension of
folk psychology.

1  T
 he Epistemic Question and the Theory of Folk–
Psychological Extension

The extension thesis is a by-product of the debates over the scientific status of psy-
choanalysis. Of course, the use of the word “science” in this context is a red-herring;
what is really at stake in that discussion is whether or not psychoanalytic explana-
tion is epistemically warranted. Ought we believe what the psychoanalysts tell us
about mental function, and if so, to what extent? This debate had roots in the behav-
iourist’s aversion to the mental (e.g., Skinner 1956), but it later became a more
exclusively epistemic disagreement over the nature of evidence and justification in
the clinical context. Karl Popper (1962), for example, stated that it is not possible to
falsify conjectures or causal inferences that are made in psychoanalysis because
there are no observable states of affairs that will count as being unambiguously
incompatible with those inferences. This inability to satisfy the principle of falsifi-
cation, claimed Popper, bars psychoanalysis from scientific status; which is to say
that its standards of justification are not high enough for it to count as an empirically-­
warranted theory. Adolf Grünbaum (1984: 104ff) later demonstrated that psycho-
analytic conjectures are indeed falsifiable, but maintained that extra-clinical testing
is required for their justification, as the clinical setting is insufficient as an autono-
mous ground of justification for those conjectures. It is in the philosophical response
to Grünbaum that the folk-psychological view of psychoanalytic explanation finds
its strongest formulation.
Probably the most seminal is that of Hopkins (1988). Hopkins’s basic argument
goes something like this: i) inference to the best explanation (IBE) is ubiquitous in
our ordinary reasoning about the world; ii) more specifically, IBE is present in our
ordinary interpretive and pre-scientific practices with respect to human behaviour;

2
 Freud also claims that self-analysis is not possible in the general case —“True self analysis is
impossible; otherwise there would be no (neurotic) illness.” (Masson 1985, p. 281) Certainly it is
correct to think that self-analysis is more prone to confusion through sorts of resistance and self-
defense. But Freud must be wrong to rule out the possibility of some self-knowledge through
third-person analysis so categorically. Not all authors in this area are so categorical (e.g., Horney
1999[1942]).
Causal Inference in the Clinical Setting: Why the Cognitive Science of Folk Psychology… 195

iii) IBE allows for weak prediction, confirmation, and disconfirmation of current
best explanation; iv) psychoanalysis is an extension of our ordinary interpretive
practices with respect to human behaviour (an extension, because it covers a wider
range of phenomena and also because it employs novel concepts in explaining those
phenomena); so v) psychoanalysis allows for weak prediction, confirmation, and
disconfirmation of current best explanation that it employs, and this warrant is suf-
ficient to independently support inference in the clinical setting.3 Here’s how he puts
it in an earlier paper:
In commonsense psychological practice we already establish causal connections (in par-
ticular concerning the role of motives) interpretively, in ways that are autonomous, cogent,
and prior to [the canons of inductive inference, on the basis of which Grünbaum objects to
the putative warrant of psychoanalytic explanation]. So it seems wrong to hold generally
that cogency in a psychology of motive must satisfy them; indeed, for motives, it is unclear
how such canons could be used, or how inductive methods could replicate commonsense
interpretation. Further, psychoanalytic theory seems an extension of commonsense under-
standing of motives, by interpretive means internal to it. So psychoanalytic theory may also
be cogent, but related to inductive methods no more closely than commonsense psychology
itself. (Hopkins 1988, p. 37)

Let me fill this out a bit. We start with the observation that in ordinary interaction
with others we have access only to their behaviour; in order to explain and predict
this behaviour, we ascribe to them mental states which stand in relations to the
behaviour that are both causal and rationally intelligible. They are rationally intel-
ligible because there is a coincidence of meaning between the mental state and our
description of the behaviour under which it is an action, and this coincidence of
meaning justifies or rationalises the action in a way that a merely causal explanation
would not. For example, we might say that Herr Lorenz is opening the car boot
because he desires a jumper cable and believes there might be a jumper cable in the
boot. This explanation rationalises the action because it makes the action a sensible
thing to do from Herr Lorenz’s point of view. Exclusively causal explanations don’t
do this — for example, the explanation that Herr Lorenz is opening the boot as a
result of mind-control by the CIA does not make it a rationally intelligible action
from Herr Lorenz’s point of view. It just makes the cause of the behaviour
apparent.
The inference can be appropriately described as causal because we take the
desire to be causally relevant to the behaviour under a counterfactual analysis. We
would not have expected Herr Lorenz to be opening the boot in the first place unless
he had that desire for a jumper cable. In other words, his behaviour would have been
different in the absence of the relevant desire. That is why the old view — for exam-
ple, in Wittgenstein (1953) and applied as a critique of psychoanalysis in MacIntyre
(1958) — that rationalising inferences cannot also be causal inferences is mistaken;
we infer causal connections on the basis of meaning whenever we infer the motives
of others on the basis of their behaviour. We can also infer desires more distant from

3
 Note however that this argument is ambiguous between two separate readings, a “partner-in-
crime” and a “substantive” reading. I discuss this in Sect. 3.
196 A. Sims

immediate behaviour by linking them with beliefs. From Herr Lorenz’s behaviour,
for example, we might attribute to him the desire to start his car by way of the belief
that if he uses the jumper cable his car will start. That is to say, the desire to start his
car, and the belief that if he uses the jumper cable his car will start, jointly generate
the further desire for a jumper cable. So folk-psychological inference can have mul-
tiple levels of depth of this kind, as in strategic reasoning.
Now, IBE is a kind of reasoning wherein we take some effect as given and posit
a cause that would best explain that effect. The given effect might be a trail of foot-
prints in the snow (a tired but illustrative example). The “best explanation” that we
infer from this effect might be that someone has walked past quite recently, before
the falling snow has had a chance to hide the footprints. This explanation will be in
competition with others that are perhaps not as likely — for example, that some-
body has faked the footprints or that they have been caused by a freak weather
phenomenon. A number of normative principles underwrite IBE (cf., Lipton 1991,
or Hopkins 2015 for a discussion of IBE in the psychoanalytic context). Importantly,
it needs to be possible for a piece of data to be capable of confirming or disconfirm-
ing the current best hypothesis (a discontinuity or crudeness in the footprints may
count as disconfirmation of the “genuine footprints” hypothesis). To confirm a
hypothesis is to increase our degree of belief in the hypothesis, and to disconfirm it
is to decrease that degree of belief. Expressed in terms of probabilities, this means
that a datum is confirming if the probability of the hypothesis given the datum is
higher than just the probability of that hypothesis alone. Likewise, a datum is dis-
confirming if the probability of the negation of the hypothesis given the datum is
higher than the probability of just the negation of the hypothesis.
On the extension thesis, folk-psychological reasoning conforms to the basic
logic of IBE. Taking as the explanandum the fact that the agent has performed some
action ϕ, we attribute to the agent the desire that p, and the belief that if he ϕ then p,
if this belief-desire pair is the best explanation of that agent’s ϕing. Davidson (1963)
calls this belief-desire pair a primary reason. It is the reason for that agent’s behav-
iour, and it is also the proximate cause of that behaviour. From the data that we
have — that data comprising the agent’s past and present behaviour and our expla-
nations for that behaviour, as well as any other relevant information — we choose
the mental states, dispositions, and processes that best explain the behaviour in
question.
And how does psychoanalytic explanation draw upon and extend this mode of
explanation? Hopkins (1988, pp. 40–44) uses two examples in sequence to illustrate
this, both of them from the Interpretation of Dreams. The first is a passage in which
Freud is trying to establish some initial plausibility for the claim that there is a
causal connection between motivational states and the content of dreams. This pas-
sage demonstrates well the extension of folk-psychological explanation to shed
explanatory light on new phenomena — in this case, dreaming:
It is easy to prove that dreams often reveal themselves without any disguise as fulfilments
of wishes […] For instance, there is a dream that I can produce in myself as often as I
like — experimentally, as it were. If I eat anchovies or olives or any other highly salted food
in the evening, I develop thirst during the night which wakes me up. But my waking is
Causal Inference in the Clinical Setting: Why the Cognitive Science of Folk Psychology… 197

preceded by a dream, and this always has the same content, namely, that I am drinking. I
dream that I am swallowing down water in great gulps, and it has the delicious taste that
nothing can equal but a cool drink when one is parched with thirst. Then I wake up and have
to have a real drink. (Freud 1900, p. 123)

