The Epistemology of Non-Visual Perception

The Epistemology of Non-Visual Perception
PHILOSOPHY OF MIND SERIES

Series Editor: David J. Chalmers, New York University
The Conscious Brain Attention is Cognitive Unison

Jesse Prinz An Essay in Philosophical Psychology
Christopher Mole
Simulating Minds
The Philosophy, Psychology, and The Contents of Visual Experience
Neuroscience of Mindreading Susanna Siegel
Alvin I. Goldman
Consciousness and the Prospects of
Supersizing the Mind Physicalism
Embodiment, Action, and Cognitive Derk Pereboom
Extension
Andy Clark Consciousness and Fundamental Reality
Philip Goff
Perception, Hallucination, and Illusion
William Fish The Phenomenal Basis of Intentionality
Angela Mendelovici
Phenomenal Concepts and Phenomenal
Knowledge Seeing and Saying
New Essays on Consciousness and The Language of Perception and the
Physicalism Representational View of Experience
Torin Alter and Sven Walter Berit Brogaard
Phenomenal Intentionality Perceptual Learning

George Graham, John Tienson and The Flexibility of the Senses
Terry Horgan Kevin Connolly
The Character of Consciousness Combining Minds

David J. Chalmers How to Think About Composite
Subjectivity
The Senses Luke Roelofs
Classic and Contemporary Philosophical
Perspectives The Epistemic Role of Consciousness
Fiona Macpherson Declan Smithies
The Epistemology
of Non-Visual
Perception
Edited by
BERIT BROGAARD
A N D D I M I T R IA E L E C T R A G AT Z IA
1
3
Oxford University Press is a department of the University of Oxford. It furthers
the University’s objective of excellence in research, scholarship, and education
by publishing worldwide. Oxford is a registered trade mark of Oxford University
Press in the UK and certain other countries.
Published in the United States of America by Oxford University Press

198 Madison Avenue, New York, NY 10016, United States of America.
© Oxford University Press 2020
All rights reserved. No part of this publication may be reproduced, stored in

a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by license, or under terms agreed with the appropriate reproduction
rights organization. Inquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above.
You must not circulate this work in any other form

and you must impose this same condition on any acquirer.
Library of Congress Control Number: 2019954225

ISBN 978–0–19–064891–6
1 3 5 7 9 8 6 4 2
Printed by Integrated Books International, United States of America

Contents
Contributors vii
About the Companion Website xi
Introduction 1
Berit Brogaard and Dimitria Electra Gatzia
1. Tasting Flavors: An Epistemology of Multisensory Perception 29
Barry C. Smith
2. Sensory Interactions and the Epistemology of Haptic Touch 53
Matthew Fulkerson
3. Multimodal Mental Imagery and Perceptual Justification 77
Bence Nanay
4. How Reliably Misrepresenting Olfactory Experiences Justify
True Beliefs 99
Angela Mendelovici
5. Hearing As 118
William G. Lycan
6. Is Tactual Knowledge of Space Grounded in Tactual Sensation? 146
John Campbell
7. “Unless I Put My Hand into His Side, I Will Not Believe”:
The Epistemic Privilege of Touch 165
Olivier Massin and Frédérique de Vignemont
8. Experiences of Duration and Cognitive Penetrability 188
Carrie Figdor
9. Experiencing Emotions: Aesthetics, Representationalism,
and Expression 213
Rebecca Copenhaver and Jay Odenbaugh
10. The Emotional Dimension to Sensory Perception 236
Lana Kühle
11. The Perception of Virtue 256
Jennifer Matey
Index 273
Contributors
Berit Brogaard is Cooper Professor of Philosophy and Director of the Brogaard Lab
for Multisensory Research at the University of Miami and Professor II at University
of Oslo. Her areas of research include philosophy of perception, philosophy of
emotions, and philosophy of language. She is the author of Transient Truths (Oxford
University Press, 2012), On Romantic Love (Oxford University Press, 2015), The
Superhuman Mind (Penguin, 2015), Seeing & Saying (Oxford University Press, 2018),
and Hate: Anatomy of an Emotion (Oxford University Press, In press).
John Campbell is Professor of Philosophy at the University of California, Berkeley.
His main interests are in theory of meaning, metaphysics, and philosophy of psy-
chology. He is working on the question whether consciousness, and in particular
sensory awareness, plays any key role in our knowledge of our surroundings. He is
also working more generally on causation in psychology. He is the author of Past,
Space and Self (MIT Press, 1994) and Reference and Consciousness (Oxford University
Press, 2002).
Rebecca Copenhaver is Professor of Philosophy at Lewis & Clark College, where
she has taught since 2001. Her main area of research is the philosophy of mind, par-
ticularly perception and memory, with special attention to modern British theories
of mind, including those of Thomas Reid, George Berkeley, and John Locke. She is
a co-author (with Brian P. Copenhaver) of From Kant to Croce: Modern Philosophy
in Italy 1800–1950 (Toronto University Press, 2012) and a number of articles on
modern British theories of mind, exploring perception, memory, consciousness, and
methodology.
Frédérique de Vignemont works at the crossroad of philosophy of mind and cogni-
tive science. She is Senior CNRS (National Centre for Scientific Research) Researcher
and Deputy Director of the Jean Nicod Institute, and Philosophy Scholar in Residence
at New York University. In 2018, she published a monograph with OUP, Mind the
Body, the first comprehensive treatment of bodily awareness and the sense of bodily
ownership, integrating philosophical analysis with recent experimental results. She
has also published in some of the best philosophy journals in the world, including
Journal of Philosophy and Mind, as well as in high-impact journals in cognitive sci-
ence, such as Trends in Cognitive Science. She is an executive editor of the Review of
Philosophy and Psychology. She was the recipient of the 2015 Young Mind & Brain
prize for her achievements in advancing our knowledge about mind and brain.
viii contributors
Carrie Figdor is Professor of Philosophy at the University of Iowa with appointments

in Psychological and Brain Sciences, the Iowa Neuroscience Institute, and the
Interdisciplinary Graduate Program in Neuroscience. She works primarily in philos-
ophy of mind, psychology, and neuroscience. Her book, Pieces of Mind: The Proper
Domain of Psychological Predicates (Oxford, 2018), is about empirically driven
conceptual change in psychology. She has published in the Journal of Philosophy,
Philosophy of Science, Mind & Language, and other leading journals.
Matthew Fulkerson is Associate Professor of Philosophy at the University of
California, San Diego. His research has focused on the sense of touch and haptic ex-
ploration. He’s especially interested in the relationship between bodily awareness and
our experience of the world, and in understanding how touch connects and interacts
with the other senses. More recently, he has been exploring the relationship between
emotions and perception, as well as writing papers on motivating sensory states and
sensory pleasure and pain. He is the author of The First Sense: A Philosophical Study of
Human Touch (MIT Press, 2014).
Dimitria Electra Gatzia is Professor of Philosophy and Department Chair at the
University of Akron. She works at the intersection of philosophy of mind and cog-
nitive science. She was a Research Fellow at the Centre of Philosophical Psychology
at the University of Antwerp (2016–2017). She is currently a Research Affiliate of the
Brogaard Lab for Multisensory Research at the University of Miami and a Research
Fellow at the Czech Academy of Sciences, Institute of Philosophy. She has published
on perception, consciousness, cognitive penetration, synesthesia, and imagina-
tion. In 2016, she co-edited a special issue on Modal Epistemology for Topoi: An
International Review of Philosophy (with Berit Brogaard); and in 2018, she co-edited
a special issue on The Rational Roles of Perceptual Experience for Inquiry: An
Interdisciplinary Journal of Philosophy (with Berit Brogaard).
Lana Kühle is Assistant Professor of Philosophy at Illinois State University. She
received her PhD from the University of Toronto. Her areas of expertise are in phi-
losophy of mind and phenomenology with close interdisciplinary connections to
neuroscience and psychology. Her overall goal is to understand how one’s bodily
awareness gives rise to the sense of self and the subjectivity of conscious experi-
ence. She has published several articles including “The Subjectivity of Experiential
Consciousness: It’s Real and It’s Bodily” in Mind & Matter, and “The Missing Pieces
in the Scientific Study of Bodily Awareness” in Philosophical Psychology.
William G. Lycan is William Rand Kenan, Jr. Professor of Philosophy Emeritus at the
University of North Carolina, and currently Distinguished Visiting Professor at the
University of Connecticut. He is author of Knowing Who (with Steven Boër) (MIT
Press, 1985), Logical Form in Natural Language (MIT Press, 1986), Consciousness,
Judgement and Justification (Cambridge University Press, 1988), Modality and
Meaning (Springer, 1994), Consciousness and Experience (MIT Press, 1996),
contributors ix
Philosophy of Language: A Contemporary Introduction, Real Conditionals (Routledge,

2018), and On Evidence in Philosophy (Oxford University Press, 2019).
Olivier Massin is Professor of Philosophy at the Institut of Philosophie of the
University of Neuchâtel. He received two PhDs in philosophy, one from the
University of Aix-Marseille and another from the University of Geneva. He has been
previously research scientist at the Institut Jean Nicod (Paris) and SNF Professor
at the Philosophy Department of the University of Zürich. He has been working
on questions such as: What are forces? What are mixtures? What are economic
exchanges? What is ownership? What is continuity? What is pleasure? What is suf-
fering? What is pain? What are desires? What are determinables? What are tryings?
What are efforts? What is touch?
Jennifer Matey is Associate Professor of Philosophy at Southern Methodist
University. Her research is largely in the philosophy of mind and the philosophical
foundations of psychology and the cognitive sciences, but often intersects with moral
psychology, aesthetics, ethics, and epistemology. Her published work has focused on
the nature of consciousness, perceptual, and emotional experience, related issues in
epistemology and metaphysics, and relevant research in the cognitive and biological
sciences.
Angela Mendelovici is Associate Professor of Philosophy at the University of
Western Ontario. She obtained her PhD from Princeton University and completed a
postdoc at the Australian National University. Her research is in philosophy of mind,
focusing on consciousness, intentionality, and the relationship between the two. Her
recent book, The Phenomenal Basis of Intentionality (Oxford University Press, 2018),
provides a systematic presentation and defense of a radically internalistic theory of
intentionality on which all represented contents are either phenomenal contents that
are “before our mind’s eye” or expressly and voluntarily singled out by us.
Bence Nanay is BOF Research Professor at the Centre for Philosophical Psychology at
the University of Antwerp and Senior Research Associate at Peterhouse, Cambridge
University. He is the author two books (Between Perception and Action, Oxford
University Press, 2013, and Aesthetics as Philosophy of Perception, Oxford University
Press, 2016) with two more under contract (Aesthetics: A Very Short Introduction,
Oxford University Press, and Seeing Things You Don’t See, Oxford University Press) as
well as more than 120 peer-reviewed articles. His work is supported by a number of
high profile grants including an ERC Consolidator Grant.
Jay Odenbaugh is Professor of Philosophy at Lewis & Clark College, where he has
taught since 2003. His main areas of research are the philosophy of biology, philos-
ophy of psychology, and environmental philosophy. He is the author of Ecological
Models (Cambridge University Press, forthcoming).
Barry C. Smith is Professor of Philosophy and Director of the Institute of Philosophy
at the University of London’s School of Advanced Study. He is also the Founding
x contributors
Director of the Centre for the Study of the Senses, which pioneers collaborative re-
search among philosophers, psychologists, and neuroscientists. He has written the-
oretical and experimental papers, publishing in Nature, Food Quality and Preference,
Chemical Senses and Flavour. In 2007, he edited Questions of Taste: The Philosophy of
Wine, Oxford University Press; in 2008, he edited The Oxford Handbook of Philosophy
of Language, Oxford University Press (with Ernest Lepore); and in 1998, he ed-
ited Knowing Our Own Minds, Oxford University Press (with Crispin Wright and
Cynthia Macdonald).
About the Companion Website
https://global.oup.com/us/companion.websites/9780190648916/
Oxford has created a website to accompany The Epistemology of Non-Visual

Perception. Material that cannot be made available in a book, namely audio
examples, is provided here. The reader is encouraged to consult this resource
in conjunction with Chapter 5: “Hearing As” by William G. Lycan. Examples
available online are indicated in the text with Oxford’s symbol
The Epistemology of Non-Visual Perception
Introduction
Berit Brogaard and Dimitria Electra Gatzia
I.1. The Epistemic Landscape
This book is an anthology of new essays by top researchers in epistemology

and philosophy of mind focused on the epistemology of non-visual percep-
tion. The focus of the volume is to highlight the many different domains in
which non-visual sensory experience, broadly construed to include multi-
modal experience associated with emotional and agential perception, plays a
rational role, for instance, as an immediate justifier of belief.
We seem to take it for granted that perceptual experience can give rise to
justified beliefs as well as knowledge. There is wide disagreement, however,
about which kinds of beliefs perceptual experience can provide justification
for and whether perceptual experience can play this role by itself or only
in combination with background belief (Davidson 1986; McDowell 1994;
Fumerton 1995; Pryor 2000; Huemer 2001; Swinburne 2001; BonJour 2001;
Burge 2003, 2010; Silins 2011; Siegel 2012; McGrath 2018). Recent empirical
results further suggest that emotions play a crucial role in decision-making
(Damasio 1994) and agential perception. While visual experience (espe-
cially, the experience of color, shape, and depth) is a well-researched topic
within the epistemology of perception, the rational roles of non-visual per-
ception for the most part have been underexplored. Non-visual perception,
as we will see, includes affective experience and agential perception, both of
which involve sensory modalities that go beyond vision. For example, it is
well established that emotional and agential perception involve multisen-
sory perception associated with, for example, interoception (the sense of the
physiological condition of the body) and other non-visual sensory modali-
ties. As a result, emotional and agential perception fit well within the topic of
non-visual perception.
Before providing a detailed account of the topics and contributions to this
volume, it will be helpful to begin with a brief description of the theoretical
Berit Brogaard and Dimitria Electra Gatzia, Introduction In: The Epistemology of Non-Visual
Perception. Edited by: Berit Brogaard and Dimitria Electra Gatzia, Oxford University Press (2020).
© Oxford University Press.
DOI: 10.1093/oso/9780190648916.003.0001
2 The Epistemology of Non-Visual Perception
landscape within epistemology. The answers to the questions relating to the

epistemology of perception turn on the nature of epistemic justification.
Different theorists provide different answers to this question. Externalists
about justification take justification to depend at least in part on external
features that need not be accessible to the subject or a feature solely of the
subject’s mental states. Reliabilism about justification, for example, grounds
justification in the reliability of belief formation. On one account, perceptual
beliefs are justified when they are the result of a reliable process that may or
may not involve reliable mind-external instruments (Goldman 1986, 1997;
Dretske 1981; Greco 2003; Sosa 2007; Kelp 2011).
Internalism about justification is the view that justification is a function
of the internal mental states of the subject. The weakest form of internalism
is mentalism (Steup 1996; Sosa 1999b). This view does not require that the
subject can access the mental states that justify the belief or is in a position to
tell when they are justified but merely that the internal features are internal
to the subject’s mental life. Weak access internalists adopt an account of jus-
tification according to which justifiers are consciously accessible to the sub-
ject, at least given suitable awareness, attention, or reflection (Alston 1989;
BonJour 1999; Audi 1998). What is also known as “strong access internalism”
holds in addition that the subject is in a position to tell when beliefs are jus-
tified (for discussion see Pappas 2017). Internalists tend to hold that the
justificatory status of a subject’s beliefs strongly supervenes on her occur-
rent and dispositional mental states or their (propositional) contents. So if
two subjects are mentally exactly alike, then they are alike justificationally
(Conee and Feldman 2001). This latter view is also sometimes referred to as
“evidentialism.”
Intentionalist theories of justification can be divided into two groups:
coherentism and foundationalism. These views differ in terms of what they
take to constitute justification. Coherentist theories do not distinguish between
foundational and non-foundational beliefs. They maintain that all beliefs are
justified in virtue of their relation to other beliefs. On this view, justification
is a result of holistic or mutual support among otherwise unsupported beliefs
(BonJour 1985; Lehrer 1990).
Advocates of foundationalism distinguish between foundational or basic
belief and non-basic belief (Descartes 1641/1984; Russell 1910–11, 1913/
1984; Fumerton 1995, 2001). The former are justified independently of their
relation to other beliefs, while the latter are justified because of their relation
to other beliefs. Given foundationalism, your belief that p is justified just in
Introduction 3
case either (a) your belief in p is basic or (b) your belief in p stands in the
right sort of relation to basic beliefs (Fumerton 1995). Beliefs directly based
on perceptual states or introspection—the capacity to inspect one’s mental
states—are often treated as basic (Fumerton 1995; Pryor 2000; Huemer
2001; McGrath 2018). The view that only beliefs that are epistemically priv-
ileged can be basic is sometimes referred to as “classical foundationalism”
(Fumerton 1995, 2001).
Belief based on introspection is sometimes thought to be epistemically priv-
ileged compared to belief based on perception (Descartes 1641/1984; Russell
1910–11, 1912; Shoemaker 1963; Alston 1971, 1986; Chisholm 1981). This
may in part be because we have immediate access to, or immediate awareness
of, our occurrent mental states in a way that we may not have immediate access
to the objects of perception (Russell 1910, 1912). For example, it may be held
that you are in a position to access your conscious desire to avoid being harmed
by Hurricane Irma in an immediate way. Perceptual experiences, by contrast,
are often thought to be indirect—at least assuming non-relationalism. On a
content theory of perceptual experience, for instance, you experience, say, the
debris on the street in the wake of Hurricane Irma, indirectly by being related
to a perceptual content that represents the debris on the street.
Belief based on introspection may also seem to be less prone to error than
belief based on perception (Chisholm 1981). For example, you are far less
likely to be mistaken about whether you are having a toothache than you are
about the exact location of the bad tooth. This may be because the belief that
you have a toothache is self-warranted, that is, it cannot be the belief it is and
fail to have some degree of justification (Alston 1976). One may also hold
a stronger view to the effect that beliefs based on introspection are indubi-
table, that is, the subject cannot have any grounds for doubting them (Alston
1971). An introspective experience of the feeling of, say, having a toothache
may be thought to eliminate all possible doubt for the subject as to whether
she has a toothache.
We may even be able to push the idea a step further and say that there is
no difference between appearance and reality when it comes to introspec-
tive experience. Beliefs based on introspection, one might hold, are infallible,
which is to say, they cannot be mistaken (Descartes 1641/1984). On this view,
if one believes that one has an experience E on the basis of introspection,
then one has the experience E. For instance, if you judge that you are having
a visual experience of redness, then you have a visual experience of redness.
This may be because the judgment that you are having a visual experience
of redness contains the red experience itself (Chalmers 2003). Infallibility