Here is an example of folk-psychological inference applied to a domain which is


commonly thought to be outside its jurisdiction. The dream is shown to be intelli-
gible in terms of the agent’ motivational states (i.e., a desire to quench thirst), and it
can be interpreted accordingly. It is rationalising, because it makes the dream sen-
sible from the agent’s point of view (it pacifies his thirst),4 and it is causal, because
in the absence of the desire to drink the content of the dream would have been dif-
ferent. Similar and more recent examples can be found in the coincidence of content
between the content of dreams in drug-addicted patients and a putative motivational
state to consume the substance of addiction (Colace 2014).
Psychoanalysis does not only apply folk-psychological inference to novel
domains, but also introduces novel concepts in order to give better explanations in
these domains. Hopkins’s demonstration of this aspect of the extension thesis is
based on Freud’s (1900, pp. 106–121) dream of Irma’s injection. This is a dream
that is far more complex and opaque than the dream of drinking, which is relatively
transparent. The dream is about a patient that he was not successful in treating for
psychogenic hysteria, and who went away after some time. He had received a letter
from a colleague (Otto) in which he detected a subtle reproach for his failure, and
the night before the dream began writing up the case in a letter to a senior colleague
(Dr. M.) in order to justify his actions both to himself and to an expert witness. Here
is Freud’s self-report in full:
A large hall — numerous guests, whom we were receiving. — Among them was Irma. I at
once took her on one side, as though to answer her letter and to reproach her for not having
accepted my ‘solution’ yet. I said to her: ‘If you still get pains, it’s really only your fault.’
She replied: ‘If you only knew what pains I’ve got now in my throat and stomach and abdo-
men — it’s choking me’ — I was alarmed and looked at her. She looked pale and puffy. I
thought to myself that after all I must be missing some organic trouble. I took her to the
window and looked down her throat, and she showed signs of recalcitrance, like women
with artificial dentures. I thought to myself that there was really no need for her to do
that. — She then opened her mouth properly and on the right I found a big white patch; at
another place I saw extensive whitish grey scabs upon some remarkable curly structures
which were evidently modelled on the turbinal bones of the nose. — I at once called in Dr.
M., and he repeated the examination and confirmed it…Dr. M. looked quite different from
usual; he was very pale, he walked with a limp and his chin was clean-shaven…My friend
Otto was now standing beside her as well, and my friend Leopold was percussing her
through her bodice and saying: ‘She has a dull area low down on the left.’ He also indicated
that a portion of the skin on the left shoulder was infiltrated. (I noticed this, just as he did,
in spite of her dress.) […] M. said: ‘There’s no doubt it’s an infection, but no matter: dys-
entery will supervene and the toxin will be eliminated.’ […] We were directly aware, too, of
the origin of the infection. Not long before, when she was feeling unwell, my friend Otto

4
 Pacification is a technical notion introduced by Hopkins (1995). A desire is pacified just in case it
is extinguished without its satisfaction conditions being brought about, and this is caused by the
belief that its satisfaction conditions obtain. In the dream of drinking the desire is pacified but it is
not satisfied, because Freud is not really drinking.
198 A. Sims

had given her an injection of a preparation of propyl, propyls … proprionic acid … trimeth-
ylamin (and I saw before me the formula for this printed in heavy type) … Injections of that
sort ought not to be made so thoughtlessly … And probably the syringe had not been clean.
(Freud 1900, p. 107)

This dream, very much unlike the dream of drinking, is not transparent with
respect to putative psychological causes; its content is much more complex. But on
the basis of other facts — amongst these its context in Freud’s writing up of the Irma
case, Otto’s reproach, and his desire to be exonerated5—  Freud is able to draw
meaningful links between some of the elements that appear in his account and
which show the dream to be likewise determined by motivational states. Roughly,
he infers that what the dream is doing is assuaging his guilt in the face of Otto’s
reproach. In other words, the way in which Irma is portrayed as suffering from an
organic illness in the dream means that he is not to be blamed for the lack of
improvement, given that he was treating her for a psychogenic illness. That is to say
that she is represented as suffering from an organic illness so that Freud need not
feel responsible for the failure of treatment. Furthermore, the dream clearly repre-
sents Otto as being the one responsible for Irma’s illness, owing to the unsafe injec-
tion he has given her with an unclean syringe. So it is the very person who has
elicited the guilt in Freud who is made to be the guilty one. In other words, Freud
projects or displaces the guilt he feels about Irma’s treatment onto Otto, so that he
may be rid of it and avenged on Otto.
Projection is a good example of a distinctively psychoanalytic concept intro-
duced in order to better explain the phenomenon of interest. Its introduction in the
theory of dreaming (but not only the theory of dreaming; projection is a concept
which can be applied to most domains to  which folk-psychological reasoning
applies) allows us to clear up a certain opacity in the dream: why is it that Otto is
represented as guilty of making Irma sick? When we see that Freud is “avenging
himself” upon Otto for the perceived slight, then this opacity becomes explanato-
rily transparent, at least in so far as the hypothesis of projection is the best explana-
tion of that dream content.
The psychoanalytic data produced in dreams and in clinically manifest symp-
toms demand explanation, and the psychoanalytic method affords a way in which an
explanation can be given for this data. It does so by ascription of unconscious men-
tal states, dispositions, and processes, both those that already exist in our folk-­
psychological vocabulary and new concepts crafted for the purpose. This is the way
in which psychoanalysis can be seen as an extension of folk psychology, and it is the
way in which the explanatory practice has typically been defended against sceptical
challenges like that of Grünbaum.

5
 It may be objected that such knowledge is not available when the inference is about another per-
son and not oneself. But the point successfully generalises to psychoanalysis in the ordinary case,
where the analyst is not also the analysand, because this kind of contextual knowledge is built up
over the course of the analysis and aids inference as it is acquired.
Causal Inference in the Clinical Setting: Why the Cognitive Science of Folk Psychology… 199

2  The Ontological Commitments of Extension Theorists

It seems clear that there is a knowledge-prioritising relationship between the ques-


tion about epistemic warrant and the question about ontological commitment. This
is clear if we understand the latter question as a question about the necessary condi-
tions of psychoanalytic explanation. When I say there is a knowledge-prioritising
relationship what I mean is that the answer to the epistemic question has an effect
on our interest in the ontological question. In particular, if it turns out that psycho-
analytic explanation is not epistemically warranted (if, for instance, Grünbaum’s
critique is correct), then the ontological commitments of psychoanalytic explainers
are more likely to be false (though they may be true by luck; lack of warrant doesn’t
make them false but it does mean that psychoanalytic explanation is in want of jus-
tification). On the other hand, there is a connection going the other way  —  an
ontology-­prioritising relationship — from the ontological question to the epistemic
question. That is to say, our answer to the epistemic question may be similarly
voided by facts about what there is. I will now explain how this could be the case.
There are two ways in which we can understand the extension thesis — or maybe
better, two ways in which the extension thesis can be made the basis of a reply to
sceptical arguments about epistemic warrant in psychoanalysis. The first of these is
a partner-in-crime strategy6 that ties the epistemic warrant of psychoanalytic expla-
nation to the epistemic warrant of folk psychology in general. This does not commit
the extension theorist empirically, because nothing need be said about psychoanaly-
sis specifically. What is doing the heavy lifting here is rather our antecedent confi-
dence in the warrant of quotidian psychological reasoning, along with the central
and mundane role that it plays in our everyday lives. In his early work on this theme,
Hopkins (e.g., 1988) appears to suggest the availability of this partner-in-crime
approach, recall:
In commonsense psychological practice we already establish causal connections (in par-
ticular concerning the role of motives) interpretively, in ways that are autonomous, cogent,
and prior to such canons [the canons of inductive inference]. So it seems wrong to hold
generally that cogency in a psychology of motive must satisfy them […] (Hopkins 1988,
p. 37)

Here’s how I read the passage: first, we make causal inferences vis-à-vis the
motives of our conspecifics all the time, and in fact we must do so cogently in order
to navigate our social environments. So they’re probably highly reliable, most of the
time. For this reason those kinds of judgments are prior to and autonomous from
science construed inductively — they’re an essential and deep part of our everyday
lives as social beings, and can be epistemically vindicated for this very reason. Now,
the kinds of inferences that are made in analytic practice have epistemic parity with

6
 A partner-in-crime argument links by entailment the rejection of some sort of fact whose exis-
tence is under dispute to the rejection of another sort of fact whose existence is much more credi-
ble. Cuneo (2007), for instance, argues that if we deny facts about moral normativity then we are
committed to denying facts about epistemic normativity.
200 A. Sims

these pre-scientific inferences. What that means is that they are of the same kind,
and have the same epistemic basis. So it’s just as inappropriate to submit then to
critique on the basis of induction as it is inappropriate to submit commonsense psy-
chological practice to a critique on the basis of induction.
In my view, this version of the extension argument is unconvincing on account of
the thoroughness of the psychoanalytic extension that is under discussion. First,
there is a serious difference of kind between some psychoanalytic and folk-­
psychological concepts that would prevent the epistemic parity which is required by
a partner-in-crime argument — and not only this, but there is a more general theo-
retical apparatus in psychoanalysis that seems to be absent from pre-scientific causal
inference of the kind that is being appealed to here. In other words, I think it’s a
simple thing for a critic to reject the parity premise that is implicit here.
First, the difference in the concepts can be seen in the fact that psychoanalytic
explanation overturns a number of key assumptions that seem to be inherent to quo-
tidian mental ascription. Probably the most widely known of these is the idea that
much of our behaviour is driven by mental states of which we are unaware and that
we are in systematic error about the true causes of our actions. This contradicts the
apparent innateness — at least in Westerners — of the quasi-Cartesian notion that
our mental states are immediate and therefore that we are reliable self-interpreters
(cf., Carruthers 2008). This may mark an important difference of kind between
mental ascriptions in psychoanalysis and those of the quotidian kind. Another such
example is the rejection of the unity of the self: on many analyses of the structure of
psychoanalytic explanation, that mode of explaining posits that the self is frag-
mented such that there can be coherent unconscious parts of the mind with their own
beliefs and desires that conflict with our conscious states. These are substantial revi-
sions of the folk view.
Second, and relatedly, such tensions with quotidian assumptions (e.g., the posit
of unconscious mental states) necessitate a larger theoretical framework which
explains the falsity of those assumptions. Take the example of unconscious states.
In order to substantiate his departures from certain folk-psychological principles,
Freud finds himself compelled to produce a metapsychology which explains how it
is that unconscious states exist and produce effects on behaviour. And metapsychol-
ogy continues to be a key component of psychoanalytic theorising, even if there are
arguments that it is not a necessary one.7 So it seems as though the partner-in-crime
argument, though it probably does not commit us empirically any more than a com-
mitment to our ordinary practices of practical reasoning does, will not be successful
in defending the epistemic warrant of causal inference in the clinical setting.
But in Hopkins’s later work (e.g., Hopkins 2012) and in the work of Lacewing
(2012a; b; in press), we can see a more substantive version of the argument put

7
 Famously, Habermas (1986 [1968], ch. 11) argues that metapsychology is a relic of Freud’s sci-
entific education and is irrelevant to the central activity of clinical interpretation. But this view and
other views of this kind assume a strict difference of kind between rational and causal explanation,
which makes psychoanalytic “causes” (and indeed, practical reason itself) epiphenomenal.
Causal Inference in the Clinical Setting: Why the Cognitive Science of Folk Psychology… 201

forward. That version imputes a particular kind of warrant to inferences of the folk-­
psychological kind which is then extended to psychoanalytic inference:
We arrive at our attribution of a motive by means of inference to the best explanation. We
form and revise our judgments in light of the evidence available to us, the scope of the
explanation (how much behavior it can explain), whether they cohere with other motives
which we attribute, the simplicity of the explanation, and how plausible it is, given our
background knowledge […] Arguably, psychoanalysis works in the same way. (Lacewing
in press, p. 8)

This substantive reading is more plausible than the former, since it does not
require epistemic parity between folk psychology and psychoanalysis. That is
because the warrant does not rest on our antecedent confidence in folk psychology,
but rather in the putative facts that folk psychology works by IBE and that psycho-
analytic inference is based on folk psychology – psychoanalysis is based on folk-­
psychological IBE, just extended to new phenomena. So evaluations of the warrant
depend on how far IBE can be stretched, and not on our pre-existing folk practices.
That is why on the substantive reading of this argument we can derive tiers of psy-
choanalytic inference in which we may have more or less confidence.
That is to say, on this variety of the argument Lacewing (2012a; b) can distin-
guish between psychoanalytic inferences that have different levels of warrant, on
the basis that the hypothesis space within which they exist is larger. There is more
competition where there the scope of the hypothesis is wider — for example, when
they posit developmental aetiologies that may enter competition with hypotheses in
developmental psychology. More simple attributions within a truncated scope – for
instance, that of a single clinical session – occupy a smaller hypothesis space and
are in less competition with competing explanations. Just as a single ascription of a
belief to another individual does not under normal circumstances come into compe-
tition with some theory of belief formation in general, it would seem that the ascrip-
tion of a projection to an individual (for instance, that his paranoia regarding
same-sex attracted males is due to his own latent same-sex attraction) should not
come into competition with theories of human development or social cognition in
general.
However, taking the substantive reading commits the theorist empirically. In
particular, it commits the theorist in the sense that it places constraints on what a
complete theory of folk psychology could be like. Lacewing (in press) himself
recognises this, and claims support from some of the relevant science: “That this is
how we proceed [in folk-psychological practice] has received considerable support
from empirical studies in social psychology […]” But it’s not clear to me that the
support that he cites (e.g., Gawronski 2004) is sufficient to establish that folk psy-
chology is literally inference to the best explanation. These studies are just behav-
ioural and therefore lack deeper information about the mechanisms that cause those
behaviours. All they can establish is that there are systematic errors in folk-psycho-
logical explanation that look like errors of inference when evaluated rationally. It
may be that these biases appear because they are the product of a sub-personal
mechanism which is not literally theory-like but which is adapted to selection pres-
202 A. Sims

sures such that it is disposed to produce behaviour that seems prone to theoretical
error.
It seems to me that if we want to consider psychoanalytic inference as extended
folk psychology, then we are committed to the following empirical hypotheses
regarding its sub-personal mechanisms: 1) the domain-generality of folk-­
psychological inference; and 2) the cognitive extensibility of folk psychology. Both
of these have some intuitive appeal, but it’s not clear that either of them are obvi-
ously true.