can come apart from indubitability. It may be impossible for a given rational
and capable subject (though perhaps not for an ideally rational subject) to
provide grounds for doubting a belief based on introspection, even if the in-
trospective belief is false (e.g., if there are operations that are subjectively in-
distinguishable from introspection but which can misfire, then one might be
certain that one is having a visual experience of red without having a visual
experience of red; for an argument along these lines, see Chalmers 2003).1
Belief based on perceptual experience, by contrast, is not indubitable or in-
fallible. For instance, you might believe on the basis of interoception that
something is wrong with your upper right wisdom tooth, even if the tooth
that is causing the pain is in fact the one next to it.
A fourth thing we might mean when we say that introspection is “episte-
mically privileged” is that introspection of a mental state puts us in a position
to know that we are in the introspected mental state (Descartes 1641/1984).
This kind of constraint is also sometimes referred to as “self-intimation” (e.g.,
Smithies 2012), “luminosity” (Williamson 2000, 95), “omniscience” (Alston
1971), and “transparency” (Shoemaker 1990).
It tends to be a variation on at least one of those four constraints that
people have in mind when they say that belief based on introspection is ep-
istemically privileged. Epistemic privilege is of special interest to classical
foundationalists because they maintain that certain kinds of epistemically
privileged beliefs are basic (Descartes 1641/1984; Russell 1910–11, 1913/
1984; Fumerton 1995, 2001). Although classical foundationalists occa-
sionally take the epistemic privilege of belief to be the result of a special
acquaintance with mind-independent entities (Fumerton 1995), most clas-
sical foundationalists treat introspection to be providing us with epistemi-
cally privileged beliefs (Descartes 1641/1984; Russell 1910–11, 1913/1984).
Not only does introspection reveal how our experiences appear to us, it may
also provide an epistemic directness that cannot be found in perception or
generate beliefs that are immune to error. So beliefs based on introspective
beliefs can provide the foundation that perceptual beliefs cannot provide.
Basic beliefs are not taken to metaphysically or logically imply non-basic
beliefs. Rather they are treated as defeasible evidence for non-basic beliefs.
To see this consider the following case. Suppose that you have an experience
1 Some also distinguish between incorrigibility and infallibility (Shoemaker 1963). A belief may
be said to be incorrigible just in case it is impossible to show that sincere introspective reports are
mistaken.
Introduction 5
of water pouring down outside your window. If your perceptual belief that
water is pouring down outside your window is basic, its justification does
not depend on other beliefs, say, your belief that your visual system is reli-
able. The basic belief that water is pouring down outside your window may
serve as defeasible evidence for the belief that it’s raining. This belief is de-
feasible, because if you were to find out that your university has installed a
new sprinkler on the rooftop, your belief that water is pouring down outside
your window would no longer count as justification for your belief that it is
raining. Your justification is then said to be “defeated.”
Richard Fumerton (1995) argues that inferentially justified beliefs must
adhere to the following principle:
Principle of Inferential Justification
To be justified in believing one proposition p on the basis of another propo-

sition q, one must be (1) justified in believing q and (2) justified in believing
that q makes p probable.
Consider again the basic belief “Water is pouring down outside the window”
and the non-basic belief “It’s raining.” In order for the former belief to jus-
tify the latter, you must be justified in holding the former, and be justified in
thinking that the former makes the latter probable.
Fumerton argues that this principle is what fuels the core of a standard
skeptical argument. The skeptic will challenge the non-skeptic to explain
how perceptual beliefs about the external world can make it probable that
there are external, mind-independent objects. The non-skeptic might reply
by referring to past correlations between our experiences and the existence
of external, mind-independent objects. In the past we have observed that
our experiences of external objects were correlated with the existence of
external, mind-independent objects. But the skeptic will then reply that in
order to be justified in believing that there are such correlations, we would
need to have direct conscious access to physical objects, a kind of access that
is independent of experience. But we have no such experience-independent
access to the physical world, so we have no justification for our belief that
there is a constant conjunction of external, mind-independent, and sensory
experiences about those objects.
Epistemic conservatism and phenomenal conservatism/ phenomenal
dogmatism are two newer versions of foundationalism that may provide the
tools for avoiding the sort of skeptical argument just outlined (Pryor 2000;
Huemer 2001; Brogaard 2018). Epistemic conservatism says that merely
believing that p gives you prima facie justification for p, provided that you
have no grounds for doubting that belief (Quine 1951; Chisholm 1980;
Harman 1986, 46; McCain 2008).2 So, on this view, if a subject believes that
p, then, in the absence of defeaters, she has some degree of justification for
believing that p. It may be argued that this view is unsustainable, as it appears
to give rise to bootstrapping cases (Foley 1983). Suppose, for example, that
you form the belief that p on purely dogmatic grounds. Since you already be-
lieve that p, it is epistemically appropriate to keep believing p in the absence
of defeaters. In fact, not only do you have evidence for believing that p, but
also the originally unjustified belief is the very source of justification in the
absence of defeaters.
Phenomenal conservatism, also known as “dogmatism” and “phenomenal
dogmatism,” aims to avoid this type of problem by proposing that non-belief
states (experiences or seemings) can provide immediate justification for belief
in the absence of defeaters (see Pryor 2000; Huemer 2001, 2007, 2013; Tucker
2010; Chudnoff 2012, 2014; Brogaard 2018). On this view, in the absence of any
defeaters, you are prima facie justified in believing that p regardless of whether
your perceptual system is reliable or you have any evidence for thinking that
your perceptual system is reliable. In other words, if it seems to you as if p, you
thereby have at least prima facie justification for believing that p.
Seemings of the kind that can justify belief are distinct from beliefs.
Suppose, for example, that you see a straight stick that is submerged in water
and you experience it as bent. On the basis of the experience, you come to be-
lieve that the stick is bent. In this case, the way the stick seems to be coincides
with your belief that the stick is bent. Now, suppose that you are presented
with perceptual evidence that shows that the stick is actually straight (say,
the stick is taken out of the water and it now looks straight) and, on the basis
of this evidence, you form the belief that the stick is straight. In this case, the
2 Roderick Chisholm’s (1980) version of conservatism is stronger than the characterization pro-
vided and hence more unintuitive. According to Chisholm, “Anything we find ourselves believing
may be said to have some presumption in its favor—provided it is not explicitly contradicted by the
set of other things we believe” (1980, 551–552). So, if one believes that it is raining—on no grounds
whatsoever—and also believes that one’s belief that it is raining likely isn’t based on any evidence,
then one still has some justification for believing that it is raining. Another problem is this, if one
already believes that p, but acquires evidence that strongly supports not-p, then you still have justifi-
cation for believing p (Foley 1983). For instance, if you already believe that it’s raining and then sub-
sequently acquire evidence that you are on the moon, and you know that it cannot rain on the moon,
then you still have justification for believing that it’s raining.
Introduction 7
way the stick seems to be does not coincide with your belief: the submerged
stick still seems bent to you but you no longer believe that it is bent.3
Phenomenal dogmatism has been thought to run into problematic
implicit-bias-driven experience and bootstrapping cases if experiences can
be cognitively or affectively penetrated (Siegel 2012, 2017). Consider the fol-
lowing core case (Siegel 2012). Jill believes unjustifiably that Jack is angry at
her. The following day Jill sees Jack’s neutral face as an angry face, owing to
cognitive penetration. Since phenomenal dogmatism holds that experiences
can prima facie justify beliefs, her unjustified belief that Jack is angry at her is
now prima facie justified. Some have argued, however, that many other epi-
stemic theories, including reliabilism, are subject to the versions of the same
problem (Tucker 2010). As we will see later, several chapters in this volume
deal with issues related to this problem (e.g., Kühle; Nanay).
Ever since Gettier, it has been held that knowledge requires more than
justified true belief. What is required besides justification and truth is still
subject of debate. But a common belief is that knowledge requires the satis-
faction of either a tracking condition (sensitivity) or a safety condition. A be-
lief that p can be said to track the truth just in case, if p had been false, then
the subject would not have believed that p (Nozick 1981). A belief that p is
said to be safe just in case, in the closest worlds where the subject believes
that p, p is true (Sosa 1999a; Williamson 2000). As we will see in Nanay’s con-
tribution, these conditions appear to be violated in cases that involve multi-
modal mental imagery.
We turn now to an overview of the literature on our sensory experiences
in the different modalities and how they present specific challenges to widely
accepted epistemic theories. Along the way, we address in further detail how
these issues come up in the contributions to this volume.
I.2. Multisensory Experience and Multimodal

Mental Imagery
The standard discussion on perception includes only five exterocep-

tive senses, that is, vision, audition, olfaction, taste, and touch, each cor-
responding to a kind of perceptual experience. However, one might also
3 Some argue that the view should be restricted (see Mackie 2005; Brogaard 2014; Chudnoff
2016) as follows: if it seems to you as if p and condition C is met, you thereby have at least prima facie
justification for believing that p.
include proprioception (our sense of the body’s position in space, move-

ment, and spatial orientation) and the vestibular system (our sense of bal-
ance) among our sensory modalities. Furthermore, there are a number
of distinct nerve endings related to several somatic senses or internal
organs, including the perception of temperature, pressure, itch, space, and
pain (nociception), and bodily sensations (e.g., the feeling of hunger or
the feeling of a quickening of the heartbeat). Bodily sensations (or bodily
feelings—also known as interoception) have not traditionally been con-
strued as sensory experiences. However, one might argue that the modality
that produces bodily sensations just is a sensory modality closely related
to proprioception, the vestibular system, and nociception (pain and space
perception), which arguably are sensory modalities unlike intuition and
introspection (Macpherson 2011; Schwenkler 2013; Briscoe 2016). Time
perception may also be associated with its own sensory modality (see, e.g.,
Brogaard and Gatzia 2015a).
Although different sensory modalities involve different mechanisms,
our sensory experiences often involve information associated with more
than one sensory modality (Auvray and Spence 2008). For example, as the
McGurk illusion (which arises when incongruent auditory and visual cues
are presented simultaneously) illustrates, what we hear depends to some ex-
tent on what we see (McGurk and MacDonald 1976). The McGurk illusion
shows that when the auditory syllable “ba” is combined with a mouth move-
ment that forms the syllable “ga,” subjects report hearing a second sound,
namely “da.” Just as in the case of the submerged stick that continues to seem
bent even after we form the belief that it is straight, the McGurk illusion
persists even after subjects form the belief (on the basis of reliable testimony)
that the auditory syllable is “ba” in the first case and “ga” in the second.
Related phenomena arise for gustation and olfaction. Although gustation
and olfaction involve distinct mechanisms, the olfactory system is vitally
important for determining food “flavors,” that is, the inseparable taste and
smell fusion that is distinguished from taste proper (Ross 2001; Stevenson
and Tomiczek 2007). For example, it is not easy to distinguish between the
flavors of an apple and a potato if you hold your nose while you are biting
into them. Nor is it easy to determine whether a jelly bean you are chewing
has, say, an apple or watermelon flavor if you hold your nose while chewing
it. Smelling the jelly bean helps you taste it as apple or watermelon. Indeed,
studies show that in many cases what we typically associate with the taste of
foods originates from the sense of smell (Rozin 1982).
Introduction 9
All of these experiences are instances of multisensory experiences (for

an overview, see, e.g., Auvray and Spence 2008; O’Callaghan 2015a, 2015b;
Brogaard and Chudnoff 2018). Our understanding of multisensory percep-
tion impacts the philosophical and neuroscientific models of perception.
But it also has a significant impact on epistemology. One of the questions
that arises is whether tastes (as in the jelly bean example) are objective prop-
erties of foods or merely features of our own experiences. Although most
philosophers are objectivists about visual properties such as colors (see
Byrne and Hilbert 2003), they tend to favor a simple subjectivist account of
taste, according to which tasting can only provide us with the awareness of a
distinctive and proprietary range of features of our own experience (Hume
1738/2012; Smith 2002). However, as noted previously, the fruity flavor of,
say, pineapple does not result from the sensory stimulation of the tongue (as
we do not have pineapple receptors on the tongue) but from the combination
of our gustatory and olfactory modalities (Stevenson and Tomiczek 2007).
So what we typically experience in tasting is a unified experience of flavor
resulting from multisensory integration.
Barry Smith (in this volume) argues that such empirical findings are con-
trary to the simple subjectivist view of tastes as simple sensations of the
tongue. Moreover, Smith discusses empirical evidence indicating that the
experiences of flavor are influenced by a wide variety of factors, including
music, lighting conditions, the colors of the plates, and so forth (Crisinel
and Spence 2011; Spence 2015; Harrar et al. 2013), which have led to a so-
phisticated subjectivism. This form of subjectivism, which is popular among
neuroscientists, suggests that flavors are constructs of the brain (Sheppard
2012, 2017; Small 2012). Smith, however, argues that sophisticated subjec-
tivism is also problematic because it fails to distinguish between flavor and
flavor perception: “while the brain creates our perceptions of flavor, it does
not create flavors” ([page no. 40]). Smith proposes an objectivist account of
tasting as a species of perception that gives us knowledge of flavors. On this
account, flavors are configurations of pleasant, odorous, and tactile prop-
erties of foods and liquids that we track by means of integrated multisen-
sory inputs. Together with our past learning and prior expectations, these
integrated multisensory inputs enable us to perceive the flavors of foods and
liquids as a way of guiding successful food choice.
A related question pertains to the impact sensory interactions have on the
epistemic quality of our perceptual experiences. The problem arises because
although in some cases the epistemic quality of our perceptual awareness is
strengthened when sensory inputs are added, it is weakened in other cases.

Matthew Fulkerson (this volume) calls the strengthening of epistemic quality
of our perceptual awareness “sensory reinforcing” and its weakening “sen-
sory discounting.” For example, it is easier to determine whether, say, a peach
is ripe when you can both see it and touch it. This is a case of sensory enfor-
cing. Certain multisensory illusions, by contrast, are examples of sensory
discounting. For example, the size-weight illusion involves larger objects
feeling heavier than smaller objects of the same mass (Jones and Lederman
2006). In this case, adding sensory inputs weakens the epistemic quality of
our perceptual awareness.
Sensory reinforcing or discounting effects involve cases in which per-
ceptual processing begins with sensory stimulation. However, perception is
influenced not only by stimulation-driven perceptual processes but also by
mental imagery (see Nanay, this volume). Consider, for example, a case in
which a visual stimulus (say, watching someone speak on a muted television
screen) gives rise to auditory mental imagery (for a review, see Spence and
Deroy 2013). The auditory mental imagery in this case depends on the visual
input as well as information about the speaker. For example, particular audi-
tory mental imagery is evoked when the speaker is Salma Hayek as opposed
to Martin Scorsese.
Bence Nanay (this volume) argues that everyday perception is, to a
great extent, a mixture of sensory stimulation-driven perceptual pro-
cessing and mental imagery. Suppose, for example, that you are sitting
in the living room as you are waiting for your partner to join you down-
stairs. Suppose now that you hear someone coming down the stairs and
that sound that triggers a visual experience of your partner. On Nanay’s
account, mental imagery involves perceptual processing that is not trig-
gered by corresponding sensory stimulation in the relevant sense mo-
dality (see also Pearson et al. 2015; Pearson and Westbrook 2015). So in
this example, your experience is a mixture of sensory stimulation-driven
perceptual processing and mental imagery. The visual experience of your
partner is the result of mental imagery because it was not triggered by cor-
responding sensory stimulation in the relevant sense modality, that is, vi-
sion, but by an auditory stimulus. The fact that perceptual experience is
influenced in this way by mental imagery presents a problem for percep-
tual justification since, unlike perception, mental imagery, by definition,
does not track truth. However, Nanay argues that as long as the underlying
(bottom-up or top-down) mechanisms that construct mental imagery are,
for the most part, reliable, the resulting perceptual experiences can justify
Introduction 11
our beliefs. For example, the (bottom-up) mechanisms involved in the “fil-
ling in” of the blind spot are very reliable since we can only fool the filling
in of the blind spot in exceptional circumstances and only in monocular
vision. So, on this view, you are justified in believing that your partner (and
not someone else) is coming down the stairs when you hear the footsteps,
as long as the underlying mechanism that constructs your visual mental
imagery is reliable. In the next section we discuss another way inaccurate
perceptual experiences can nevertheless be reliable.
I.3. Representation and Accuracy
Sensory reinforcement and discounting occur not only between sense

modalities but also within a single modality. For example, with regard to
haptic perception (which is achieved through the active exploration of
the world by a moving subject) an object that feels larger than another
object also feels heavier. One way to frame the impact non-visual sen-
sory experiences have on the epistemic status of our perceptual beliefs is
in terms of the accuracy of our sensory experiences as opposed to other
equally important epistemic properties of sensory experiences such as co-
herence or consistency. This is indeed the approach taken by Fulkerson
(this volume), whose discussion focuses on the following question: What
impact do sensory interactions have on the epistemic quality of our per-
ceptual experiences? The kind of “epistemic quality” Fulkerson is primarily
concerned with in his chapter is the likelihood of our basic beliefs being
true given the nature of the perceptual experiences that ground them.
Another way to approach the epistemic status of perceptual belief based on
non-visual sensory experiences is by looking closer at what experiences rep-
resent (for a detailed discussion of representationalism and representational
content, see Chalmers 2004). Consider, for example, olfactory experiences.
Do olfactory experiences represent properties of ordinary objects (e.g., trash
cans, cloves of garlic, or a carton of milk that has gone sour)? Or do they
simply represent properties (if they represent at all) (see, e.g., Tye 1995)?4
4 Burge (2010) argues that chemical senses like olfaction and gustation do not represent at all.
According to Burge, this is because the chemical senses do not compute perceptual constancies (the
equivalent of size, shape, and color constancies in the case of vision). Smell chemicals appear to mix
before being detected by the olfactory bulb. So, unlike vision that adjusts for, say, differences in illu-
mination, olfaction doesn’t seem to adjust for smell contamination by background smells. The same
goes for taste. If you just brushed your teeth, the chemical residue in your mouth is going to heavily
influence your taste perception of, say, wine.
By comparison to visual experiences, which are typically thought to repre-