2.1  Domain-Generality

Domain-generality is a concept in cognitive psychology which finds great utility in


the discussions about modularity (Fodor 1983). A cognitive module is a sub-­
personal mechanism which is the explanation (or part of an explanation) for some
personal-level capacity — for example, visual perception. Modular mechanisms are
distinguished from domain-general capacities in a number of ways. For the pur-
poses of this discussion, the most important distinguishing feature is that of infor-
mational encapsulation. A mechanism or sub-mechanism is informationally
encapsulated if it performs its function in a way that is highly stereotyped and which
is impenetrable to information which is not the ordinary input of that mechanism.
So a module takes some highly specific input and performs a function on it in some
way that is informationally encapsulated, producing an output for consumption by
either other modules or domain-general processes. Processes that work in this way
typically do so in a way that is fast and automatic. It is the informational encapsula-
tion of the process that is the cause of this speed and automaticity.
A putative example of this in action can be seen in the early stages of visual per-
ception. It is thought that the processes at these early stages are modular. They
exclusively take a highly specific or “proprietary” input (patterns of retinal stimula-
tion), perform a specific function on that input (produce a simple two-dimensional
image out of that stimulation), and output that information to specific consuming
systems (upstream stages of visual perception which take place in the visual and
association cortices) (Marr 1982). Fodor (1984) has argued that it is on account of
the informational encapsulation of visual perception (or more plausibly, some part
of it) that we fail to correct visual illusions on the basis of a justified true belief that
what we are seeing are visual illusions. After measuring the lines in the Müller-Lyer
illusion (Fig. 1) and seeing that they are of the same length, we nonetheless continue
to see one as being longer than the other.
Domain-general capacities are not informationally encapsulated in this way. In
IBE, for example, we need to have access to many different pieces of information of
different kinds in order for that sort of inference to be effective. We must be aware
of multiple competing hypotheses, and we must be aware of the data that will allow
us to evaluate these hypotheses. It is a domain-general capacity because it is not
encapsulated in a way that would restrict its inputs and outputs. But if folk psychol-
Causal Inference in the Clinical Setting: Why the Cognitive Science of Folk Psychology… 203

Fig. 1 Müller-Lyer
Illusion

ogy is not a domain-general process, or if it includes significant automatic or


informationally-­ encapsulated components, then this commitment to domain-­
generality may be a liability. There are at least two reasons why this would be so.
The first has to do with the importance of transparency to causal inference. When
we reason from an effect to the most probable cause, it is important that each part of
the inference is transparent to us, because it is only in this way that we can evaluate
the validity of these steps. But if causal inference is not domain-general — if it is
informationally encapsulated — then it cannot be transparent in this way. So if folk
psychology is not entirely domain-general, then it may not proceed by IBE in the
explicit sense that — for instance — Lacewing (in press) is appealing to in his treat-
ment of psychoanalytic explanation, and the extension thesis may be undermined.
The second reason has to do with the link between informational encapsulation
and bias. Informationally-encapsulated processes perform a function on their inputs
in a way that is not necessarily optimal according to some normative standard.
Rather, it may be that there are deep and systematic biases that are present in those
processes. Such biases can be conceived of on analogy with visual illusion. The way
in which early stages of visual processing perform their function on patterns of reti-
nal stimulation is such that they produce systematic misperceptions in the way we
perceive objects; the Müller-Lyer illusion is one example of this. A bit closer to
home, there are fairly robust results in cognitive psychology that show that statisti-
cal reasoning is prone to widespread and systematic bias (Tversky and Kahneman
1974). Such bias can be corrected in domain-general processes, because they are
accessible to normative standards of good reasoning. We can apply the norma-
tive standards in order to correct for bias. But such bias is more likely to occur in
informationally-encapsulated processes, and would not be apt for correction in
those processes (for example, we cannot correct the way we see the Müller-Lyer
lines). So if folk psychology is not domain-general, then psychoanalytic inference
is open to sceptical arguments that leverage this vulnerability to bias.

2.2  Cognitive Extensibility

Cognitive extensibility is a notion that I am introducing in this paper; it is not already


widespread. What I have in mind is the way in which psychoanalytic explanation
supplements the pre-psychoanalytic folk-psychological framework by formulating
new concepts, such as PROJECTION and REPRESSION. A theoretical framework
204 A. Sims

that can be extended in this way is cognitively extensible. Extension theorists are
committed to the idea that folk psychology is cognitively extensible. That is because
psychoanalytic inference utilises many novel concepts, and could not adequately
make intelligible its explananda without those concepts. Likewise, psychoanalytic
explanation implies significant revisions in ordinary folk psychology, as discussed
above. It is an open question whether folk psychology is cognitively extensible, or
revisable. We should be less likely to think that it is so under one or more of the fol-
lowing conditions.
First, we may think it unlikely that folk psychology is not extensible if it turns
out that it is not theory-like. Much of the early work on folk psychology in the phi-
losophy of action — work that significantly informs the theory of extension — was
premised on the idea that it was indeed a sort of theory. Davidson (1973), for exam-
ple, implicitly characterises folk psychology in its role as a tool for the explanation
and prediction of behaviour, and Lewis (1972, p.  257) quite explicitly outlines a
way in which we can construe mental concepts as theoretical terms: “[…] names of
mental states are like functionally defined theoretical terms […]” If folk psychology
is theory-like, then it seems true that it must be extensible, since theories are exten-
sible: we can add new theoretical terms to them in order to explain and predict
observable phenomena. Similarly, we can revise them on the basis of new discover-
ies. But if folk psychology is not theory-like, then it is up to the extension theorist
to explain how it is extensible, if not in the way in which ordinary theories are.
Secondly, and additionally, we may think cognitive extensibility unlikely if folk
psychology is somehow innate or fixed in its development. For example, it may be
that the capacity to interpret the mental states of others on the basis of their behav-
iour develops during infancy and is relatively inflexible after its development. That
would mean that the concepts (if they are concepts at all) which we use to under-
stand the mind of the other in interpersonal contexts are fixed, and so not apt for
extension or revision through theoretical development.
There are results in the cognitive neuroscience of mindreading that suggest that
folk psychology is in fact at least partly composed in such a way that it is not
entirely domain-general nor cognitively extensible. Let’s survey these.

3  C
 ontemporary Research on the Sub-Personal Mechanisms
of Mindreading

I have been arguing that there is an ontology-prioritising relationship between the


epistemic and ontological questions that are two of the central orienting questions
in the philosophy of psychoanalysis. In particular, I have been arguing that constru-
ing psychoanalytic explanation as an extension of folk psychology commits one to
the notions that folk psychology is domain-general and that it is cognitively exten-
sible. In this section of the paper I wish to make a start on thinking about these
commitments in the context of contemporary research on mindreading from other
Causal Inference in the Clinical Setting: Why the Cognitive Science of Folk Psychology… 205

disciplines. The discipline I have in mind is cognitive science. Cognitive science


aims to explain psychological effects by decomposing some capacity into simpler
functions and positing sub-mechanisms which could perform these functions (e.g.,
Cummins 2000, see also Craver 2007 for a view on how this works in cognitive
neuroscience). So a complete theory of folk psychology in cognitive science will
show how the capacity of understanding the minds of others can be explained in
terms of sub-capacities, and ultimately how these sub-capacities can be explained in
terms of mechanisms that are biologically plausible.8
There are in fact a number of competing sub-personal models of mindreading
that are extant in the literature.9 So it’s not possible to say anything about the com-
mitments of the extension-theorist that is not defeasible on the basis of subsequent
findings. But we may be able to show something interesting in the way in which
empirical findings in this area have consequences for the theory of folk-­psychological
extension, and to do so on the basis of an “as-if” model of mindreading. What I
mean by this is that we treat a model as if it is the correct one. That is what I hope
to do here. The reason I think that this is a worthwhile activity is that it will be able
to demonstrate how work in this area will be able to make a positive contribution to
work in the epistemology of psychoanalysis, and not just produce sceptical con-
cerns. For example, it may be able to demonstrate that some sorts of inference are
more reliable than others by virtue of some parts of the total mechanism being
domain-general, and some not — or it may be able to show that some sorts of con-
cepts are more plausibly introduced in different contexts in virtue of differences in
the cognitive extensibility of different components of the total mindreading system.
In this section I intend to give a proof of concept of these possibilities by assuming
the truth of the most widely-accepted model of mindreading and showing how it
entails such consequences in the epistemic context.
Now, there is a theoretical puzzle which is central to contemporary debates over
the nature of mindreading. This puzzle is how to accommodate two sets of empirical
results which seem to be in tension. On the one hand, there are results which suggest
that mindreading is fast, automatic, inaccessible to domain-general reasoning, and
that it develops very early in infancy. One the other hand, there are results which
suggest that it is slow, deliberative, and develops only later in childhood. The results
which suggest the former come from both animal studies that demonstrate complex
social behaviours as well as studies with very young infants. Infants aged 15 months
appeared to understand that others could have false beliefs — a central capacity of
theory of mind — when this capacity was tested on the basis of a non-verbal task
(Onishi and Baillargeon 2005). These results contradict evidence that infants in fact
do not understand that others can have false beliefs. This evidence is that children

8
 There is room for the reader to demur on this point; explanation in cognitive science is a busy area
of research. But it seems fair to say that the dominant framework for making sense of such expla-
nation is the “systems” approach that I’ve just briefly described. Craver (2007) is essential reading
in this area if one is interested in the details.
9
 For example, the one-system models of Heyes (2014) and Carruthers (2016), and the more outré
alternatives discussed in Hutto et al. (2011).
206 A. Sims

fail verbal tasks to test this capacity when they are as old as 3 or 4 years — for
example, they fail in correctly answering questions about the beliefs of characters in
a vignette they have just witnessed. That suggests that the capacity for mindreading
develops only after this time (Wellman et al. 2001).
The evidence for automaticity comes from studies which show that the existence
of cues that trigger mindreading capacities interferes with unrelated tasks. For
example, Kovács et al. (2010) have found that adult participants are slower to iden-
tify the location of an object when they are aware of a spectator who has a false
belief about the location. The irrelevance of that fact suggests that automatic activity
of the mindreading capacity is being triggered in a stereotyped manner and thus
interfering with domain-general reasoning. On the other hand, there is evidence to
suggest that mindreading can consume significant domain-general resources, such
as executive function, attention, and working memory (cf., Butterfill and Apperley
2013). And on a more anecdotal level, we all have the experience of reasoning
explicitly about what is going on in the minds of others.
It’s in order to accommodate these contradictory results that Apperley and
Butterfill (2009) have proposed a two-system theory of mindreading. On their view,
these results can be explained in terms of two systems – one that is present from very
early on in infancy, and which is conceptually very simple; and one that develops
later, and which is conceptually more sophisticated. The simpler — or “minimal”
(Butterfill and Apperly 2013) — mindreading capacity does not include fully-fledged
mental concepts with semantic contents. Instead, it contains simpler mental concepts
that are merely teleological. That is to say that they allow that those states have cer-
tain functions but not any intentionality proper — they allow that agents are directed
towards objects in certain ways but not that they represent those objects.
The idea behind minimal theory of mind is to reconstruct a mechanism that could
explain success on false belief tasks without appealing to representational content.
This is important because the presence of representational content may imply lin-
guistic capacities, which infants lack. Infants may represent goals and bodily move-
ments (in a teleological manner), but they cannot represent beliefs or actions (which
are semantically loaded). That is because to do so requires sophisticated conceptual
abilities, which seems unlikely in the absence of some linguistic ability. Therefore,
Butterfill and Apperly (2013, p. 9) say that they:
[…] aim to provide the core elements of a computational theory […] where our computa-
tional theory […] unlike the standard full-blown theory of mind which hinges on beliefs,
desires and other propositional attitudes, is one that could be realised in a cognitively effi-
cient manner without requiring conceptual sophistication.

The principles which Butterfill and Apperly (2013) posit in the reconstruction of
this mechanism are fourfold.
First, they re-introduce in this context the notion of goal-directed action. The
notion they have in mind is a rather weak one: it is one specified in a way that lacks
all intentional content. Rather, it is merely teleological. A goal-directed action in the
relevant sense is a set of events of bodily movements which occur in order to fulfil
some particular function. This is something that can be grasped without grasping
Causal Inference in the Clinical Setting: Why the Cognitive Science of Folk Psychology… 207

the notion of “full-blown” mind. It is just the concept of a set of movements that are
directed towards some end, akin to Aristotelian “formal causation.”
Secondly, they introduce the two concepts of field and encountering. A field is a
relation between an agent (understood in a deflationary way, as an object which
performs goal directed actions) and a set of objects. The objects that are picked out
for the agent by the field are determined by proximity, lighting conditions, posture,
and whether there are opaque barriers between the agent and the object. When an
object is in the field of an agent, the agent is said to have encountered that object.
That leads to the second principle: “[…] if an outcome involves a particular object
and the agent has not encountered that object, then that outcome cannot be a goal of
her actions.” (Butterfill and Apperly 2013, p. 11)
In their specification of the third principle they introduce the notion of registra-
tion. This is another relational notion which holds between an agent, an object, and
a location. An agent “[…] registers an object at a location if and only if she most
recently encountered it at that location.” (Butterfill and Apperly 2013, p.  13)
Registration has conditions of propriety or correctness; there can be inappropriate
registrations. That is because an agent may have registered an object prior to the
object being moved to a different location. In this case, the agent registers the object
at a specific location where it is not. Registration that is appropriate (or correct) is a
necessary condition for a successful goal-directed action concerning some object.
Without correct registration the action would fail — for instance, the taking of a
food item cannot occur unless an agent’s registration of the item in a location is
correct.
The fourth principle adds more detail to the notion of registration, and the role
that it plays in minimal theory of mind. While the third principle outlines registra-
tion as a success condition for goal-directed action, the fourth states that a registra-
tion can be a causal factor in the behaviour of an agent. So registrations can cause a
goal-directed action, even if the incorrectness of some registration that is causally
efficacious means that the goal-directed action fails. For example, an incorrect reg-
istration may cause an agent to collect a food item from a location where it is no
longer situated, leading to the failure of that action.
These four purely teleological principles comprise the basis of a potential mecha-
nism by which young infants and some animals (for example, crows and chimpan-
zees) could successfully pass false-belief tests without having a concept of belief (or
any other propositional attitude). The mechanism allows the tracking of belief, but it
does not do so by means of representing beliefs. Instead it does so by tracking reg-
istrations. Registrations track beliefs, but the concept of registration is solely teleo-
logical (it is without intentional content) and so it is distinct from the concept of
belief. That means that registrations — on account of their mere teleology — cannot
track more sophisticated beliefs having to do with identity and quantification. For
example, mere tracking of registrations will not be able to predict the behaviour of
an agent who has a false belief owing to the failure to recognise an identity between
two objects (e.g., that Harry is Sally). That is because understanding false beliefs of
this kind requires tracking them as propositional attitudes with intensional contexts.
Such a process can explain the apparent automaticity and very early appearance of
208 A. Sims