sent properties of ordinary objects such as the squareness of an end table
(Schellenberg 2014), it is not clear whether olfactory experiences represent
properties of ordinary objects.
One factor that may make us inclined to treat visual and olfactory experi-
ence differently is that when we visually misrepresent a property or an object,
the experience is typically thought to be illusory or hallucinatory. But this is
not the case for olfactory experience. Unlike visual properties, smells need
not be bound to objects for the experience to be accurate. We can, and often
do, represent a smell without representing it as being a property of an ordi-
nary object. When you see lavender, for example, you represent the visual
properties of lavender. But when you smell lavender in the air, your olfac-
tory experience need not represent the properties of any particular object.
Suppose, for example, that you used a lavender spray to freshen up the room.
In this case, there is no particular object in the room that smells like lavender;
it is the air that smells like lavender. So it may seem that, at least in such cases,
olfactory experience does not represent an object (e.g., in the sense that it
does not represent a portion of space as being lavender-y). However, it may
be argued that olfactory experiences in this case represent “bare olfactory
objects,” that is, objects that are represented as having no properties other
than the olfactory properties and locations (see Mendelovici, this volume).
There are, of course, non-bare objects that have olfactory properties.
Consider, for example, a lavender candle. In this case, we might want to at-
tribute the olfactory property to the candle and say that the candle smells
like lavender (see, however, Batty 2010b). But if this is the case, then the
olfactory experience does indeed represent an object. Nevertheless an im-
portant question arises here: What is the nature of olfactory experience?
Representationalists about olfaction maintain that olfactory experiences
represent mind-independent physical properties (such as molecular prop-
erties of objects), while relationalists about olfaction maintain that olfactory
properties are relations between olfactory objects and perceivers (for discus-
sion, see Batty 2010a, 2010b, 2010c).
Angela Mendelovici (this volume) defends a primitivist account, ac-
cording to which olfactory experiences represent their objects as having
primitive olfactory properties, which are sui generis, categorical, non-
relational, and non- mental. Mendelovici’s account is an error theory
because it maintains that objects do not have the primitive properties
that olfactory experience normally represents. On this view, olfactory
Introduction 13
experience systematically misrepresents the world around us, but it does so

reliably, that is, olfactory experiences misrepresent in the same way in sim-
ilar circumstances. As a result, olfactory experiences can play an important
justificatory role: they can be used to form true and justified beliefs. In other
words, although the content of the olfactory experience5 is false, the content
of the belief that is inferred on the basis of the falsidical olfactory experience
can be both true and justified. For example, we often use olfactory experi-
ence to determine the freshness of food. In these cases, an olfactory experi-
ence can lead to true and even justified beliefs about the freshness of food
despite the fact that it does not have the primitive olfactory properties we
misrepresent it as having.
Another related question is that of whether the knowledge that an object is
perceived through a certain modality is primitive or derivative. For example,
we typically seem to know whether we have seen or heard something. The
question is whether we know that we have heard something immediately or
we know it because loudness and not brightness is presented in our experi-
ence. How we answer these questions have consequences not only with re-
spect to the criteria we typically use to distinguish among different sensory
modalities, but also with respect to epistemic justification.
I.4. Audition and Aspect Perception
Auditory experiences pertain to the sense of hearing. As with visual experi-

ence, auditory experience seems to be localized. For example, we typically
hear sound as coming from a certain direction. One hears the ringing of one’s
phone as coming from a certain location, say the table on which the phone has
been placed. One question that arises is whether we hear sounds or merely
their properties, echoes, or sound sources. Like most questions in philos-
ophy, this is a controversial one. While some argue that auditory experience
represents sound events, which are taken to be worldly events as opposed to
sounds waves, objects, or properties (O’Callaghan 2007, 2010, 2011), others
argue that empty space is also represented in auditory perceptual content
5 It has been argued on previous occasions that olfactory experiences have contents (Batty 2010a,
2010b, 2010c). Mendelovici (this volume) argues that the contents of olfactory experience are sin-
gular and propositional. A state is said to have singular contents just in case it involves a particular
object, whereas a state is said to have propositional contents just in case it states that something is
the case.
(Young 2016),6 or that sound absences can be perceived as such (Sorensen

2009; Farennikova 2013).
A topic that has received little attention in the philosophical literature, de-
spite its pervasiveness, is aspect perception. Historically, a distinction has
been made between sensation and perception and the corresponding sen-
sational properties and representational properties of experience (Peacocke
1984). Representational properties are thought to be properties an expe-
rience has in virtue of features of its representational content (Tye 1995;
Chalmers 2004); sensational properties are thought to be properties an expe-
rience has in virtue of some aspect (other than its representational content)
of what it’s like to have that experience (Peacocke 1984; see also Block 2003;
Chalmers 2004; Burge 2010).
In vision, aspect perception is typically discussed in relation to, among
other things, ambiguous figures and gestalt grouping phenomena (Peacocke
1984; Merleau-Ponty 2012; for a discussion on gestalt phenomena, see also
Koffa 1935). Ambiguous figures can be experienced as different things
depending on the aspects that are attended. For example, an ambiguous
figure of a duck-rabbit can be experienced as a duck but also as a rabbit. Here
the representational properties of our experience are the same regardless
of whether we experience a duck or a rabbit, but the sensational properties
seem to differ, which is supposed to account for why we can experience the
same figure either as a duck (at one time) or as a rabbit (at another time).
Although aspect perception has been associated with sensational properties,
we can avoid the contentious topic (e.g., questions about the existence or na-
ture of sensational properties) by contrasting aspect perception with repre-
sentation in the sense that seeing-as (e.g., a duck or a rabbit) “outruns,” so to
speak, what is visually represented.
The duck-rabbit example illustrates that aspect-flipping is fairly straight-
forward in visual experience: we experience the image as “flipping” from
a duck to a rabbit. However, it is not entirely clear how this aspect-flipping
translates to audition. Is aspect-flipping in audition similar to aspect-flipping
in vision? Do we hear something sound-as one thing rather than another? Or
is aspect-flipping different in the case of audition?
William Lycan (this volume) examines whether the Wittgensteinian
platitudes we accept for the case of seeing- as hold for the auditory
6 Young (2016) argues that experiences of reverberation cannot be reduced to the perception of
hearing a distinct sound, hearing echoes, or hearing a property of a sound, and, therefore, empty
space is represented in auditory perceptual content.
Introduction 15
counterpart. He argues that while the platitudes do, there are two differences
in the plausibility of theoretical claims. First, what Lycan calls, expectation
is the claim that perceiving something under a particular aspect is, at least
in large part, a matter of what the perceiver expects or projects. Second, at-
tention is the claim that perceiving something under a particular aspect is
entirely or at least primarily a matter of selective attention. To test these for
the case of audition, Lycan examines “hearing as” in listening to music and
“hearing as” in understanding speech.
Lycan uses the example of tonal ambiguity in music, the source of all key
modulations, which are crucial to understanding any interesting music.
Tonality is based on divisions of the octave, and the most dramatic case of
ambiguity is the diminished seventh chord. Since a diminished chord can be
resolved in any of four different ways and the resolution is not determined
by what is strictly represented in hearing the chord itself, one’s experience
of it is solely a matter of hearing-as. The tritone illusion, is another example
of auditory ambiguity (Deutsch 1991). When an identical tritone interval
is played in one key, a listener may hear a descending pattern; but when it
is played in another key, the same listener may hear an ascending pattern,
even though the melody does not change when it is transposed from one
key to another.
Lycan argues that expectation is clearly applicable to aspect perception in
music, indeed much more literally than in the visual arts. Music is temporal
and is a performance art. As such, it involves expecting events that have not
yet occurred, and it is precisely those expectations that underlie harmonic
hearing-as. Attention, by contrast, is not applicable to musical aspect percep-
tion. Although empirical studies suggest that seeing-as is an attentional phe-
nomenon, Lycan argues that hearing-as in music is not: we don’t normally
focus on different aspects of notes or chords even when a particular chord is
sustained for a long period of time.
Lycan then considers how aspect perception extends to listening to
speech, specifically in cases of routine, automatic disambiguation and also
in interpreting illocutionary force. As an unusual case, Lycan argues that al-
though the illocutionary forces conveyed by actors in a play are not real but
only pretended, we nevertheless indelibly hear their utterances as having
those forces. Force ambiguities cannot be explained entirely in terms of lin-
guistic representation; aspect perception is irreducible there. Attentional
differences may play a role in explaining the linguistic ambiguities, but,
Lycan argues, they cannot be the whole story.
Vision also differs from touch. John Campbell (this volume) argues that
while vision has only an objective aspect (e.g., we see the objects out there),
touch has both a subjective and an objective aspect. For example, one can
feel a prickling sensation as well as a pointed object out there. The former
is a subjective aspect, while the latter is an object aspect of touch. In light of
this distinction, how can the claim that experience gives knowledge of our
surroundings be understood in the case of touch experience? Unlike vision,
one can hardly deny that there are tactile sensations, that is, a bodily sensa-
tion one gets when a needle is injected in your arm. The question then is: Do
tactile sensations play an essential role in grounding our knowledge of the
world? Campbell argues that even if we were to accept that the subjective as-
pect of touch is intrinsically spatial, we can still ask whether the knowledge
we have from the objective aspect of touch is grounded in the knowledge that
we have from its subjective aspect.
I.5. Epistemic Privilege and Perceptual Experience
As noted in section I.1, introspective beliefs are typically treated as being ep-
istemically privileged. There is, however, a sense in which we can talk about
certain types of perceptual experience as being epistemically privileged
compared to other types. For example, tactile experience seems to enjoy an
epistemic advantage over the other sensory experiences. Descartes (1998,
chap. 1, para. 6), for example, famously said, “Of all our senses, touch is
the one considered least deceptive and the most secure.” Why is that? Does
the difference lie in the fact that in tactile experience the object is in con-
tact with the body? Not necessarily. Physical contact can give rise to illusory
experiences. For example, when you are touched on two spots on your finger
after seeing an image of a smaller version of your hand, you experience the
distance between the two tactile stimuli as relatively smaller (Taylor-Clarke
et al. 2004).
An alternative explanation of what makes tactile perception epistemically
privileged is that tactile perception gives us special access to the objects and
properties in the world (Vignemont and Massin 2015). Olivier Massin and
Frédérique de Vignemont (this volume) argue that what gives tactile experi-
ence its epistemic privilege is not that it gives us better access to objects and
their properties than, say, vision. Rather, its epistemic privilege is attributable
to the fact that tactile experience presents us, under certain conditions, with
Introduction 17
the objects as being real, meaning that the objects of tactile perception are
presented to us as existing independently of us.
Massin and Vignemont (this volume) defend what they call the muscular
thesis. They characterize this thesis as the claim that “only the experience
of resistance to our motor efforts, as it arises in effortful touch, presents us
with the independent existence of some causally empowered object” ([page
no. 165]). The idea is that the epistemic privilege of tactile perception is typ-
ically described as involving some voluntary effort and that resistance to
our motor effort makes us aware of objects as existing independently of us.
This sort of effortful tactile perception provides us with exclusive access to
mind-independent physical objects. Compare, for example, the resistance
to motor effort you typically experience when you come into contact with
physical objects as opposed to holograms: in the latter case the lack of re-
sistance indicates the absence of a physical object. Massin and Vignemont
then consider two potential counterexamples to the muscular thesis. First,
they ask whether tactile perception provides the appearance of the mind-
independency not only of external reality, but also of the body, thus leading
to a breaking down of the unity of the body and the self. However, they show
a fundamental difference between the two: contrary to the sense of external
resistance, the sense of bodily resistance is rare. Second, they analyze the syn-
drome of depersonalization, in which patients no longer feel the world and
their body as being real and yet still experience resistance. They propose that
what is missing may rather be a fundamental sense of the self or a sense of
presence, to be distinguished from the sense of mind-independency.
I.6. The Epistemology of Event Perception
In ordinary experience, objects remain stable over time. For example, when
you stare at a banana for a few minutes, the banana appears to remain stable
over time. Events, by contrast, unfold in time. For example, the meeting with
your colleagues seems to unfold in time as we move through the items on
the agenda. In both cases, however, we seem to perceive duration. For ex-
ample, when you hear the soprano at the opera who hits a high C and holds
it, it comes to seem to you that you are experiencing its duration (Kelly 2004;
Brogaard and Gatzia 2015a). So how do we perceive durations?
According to the specious present theory, we experience durations be-
cause the present of which we are directly aware is somewhat extended in
time. It, therefore, requires that we are either directly aware of the duration
or that, at least for moments close to the present, we have direct perceptual
access to the recent past and near future (James 1890/1981; Treisman 1963;
Phillips 2008). The problem with this proposal is that while we seem to have
direct awareness of present events, for example, the sound of your phone
ringing, we do not have direct awareness of recent past events, for example,
the sound of your phone a few minutes ago, or recent future events, for in-
stance, the sound of your phone before it rings (Kelly 2004).
According to the retention theory, duration is experienced as a series of
snapshots (Husserl 1964; Kelly 2005). The snapshots of which we are directly
aware are augmented by states of retention and protension, which are not
memories in the standard sense. That is, we do not infer the perception of
duration from long-term or short-term memory. One objection to the re-
tention theory is that it does not explain the perception of duration; it merely
provides labels for that which we need to explain (Kelly 2004).
Carrie Figdor (this volume) defends an empirically based account of
experiences of duration according to which these experiences are derived
from event segmentation processes and the subsequent triggering of mid-
range interval timing mechanisms at a variety of timescales. Her account
augments the event segmentation theory, a leading theory of how we seg-
ment continuous activity to generate perceptions of discrete events (e.g.,
Zacks and Tversky 2001), with the prominent pacemaker- accumulator
model of interval timing. This account provides fresh insight into the phe-
nomenology of experiences of duration in terms of change in quality as well
as change in subjectively experienced speed of the passage of time. It also
entails that duration experiences are infused with top-down influences.
Nevertheless, Figdor argues that the problem of cognitive penetrability
is not a problem once the complexity of information processing is fully ac-
knowledged. Event duration processing violates a routing presupposition of
the debate: while tradition holds that information and justification flows from
perception to perceptual belief, cognitive penetratability holds that informa-
tion flow can also go the other way, creating a vicious circle of justification
flow. Duration perception and duration beliefs appear to have endogenous
interval timing mechanisms as their common cause, with top-down infor-
mation influencing the timing mechanisms. Figdor suggests that the viola-
tion of the required routing assumption reveals a general problem: the more
we discover about complex information-processing systems, the less reason
we have to think epistemic features can or should be assigned to processing
Introduction 19
stages. This suggests a need to distinguish epistemic (personal) and cognitive

(subpersonal) notions of belief and perception.
I.7. Emotional Experience and Interoception
Emotions seem to affect how we experience things. For example, if you had
an exceptionally busy and difficult day at work and you are irritated, you may
experience the loud voices outside your office in a negative way; but if you
had a great day at work and you are feeling happy, you may experience the
same loud voices in a positive way. As in the case of aspect perception, the
stimuli, that is, the voices, in these two scenarios are the same but your au-
ditory experience of them differs in each case. What accounts for the two
auditory experiences is not the auditory input but your emotional state. The
putative effects of emotions on perceptual experience present a challenge to
the notion of perceptual justification. Can our perceptual experiences pro-
vide justification for our corresponding beliefs if they are influenced by our
emotional states?
In order to answer these questions it may be helpful to step back and
have a closer look at the nature of emotions and their expression. Rebecca
Copenhaver and Jay Odenbaugh (this volume) argue that basic emotions
such as joy, sadness, surprise, anger, disgust, and fear are experiences with
representational content that represents both exteroceptive and interocep-
tive properties.
This view can be contrasted with the somatic theory of emotions, ac-
cording to which emotions are simply perceptions of one’s bodily states, pri-
marily because the phenomenology of emotions is, at least in part, directed
to the world outside our bodies (James 1884; Prinz 2004a, 2004b; Brogaard
2015). For example, the anger you feel when you find out that someone
dented your new car is directed outward toward whoever dented your new
car. Copenhaver and Odenbaugh reject the somatic theory on the grounds
that it commits us to saying that “emotions are representations of bodily states
that are representations of core relational themes” ([page no. 219]), such as
wanting what someone else has (the theme) for envy (the emotion) or having
experienced an irrevocable loss (the theme) for sadness (the emotion); but
since transitivity is not guaranteed in intensional contexts, an emotion may
represent a bodily state, and that bodily state may represent properties in the
environment without the emotion itself representing these properties.
Emotions, on Copenhaver and Odenbaugh’s account, are experiences

that are expressed in humans and other animals through various embodied
states. In this sense, emotions are objects of experience. As they put it, “Our
perceptual systems are more or less sensitive to the expression of emotion
in our environment by features that indicate and have the function of indi-
cating emotions” ([page no. 213]). Fear, for example, may be expressed in
anxious behavior or trembling. An embodied state such as trembling, for in-
stance, expresses an emotion that is experienced as a feature of the world. For
example, you know that your cat is excited when you let her play with her
toy mouse because of the way she plays with it. Your experience of your cat
playing with her toy mouse represents her as excited. In other words, you ex-
perience your cat’s excitement as a feature of the world.
One application of Copenhaver and Odenbaugh’s account is in artworks.
Like humans and animals, artworks can express emotions. Given the inherent
differences between organisms such as humans and artworks, the question
that arises is how artworks express emotions, given that they cannot, say, be
excited or afraid. Copenhaver and Odenbaugh defend what they call “minimal
contour theory,” that is, an impersonal account of expression in artwork, ac-
cording to which an artwork may express an emotion in the absence of a (real
or imagined) person to whom the emotion is attributable and whose emotion
constitutes successful expression. Artworks, on this account, express emotions
because they have the function of indicating these emotions. The mechanisms
that ground indication (the function of indicating certain emotions) in art-
work tend to be cultural. For example, the color yellow is expressive of hap-
piness in the United States but expressive of sadness in Egypt. This account
does not require that artworks express emotions in virtue of a viewer, and, it is
argued, it therefore has more explanatory power than alternative theories.
One way to discuss the putative effects of emotions on our perceptual
experiences is in relation to interoception, our sense of the inner or vis-
ceral body, such as feeling a tension in the throat. Emotions depend on
interoception. For example, if you are angry about losing an election to an
inferior candidate, your experience includes a variety of bodily and behavior
changes as well as a cognitive understanding. Your heart rate increases, blood
flushes to your face, your breathing becomes more difficult, you raise your
voice or use harsh words. It is on the basis of interoception that you are able
to classify your emotional state as one of anger rather than of surprise or joy.
All these elements form the emotion of anger. Neither one of them on its own
is sufficient to evoke a feeling of anger.
Introduction 21
Given that emotions depend significantly on interoception, the question