mind-reading by positing a mindreading module that produces correct answers on


false-belief tests by way of this process.
Sophisticated mindreading or folk psychology  —  the explanatory apparatus
described by Davidson (1980 [1970]), for instance — is presumably built on top of
this simpler mechanism and others like it, though Butterfill and Apperly are coy
about what they take the relationship between the two sorts of system to be. One
plausible suggestion is that the first system provides automatic inputs to domain-­
general reasoning, such that the full-blown theory of mind that can be observed in
older children is composed of the outputs from the minimal system and a domain-­
general reasoning about these outputs that may potentially proceed according to the
norms of IBE, or at least are apt for correction on the basis of those norms.
It is at present an open question whether the two-system model of mindreading
will turn out to be correct; there is simply not enough data to confirm this view, so
it is currently an “as if” model. It is able to explain how the task is performed, but it
is just one amongst other plausible models. But that theory, and any competitor, will
need to explain how the implicit and automatic performances of the capacity are
related to those that are explicit and deliberate. That is the puzzle which drives cur-
rent theorisation in this area.

4  S
 ome Potential Consequences for the Theory of Folk-­
Psychological Extension

In the end, it doesn’t seem to me that we can say anything too certain at this point
with respect to what this model means for epistemic warrant. Our knowledge
regarding the sub-personal mechanisms of folk psychology is just too incomplete to
make inferences about the extension thesis that will be robust in the face of future
findings. But it nonetheless seems clear that there will be some sort of ontology-­
prioritising relationship between the two, and here’s what I think that we can say at
present. Let me start out by reviewing what is at stake in this discussion.
The extension thesis aims to demonstrate how it is that causal inference can be
epistemically warranted in the context of clinical treatment. In other words, it aims
to show that inferences like the following are epistemically warranted in that
context:
The patient’s obsessive fears that his father are being tortured, and the compulsive and self-­
punitive actions that are performed in order to ward off these obsessive fears, are caused by
his ambivalence towards his father – the torture fantasy is a wish-fulfilment and his compul-
sions are produced by his own guilt and anxiety in the face of these wishes.

To this end, it includes the following claims: i) that causal inference in the clini-
cal setting is an extension of folk psychological practices; ii) that folk psychology
proceeds by IBE; iii) that causal inference in the clinical setting proceeds by IBE;
iv) and so causal inference in the clinical setting is justified insofar as one can justify
causal inferences by IBE. These claims make demands on folk psychology. Those
Causal Inference in the Clinical Setting: Why the Cognitive Science of Folk Psychology… 209

demands are i) that folk psychology is domain-general — that it is not information-


ally encapsulated; and ii) that folk psychology is cognitively extensible— that it is
apt for revision and augmentation with new concepts.
And so it does, in fact, make a difference to the extension thesis how our folk-­
psychological capacities are sub-personally composed. That’s because facts about
these sub-personal underpinnings have the potential to falsify the extension-­
theorist’s commitments vis-à-vis folk psychology. Take for example Carruthers’s
(2013, 2016) view on which mindreading is a single system that becomes progres-
sively conceptually elaborated over the course of development. For Carruthers, the
apparently domain-general capacity of folk psychology is in fact served by a coali-
tion of sub-personal modules, and the development of mindreading is in fact the
development of stronger connections between these modules and capacities such as
working memory and other systems. If this is the case, then it’s not so clear how
theoretical extension of folk psychology would be possible. It is part of the theory
of modularity that modules operate on their inputs in a fixed and proprietary man-
ner: they do not change over time through learning. If that’s the case, then it seems
that folk psychology is not extensible. We are stuck with what is given to us in
development.
However — and this is crucial — this contribution is not limited to potential fal-
sification. In particular, I wish to suggest two potential positive contributions here:
one regarding the theory of counter-transference, and the other regarding the notion
of theory-neutral observation in the clinical setting. Both of these contributions
would depend on the theory of Apperly and Butterfill (2009) being more or less
accurate, with a picture of the relationship between the two systems where outputs
from the minimal mindreading system are given to domain-general and more com-
plex psychological reasoning. If a picture like this turns out to be correct, then it
appears that we can say the following about causal inference in the clinical setting.
First, and apropos Fodor’s (1984) remarks about modularity, if the mechanisms
that subserve mindreading are partially modular, then the outputs that are given to
domain-general reasoning are likely to be theory-neutral. What that means is that
they provide a source of data that are immune to interpretation on the basis of prior
belief. They are so immune because they are informationally encapsulated; they
perform a fixed function on proprietary inputs without being accessible to other
modules or to domain-general reasoning. An illustrative example of this is Fodor’s
own example of the Müller-Lyer lines. The lines appear to be of different lengths
even after we acquire knowledge that they are the same length, perhaps after care-
fully measuring them. Our visual experience of them is inaccessible to domain-­
general reasoning.
This is an important point because it provides one consideration against the criti-
cism that psychoanalysis generates its own (contaminated) data, data which say
more about the theoretical background of the analyst than they do about any real
phenomenon in the clinical setting. Look at what Macmillan (1991, p. 576) says
about this: “Everything is consistent with the method of free association creating its
own data and nothing is consistent with those data being obtained by means of an
objective method and then being interpreted in a partial or biased way.” Macmillan
210 A. Sims

puts this a bit too strongly but the point is clear. In any case, if psychoanalytic expla-
nation is an extension of folk-psychological explanation, and if the data of folk-­
psychological explanation are processed by means of a partly modular process, then
it would appear that knowledge of the sub-personal mechanisms of folk psychology
provides one sort of answer to this concern: it makes sense to say that there is a
theory-neutral basis to the observations that are made in the clinical context. Even
if there are biases in this context, any such bias will remain stable across practitio-
ners and will not therefore be contaminated by the theoretical background or educa-
tion of the clinician. Furthermore, since any such bias is likely to be a product of
natural selection rather than learning, then it’s not unreasonable to suppose that
those biases are likely to be heuristics that produce reliable judgments in most
situations.
Second, and relatedly, I suggest that this provides a consideration in aid of resolv-
ing a controversy about the trustworthiness of the so-called “countertransference”
phenomenon in clinical practice. The notion first appears in Freud (1910, pp. 144–
145), where he mentions it as an obstacle to successful treatment:
We have become aware of the ‘counter-transference’, which arises in him as a result of the
patient’s influence on his unconscious feelings, and we are almost inclined to insist that he
shall recognize this counter-transference in himself and overcome it. Now that a consider-
able number of people are practising psycho-analysis and exchanging their observations
with one another, we have noticed that no psycho-analyst goes further than his own com-
plexes and internal resistances permit; and we consequently require that he shall begin his
activity with a self-analysis and continually carry it deeper while he is making his observa-
tions on his patients. Anyone who fails to produce results in a self-analysis of this kind may
at once give up any idea of being able to treat patients by analysis.