arises whether interoception can influence (exteroceptive) perceptual con-
tent, and if so, what consequences this may have with respect to perceptual
justification. Suppose that you hear loud noises outside your office while you
are trying to finish a paper that is due in a few hours. You feel pressed and
anxious because of the impending deadline (case A). Suppose now that you
hear the same loud noises after you received a prestigious award and you are
feeling happy and excited (case B). Could the differences in interoception af-
fect the content of your auditory experience, that is, might it affect how you
experience the loud noises in cases A and B?
Lana Kühle (this volume) argues that empirical evidence supports the
claim that moods and emotions influence the content of conscious per-
ception. For example, depression causes a significant reduction in contrast
grain, which correlates with subjective reports by depressed individuals who
describe the world as looking gray (Bubl et al. 2010). Anxiety has also shown
to affect color contrast perception (Hosono et al. 2014), while feelings of love
make distilled water taste sweeter (Chan et al. 2013), and so on. Kühle argues
that such empirical evidence suggests that various interoceptive states, other
than those associated with moods and emotions, have been shown to af-
fect how we perceive the world. For example, feeling thirsty can cause you
to perceive a glass of water as being taller than it is, and being in a state of
fatigue can cause you to perceive a hill as being steeper than it is (Zadra and
Clore 2011).
The two main types of influences on sensory perception are cognitive pen-
etration and multisensory integration (Brogaard and Gatzia 2015b; Brogaard
and Chomanski 2015; Brogaard and Chudnoff 2018). Cognitive penetration
involves putative semantic influences of cognitive states such as beliefs on
perception (Pylyshyn 1999). Suppose, for example, that Kelly has adopted
a variety of attitudes associated with religious bigotry, but Steven has not.
The question of cognitive penetration is whether Kelly’s perceptual experi-
ence of, say, a Muslim woman wearing a hijab might differ from Steven’s as
a result of Kelly’s religious bigotry. If cognitive penetration occurs, then we
should expect Kelly’s perceptual experience of a woman wearing the hijab
to differ qualitatively from Steven’s. The question is whether emotions could
have similar influences. Kühle (this volume) argues that it is intuitive to think
of emotions (as opposed to moods, for example) as cognitive states and to
accept that emotions influence our perceptual experience. However, unlike
other cognitive states such as beliefs, emotions are felt. Since one may not
realize that one is feeling stressed or anxious, the influence of emotions is

best described, not as cognitive, but as interoceptive. So the influence in this
case is best described as interoceptive.
Multisensory integration involves a representational integration across
different modalities. Multisensory integration occurs when the experi-
ence associated with one sense (say, auditory experience) is influenced by
input from another (say, a visual input), as is the case of the McGurk illusion
(McGurk and MacDonald 1976). If emotions are understood as states of the
inner body along with certain sensations and state changes, then instances
of emotions affecting perceptual experience may be explainable on the basis
of multisensory integration (assuming that interoception is itself a sensory
modality).
Both of these theoretical possibilities—emotional penetration and mul-
tisensory integration—indicate that interoception might have epistemic
consequences. For example, our perceptual beliefs may not be veridical if the
experiences upon which they are based are influenced by emotions (Siegel
2017). In her contribution to this volume Kühle suggests that if the current
view of perception (as a relation between sensory systems and the environ-
ment) needs to be revised, we may want to rethink how to rationally ground
perception and perceptual justification. If interoception contributes affective
aspects to perception, a theory of perceptual justification should account for
this contribution in some way.
Emotions can play a role similar to that of perceptual experience insofar
as they may give rise to evaluative beliefs. For example, the fact that moral
beliefs are typically accompanied by motivations to act in certain ways can be
explained on the assumption that moral beliefs are influenced by emotions,
which unlike beliefs can be inherently motivational. Psychopathic individuals,
for instance, who lack the ability to feel empathy have difficulties forming their
own moral beliefs and grasping moral concepts such as “right” or “wrong.”
When they do, they rely on the testimony of others (Casey et al. 2013).
Esteeming emotions, such as fondness, loving, or caring for, are directed
toward those we admire or respect. Such emotions have traditionally been
taken to represent the objects they are directed at as being valuable, perhaps
by increasing our well-being or the well-being of those who are close to us
(Prinz 2004a, 2004b). Jennifer Matey (this volume) argues that one’s moral
character is a trait that allows one to become the object of esteem. For ex-
ample, we like caring people more than we like uncaring, selfish people. One
can, therefore, become a better perceiver of virtuous character by becoming
Introduction 23
better at seeing it as having the qualities that are distinctive to it. Since virtues
are distinguished by the fact that they have positive value, representing vir-
tuous traits as valuable is one way of perceiving them as being valuable.
Assuming that virtuous character traits are grounded in the material features
of a person and the behavioral manifestation of these traits, one way of be-
coming a better perceiver is to develop our responses of esteem in such a way
as to track these material features. Matey further argues that cultivating a
virtuous character can make us more sensitive with respect to tracking such
material features, which in turn become the cognitive base for the appro-
priate emotional response of esteem. On Matey’s account, therefore, having
a virtuous character makes us both better perceivers (since we become more
sensitive in tracking the material features that ground virtuous character
traits) and better knowers of persons.
I.8. Conclusion
This volume provides novel approaches to questions about justification with

regard to non-visual perception. Our hope is that it will help further the dis-
cussion of, and generate greater interest in the, epistemology of non-visual
perception, which is currently an underexplored topic in philosophy. The va-
riety of different domains in which conscious non-visual experience plays a
rational role that this volume explores provides a platform for a promising
inquiry within the fields of epistemology and perception.
References
Alston, W. 1971. Varieties of privileged access. American Philosophical Quarterly
8: 223–241.
Alston, W. 1976. Self- warrant: A neglected form of privileged access. American
Philosophical Quarterly 13: 257–272.
Alston, W. 1986. Internalism and externalism in epistemology. Philosophical Topics,
14: 179–221.
Alston, W. 1989. Epistemic Justification: Essays in the Theory of Knowledge. Ithaca,
NY: Cornell University Press.
Audi,R.1998.Epistemology:AContemporaryIntroductiontotheTheoryofKnowledge.London:
Routledge.
Audi, R. 2014. Doxastic innocence: Phenomenal conservatism and grounds of justi-
fication. In C. Tucker, ed., Seemings and Justification: New Essays on Dogmatism and
Phenomenal Conservatism. Oxford: Oxford University Press. 181–201.
Auvray, M., and C. Spence. 2008. The multisensory perception of flavor. Consciousness
and Cognition 17: 1016–1031.
Batty, C. 2010a. A representational account of olfactory experience. Canadian Journal of
Philosophy 40(4): 511–538.
Batty, C. 2010b. Olfactory experience I: The content of olfactory experience. Philosophy
Compass 5(12): 1137–1146.
Batty, C. 2010c. Olfactory experience II: Objects and properties. Philosophy Compass
5(12): 1147–1156.
Block, N. 2003. Mental paint. In M. Hahn and B. Ramberg, eds., Reflections and
Replies: Essays on the Philosophy of Tyler Burge. Cambridge, MA: MIT Press. 165–200.
BonJour, L. 1985. The Structure of Empirical Knowledge. Cambridge, MA: Harvard
University Press.
BonJour, L. 1999. The dialectic of foundationalism and coherentism. In J. Greco and E.
Sosa, eds., The Blackwell Guide to Epistemology. Malden, MA: Blackwell. 117–144.
BonJour, L. 2001. Toward a defense of empirical foundationalism. In Michael R. DePaul,
ed., Resurrecting Old- Fashioned Foundationalism. Lanham, MD: Rowman and
Littlefield. 21–38.
Briscoe, R. E. 2016. Multisensory processing and perceptual consciousness: Part I.
Philosophy Compass 11(2): 121–133.
Brogaard, B. 2014. Phenomenal seemings and sensible dogmatism. In C. Tucker, ed., Seemings
and Justification: New Essays on Dogmatism and Phenomenal Conservatism. Oxford:
Oxford University Press. 270–289.
Brogaard, B. 2015. On Romantic Love. New York: Oxford University Press.
Brogaard, B. 2018. Phenomenal dogmatism, seeming evidentialism and inferential justifi-
cation. In Kevin McCain, ed., Believing in Accordance with the Evidence: New Essays on
Evidentialism. Cham: Springer.
Brogaard, B., and B. Chomanski. 2015. Cognitive penetrability and high-level properties
in perception: Unrelated phenomena? Pacific Philosophical Quarterly 96: 469–486.
Brogaard, B., and D. E. Gatzia. 2015a. Time and time perception. Topoi 34(1): 257–263.
Brogaard, B., and D. E. Gatzia. 2015b. Is the auditory system cognitively penetrable?
Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2015.01166.
Brogaard, B., and E. Chudnoff. 2018. Multisensory consciousness and synesthesia. In R.
Gennaro, ed., Routledge Handbook of Consciousness. New York: Routledge.
Bubl, E., E. Kern, D. Ebert, M. Bach, and L. T. van Elst. 2010. Seeing gray when feeling
blue? Depression can be measured in the eye of the diseased. Biological Psychiatry
68: 205–208.
Burge, T. 2003. Perceptual entitlement. Philosophy and Phenomenological Research
67(3): 503–548.
Burge, T. 2010. Origins of Objectivity. Oxford: Oxford University Press.
Byrne, A., and D. R. Hilbert. 2003. Color realism and color science. Behavioral and Brain
Sciences 26: 3–21.
Casey H., R. D. Rogers, T. Burns, and J. Yiend. 2013. Emotion regulation in psychopathy.
Biological Psychology 92: 541–548.
Chalmers, D. J. 2003. The content and epistemology of phenomenal belief. In Quentin
Smith and Aleksandar Jokic, eds., Consciousness: New Philosophical Perspectives.
Oxford: Oxford University Press. 220–272.
Chalmers, D. J. 2004. The representational character of experience. In B. Leiter, ed., The
Future for Philosophy. Oxford: Oxford University Press. 153–181.
Introduction 25
Chan, K. Q., E. M. Tong, D. H. Tan, and A. H. Koh. 2013. What do love and jealousy taste
like? Emotion 13(6): 1142–1149.
Chisholm, R. 1977. Theory of Knowledge. Englewood Cliffs, NJ: Prentice-Hall.
Chisholm, R 1981. The First Person. Minneapolis: University of Minnesota Press.
Chisholm, R. 1980. A version of foundationalism. In P. A. French, T. E. Uehling Jr., and
H. K. Wettstein, eds., Midwest Studies in Philosophy 5. Minneapolis: University of
Minnesota Press. 543–564.
Chudnoff, E. 2012. Presentational phenomenology. In S. Miguens and G. Preyer, eds.,
Consciousness and Subjectivity. Frankfurt am Main: Ontos Verlag. 51–72.
Chudnoff, E. 2014. Review of C. Tucker, ed., Seemings and Justification. Notre Dame
Philosophical Reviews. https://ndpr.nd.edu/news/seemings-and-justification-new-
essays-on-dogmatism-and-phenomenal-conservatism/.
Chudnoff, E. 2016. Epistemic elitism and other minds. Philosophy and Phenomenological
Research. https://doi.org/10.1111/phpr.12308.
Conee, E. 2014. Seeming evidence. In C. Tucker, ed., Seemings and Justification: New
Essays on Dogmatism and Phenomenal Conservatism. Oxford: Oxford University
Press. 52–70.
Conee, E., and R. Feldman. 2001. Internalism defended. American Philosophical Quarterly
38(1): 1–18.
Crisinel, A. S., and C. Spence. 2012. A fruity note: Crossmodal associations between odors
and musical notes. Chemical Senses 37(2): 151–158.
Damasio, A. R. 1994. Descartes’ Error: Emotion, Reason, and the Human Brain. New York:
Putnam Berkley Group.
Davidson, D. 1986. A coherence theory of truth and knowledge. In E. LePore, ed., Truth
and Interpretation: Perspectives on the Philosophy of Donald Davidson. Oxford: Basil
Blackwell. 307–319.
Deroy, O., and C. Spence. 2013. Crossmodal mental imagery. In S. Lacey and R. Lawson, eds.,
Multisensory Imagery. New York: Springer. https://doi.org/10.1007/978-1-4614-5879-1_9.
Descartes, R. 1641/1984. Meditations on First Philosophy. In The Philosophical Writings
of Descartes. Trans. J. Cottingham, R. Stoothoff, and D. Murdoch, vol. 2. Cambridge:
Cambridge University Press.
Descartes, R. 1998. Descartes: The World and Other Writings. Ed. S. Gaukroger.
Cambridge: Cambridge University Press.
Deutsch, D. 1991. The tritone paradox: An influence of language on music perception.
Music Perception 8: 335–347.
Dretske, F. 1981. Knowledge and the Flow of Information. Cambridge, MA: MIT Press.
Farennikova, A. 2013. Seeing absence. Philosophical Studies 166: 429–454.
Foley, R. 1983. Epistemic conservatism. Philosophical Studies 43: 165–182.
Fumerton, R. A. 1995. Metaepistemology and Skepticism. Lanham, MD: Rowman and
Littlefield.
Fumerton, R. A. 2001. Classical foundationalism. In M. DePaul, ed., Resurrecting Old-
Fashioned Foundationalism. Lanham, MD: Rowman and Littlefield. 3–20.
Goldman, A. I. 1979. What is justified belief? In G. Pappas, ed., Justification and Knowledge.
Dordrecht: Reidel. 1–25.
Goldman, A. I. 1986. Epistemology and Cognition. Cambridge, MA: Harvard University Press.
Greco, J. 2003. Knowledge as credit for true belief. In M. DePaul and L. Zagzebski,
eds., Intellectual Virtue: Perspectives from Ethics and Epistemology. Oxford: Oxford
University Press.
Harman, G. 1986. Change in View. Cambridge, MA: MIT Press.

Harrar, V., B. Smith, O. Deroy, and C. Spence. 2013. Grape expectations: How the propor-
tion of white grape in champagne affects the ratings of experts and social drinkers in a
blind tasting. Flavour 2: 25. https://doi.org/10.1186/2044-7248-2-25.
Hosono, Y., K. Kitaoka, R. Urushihara, H. Séi, and Y. Kinouchi. 2014. Anxiety affects the
amplitudes of red and green color-elicited flash visual evoked potentials in humans.
Journal of Medical Investigation 61: 79–83.
Huemer, M. 2001. Skepticism and the Veil of Perception. Lanham, MD: Rowman and
Littlefield.
Huemer, M. 2007. Compassionate phenomenal conservatism. Philosophy and
Phenomenological Research 74(1): 30–55.
Huemer, M. 2014. Phenomenal conservatism uber alles. In C. Tucker, ed., Seemings and
Justification: New Essays on Dogmatism and Phenomenal Conservatism. Oxford: Oxford
University Press. 328–350.
Hume, D. A. 1738/2012. Treatise on Human Nature. Project Gutenberg Ebook: http://
www.gutenberg.org/files/4705/4705-h/4705-h.htm.
Husserl, E. 1964. The Phenomenology of Internal Time Consciousness. Trans. J. S. Churchill.
Bloomington: Indiana University Press.
James, W. 1890/ 1981. The Principles of Psychology. Cambridge, MA: Harvard
University Press.
James, W. 1884. What is an emotion? Mind 34: 188–205.
Kelly, S. 2004. The puzzle of temporal experience. In A. Brook and K. Akins, eds.,
Philosophy and Neuroscience. Cambridge: Cambridge University Press. 208–238.
Kelly, S. 2005. The puzzle of temporal experience. In A. Brook and K. Akins, eds.,
Cognition and the Brain. Cambridge: Cambridge University Press. 208–238.
Kelp, C. 2011. In defence of virtue epistemology. Synthese 179: 409–433.
Koffa, K. 1935. Principles of Gestalt Psychology. London: Routledge and Kegan Paul.
Lehrer, K. 1990. Theory of Knowledge. Boulder, CO: Westview.
Macpherson, F., ed. 2011. The Senses: Classic and Contemporary Philosophical Perspectives.
Oxford: Oxford University Press.
Mackie, P. 2005. The mystery of direct perceptual justification. Philosophical Studies
126: 347–373.
Merleau-Ponty, M. 2012. Phenomenology of Perception. Trans. D. A. Landes. New York:
Routledge.
McCain, K. 2008. The virtues of epistemic conservatism. Synthese 164 (2): 185–200.
McDowell, J. 1994. Mind and World. Cambridge, MA: Harvard University Press.
McGrath, M. 2018. Looks and perceptual justification. Philosophy and Phenomenological
Research 96(1): 110–133. https://doi.org/10.1111/phpr.12289.
McGurk, H., and J. MacDonald. 1976. Hearing lips and seeing voices. Nature
264: 746–748.
Nozick, R. 1981. Philosophical Explanations. Cambridge, MA: Harvard University Press.
O’Callaghan, C. 2007. Sounds: A Philosophical Theory. Oxford: Oxford University Press.
O’Callaghan, C. 2010. Perceiving the locations of sounds. Review of Philosophy and
Psychology 1: 123–140.
O’Callaghan, C. 2011. Lessons from beyond vision (sounds and audition). Philosophical
Studies 153(1): 143–160.
O’Callaghan, C. 2015a. The multisensory character of perception. Journal of Philosophy
112(10): 551–569.
Introduction 27
O’Callaghan, C. 2015b. Not all perceptual experience is modality specific. In D. Stokes, M.

Matthen, and S. Biggs, eds., Perception and Its Modalities. Oxford: Oxford University
Press. 133–165.
Pappas, G. 2017. Internalist vs. externalist conceptions of epistemic justification. In E. N.
Zalta, ed., Stanford Encyclopedia of Philosophy, Fall 2017 ed. https://plato.stanford.edu/
archives/fall2017/entries/justep-intext/.
Peacocke, C. 1984. Sense and Content: Experience, Thought, and Their Relations. Oxford:
Oxford University Press.
Pearson, J., and F. Westbrook. 2015. Phantom perception: Voluntary and involuntary
nonretinal vision. Trends in Cognitive Sciences 19: 278–284.
Pearson, J., T. Naselaris, E. A. Holmes, and S. M, Kosslyn. 2015. Mental imagery: Functional
mechanisms and clinical applications. Trends in Cognitive Sciences 19: 590–602.
Phillips, I. 2008. Perceiving temporal properties. European Journal of Philosophy
18(2): 176–202.
Prinz, J. 2004a. Gut Reactions: A Perceptual Theory of Emotion. New York: Oxford
University Press.
Prinz, J. 2004b. Embodied emotions. In R. E. Solomon, ed., Thinking about Feeling:
Contemporary Philosophers on Emotions. Oxford: Oxford University Press. 44–60.
Pryor, J. 2000. The skeptic and the dogmatist. Noûs 34(4): 517–49.
Pylyshyn, Z. W. 1999. Is vision continuous with cognition? The case for cognitive impene-
trability of visual perception. Behavioral and Brain Sciences 22: 341–423.
Quine, W. V. O. 1951. Two dogmas of empiricism. In From a Logical Point of View. 2nd ed.
New York: Harper and Row. 20–46.
Ross, P. W. 2001. Qualia and the senses. Philosophical Quarterly 51: 495–511.
Rozin, P. 1982. Taste-smell confusions and the duality of the olfactory sense. Perception
and Psychophysics 31: 397–401.
Russell, B. 1910– 11. Knowledge by acquaintance and knowledge by description.
Proceedings of the Aristotelian Society 11: 108–128.
Russell, B. 1912. The Problems of Philosophy. New York: Oxford University Press.
Russell,B.1913/1 984.TheoryofKnowledge:The1913Manuscript.Ed.ElizabethEames.London:
Allen and Unwin.
Schellenberg, S. 2014. The relational and representational character of perceptual experi-
ence. In B. Brogaard, ed., Does Perception Have Content? New York: Oxford University
Press. 51–75.
Schwenkler, J. 2013. The objects of bodily awareness. Philosophical Studies 162(2): 465–472.
Sheppard, G. 2012. Neurogastronomy: How the Brain Creates Flavor and Why It Matters.
New York: Columbia University Press.
Sheppard, G. 2017. Neuroenology: How the Brain Creates the Taste of Wine. New York:
Columbia University Press.
Shoemaker, S. 1963. Self-Knowledge and Self-Identity. Ithaca, NY: Cornell University Press.
Shoemaker, S. 1990. First person access. Philosophical Perspectives 4: 187–214.
Siegel, S. 2012. Cognitive penetrability and perceptual justification. Noûs 46(2): 201–222.
Siegel, S. 2017. The Rationality of Perception. Oxford University Press.
Silins, N. 2011. Seeing through the “veil of perception.” Mind 120(478): 329–367.
Small, D. 2012. Flavour is in the brain. Physiology and Behavior 107(4): 540–552.
Smith, A. D. 2002. The Problem of Perception. Oxford: Oxford University Press.
Smithies, D. 2012. A simple theory of introspection. In D. Smithies and D. Stoljar, eds.,
Introspection and Consciousness. New York: Oxford University Press. 259–294.
Sorensen, R. 2009. Hearing silence: The perception and introspection of absences. In M.

Nudds and C. O’Callaghan, eds., Sounds and Perception: New Philosophical Essays.
Oxford: Oxford University Press. 126–145.
Sosa, E. 1999a. How to defeat opposition to Moore. Noûs, Supplement, Philosophical
Perspectives, Epistemology, 33(s13): 141–153.
Sosa, E. 1999b. Skepticism and the external/internal divide. In J. Greco and E. Sosa, eds.,
The Blackwell Guide to Epistemology. Malden, MA: Blackwell. 145–157.
Sosa, E. 2007. A Virtue Epistemology: Apt Belief and Reflective Knowledge. Vol. 1.
Spence, C. 2015. On the psychological impact of food colour. Flavour 4: 21. https://doi.
org/10.1186/s13411-015-0031-3.
Stevenson, R., and C. Tomiczek. 2007. Olfactory-induced synesthesias: A review and
model. Psychological Bulletin 133: 294–309.
Steup, M. 1996. An Introduction to Contemporary Epistemology. Upper Saddle River,
NJ: Prentice Hall.
Swinburne, R. 2001. Epistemic Justification. Oxford: Oxford University Press.
Taylor-Clarke, M., P. Jacobsen, and P. Haggard. 2004. Keeping the world a constant
size: Object constancy in human touch. Nature Neuroscience 7(3): 219–220.
Tooley, M. 2014. Michael Huemer and the principle of phenomenal conservatism. In C.
Tucker, ed., Seemings and Justification: New Essays on Dogmatism and Phenomenal
Conservatism. Oxford: Oxford University Press. 306–327.
Treisman, M. 1963. Temporal discrimination and the difference interval: Implications for
a model of the internal clock. Psychological Monographs 77: 1–31.
Tucker, C. 2010. Why open-minded people should endorse dogmatism. Philosophical
Perspectives 24(1): 529–545.
Tye, M. 1995. Ten Problems of Consciousness: A Representational Theory of the Phenomenal
Mind. Cambridge, MA: MIT Press.
Vignemont, F. de., and O. Massin. 2015. Touch. In M. Matthen, ed., Oxford Handbook of
Philosophy of Perception. Oxford: Oxford University Press. 294–313.
Williamson, T. 2000. Knowledge and Its Limits. New York: Oxford University Press.
Yuki, H., K. Kitaoka, R. Urushihara, H. Séi, and Y. Kinouchi. 2014. Anxiety affects the
Young, N. 2016. Hearing spaces. Australasian Journal of Philosophy 95(2): 242–255.
Zacks, J. M., and K. M. Swallow. 2007. Event segmentation. Current Directions in
Psychological Science 16(2): 80–84.
Zacks, J. M., B. Tversky, and G. Iyer. 2001. Perceiving, remembering, and communicating
structure in events. Journal of Experimental Psychology: General 130(1): 29–58.
Zadra, J. R., and G. L. Clore. 2011. Emotion and perception: The role of affective informa-
tion. Wiley Interdisciplinary Review of Cognitive Science 2(6): 676–685.
1
Tasting Flavors
An Epistemology of Multisensory Perception
Barry C. Smith
1.1. The Subjective Nature of Tastes and Tasting
Our fundamental contact with the world comes through our senses. It is by
means of the senses that we perceive the world around us and ourselves. We
see objects in our immediate vicinity: their size, their shape, their colors. We
hear the sounds around us: the sound of the traffic, music from the street,
or the planes overhead. We feel the surfaces of objects, the stiffness in our
joints, water running through our fingers. In the morning we smell soap, and
the freshly brewed coffee. In this way, seeing, hearing, touching and smelling
are all cases where we exercise our senses to perceive parts of our immediate
environment. So what happens when we add tasting to this list? What sort of
knowledge does tasting provide?
We might say that tasting enables us to discover the tastes or flavors of foods
and drinks. We taste how sweet the strawberries are, whether there is sugar in
the coffee, and how spicy the soup is. Tasting reveals the sharp tang of lemon,
the creaminess of the hollandaise sauce, the coolness of the menthol lozenge
in the mouth. But what are these taste qualities? Unlike shapes, which can
be both seen and felt, tastes seem to be known to us through the exercise of
a single sensory modality, which makes them secondary qualities par excel-
lence: qualities that are dependent not on objects but on subjects, constituted
and exhausted, perhaps, by what we are aware of in experience. They can be
known only in a direct experiential way and not through testimony or scien-
tific description. As David Hume noted: “We cannot form to ourselves a just
idea of the taste of a pine-apple, without having actually tasted it” (Treatise
on Human Nature, book 1, pt. 1, sec. 1). There is no chemical formula by
means of which to specify to someone who hasn’t tasted a pineapple what a
Barry C. Smith, Tasting Flavors In: The Epistemology of Non-Visual Perception. Edited by: Berit
Brogaard and Dimitria Electra Gatzia, Oxford University Press (2020). © Oxford University Press.
DOI: 10.1093/oso/9780190648916.003.0002
pineapple tastes like.1 These days, it may be objected, pace Hume, that syn-
thetic flavorings could provide us with an idea of what a pineapple tastes like.
But this is only possible by synthesizing odor compounds in pineapple and
adding sweetness and texture to give someone something that tastes like a
pineapple. So you have to either taste a pineapple or taste something that
tastes like a pineapple. The point about direct experience still holds. Tastes
are given to us in experience.
But what are tastes? This is a question many philosophers and sensory
scientists answer by talking about inherent features of our experience: tastes
just are the distinctive kinds of experiences we undergo when we eat and
drink. As sensory scientist Gordon Sheppard puts it: “When we use the word
‘taste’ to apply to wine, we think we know what we are talking about: the ex-
perience of the wine in our mouths” (2017, 106). Typically, we associate these
experiences with a single modality, taste. According to this line of thought,
foods and drinks don’t themselves have tastes; rather they give rise to tastes
in us. We may associate particular food items, a lemon, or a piece of choco-
late, or particular wines, like a sauvignon blanc, with the tastes we get when
consuming them: that’s how they taste to us.
This makes tasting very different from the other senses. For while many
philosophers are willing to accept that sight, hearing, touch (and maybe
even smell) are perceptual senses providing us with knowledge of the enti-
ties around us in space, tasting is here relegated to a bodily sense, telling us
about nothing but ourselves. It involves no engagement with the world and
no perception of it; tasting is just a matter of sensations occurring in us.
For example, A.D. Smith declares: “Taste, as such, is a bodily sensation,
one entirely lacking the kind of spatiality we are concerned with here.
Tastes are in the mouth as headaches are in the head” (2002, 139). The par-
adigm of a taste, for A. D. Smith, is having a taste in one’s mouth, without
any object being in the mouth or felt by the tongue; an aftertaste, for ex-
ample. In this way tastes are not bound to foods and liquids and are only
vaguely localized to the mouth. Although ordinary experience tells us that
tasting is done with tongue, as seeing is done with the eyes, and hearing
with the ears, and it’s from the tongue’s contact with foods that we get taste
sensations. A taste in one’s mouth when there’s nothing in there is a kind of
1 Steven Shapin notes the inability of 18th-century science to capture in words or symbols how a
pineapple tasted in order to convey it to those who wanted to know how this newly discovered, rare,
exotic fruit tasted. This is one of the main reasons why tastes came to be excluded from the scientific
inventory of the world and relegated to subjective experiences.
Tasting Flavors 31
afterimage, an aftertaste as we say, and that’s not a paradigm case of tasting

so much as a result of it.
Although philosophers often take themselves to be defending our common-
sense way of thinking, often as a way of resisting scientific descriptions of the
same phenomena, the current line of thought is at odds with our more ob-
jective, ordinary ways of talking about tastes. We say the soup tastes salty, the
curry is spicy, and the pudding is sweet, thus attributing these qualities to the
foods themselves. However, the supposedly common-sense philosopher may
wish to explain this away by saying that what is really going on is that in tasting
the soup we bring about a kind of experience in us—or most of us do—that
we are immediately aware of, and which we call a salty taste. Similarly, sweet,
bitter, and sour would be identifiable features of our experience and not some-
thing food items can have independently of experience.
If tastes are inherent features of our own experiences, then the tastes I am
getting from a dish are how the dish tastes to me, and there is no more to tastes
that that. The qualification, how something tastes ‘to me,” is what gives rise to
the subjectivity of taste. Tastes are said to be characteristic features of individ-
uals’ experiences that cannot be directly compared. For each of us knowledge
of how something tastes to us is supposed to be unimpeachable. It is consti-
tuted by taste sensations present in consciousness as we sip, chew, and swallow.
They cannot be overlooked and they are just as they appear: occurrences of
distinctive kinds of experience we undergo while eating or drinking.
That’s the view that many philosophers have subscribed to, largely as a
result of an unshakeable faith in reflections on their own experiences as a
proper guide to the nature of tastes and tasting without actually knowing an-
ything about how tasting how works.
As I shall argue, subjectivity about taste is wrong about both the nature of
tastes and our experience of tasting. Rather than being incorrigible, our know-
ledge of what we experience in tasting can be usefully revised by a mixture of
empirical investigation and reflection of cases. I shall set out a very different
epistemology and metaphysics of tastes and tasting, showing how tasting
provides us with the perceptual knowledge that guides successful food choice.
1.2. Taste Sensations Alone?
The simple subjectivism about taste as sensations amounts to the idea that
all tasting provides us with is awareness of a distinctive and proprietary
range features of our own experience. But is this all the knowledge that
tasting provides us with? And how well do we even know our own tasting
experiences?
The trouble with tasting is that it is ephemeral and fleeting. We eat, we
chew, we swallow and the sensations disappear fast, making them hard for us
to concentrate on. Even as we attend, things change rapidly during tasting.
The moment food enters our mouths and makes contact with the tongue we
begin to experience how something taste to us. The distinctive sensations of
tasting appear to be localized to the tongue—an impression that lingers even
when contemplating aftertastes. When we pop a piece of bran muffin into the
mouth and begin chewing it, we experience sweetness but also a malty flavor
and a springy texture. How much of this flavor is released as we chew and
swallow, and in what way does swallowing alter the experience of tasting?
How does the flavor change as the texture is altered by chewing?
The slowing down and attending to the temporal dynamic of tasting
reveals more than is usually picked up on, or noticed, by philosophers talking
about taste. By engaging in careful exploration, it is possible to demonstrate
to oneself that the experiences we usually call tastings involve far more than
the simple sensations from the tongue like sweetness or sourness. Tasting
involves touch and texture, how sticky, creamy, slimy, or crunchy a food is,
and this contributes to what is usually called its taste, or how it tastes to us.
Only in the case of an aftertaste, which is not a central case of tasting some-
thing, is the taste we get from a food not accompanied by the feel of it. Taste
and touch accompany one another in normal tasting experiences. And al-
though we can phenomenologically parse them into their different sensory
experiences or sensations, the experience of tasting doesn’t reveal how taste
sensations are affected by touch sensations. The interaction effects of taste
and touch can be probed experimentally by altering the texture of a food
item. Melted chocolate will taste sweeter to us than it does unmelted, and re-
set melted chocolate will taste different still, even though the flavor profile of
the particular chocolate brand we are using is recognizable throughout.
The next important challenge to recognize—again by experiment—to
the simple subjectivism of taste as sensations on the tongue is how much the
sense of smell contributes to how a food or drink “tastes” to us. If we block
our nose by putting nose clips on and then taste a food—a fruit, or a piece
of cheese—all we will experience are the sensations that give us experiences
of sweetness or sourness, saltiness or bitterness. And it is only when we
remove the nose clip that we experience the full flavor, like the fruitiness,
Tasting Flavors 33
say, of what’s in the mouth. As soon as odors are released in the mouth,
they rise up to the receptors in the olfactory epithelium at the bridge of the
nose. This is retronasal olfaction, which contributes the largest part of the
distinctive experiences we call “tasting” something. What is more, when we
remove the nose clipe, and after that fraction of a second’s delay, the odors
rush in and fuse with the tastes from the tongue, and it suddenly becomes
impossible, phenomenologically, to parse our tasting experience into the
part played by sensations from the tongue and the part played by smells
from the nose: “For quite some time now and in a variety of ways, food
scientists and sensory scientists have noted that people conflate taste and
smell stimuli when presented simultaneously in the mouth and that they
are unable to decompose them even with training, and when instructed
to pay attention to each component” (Stevenson and Tomiczek 2007,
296). They become part of an inseparable fusion that sensory scientists
call flavor, to distinguish it from taste proper. Taste proper—gustation—is
restricted to what the tongue provides, namely sensations of sweet, sour,
salty, bitter, savory (umami), along with metallic and, perhaps, fattiness.
That’s all we get from as a result of dedicated taste receptors on the tongue,
in the epiglottis and in the gut. The different, dedicated taste receptors on
the tongue respond selectively to substances in our foods, giving rise to the
different taste qualities of sweet, sour, salty, bitter, savory, and metallic. But
that’s all they give us. The fruity flavors of melon, mango, raspberry, straw-
berry, pear, peach are due not to the tongue’s response to these foods—
we don’t have peach receptors on the tongue—but are due to smell. It’s the
fruity aroma of peach that combines with the taste of sweetness and slight
sourness of acidity to give us the flavor of a peach.
In fact, our tasting of flavors always involves touch, taste, and smell, and
the failure to recognize this, and the insistence that the relevant experiences
are due to sensations on the tongue alone and due to a single sense modality,
is due to a failure to know our own experiences in full; though the multisen-
sory character of flavors has been recognized from time to time by a few per-
ceptive philosophers, psychologists, and physiologists. Aesthetician Frank
Sibley noted that there was no arithmetic of the basic tastes from the tongue
that could add up to the flavors of coconut or lemon:
Coconut may be somewhat sweet, and lemon sour or acid, but what other
tastes combine with sweetness to give coconut, or with sourness or acidity to
give lemon? How could one construct a blend of distinguishable tastes . . . to
yield that of coconut, or lemon, or mint? Try to imagine a recipe: “To make
the flavour of onion (or pepper, or raspberries, or olives), add the fol-
lowing flavours (not substances) in the following proportions . . . ” (Sibley
2001, 217)
The early 20th-century psychologist E. B. Titchner beautifully captures the

complex sensations we have when eating a peach:
Think, for instance, of the flavour of a ripe peach. The ethereal odor may be
ruled out by holding the nose. The taste components—sweet, bitter, sour—
may be identified by special direction of the attention upon them. The
touch components—the sourness and stringiness of the pulp, the pucker
feel of the sour—may be singled out in the same way. Nevertheless, all these
factors blend together so intimately that it is hard to give up one’s belief in a
peculiar and unanalyzable peach flavour. (Titchner 1909, 135)2
And the nineteenth- century physiologist Jean- Anthelme Brillat- Savarin

declared himself to be “tempted to believe that smell & taste are but a single
sense; whose laboratory is the mouth and whose chimney is the nose” (1825,
41). These authors are aware of the complexity of tasting, even though or-
dinary tasters mostly fail to recognize this experience as a complex interac-
tion effect involving touch, taste, and smell. Ordinary tasters tend to focus,
in the way Titchner indicates, on the whole, unified experience of flavor.
Meanwhile, philosophers often classify these experiences, erroneously,
as unisensory experiences of taste: something they experience as coming
from the tongue. Why so? Ordinary tasters are fully conscious of an array of
flavors, however they suppose these to be detected by the tongue; although,
A. D. Smith, regards paradigm taste experiences as occurring in the mouth
without any food or liquid present. (Intriguingly, these cases of aftertastes
are actually mis-characterizations of what are really after-smells: odors rising
from the gullet to the nose.)
One mistaken reason for supposing that all the flavor components we can
taste come from the tongue may be due to phenomenology and the fact that
it is hard to separate taste and smell. Tasting flavors with the tongue is how
things appear in experience. So are these experiences perceptual illusions?
Some believe they are.
2 Thanks to Janice Wang for drawing this passage to my attention.