On this view the counter-transference, defined as the entirety of the analyst’s


feelings towards the analysand, should be understood in terms of the idiosyncratic
history of the analyst and put aside as far as possible in the interpretation of the
patient.10 However by the mid-twentieth century there is a disagreement about this;
some analysts (particularly within the object-relations school) are claiming that
counter-transference can be a reliable source of knowledge about the mind of the
patient. Heimann (1950), for instance, characterises counter-transference as a useful
instrument with which to complement explicit inference:
[…] often the emotions roused in [the analyst] are much nearer to the heart of the manner
than his reasoning, or, to put it in other words, his unconscious perception of the patient’s
unconscious is more acute and in advance of his conscious conception of the situation.
(Heimann 1950, p. 82)

Practitioners still disagree about this. My claim now is that one or another sub-­
personal model of folk psychology would provide a consideration relevant to this
epistemic controversy. In particular, if the Apperly-Butterfill model is correct, then
we have a reason to think that counter-transference may be a reliable source of

 Some analysts prefer to define counter-transference as just those feelings that the analyst experi-
10

ences in relation to the transferential love or hatred of the patient. See the entry in Laplanche and
Pontalis (1973).
Causal Inference in the Clinical Setting: Why the Cognitive Science of Folk Psychology… 211

knowledge. That is because we may conceive of the counter-transferential feelings


towards the patient as being given to domain-general reasoning through modular
processes. If they are given as outputs from modular processes, then they must be
somewhat fixed across individuals; they always take the same proprietary inputs and
perform the same function on those inputs. For example, it may be that feelings of
counter-transferential hatred (Winnicott 1949) are produced by modular process
that reliably respond to real patterns (cf., Dennett 1991) in the patient’s behaviour;
patterns which are knowledge-conducive in the clinical context.
We should be careful at this point, however. First, for this to establish the epis-
temic reliability of counter-transference in a particular emotional context (like
hatred), more would have to be known about how the relevant emotional sub-­
systems interact with the outputs of the modular components of mindreading (cf.,
Panksepp and Biven 2012). Furthermore, the sub-personal model would comprise
just one consideration for or against the epistemic reliability of counter-transference
in our overall reflective equilibrium about such issues. It wouldn’t resemble any
kind of conclusive proof. But here’s my point: it’s nonetheless a relevant consider-
ation. There is an ontology-prioritising relationship between the question of reli-
ability and questions about deeper mental structure.
In closing, I’d like to consider two objections to the general notion that the sorts
of inferences I’ve just made regarding the epistemic basis of psychoanalysis are
sound. The first objection is that I have overstated the importance of folk psychol-
ogy to psychoanalytic inference. Perhaps, the critic will say, it is wrong to say that
causal inference in the clinical setting is in fact an extension of folk psychology.
Perhaps it is rather more appropriate to characterise it in terms of IBE alone. On this
picture, although extension-theorists often refer to folk psychology in order to moti-
vate their view, it is in fact IBE that is doing the heavy epistemic lifting. That means
that it is IBE that we need to focus on, and not folk psychology. If that is the case,
then an examination of the sub-personal mechanisms of folk psychology will be
missing the point entirely; it is IBE, not folk psychology, that produces causal infer-
ence in the clinical setting.
It strikes me that the appropriate reply here is twofold. Firstly, we may want to
point out that divorcing psychoanalytic inference from folk psychology comes at a
cost. The cost is that we lose both some trueness to the phenomena and some inde-
pendently attractive features of a folk-psychologically based description of clinical
practice. The phenomena which aren’t accounted for by a pure-IBE account are
situations in which the clinician relies on her intuitions regarding the states of the
other person. Analysts often describe insights into the patient as something that they
receive quite passively, not as something that they mechanically infer by comparing
various hypotheses against the data. Look at what Freud says in his treatment of the
patient who obsesses about the torture of his father:
At all the more important moments while he was telling his story his face took on a very
strange, composite expression. I could only interpret it as one of horror at pleasure of his
own of which he himself was unaware. (Freud 1909, pp. 166–167)
212 A. Sims

If a pure-IBE (not folk-psychological) picture of psychoanalytic inference is


true, then it’s hard to make sense of such episodes. The independently attractive
features of a folk-psychologically based description are the fact that it means a con-
tinuity between psychoanalytic theory and our ordinary ways of understanding oth-
ers (this is probably why it has found great utility in literary criticism, for instance).
Secondly, if it is the case that the critic is suggesting that IBE is the normative
standard against which clinical inference should be measured, and therefore that it
does not rely upon the mechanisms of folk psychology, then we may wish to ask just
how widespread the practice of following those norms actually is. If it is the case, as
I think it is, that explicit conformity to IBE is not widespread, then for the time
being our knowledge of the sub-personal mechanisms of folk psychology will none-
theless allow us to make inferences about the epistemic basis of psychoanalytic
theorising.
The second objection is that the foregoing argument ends up overreaching – it
undermines far too much of our quotidian psychological practices. That is to say, if
it is the case that the cognitive science of folk psychology has the potential to under-
mine the notion of folk psychology as domain-general and cognitively extensible,
then this does not just undermine an extension view of causal inference in psycho-
analysis, but also undermines the phenomenology of practical reason itself. It seems
to us, that is, when we reason about the motives of other people, that we are doing
so in a way that allows for us to posit new concepts (maybe concepts for emotions
that have a socially constructed component (Russell 1991), like nostalgia or
schadenfreude), and in a way that is domain-general. If our sub-personal models of
folk psychology contradict the lived experience of folk-psychological practice, one
might say, then so much the worse for those sub-personal models.
To this objection one can give two replies. The first is that this is in fact a live
issue in the philosophy of cognitive science. That is to say that there is an extant
controversy over how robust we ought to take our pre-theoretical experiences to be
in the face of evidence about the mechanisms which subserve our mental lives.
Although out-and-out eliminativism about folk psychology is not so common
(though it exists, notably in Churchland (1981) and Stich (1983)), the claim that
some of our phenomenology may be shown to be systematically mistaken on the
basis of scientific findings is a position that is not obviously absurd (e.g., Carruthers
(2008) on the putative innateness of Cartesian epistemology). So being committed
to a view like this does not itself make the argument of this essay implausible.
Secondly, it’s not even clear that what I am suggesting goes as far as partial elimi-
nativism. My comments on counter-transference, for example, should make it clear
that work on the sub-personal mechanisms of folk psychology in fact has the poten-
tial to vindicate the phenomena of clinical experience.
One last word. Granting that I may seem to be highly critical of extension-­
theorists in this essay, in fact I see my line of thinking to be a continuation of their
own. As Lacewing (2012b, p.  223) himself states: “[…] my argument regarding
causal inferences cannot be taken to support any form of isolationism regarding
psychoanalytic theory.” Just so: my claim here is that this denial of any sort of
autonomy regarding psychoanalytic theory needs to be extended to the process of
Causal Inference in the Clinical Setting: Why the Cognitive Science of Folk Psychology… 213

psychoanalytic theorising itself. Certainly in the past many analysts have not hesi-
tated to do so. Psychoanalytic theorising about transference and counter-­transference
is one salient example of a way in which theory has informed method in this way
(e.g., Winnicott 1949; Heimann 1950). I have been trying to show that that a consis-
tent naturalism in the philosophy of psychoanalysis would require that the epistemic
basis of causal inference in the clinical setting is submitted to the same interdisci-
plinary openness, and on the basis of sub-personal theories of our folk-­psychological
capacities.