Tasting Flavors 35
In a now-classic 1982 article, psychologist and biologist Paul Rozin first

proposed the idea of smell as a dual sense with odors inhaled from the
environment—orthonasal olfaction—and odors arising from the oral cavity
exhaled through the nose—retronasal olfaction. The same molecules “may
be perceived and evaluated in two qualitatively different ways, depending
on whether it is referred to the mouth or the external world” (Rozin 1982,
397). It is retronasal olfaction that contributes so prominently to our expe-
rience of the flavors of the foods and drinks we ingest and imbibe. And yet
the sense of smell seems to go missing in our flavor experiences, at least as far
as recognizing or categorizing the modality of the experience is concerned.
We are well aware of the qualities that smell contributes (though retronasal
olfaction) to flavor experiences; we just don’t recognize that contribution as
originating through smell.3 We conceive what smell provides as a taste (gus-
tation) arising from the tongue. Rozin calls this the “location illusion.” We
experience an illusion in that what is in fact an olfactory experience is expe-
rienced as a taste occurring on the tongue. This is oral referral: the referral
of sensations from one part of the body to another. Referred pain provides a
similar example. Referred sensations “are sensations perceived as emanating
from a body site other than the one stimulated” (Solar et al. 2010, 192). For
example: “Stimuli to the upper arm were sometimes felt in more distal parts
of the limb. Stimuli to the face were localized to the arm and not the hand”
(Turton and Butler 2001, 397). These cases are relatively stable, reflecting
“some change in [the patient’s] somatosensory representation that occurred
with experience” (Turton and Butler 2001, 397). So are cases of oral referral
perceptual illusions? Yes and no. Although patients with referred pains are
certainly feeling pain, which we can take to be an indication of damage to
the body, our perceptions of where the pains are don’t reveal where the actual
damage is taking place. In the case of oral referral, we regularly and in a stable
way perceive fruitiness as coming from the mouth and the tongue, although
we pick up on it via olfactory stimulation in the nose. What is different about
this case is the way that what we are aware of turns up not just in a different
location but a a different modality, classified as a taste, not a smell. Whatever
mechanism explains the location illusion—referral to the tongue—may
also explain why we end up treating the resultant state of awareness as a
taste: that’s what the tongue provides. However, we are still entitled to say,
3 Gordon Sheppard believes that “the smell that contributes to taste is entirely unconscious” (2017,
30). That’s false. We ate entirely conscious of the fruitiness of a peach, which is due to smell. We are
simply not aware that it is smell.
despite this erroneous localization (and classification), that we are aware of a

flavor and count as veridically perceiving it if these really are the flavors the
food has. It is this objectivist view that I shall argue for subsequently. But for
the present, it is important to consider what mechanisms might be respon-
sible for oral referral.
Several sensory scientists believed that oral referral was entirely due to
touch. Since sensations of salty or sour occur when the tongue is in contact
with a food source, touch and taste seem to occur in the same place. The
brain then shifts the perceived location of the flavor dimension provided
through retronasal olfaction so that all dimensions of flavor appear to occur
in the same place, where the food source is. In effect touch binds the olfactory
dimensions of flavor, anchoring it to the tongue.
This is the explanation championed by Linda Bartoshuk (Bartoshuk and
Duffy 2005) and others, but we now know that touch is not enough. The key
to the integration of retronasal olfaction and taste proper (gustation) is the
presence of a tastant—salt, sweet, sour, and so on—that is congruent with the
retronasally sensed odor. Juyun Lim and colleagues have shown that when
odors of lemon are presented to the olfactory epithelium and a sour tastant is
on the tongue, participants experience a flavor; the same is true for chicken
odor with salt or umami. However, if the lemon odor is presented with a salt
tastant in the mouth, or chicken odor with a sweet tastant, participants get
a smell and a taste but not a unified flavor (Lim and Johnson 2012). It is in-
teresting to note that taste and retronasal smell are only congruent when
these combinations of tastants and odors are ecologically valid, belonging
to foodstuffs we encounter and consume, strengthening the idea that flavors
may be out there to be tracked by the multisensory integrated experiences of
flavor. I’ll return to this later.
First, it’s worth noticing how many other sensory systems are involved in
our multisensory experiences of flavor. The “taste” of black pepper or mint
or chili is due not to our taste buds but to the irritation of trigeminal nerve
endings. The trigeminal nerve is the fifth cranial nerve that serves the eyes,
the nose, and the mouth, and it’s the chemical stimulation of trigeminal
nerve endings that makes mustard taste hot and peppermint taste cool in the
mouth even though they are at the same temperature. Eating too much wa-
sabi leads to pain at the bridge of the nose, whereas the capsaicin in chili also
activates local pain receptors on the tongue, leading to a feeling of burning.
These effects are due to chemesthesis: chemical irritation of trigeminal nerve
endings that produce sensations of burning, cooling, or stinging.
Tasting Flavors 37
To these inputs we must add temperature, which has an effect on the per-
ception of sweet and bitter. The colder your coffee gets the more bitter it
tastes. Slight warming of a liquid can enhance its perceived sweetness. Some
people are thermal tasters who perceive ice cubes as sour or salty, and warm
liquids or spoons as sweet. This affects about 25% of the population (Green
and George 2004).
In addition to chemesthesis, somatosensation, temperature, olfaction,
and gustation, multisensory flavor experience can also involve vibration.
Szechuan pepper has an active ingredient called sanshool that activates
the mechanoreceptors in the tongue, leading to the strange feeling where
the tongue feels both numb and tingly. People experience this as a kind of
electricity on the tongue and think it is due to spice. However, the effect of
sanshool is not due to chemesthesis (although hydroxyl alpha sanshool in
Szechuan cuisine does make the mouth feel cool as a result of alpha’s effect on
the trigeminal system); instead it makes the mechanoreceptors in the tongue
vibrate at 50 Hz. This is what gives rise to the tingling sensation in foods sea-
soned with Szechuan pepper (Hagura et al. 2013).
It is not enough just to point to the number of senses that contribute to our
awareness of flavors. The different sensory inputs don’t just combine, they
often affect one another’s workings, giving rise to interesting cross-modal
effects. How things smell can affect how they taste. The aroma of vanilla ex-
perienced outside the mouth orthonasally can affect how sweet something is
perceived to taste on the tongue. This is called the sweetness enhancement ef-
fect (see Cliff and Noble 1990; Prescott 2004; Stevenson and Tomiczek 2007;
Hort and Hollowood 2004; Dalton et al. 2000). Smell can also affect how
creamy something is perceived to be (Bult et al. 2007). How creamy some-
thing is can affect how sweet it tastes to us (Rolls et al. 1999; Rolls 2014).
What we experience in tasting is a unified experience of flavor due to mul-
tisensory integration. Normally, we receive inputs from our senses and expe-
rience the flavor of a dish or a wine as a single, unified whole. We don’t realize
it as an interaction effect, the upshot of multisensory integration, where in-
formation from different senses get fused together.
In tasting, we are usually tasting flavors, not just tastes. If tastes are the
proper sensible of the sense of taste (gustation), and we need to exercise more
than one sense in order to experience flavors, are they common sensibles,
in Aristotle’s sense? No, since we cannot experience flavors by either smell
or taste or touch alone; rather, we need to exercise all three senses at once
in order to experience flavors. Flavors are qualities we need many senses to
experience. No such possibility is available on Aristotle’s framework, and

some resist the postulating of such complex sensory qualities.
Scientists like Charles Spence and Malika Auvray have entertained the
idea that we have a dedicated flavor sense whose proper sensible is flavors
(Auvray and Spence 2008). While some philosophers still cling to the idea
that the unified experiences we are aware of when tasting can be thought of
as the result of exercising the single sense of taste; where the sense of taste
invoked here is identified on purely phenomenological grounds. This uni-
modal idea of tasting loses all connection with underlying explanations of
those experiences as being shaped by different kinds of gustatory, olfactory,
somatosensory stimuli and the activation of the associated receptors.4 This
last-ditch attempt to explain how we ordinarily talk before learning any of
the facts previously mentioned (although not how ordinary people have no
problem accommodating the scientific findings) assumes that our awareness
of a “taste” (flavor) comes to us as a unified, undifferentiated kind of sensory
experience not even associated with the tongue. But such tastings are not
always experienced as unisensory, as more careful reflection on our tasting
experiences shows. There are examples, albeit rare, where we can phenome-
nologically parse our experiences of “taste” (flavor) into its multisensory sen-
sory components.
Consider the “taste” (flavor) of a menthol lozenge. It comprises at least
three things: a minty aroma, a slightly bitter taste, and a cool sensation in
the mouth. When our attention is drawn to these aspects, we are, as a matter
of phenomenology, able to parse this multisensory experience of “taste” (or
flavor) into its unisensory parts. We can easily recognize that were we to
take away any of these components, we simply would not have the “taste”
(flavor) of menthol.5 Therefore, a more comprehensive account of our flavor
experiences is needed: one that accommodates cases like our experience of
menthol, which is congenial to the best scientific explanations of the under-
lying processes that give shape and content to our experiences of flavor, and
it will have to be one that acknowledges the multisensory nature of tasting
experiences.
From what has been said so far, we are able to rule out a simple subjec-
tivism of taste where all that tasting provides us with is knowledge of our
4 This appears to be the position adopted by Richardson 2013.

5 In this case, the activation of the trigeminal nerve by the compound I-menthol, seems to localize
the minty aroma to the airwaves, enabling us to parse the experience into its taste, touch and smell
components.
Tasting Flavors 39
own taste sensations. However, this is not enough to secure the objectivity of
tasting. Complexity in our experience does not in and of itself amount to ob-
jectivity. Further arguments are needed to establish the act of tasting as a case
of perception that can give us knowledge of a food’s or drink’s flavors and not
just knowledge of ourselves.
1.3. Flavors as Constructs of the Brain
A more sophisticated and scientifically respectable subjectivism about taste

now emerges; one that treats tastes or flavors as a complex set of sensations
or rather a construct of the brain that comes into play only when inputs from
taste, touch, smell, and other systems have been integrated into a unified per-
cept. This view of taste or flavor as a psychological construct of the brain is
popular among psychologists and neuroscientists.
As John Prescott puts it in his 2015 article: “The idea of flavours as the out-
come of the integration of tastes, odours and oral somatosensory (tactile)
qualities has a long pedigree. In recent years this concept has received sup-
port from the identification of the brain’s network of neural structures that
function together to uniquely encode flavours.” But notice that in the same
article the author says: “Flavour perception reflects the integration of distinct
sensory signals, in particular odours and tastes, primarily through the action
of associative learning” (italics mine). He is not alone in running together
flavor and flavor perception. Physiologist Gordon Sheppard in his popular
books Neurogastronomy: How the Brain Creates Flavor and Why It Matters
(2012) and Neuroenology: How the Brain Creates the Taste of Wine (2017) tells
us that flavor is not in the food; flavor is in the brain. He is confident enough
to claim, “We now know that the flavor of a wine is created in a consumer’s
brain” (Sheppard 2017, 168). But elsewhere he is more careful: “The brain
creates the perception of taste within the context of interactions between all
the sensory systems to produce the overall perception of flavor” (116, italics
mine). Here is neuroscientist Dana Small, approving of Sheppard’s view but
making exactly the same conflation: “Flavor is in the brain, not the food. It is
the brain that unites the discrete sensory inputs from the food and drinks we
ingest to create flavor perceptions” (Small 2012, 540; italics mine). Notice that
the second sentence, which is unexceptional, does not support the strong
conclusion stated in the first. Flavor is one thing, flavor perception is another.
Sheppard even uses a useful analogy between flavor and faces to point out
that our perceptions of both of them are extremely discriminating though we

struggle to describe them in words. The analogy is useful too in illustrating
the distinction between a perception and what it is a perception of; the brain
creates our perception of faces, but it does not create faces. This is a distinc-
tion to which some neuroscientists seem entirely oblivious.
We could easily say that while the brain creates our perceptions of flavor,
it does not create flavors. Sheppard, himself, even acknowledges that al-
though for him the brain creates the flavor of a wine, “everything in the brain
depends on what is going on in the wine in the mouth” (Sheppard 2017, 2).
That’s right and that’s why there is something we are getting right (or wrong)
when we perceive the flavor of the wine.
Resistance to objectivism about perceived flavors comes from another
analogy. It’s not faces, neuroscientists will say, but colors that provide the best
analogy. Sheppard, like many of his colleagues, is convinced that “color is not
in the object, color is created by the brain” (2017, 65). This, however, is a phil-
osophical thesis, not a scientific finding, and it has been strongly contested
(see Kalderon 2003). However that debate turns out, it is important in the
context of perception to maintain a distinction between colors and color
perceptions, sounds and sound perceptions, flavors and flavor perceptions,
whatever the ontological status entities of the former kind is thought to be.
To sum up so far, if tasting is a case of genuine perception, we need an
account of the objects of perception and how they can be known to us. The
“flavors are in the brain” school of thought owes us an explanation of how its
proponents can distinguish flavors from flavor perceptions, while objectivists
owe an explanation of what objective flavors are. Of course, “flavors in the
brain” theorists could give up on the claim that we perceive flavors and con-
cede that we only ever have multisensory flavor experiences: experiences
whose complexity and multisensory nature is usually hidden from us. The
objection I have to this line of reasoning is that it gives up on an important
feature of the epistemology of tasting: namely, that there can be illusions
where we appear to taste something that isn’t there, or fail to taste a flavor the
foods actually has. It is possible to treat some cases of tasting as more reliable
than others, and certain conditions for tasting as better than others.
The thesis that tasting is a form of perceptual knowledge enables us to
say that tasting sometimes gives rise to veridical perception and at other
times is subject to perceptual illusions. No such distinction is available to
experientialists: for them these are all equally flavor experiences. In the next
section I shall argue for a perceptual epistemology of tasting.
Tasting Flavors 41
1.4. Multisensory Perception and

Sophisticated Subjectivism
It is easy to understand the motivation of neuroscientists who see flavors

as constructs of the brain, for it can seem tempting to say that flavors aren’t
out there and only come into being when the brain puts together the sepa-
rate signals of smell, touch, and taste to create a unified experience of flavor.
Flavor could be seen as synthesis of the information from different sensory
systems: a multisensory experience we have when eating or drinking. But if
that is right, we need to know the extent of those flavor experiences and what
makes them flavor experiences. Just which sensory inputs are combined or
integrated to create flavor experiences: where do they begin and end? This
is an instance of the binding problem: which neural signals get combined
and how do they get combined? For the perceptualist, this amounts to the
problem of saying what constitutes flavor perception and what is merely aux-
iliary and perhaps likely to affect it causally. Let’s look at the sharp end of this
problem.
Psychologist Charles Spence has done more than most to show how many
factors can influence our experiences of flavor: these factors belong to the
external context as well as to the body’s internal milieu. In a series of imag-
inative experiments, Spence has shown how ambient sounds (Spence et al.
2011), music (Crisnel and Spence 2011), changes in lighting conditions
(Spence 2015), color of plates (Harrar et al. 2011), weight of cutlery (Michel
et al. 2015, weight of bowls (Piqueras-Fiszman and Spence 2012), and the
feeling of fabrics in the hand (Spence et al. 2013) can affect how we perceive
flavors. To give just a few examples, Zampini and Spence (2004) showed how
Pringles potato chips that had been left out of the box long enough tasted
stale to participants, although when they ate them while receiving through
headphones the amplified sound of their own crunching, the potato chips
tasted fresh. So should we say there is an auditory component to flavor?
Similarly, in another experiment participants are handed a bowl of yogurt
to taste, and a moment later handed another bowl to try and asked to say
which yogurt is creamier, richer, and thicker. Unbeknownst to them, the yo-
gurt in the two bowls is the same, but at the bottom of one is a small weight.
Participants report one of the bowls as containing creamier, richer, thicker
yogurt, and this is the bowl with the weight in it. Color, too, affects tasters
perceptions of flavor. When liquids are colored with a tasteless, odorless
dyes that are either red or lime green, participants report the red liquid to
be sweeter and the lime green one to be more sour, although they contain
the same amount of sugar or sometimes have no flavor at all (Zampini et al.
2007). So should we say that there is such a thing as visual flavor? What goes
into flavor experiences and how can they be defined? Most psychologists are
maximalists, as this quotation from Martin Yeomans shows:
Arguably, multi-sensory integration may be at its most extreme in the case

of flavour perception since few other experiences offer the opportunity for
concomitant stimulation of all the major senses: gustation through the five
primary tastes, olfaction through both ortho-and retronasal stimulation
of olfactory receptors by volatile compounds released from food, mecha-
noreception contributing to our perception of texture and providing infor-
mation on temperature, pain arising from oral irritants and hearing that
results from sounds and vibrations coming from the mouth contributing to
our perception of aspects of texture. (Yeomans et al. 2008, 565)
But many of us would want to say that Spence’s carefully constructed exper-
imental setups are cases of perceptual illusions: ways in which our percep-
tion of flavor is being influenced or distorted by the impact of other sensory
impacts. Such a strategy is not open to experientialists. They are all just cases
of multisensory flavor experiences. What is more, when we are not in a good
condition to taste because of ill health, a depressed mood, or low affect, what
we experience of our foods counts as just as much of a flavor experience as
under any other condition. There is no sense in which our experiences of
flavor are getting things right or wrong; they are just whatever experiences
result from the multisensory factors in play when we are eating or drinking.
Of course, many things go on when we eat and drink, and the overall ex-
perience that results will be shaped by what we see and hear, taste, feel, and
smell, as well as the internal physiological factors that have an influence on
emotion and cognition. But intuitively, at any rate, we would not want to
count every dimension of our overall experience while eating and drinking
as part of a flavor experience. So where do we draw the line? Our conscious
experience does not parcel up these events of tasting, hearing, seeing, and
so on as belonging to different zones of consciousness: they are all part of a
single unified conscious experience. One reply to this point might be that we
can still tell, on purely phenomenological grounds, which qualities of expe-
rience count as experiences of tasting. But this still leaves many factors that
affect what we smell or taste, such as sound and vision, as contributing to
Tasting Flavors 43
flavor experience, and it doesn’t distinguish between veridical experiences

and illusions. The key question is why the brain doesn’t bind all the sensory
inputs that can affect our perceptions of a food’s flavor or that are taking place
simultaneously, including the ambient temperature of the room, into what
counts as a flavor experience. So it appears as though flavor experiences will
have a visual, auditory, and thermal dimension.
And what of the hedonic aspect of eating and drinking? Is the hedonic
value we associate with a food part of its flavor; part, that is, of flavor experi-
ence? If I like broccoli and you don’t, is liking part of the flavor of broccoli for
me but not for you? When I acquire a taste for something that I didn’t previ-
ously like, have I really come to like that flavor, or does it now have a different
flavor, one that includes my liking? If individual liking is part of flavor, then it
is hard to resist some version of a sophisticated subjectivism about taste and
flavor. But must we accept the merging of tasting and liking?
I will argue that there is no need to collapse tasting and liking if we can
distinguish flavor perceptions from the hedonic reactions to foods that are
based on them.
To begin with, different cortical and subcortical regions of the brain are
responsible for recognizing the flavor qualities of a stimulus, on the one hand
and processing hedonic reactions to it, on the other. But this may not move
philosophers who consider such neurological talk as mere implementa-
tion details. So let’s turn to a behavioral test in which we can experience the
separation of flavor perception from the hedonic reaction to the stimulus.
Suppose that you like chocolate and I feed you piece. You like the experi-
ence of the chocolate’s flavor. Now I feed you another piece, and another, and
I go on feeding you chocolate until the chocolate no longer has any appeal
and the hedonic value of the experience has been reversed. You no longer
want to eat chocolate even though you are not full. This is called stimulus-
specific satiety (Kringelbach et al. 2003). What is important to note in this
case is that if during the time when the hedonics were varying so much, I fed
you a different make of chocolate, you would instantly recognize that some-
thing had changed. In other words, despite the swing in hedonic value, you
are maintaining a baseline fix on the flavor properties of the chocolate. So it
is not correct to say that flavor must change as a function of a change in our
liking. The hedonic value of a food is not part of its flavor. Rather, we need to
identify the flavor in order to react to it is with liking or disliking.6
6 For more on the difference between tasting and liking see Smith 2017.
The experientialist view of flavors faces some challenges and may oblige
us to say some strange things about what makes up flavor experience. If
we embrace experientialism about flavor, we may be forced to accept these
commitments. However, if there is a satisfactory account of flavor perception
that steers away from these puzzling cases and provides a better explanation
of tasting flavors, we should embrace it. I will now endeavor to sketch a more
satisfactory account of the epistemology of multisensory flavor perception.
1.5. The Multisensory Perception

of Objective Flavors
To be entitled to claim that tasting provides us with genuine knowledge of