References

Apperley, I.A., and S.A. Butterfill. 2009. Do humans have two systems to track beliefs and belief-­
like states? Psychological Review 116: 953–970.
Beck, A.T., and B.A.  Alford. 2009. Depression: Causes and treatment. 2nd ed. Philadelphia:
University of Pennsylvania Press.
Butterfill, S.A., and I.A.  Apperley. 2013. How to construct a minimal theory of mind. Mind &
Language 28: 606–637.
Carruthers, P. 2008. Cartesian epistemology: Is the theory of the self-transparent mind innate.
Journal of Consciousness Studies 15: 28–53.
———. 2013. Mindreading in infancy. Mind & Language 28: 141–172.
———. 2016. Two systems for mindreading? Review of Philosophy and Psychology 7: 141–162.
Churchland, P.M. 1981. Eliminative materialism and the propositional attitudes. Journal of
Philosophy 78: 67–90.
Colace, C. 2014. Drug dreams: Clinical and research implications of dreams about drugs in drug-­
addicted patients. London: Karnac.
Craver, C.F. 2007. Explaining the brain: Mechanisms and the mosaic unity of neuroscience.
Oxford: Clarendon Press.
Cummins, R. 2000. “How does it work?” versus “what are the laws?”: Two conceptions of psy-
chological explanation. In Explanation and cognition, ed. F. Keil and R.A. Wilson, 117–144.
Cambridge: MIT Press.
Cuneo, T. 2007. The normative web: An argument for moral realism. Oxford: Oxford University
Press.
Davidson, D. 1963. Actions, reasons, and causes. Journal of Philosophy 60 (23): 685–700.
Davidson, D. 1973. Radical interpretation. Dialectica 27: 313–328.
———. 1980 [1970]. Mental events. In Readings in the philosophy of psychology, ed. N. Block,
vol. 1, 107–119. Cambridge: Harvard University Press.
Dennett, D.C. 1991. Real patterns. Journal of Philosophy 88: 27–51.
Fodor, J.A. 1983. The modularity of mind: An essay on faculty psychology. Cambridge: MIT Press.
Fodor, J. 1984. Observation reconsidered. Philosophy of Science 51: 23–43.
Freud, S. 1900. The interpretation of dreams (first part). In The standard edition of the complete
psychological works of Sigmund Freud, ed. J. Strachey, vol. 4. London: The Hogarth Press/The
Institute of Psychoanalysis.
Freud, S. 1909. Notes upon a case of obsessional neurosis. In The standard edition of the com-
plete psychological works of Sigmund Freud, ed. J. Strachey, vol. 10, 151–318. London: The
Hogarth Press and The Institute of Psychoanalysis.
———. 1910. The future prospects of psycho-analytic therapy. In The standard edition of the
complete psychological works of Sigmund Freud, ed. J. Strachey, vol. 11, 139–152. London:
The Hogarth Press and The Institute of Psychoanalysis.
214 A. Sims

Gardner, S. 1993. Irrationality and the philosophy of psychoanalysis. Cambridge: Cambridge


University Press.
Gawronski, B. 2004. Theory-based bias correction in dispositional inference: The fundamen-
tal attribution error is dead, long live the correspondence bias. European Review of Social
Psychology 15: 183–217.
Grünbaum, A. 1984. The foundations of psychoanalysis: A philosophical critique. Berkeley:
University of California Press.
Habermas, J.  1986 [1968]. Knowledge and Human Interests. Trans. J.  J. Shapiro. Cambridge:
Polity Press.
Heimann, P. 1950. On counter-transference. International Journal of Psycho-Analysis 31: 81–84.
Heyes, C. 2014. Submentalizing: I am not really reading your mind. Perspectives on Psychological
Science 9: 131–143.
Hopkins, J.  1988. Epistemology and depth psychology: Critical notes on The Foundations of
Psychoanalysis. In Mind, psychoanalysis and science, ed. P.  Clark and C.  Wright, 33–60.
Oxford: Blackwell Publishing.
———. 1995. Irrationality, interpretation and division. In Philosophy of psychology, ed.
C. McDonald and G. McDonald, 409–432. Oxford: Blackwell.
———. 2012. Psychoanalysis, representation, and neuroscience: The Freudian unconscious and
the Bayesian brain. In From the couch to the lab: Trends in psychodynamic neuroscience, ed.
A. Fotopoulou, D. Pfaff, and M.A. Conway, 230–265. Oxford: Oxford University Press.
———. 2015. The significance of consilience: Psychoanalysis, attachment, neuroscience and evo-
lution. In Psychoanalysis and the philosophy of mind: Unconscious mentality in the twenty-first
century, ed. S. Boag, L.A.W. Brakel, and V. Talvitie, 47–136. London: Karnac.
Horney, K. 1999/1942. Self-analysis. London: Routledge.
Hutto, D.D., M. Herschbach, and V. Southgate. 2011. Social cognition: Mindreading and alterna-
tives. Review of Philosophy and Psychology 2 (375).
Kovács, A.M., E. Téglás, and A.D. Endress. 2010. The social sense: Susceptibility to others’ beliefs
in human infants and adults. Science 330: 1830–1834.
Lacewing, M. 2012a. Inferring motives in psychology and psychoanalysis. Philosophy, Psychiatry
& Psychology 19: 197–212.
———. 2012b. Statistics, desire and interdisciplinarity. Philosophy, Psychiatry & Psychology 19:
221–225.
Lacewing, M. 2018. The science of psychoanalysis. Philosophy, Psychiatry & Psychology 25:
95–111.
Laplanche, J., and J.B. Pontalis. 1973. The Language of Psycho-analysis, Trans. by D. Nicholson-­
Smith. London: Karnac.
Lewis, D. 1972. Psychophysical and theoretical identifications. Australasian Journal of Philosophy
50: 249–258.
Lipton, P. 1991. Inference to the best explanation. London: Routledge.
MacMillan, M. 1991. Freud evaluated: The completed arc. Amsterdam: North Holland Publishing.
Marr, D. 1982. Vision. San Francisco: W. H. Freeman.
Masson, J.M. 1985. The complete letters of Sigmund Freud to Wilhelm Fliess, 1887–1904.
Cambridge: The Belknap Press of Harvard University Press.
McIntyre, A. 1958. The unconscious: A conceptual analysis. London: Routledge/Kegan Paul.
Onishi, K.H., and R. Baillargeon. 2005. Do 15-month-old infants understand false beliefs? Science
308: 255–258.
Panksepp, J., and L. Biven. 2012. The archaeology of mind: Neuroevolutionary origins of human
emotion. New York: W.W. Norton & Company.
Pataki, T. 2014. Wish-fulfilment in philosophy and psychoanalysis: The tyranny of desire. London:
Routledge.
Popper, K. 1962. Conjectures and refutations: The growth of scientific knowledge. New  York:
Basic Books.
Causal Inference in the Clinical Setting: Why the Cognitive Science of Folk Psychology… 215

Rosenzweig, S. 1936. Some implicit common factors in diverse methods of psychotherapy.


American Journal of Orthopsychiatry 6: 412–415.
Russell, J.A. 1991. Culture and the categorization of the emotions. Psychological Bulletin 110:
426–450.
Skinner, B.F. 1956. Critique of psychoanalytic concepts and theories. In Minnesota studies in the
philosophy of science, volume 1: The foundations of science and the concepts of psychology
and psychoanalysis, ed. H. Feigl and M. Scriven, 77–87. Minneapolis: University of Minnesota
Press.
Stich, S. 1983. From folk psychology to cognitive science. Cambridge: MIT Press.
Tversky, A., and D. Kahneman. 1974. Judgment under uncertainty: Heuristics and biases. Science
185: 1124–1131.
Wellman, H., D. Cross, and J. Watson. 2001. Meta-analysis of theory of mind development: The
truth about false-belief. Child Development 72: 655–684.
Winnicott, D.W. 1949. Hate in the counter-transference. International Journal of Psycho-Analysis
30: 69–74.
Wittgenstein, L. 1953. Philosophical investigations (trans: Anscombe, G.E.M.). Oxford: Blackwell.

You might also like