flavors in our food and drink, there must be flavors that we count as being
perceived correctly or incorrectly, and some way, in principle, of telling
whether we are perceiving them veridically or not. That is what I shall now
attempt to spell out.
We need a way to motivate the distinction between accurate flavor
perceptions and perceptual illusions, and for this we need to establish what
the object of perception is. Relatedly, we need to understand which senses are
constitutively involved in flavor perceptions and which merely causally affect
it. Though making out the distinction between what is constitutive and what
is merely causally affecting is not straightforward. There is no agreement
among sensory scientists about which senses flavor perception involves.
Some researchers, like Bartoshuk and Duffy (2005), believe that only taste
and retronasal smell are involved, although touch helps to ensure the inte-
gration of their inputs. Others include trigeminal stimulation, and some, like
Spence, are tempted to include visual flavor and sonic seasoning. All agree on
what is taking place as we eat, but there appears to be no principled way to
decide which sensory systems and inputs of all those that are actively going
on should be included in the formation of flavor experiences.7
My proposal is to start the other way round with the flavors in foods and
liquids. Insofar as there are flavors to be perceived, we use our perceptual
apparatus to keep track of them and guide successful food choice. At its sim-
plest, the organism needs to ensure intake of protein, salt (which the body
7 Note the science won’t settle this questions for us: we need a philosophical means to provide the
answer.
Tasting Flavors 45
doesn’t produce), avoid bitter toxins and monitor levels of acidity. So the taste
receptors on the tongue play a role in detecting these substances, in order to
ensure homeostasis. We almost never experience pure tastes but only tastes
as part of flavors (see Spence, Auvray and Spence 2014). Beyond homeostasis
we are creatures facing a much wider range of food choices, and we have in-
dividual preferences. These preferences are subjective and idiosyncratic (al-
though many are formed in utero; see Mennella et al. 2001 and Schaal and
Orgeur 1992), and our multisensory perceptions of flavor are able to track
the flavors of foods we like or don’t like, which guides food behavior.
When it comes to which senses are involved in creating multisensory
flavor perceptions, the answer is it depends what we are eating or drinking.
Spices will require the activation of the trigeminal nerve, through since not
all foods are spicy and not all odors have a trigeminal component (though
most do), chemesthesis is not always involved. Similarly, vibration is only
part of flavor perception when an active ingredient like sanshool in Szechuan
pepper or Brazilian jambue is eaten. Taste and retronasal smell are always in-
volved, and touch is usually involved when they are combined; although we
know that aftertastes can be created by retronasal olfaction and the image of
taste. And of course, as we know from Juyun Lim’s research, previously men-
tioned, tastes and retronasal olfaction combine to form perceptible flavors
only when they are congruent, the most likely explanation for why there is
sensory congruency between taste and smell due to their statistical regularity
in being environmentally valid combinations of tastants and odors found in
food sources.
When it comes to the environmental contextual effects of sounds and
lighting that can impinge on and change our perceptions of a food or wine,
it is important to note that when these effects are removed and our eyes and
ears are not engaged, we still perceive flavors.
We still need to explain why we bind the sensory elements we do and not
others. Why does information from the gustatory, retronasal olfactory, so-
matosensory system and/or the trigeminal system get bound together by
multisensory integration? One answer is that multisensory integration typi-
cally takes place under an assumption of unity, where there is a common en-
vironmental source that prompts the different sensory responses. The sound
of a plastic water bottle being crushed and the simultaneous sight of a plastic
water bottle being crushed do not just lead to the combined activation of
auditory and visual information but instead produces a greater activation
when combined; namely, a superadditive effect. Superadditivity is a mark
of multisensory integration, and there are signs of just such superadditivity

when taste, smell, and touch combine. So the thought here is that a common
environmental source, the food or liquid we consume with the flavor it has,
provides the unity that leads to such multisensory combinings.
This strategy relies on the idea of flavors as properties of food substances
and liquids. So what are they? Attempts to reduce flavors to chemical
compounds and relate these to perceptible properties in flavor experiences
have been unsuccessful, but that is because we are trying to connect chem-
istry and perception directly. What’s needed is the postulation of flavors as
emergent properties dependent on but not reducible to the underlying chem-
ical properties. In this way flavors occupy an intermediate level between the
chemical compounds in foods and drinks and the individual and variable
flavor perceptions of tasters. We can characterize flavors as configurations
of sapid, odorous, and tactile properties of the foods and drinks we con-
sume, some of which give rise to irritations of trigeminal nerve endings or
vibrations of mechanoreceptors in the tongue, and which we track by means
of integrated, multisensory flavor perceptions.
Now we have two challenges instead of one: we must relate flavors to the
underlying chemistry of foods and liquids, on the one hand, and relate mul-
tisensory flavor perceptions to the flavors they attempt to track, on the other:
Flavor perceptions
↓
Flavors
↑
Chemical compounds
For our purposes, in spelling out the epistemology of tasting, we are inter-
ested in the connection between the intermediary level and the top level in
this model, and the first things to say is that an individual’s flavor perception
at a time can only be a snapshot of the flavor of the food or wine we are trying
to detect. So contrary to the simple sensations view of taste, we do not get all
aspects of flavors all at once. We often have to taste and retaste a dish or a
wine to pick out and note all of the flavors it has.8 A wine worthy of our atten-
tion does not give up all its secrets at once, and each sip provides a glimpse
of what we try to get to grips with. The wine example is helpful when trying
8 For more on this point see B. Smith 2007.

Tasting Flavors 47
to explain how we can tell whether there is something there in the flavor of a
wine for us to get right of wrong. Only if this is the case can we claim there are
objective flavors and that tasting can, all being well, provide us with know-
ledge of those flavors.
When we consider a well-made wine that develops in the bottle and in
the glass over time, we should think not of a static flavor but a dynamic
and evolving flavor profile. In some cases, wines are made to develop in the
bottle for twenty years or more and may be less approachable when young.
Nevertheless, in the cases of grand cru wines from Bordeaux, proficient
tasters including the winemakers themselves may be able to taste the wines in
their infancy and predict reasonably well how they will develop as they age in
the bottle. Similarly, when pouring a wine and tasting it, one may realize that
it needs time to open up through the effects of oxygen on the liquid. It may
be that the wine would benefit from being aerated by decanting it and would
improve by being a degree or two warmer or colder. In all these cases, know-
ledge and experience give practiced tasters an understanding of the stage of
evolution or condition of the wine they are tasting, that is, what part of the
flavor profile they are sampling, and this gives rise to expectations about how
the wine will taste after time in the bottle or the glass, one or two degrees
warmer or colder.
This model of flavor perceptions as a series of snapshots by which to sample
and learn about the flavor profile of an age-worthy wine fits in well to a cur-
rently popular theory of neural processing of sensory signals; namely, pre-
dictive coding based on the workings of the Bayesian brain, which requires
a balance between top-down expectations or priors and bottom-up input
sensory signals. We don’t just process the incoming sensory signals, passing
them up to higher cortical regions of the brain in order to figure out what is
going on around us. Instead, the brain makes top-down predictions of the
sensory signals it receives on the bases of prior expectations and knowledge
of the external world built up on the basis of experience gradually refined by
adjusting priors to match the inputs as a result of repeated prediction errors.
We sample the world by acting to generate sensory signals that are predicted
by our priors and learn by prediction errors that lead to finer tuning of our
priors. Neuroscientific and neurobiological evidence has given increased
credence to this model of perception and action (see Feldman and Friston
2010). And it fits well with our account of how practiced tasters, through
years of experience of drinking particular kinds of wines young and old, gen-
erate predictions of how the wine will behave in the future, predictions that
can be confirmed or refuted by subsequent perceptions. These predictions

help tasters build up not so much a series of snapshots but an understanding
of the flavor profile of the wine. In this way, expert tasters are able to tell by
tasting where they are in the flavor profile of age-worthy wine. The color of
the wine in the glass, the aromas and flavors perceived, can tell us whether
the wine is nearing the end of its life or has some time to go until it reaches
maturity. When gauging how the wine is doing and whether we should drink
it now or wait, tasters are able to test their predictions and find out how accu-
rate they are.9
Expectations are important in flavor perception, and they can be either
leading or misleading. In a famous experiment by Martin Yeomans and
Heston Blumenthal, participants were given a novel food to try: a smoked
salmon ice cream created by Blumenthal. One group of participants was told
it is novel flavor of ice cream and asked to rate it for sweetness, saltiness, and
liking. The other group, who were also asked to rate it, were told it was a
frozen savory mousse. The former group disliked it, while the latter group
found it acceptable. But the key finding is that when expecting to taste an ice
cream, participants not only disliked it, they rated it as saltier and less sweet
than the other group, who were expecting a frozen savory mousse. Our prior
expectation and memory for food can distort our experience of the food’s
flavor. How we perceive a food can be affected by the description we give it
and the prior associated with it (Yeomans et al. 2008).
However, expectations based on the look, the feel, and the smell of the
food or wine we are about to taste can provide useful information that leads
to the right predictions and helps minimize prediction errors.
But what happens when the aroma of a food affects how it subsequently
tastes in such a way that it is inseparable from the flavor we usually experi-
ence? This happens in the case of sweetness enhancement, where an exter-
nally inhaled odor can result in the enhancement of the perceived sweetness
of the food. Take vanilla ice cream. Vanilla odor is experienced as sweet
smelling, although sweetness is a taste not a smell, and vanilla pods have no
sweetness and taste bitter. It is because we combine vanilla with things that
are sweet, such as ice cream, chocolate, and custard, that we have learned to
associate the aroma of vanilla with a taste property of sweetness. (In cultures
where vanilla is combined with salt and fish, vanilla will “smell” salty.) Thus,
9 For more on this see B. Smith 2011.

Tasting Flavors 49
when we taste the sweetness of ice cream, it will taste sweeter in the presence
of vanilla aroma, and this effect remains even when we know what is going
on. So should we say that sweetness enhancement by orthonasal odors like
vanilla is a case of perceptual illusion? Do we always misperceive the sweet-
ness of vanilla ice cream? We don’t need to say this when we are committed to
the multisensory perception of flavor. Orthonasal odors like vanilla can en-
hance the perceived sweetness of what we taste as a result of the cross-modal
influence of one sense modality on the working of another, and it is that
enhanced sweet taste together with the retronasally sensed odors that get in-
tegrated into the multisensory perception of vanilla flavor. Hence, this regu-
larly experienced effect counts as veridical perception and a case of knowing
the flavor of vanilla ice cream.
1.6. Conclusions
We have seen that tasting is not merely a matter of experiencing simple taste
sensations. In tasting we receive inputs from taste, smell, touch, and other
sensory modalities to produce a multisensory experience of flavor. This has
led some to suppose that flavors are constructs of the brain, which would
support a sophisticated subjectivism about taste. But the idea of a construct
cannot support the idea of tasting as perceptual because it fails to distin-
guish flavors and flavor perceptions. What is more, the sophisticated subjec-
tivist about flavor struggles to explain which sensory inputs get bound into
flavor experiences and has no principled way to keep the hedonic reactions
of tasters separate from flavors even though they can come apart. Instead of
this account, I have proposed an objectivist view of flavor and restored the
idea of tasting as a species of perception that gives us knowledge of flavors.
The latter can be seen as configurations of sapid, odorous, and tactile prop-
erties of foods and liquids that we track by means of integrated multisensory
inputs, which together with our past learning and prior expectations enable
us to perceive the flavors of food and drink as a means of guiding successful
food choice. The arguments are not conclusive, but the picture on offer here
provides an account of tasting as perceptual sense, which, alongside the other
senses and working in tandem with them, provides us with knowledge of the
objective features of the world around us and accounts for much of the data
we have from both science and common sense about the everyday experi-
ence of eating and drinking.
References
Auvray, M., and C. Spence. 2008. The multisensory perception of flavour. Consciousness
and Cognition 17(3): 1016–1031.
Bartoshuk, L. M., and V. B. Duffy. 2005. Chemical senses: Taste and smell. In C. Korsmeyer,
ed., The Taste Culture Reader: Experiencing Food and Drink. Oxford: Berg. 25–33.
Brillat-Savarin, J. 1825. The Physiognomy of Taste. New York: Penguin.
Bult, J., R. Wijk, and T. Hummel. 2007. Investigations on multimodal sensory integra-
tion: Texture, taste, and ortho-and retronasal olfactory stimuli in concert. Neuroscience
Letters 411(1): 6–10.
Cliff, M., and Noble, A. 1990. Time-intensity evaluation of sweetness and fruitiness and
their interaction in a model solution. Journal of Food Science 55(2): 450–454.
Crisnel, A. S., and C. Spence. 2012. A fruity note: Crossmodal associations between odors
and musical notes. Chemical Senses 37(2): 151–158.
Dalton, P., N. Doolittle, H. Nagata, and P. A. S. Breslin. 2000. The merging of the
senses: Subthreshold integration of taste and smell. Nature Neuroscience 3: 431–432.
Feldman, H., and K. Friston, 2010. Attention, uncertainty, and free energy. Frontiers in
Human Neuroscience 4(215): 1–23.
Gallace, A., and C. Spence. 2008. The cognitive and neural correlates of “tactile conscious-
ness”: A multisensory perspective. Consciousness and Cognition 17: 370–407.
Green, B., and P. George. 2004. “Thermal taste” predicts higher responsiveness to chem-
ical taste and flavour. Chemical Senses 29: 617–628.
Hagura, N., H. Barber, and P. Haggard. 2013. Food vibrations: Asian spice sets lips trem-
bling. Proceedings of the Royal Society B 280: 20131680.
Harrar, V., Piqueras-Fiszman, B., Spence, C. 2011. There’s More to Taste in a Coloured
Bowl. Perception Jan 1,
Hort, J., and T. Hollowood. 2004. Controlled continuous flow delivery system for
investigating taste-aroma interactions. Journal of Agricultural and Food Chemistry
52(15): 4834–4843.
Kalderon, M. 2007. Colour Pluralism. The Philosophical Review 116(4): 563–601.
Kringelbach, M., O’Docherty, J., Rolls, E., and Andrews, C. 2003. Activation of the human
orbitofrontal cortexto a liquid food stimulus is correlated with its subjective pleasant-
ness. Cerebral Cortex 13(10): 1064–1071.
Lim, J., and M. Johnson. 2012. The role of congruency in retronasal odor referral to the
mouth. Chemical Senses 37: 515–521.
Mennella, J. A., J. A. Jagnow, and G. K. Beauchamp. 2001. Prenatal and postnatal flavor
learning by human infants. Pediatrics 107: E88.
Michel, C., C. Velasco, and C. Spence. 2015. Cutlery matters: Heavy cutlery enhances
diners’ enjoyment of the food served in a realistic dining environment. Flavour 4: 26.
Piqueras-Fiszman, B., and C. Spence. 2012. The weight of the bottle as a possible extrinsic
cue with which to estimate the price (and quality) of the wine? Observed correlations.
Food Quality and Preference 25: 41–45.
Prescott, J. 1999. Flavour as a psychological construct: Implications for perceiving and
measuring sensory qualities of foods. Food Quality and Preference 10: 349–356.
Prescott, J. 2004. Odor-taste interactions: Effects of attentional strategies during exposure.
Chemical Senses 29: 331–340.
Prescott, J. 2015. Flavours: The pleasure principle. Flavour 4: 15.
Tasting Flavors 51
Richardson, L. 2013. Flavour, taste and smell. Mind and Language 28(3): 322–341.
Rozin, P. 1982. “Taste-smell confusions” and the duality of the olfactory sense. Perception
and Psychophysics 31(4): 397–401.
Rolls, E, Critchley, H., Browning, A., Hernadi, I., Leonard, L. 1999. Responses to the
sensory properties of fat of neurons on the primate orbitofrontal cortex. Journal of
Neuroscience 19(4): 1532–1540.
Rolls, E. 2013. Emotion and decision- making explained. Oxford: Oxford
University Press.
Rolls, E. Taste. 2015. Olfactory and food reward processing in the brain. Progress in
Neurobiology 127(8): 64–90.
Schaal, B., and P. Orgeur. 1992. Olfaction in utero: Can the rodent model be generalized?
Quarterly Journal of Experimental Psychology 44: 254–278.
Sheppard, G. 2012. Neurogastronomy: How the Brain Creates Flavor and Why It Matters.
New York: Columbia University Press.
Sheppard, G. 2017. Neuroenology: How the Brain Creates the Taste of Wine. New York:
Columbia University Press.
Sibley, F. 2006. Approaches to Aesthetics: Collected Papers on Philosophical Aesthetics. Ed. J.
Benson, B. Redfern, and J. R. Cox. Oxford: Oxford University Press.
Small, D. 2012. Flavor is in the brain. Physiology & Behavior 107(4): 540–552.
Smith, A. D. 2002. The Problem of Perception. New York: Oxford University Press.
Smith, B. 2007. The objectivity of tastes and tasting. In B. C. Smith, ed., Questions of
Taste: The Philosophy of Wine. Oxford: Oxford University Press.
Smith, B. 2011. The True Taste of a Wine. In D. Chassagne and R. Gougons, eds., Wine
Active Compounds. Chantré, France: Oenopluria Dicilour Groupe. 283–286.
Smith, B. 2017. Tasting and liking: Multisensory flavour perception and hedonic eval-
uation. The Taste Culture Reader, 2nd ed. Carolyn Korsemeyer Bloomsbury Press.
250–261.
Solar, M., H. Kumru, J. Vidal, R. Pelayo, J. Tormos, F. Fregni, X. Navarro, and A. Pascual-
Leone. 2010. Referred sensations and neuropathic pain following spinal cord injury.
Pain 150(1): 192–198.
Spence, C. 2015. On the psychological impact of food colour. Flavour 4: 21.
Spence, C., M. Auvray, and B. Smith. 2014a. Confusing tastes and flavours. In D. Stokes, M.
Matthen, and S. Briggs, eds., Perception and Its Modalities. Oxford: Oxford University
Press. 247–277.
Spence, C., C. Michel, and B. Smith. 2014b. Airplane noise and the taste of umami flavour.
Flavour 3(2).
Spence, C., M. Shankar, and H. Blumenthal. 2011. “Sound Bites”: Auditory Contributions
to the Perception and Consumption of Food and Drink in Art and the Senses. New York:
Stevenson, R., and C. Tomiczek. 2007. Olfactory-induced synesthesias: A review and
model. Psychological Bulletin 133: 294–309.
Titchner, E. 1909. A Textbook of Psychology. New York: Macmillan.
Turton A. J., and S. R. Butler. 2001. Referred sensations following stroke. Neurocase
7(1): 397–405.
Yeomans, M., L. Chambers, H. Blumenthal, and A. Blake. 2008. The role of expectancy in
sensory and hedonic evaluation: The case of smoked salmon ice-cream. Food Quality
and Preference 19: 565–573.
Zampini, M., and C. Spence. 2004. The role of auditory cues in modulating the perceived
crispiness and staleness of potato chips. Journal of Sensory Studies 19(5): 347–363.
Zampini, M., D. Sanabria, N. Phillips, and C. Spence. 2007. The multisensory perception
of flavor: Assessing the influence of color cues on flavor discrimination responses. Food
Quality and Preference 18: 975–984.
2
Sensory Interactions and
the Epistemology of Haptic Touch
Matthew Fulkerson
2.1. Introduction
We possess multiple senses. This should put us in a superior epistemic posi-

tion relative to simpler creatures with fewer sensory resources. The situation
is a bit more complicated, however. Simply adding sensory resources doesn’t
always bring with it epistemic improvement. Whether it does crucially
depends on the nature of the addition and the context in which it occurs.
Consider two examples. First, an example where using multiple senses
works as we might expect:
Avocado: Desiree is making guacamole, and she wants to use a ripe avo-
cado. She finds one with a deep green color, knowing an avocado with such
a color is more likely to be ripe. Her visual experience seems to justify her
belief that the avocado is ripe, but she can improve her epistemic situation
by picking up and squeezing the fruit, checking that her choice also has
enough give. Seeing and squeezing the fruit seems to put her in a superior
epistemic situation with respect to the belief that the avocado is ripe.
The lesson of this example seems to be that it is sometimes epistemically

better to use more senses.
The addition works in Avocado because the senses are providing infor-
mation about two distinct sensory features, and both of these features are
strongly correlated with the property that figures in the content of the be-
lief. For this reason, the combined evidence from both senses reinforces
her belief and makes it more likely to be true. Call these reinforcing cases of
Matthew Fulkerson, Sensory Interactions and the Epistemology of Haptic Touch In: The Epistemology of
Non-Visual Perception. Edited by: Berit Brogaard and Dimitria Electra Gatzia, Oxford University Press
(2020). © Oxford University Press.
DOI: 10.1093/oso/9780190648916.003.0003
sensory interaction: the addition of sensory resources increases the accuracy

of beliefs derived from them.
Adding senses doesn’t always work this way. Consider another case:
Squash: The market is selling squash of any weight for $1.99 each. George,
who likes a good deal, is comparing two different squash, wanting to buy
the heavier of the two. He places each in a plastic bag and holds them up by
the handle to determine their weight. The smaller squash feels heavier to
him. He buys the smaller one, thinking he’s getting a better deal. Turns out
the smaller squash actually weighs less than the larger squash.
Call this a case of sensory discounting. This occurs when adding sensory re-
sources weakens the epistemic quality of our perceptual awareness. In this
case, George falls victim to the well-known size-weight illusion, described by
Jones and Lederman (2006): “A large and small object of the same mass do
not appear to be equally heavy; rather, the large object is judged to be lighter
than the small one. The classic form of this illusion involves both visual and
haptic inputs, that is, observers heft the object while looking at it” (83). In
this case, George would have been in a much better epistemic condition con-
cerning the weight of the squash if he had kept his eyes shut.
We can already see that coordinating evidence from different senses can
influence the epistemic quality of our experiences. Sometimes the coordi-
nation will improve our epistemic situation, and sometimes it will make it
worse. The situational and sensory features of perceptual contact that help
determine the epistemic quality of our perceptual beliefs are relatively well
known in cognitive psychology (especially from work on multisensory
interactions). How our sensory systems coordinate and interact would thus
seem to have important epistemological consequences.
Sensory reinforcing and discounting frequently occur entirely within
haptic perception. A large felt object will feel lighter than a smaller felt
object of the same mass. In the initial example, George’s use of a bag
avoids this worry, and so he really would have been better off with his
eyes closed. But even if he were to use haptic perception alone, he would
still be subject to the same discounting effect. Other similar interaction
effects in haptic touch include the material-weight illusion, where metal
and stone objects feel heavier than objects of the same size and mass made
of wood, plastic, or foam; and the thermal-weight illusion, where cold
objects feel heavier than warmer objects of the same size and mass. These
The Epistemology of Haptic Touch 55
intra-modal effects seem to result from the same sorts of interactions that
occur between the senses.1
Sensory interactions are highly varied and occur at all levels of sensory
processing. In addition, there are important elements of interaction that
occur between perceptual and other systems, in particular systems respon-
sible for emotional processing and motor control. To simplify and focus the
discussion, I will focus on the unique epistemic features of human haptic
touch. The idea is that, even when we look at the individual sensory modal-
ities, we see evidence of competitive, interactive exchanges that bear epi-
stemic consequences. My goal is to investigate some of the details of human
haptic perception and see how they influence the epistemic quality of our
perceptual beliefs. These insights will then be used as a model for the more
general claim that, based on our best current understanding of the sense of
touch and perception generally, we should be adopting a more nuanced un-
derstanding of how perceptual interactions influence the epistemic quality of
our perceptual beliefs.
Some caveats before we begin. I am not in what follows presupposing
any framework or theory of perceptual justification or perceptual know-
ledge. I am attempting to engage with epistemic questions at a more basic
level: what impact do sensory interactions have on the epistemic quality of
our perceptual experiences? By “epistemic quality” here I primarily intend
to focus on the accuracy of our senses; on how likely our basic perceptual
beliefs are to be true, given the nature of the perceptual experiences on which
they are grounded. I do want to leave open, for reasons that will become ap-
parent, that there may be other important epistemic properties of our sen-
sory experiences besides accuracy, including coherence, consistency, speed,
and defense.2 The goal is to look at some general, fundamental epistemic
features of sensory interactions and haptic perception, in ways that could be
applicable to theorizing across a wide range of more sophisticated epistemic
theories at different levels of generalization. When I talk about epistemic
merit in what follows, then, I have in mind the property or set of properties
1 Though not exactly the same. While the general principles of coordination apply throughout all
levels of cognition and experience, they do occur at different levels of processing and often involve
proprietary coordination rules and functions.
2 By “defense” here I mean something along the lines of the risk-averse, organism-protecting goals
of some sensory systems. Such a system, for instance, might err on the side of treating ambiguous
visual inputs as full-blown tiger or spider representations. This of course incurs a local epistemic cost
(the perceptual beliefs will be less accurate or reliable) but may serve longer-term epistemic goals (a
living organism will go on to generate more true beliefs than a dead one). Of course, all of this is a bit
controversial. For some discussion see of these considerations see Akins 1996 and Craig 2002.
of a perceptual belief which functions to ensure truth. This property serves

an important function, even if it ultimately plays little role in the person-
level theorizing about more robust epistemic properties. For instance, many
internalists about justification would deny that the accuracy or reliability
of a sensory interaction is a crucial element of genuine justification (which
involves only features to which a subject has access). Nevertheless, I hope
that such an internalist can agree that accuracy, reliability, coherence, and
consistency are features that can make an epistemic difference in the ways
just specified (that is, can influence the likelihood that a perceptual belief is
true). Since I’m hoping for maximum applicability in what follows, I’m going
to eschew some critical elements of these discussions that would be part of
any well-worked-out theory of perceptual epistemology. For instance, while
for a full-blown epistemic account it will be critical, I will not provide any
view here about the precise contours that separate perception proper and the
automatically inferred beliefs that follow from them. Similar issues arise for
the precise distinctions between perception and emotion and perception and
action.
2.2. Interaction Effects
An extensive empirical literature exists on cross-modal interactions, and

it seems clear that adding sensory resources can often lead to degraded,
distorted, and outright illusory perceptual experiences.3 Such effects are
ubiquitous: studies in recent years (often using ingenious experiments)
have demonstrated that our awareness, in nearly all modalities, of self-mo-
tion, self-location, distances, sizes, weights, velocities, acceleration, orien-
tation, and much else, depends on (and is therefore partially influenced by)
proprioception, kinesthesis, and vestibular awareness, among other things
(Frissen et al. 2011; Campos et al. 2012; Graziano 1999; Knox et al. 2006;
Maravita et al. 2003). Of particular importance for my purposes here is the
3 Some of these findings are surveyed by Ernst et al. (2007, 243). One doesn’t need elabo-
rate experiments to believe that adding sensory resources doesn’t always increase our epistemic
circumstances. Consider some of the many commonly cited examples: sensory deficits can often
seem to strengthen the accuracy and dependability of remaining senses. Removing extraneous, pos-
sibly distracting, sensory signals can aid the accuracy and usefulness of many senses, from hearing
a musical performance to tasting a fine wine. And for many tasks, adding a modality doesn’t help
at all: if we want to be sure about an object’s color, it often doesn’t help to add unrelated smells and
sounds to the mix. Still, I believe the best evidence here comes from our current best science of
the mind.
fact that both proprioception and kinesthesis are central components of

haptic touch (see Fulkerson 2014). In addition, haptic touch is itself a com-
plex form of awareness combining inputs from many otherwise distinct
systems, including thermal awareness, motor control and feedback, stretch
receptors in the muscles and skin, and a variety of tactile receptors coding
for pressure, vibration, and shape. Our experience of grasping a hot mug
of coffee is itself the result of integration and coordination between many
distinct sensory systems. In other words, epistemic effects have been found
connecting awareness in all modalities and, importantly, also often within
modalities).4
Why does sensory facilitation and discounting occur? It might seem rea-
sonable to think that our senses always work well together, and that adding
sensory resources would always be a good thing, at least epistemically. In
fact, human perceptual experience involves many distinct systems with their
own preferred ways of doing things.5
Different senses often carry redundant but conflicting information, and
when this happens various processes of integration must resolve the initial
disagreements. These integration effects reveal themselves frequently in our
perceptual experiences, most notably in cross-modal illusions.
There are many interesting details about how these distortions occur.
To avoid getting bogged down, we’ll just consider a single representative
case: What happens when the senses disagree on a feature about which
both provide information? Call these cases of “redundant overlap.” Well, it
depends. While on their own, vision and haptic perception tend to give the
same verdicts about real-world texture (see Bergmann Tiest and Kappers
2007), when they operate together their outputs get combined into a
weighted average that reflects the relative accuracy of each individual mo-
dality on that task. One popular model for this weighting is the maximum
likelihood estimation (MLE) model (see Ernst and Banks 2002; Helbig and
Ernst 2007). According to this model, each sensory system performs an
estimation operation on an environmental property. Assuming the noise
4 Philosophers in recent years have discussed a small but representative subset of sensory illusions,
especially focusing on the McGurk effect, the motion-bounce illusion, and the flash-induced sound
effect (see especially O’Callaghan 2008, 2012). These works have been immensely influential in our
understanding of perception.
5 The multisensory nature of perceptual experience has recently received a lot of attention, in a
variety of disciplines. See (as a start) Deroy et al. 2014; Driver and Spence 2000; O’Callaghan 2008,
2011; Bennett and Hill 2014; Stein 2012; Naumer and Kaiser 2010; Klatzky and Lederman 2010b;
Macpherson 2011.
in each modality is independent with a uniform Bayesian prior, then we

can weigh the contributions of each modality to minimize variance (and
thereby maximize the accuracy of the estimate): “The MLE rule states that
the optimal means of estimation (in the sense of producing the lowest-
variance estimate) is to add the sensor estimates weighted by their nor-
malized reciprocal variances” (Ernst and Banks 2002, 430). According to
this model, vision will dominate (that is, have greater weight) in contexts
where pattern is most salient (since vision is much better at detecting pat-
tern). In contexts where uniform intensity is more salient (say, for wool),
haptic touch will dominate. All available evidence suggests that this pro-
cess is attentionally modulated and task sensitive and occurs very late
in perceptual processing (Klatzky and Lederman 2010, 222). This is not
the only model, of course, and it’s not clear that there is going to be any
single algorithm that can capture all sensory interactions. The multisen-
sory interactions that produce the experience of flavors, for instance, do
not seem to fit anything like the MLE model. Still, in general something
like this process occurs in a wide range of sensory interactions, especially
those involving redundant overlap.
The key takeaway here is that proprioception influences self-motion
awareness in an MLE-friendly manner, across all modalities. But it is espe-
cially important in self-motion awareness in haptic touch. This is because
haptic touch involves exploratory movements over time. This is what allows
us to build spatial representations of tangible objects, segment and isolate
parts of complex objects, and identify and recognize objects through touch.
We often use specific exploratory procedures (EPs) to engage with tangible
objects, and these complex movements unify and connect the activations
of many different tactual areas into coherent object representations (see
Lederman and Klatzky 1987 for the initial discussion of EPs and their role in
selective awareness of specific tangible features). Even entirely within touch,
we get interesting interactions and resultant illusions. We’ve all experienced
the illusion of wetness that can occur when clothes feel cold (our experience
of dampness often involves a feeling of cold mediated by the moisture in an
absorbent object). And, of course, the same could be said for all the other
modalities as well.
The multisensory interactions previously discussed are prime examples of
how adding senses actually works in context. For instance, as we’ve seen, the
senses sometimes work together cooperatively to generate facilitation and
other superadditive effects. As Calvert and Thesen (2004) describe:
The observed facilitation of the neuronal response is often maximal when

the responses to the individual inputs are weakest, a principle known as
inverse effectiveness. In contrast, crossmodal stimuli that show spatial or
temporal disparity can induce profound response depression. This means
that the response to an unimodal stimulus can be severely lessened, even
eliminated, by the presence of an incongruent stimulus from another
modality. (194)
There is a way in which such facilitation mimics the influence of cogni-

tive penetration. Prior beliefs can boost the likelihood that what we are
seeing is A rather than B, and so can influence us to see A even when the
visual stimulus is neutral between A and B.6 This is epistemically bad when
a cognitive state does it. But notice now that an auditory state can play the
exact same role, putting a finger on the scales even in cases where the visual
input is underdetermined. In these cases, it should also be epistemically
problematic.7 Interaction effects like dominance, inverse effectiveness,
and response depression are several important features of sensory interac-
tion, and they all seem to make an epistemic difference in our perceptual
experiences.
Consider another well-known case of sensory discounting:
Movie: Slatka is sitting in the old movie theater downtown, waiting for the
lights to go dim and for her favorite classic film to start. Finally, the screen
lights up, and she is presented with an opening scene of two characters
talking over a cup of coffee. She is watching the screen while listening to
their quick-witted banter back and forth.
Slatka hears the voice of the character on the left coming from her mouth,
and the voice of the second character coming from his mouth on the right.
This despite the fact that, in this old theater, the sound in both instances is
coming from speakers off to the side of the screen.8 This is the ventriloquism
effect (Bertelson 1999). It is, in our modern context, incredibly common.
6 Cf. Siegel 2011.

7 That sensory interactions sometimes boost our epistemic situation is no help here; all parties to
the debate agree that prior cognitive states can also often boost our epistemic situation.
8 Of course, modern cinemas often have advanced speaker systems, sometimes built right into the
screen itself, to increase cinematic immersion. Slatka is not so lucky. Neither are most of us, since our
home televisions, laptops, tablets, and phones often have small, fixed speakers below or to the sides of
the screen.
Some might hold that films are inherently illusory, since there are only
apparent motions, apparent depths, apparent objects, and so on. Indeed, in
modern films most of what we see may be constructed (indeed, there may
not be any film). But the ventriloquism effect is one form of illusory error that
does not arise from the artifice or skill or the filmmaker, but from the way our
senses interact (of course, filmmakers, or more accurately, cinema owners
and designers, often exploit this fact). When we both see a person speaking
and hear audio temporally and thematically connected to that sound, our
sensory systems associate the sound with that source. Slatka is, in a straight-
forward sense, misperceiving the location of the sound source. If she simply
closed her eyes, she would be much more likely to experience the sound as
coming from its actual source. On the other hand, this systematic misper-
ception is actually the result of a very useful process that contributes to the
coherence and immersion of the cinematic experience.9 The interactions in-
volved here seem to undermine the accuracy of the auditory signal in order
to provide a more cohesive overall experience. In other contexts, of course,
these very same interaction effects facilitate our awareness. In a noisy room,
seeing lip movements helps us better determine what someone is saying. As
we’ll see, my interest isn’t simply in cases that lead us astray. It’s what these
cases reveal about sensory interactions generally.
The first step forward is to draw an important, epistemically salient dis-
tinction between sensory interactions.10
There are two kinds of sensory addition that can occur when we direct
multiple senses on a single epistemic target. First, there are cases of redun-
dant overlap. This occurs when the senses provide awareness of the same sen-
sible feature or property. For instance, we can use vision and touch to assess
the texture of a surface. Second, there are cases of complimentary overlap,
where each additional sensory system brings awareness of information not
available in the other modality. For instance, one might see the beautiful red
of a fresh-picked rose while simultaneously smelling its rosy scent. These
percepts are not available in the other modality.11
9 We have all experienced the disconnect and disruption caused by audio tracks that are not prop-
erly synced up with the visuals.

10 There are many ways of classifying sensory interactions; this is just one useful way of carving up
the space (Fulkerson 2014a).

11 Aristotle noticed a distinction very much like this one (De anima, book. 2, chap. 11). He called
those features capable of redundant overlap “common sensibles,” and those only capable of compli-
mentary overlap “proper sensibles.” Many philosophers have worried about these sorts of interactions
between sensory features given through different sensory modalities (Grice 1962/2002; Evans 1985).
But we are now in a position to say a lot more about the nature of these interactions.
Let’s start with cases of redundant overlap. Consider the experience of tex-
ture we get when both looking at and touching a fabric. Vision and touch
here provide redundant information about the texture of the fabric, but each
sense determines the texture in its own way, and as a result they can some-
times disagree about the actual texture of the fabric (Klatzky and Lederman
2010a). Touch determines texture by appeal to surface roughness and other
haptically available features (friction, stickiness, etc.). Rough wool possesses
many of these features, and so feels rough even when it looks smooth. Vision,
on the other hand, determines texture by appeal to “the pattern of brightness
of elements across a surface” (Klatzky and Lederman 2010a, 212). A shiny
wool, with a uniform pattern of brightness, will not look rough at all. It will
seem to vision to have a smooth texture. Audition also provides critical infor-
mation concerning the texture of a surface, using features like “crackliness”
and “scratchiness” when we move or explore the object with our hands. For
wools, it might be that distinctive rustling sound we hear as it moves and
rubs against a surface.12
This kind of integration is not unique to texture perception, of course. The
process of weighting and averaging the inputs of different modalities in a
context-sensitive manner seems to be a common method used by our per-
ceptual systems to handle disagreement, both within and between modali-
ties. In representational terms, this process is usually thought to result in a
new representational content that expresses the consensus view of the sen-
sory systems involved:
Because the output no longer resembles the response obtained to either

input, there is a de facto assumption that the information obtained from
two sources has been combined to form a single (new) output signal. This
process is referred to as multisensory integration. (Clavert and Thesen
2004, 194)
Dominance is just one outcome of sensory disagreement. Integration

occurs in many ways, as the senses handle different forms of disagreement
in different ways in different contexts. Sometimes one signal is completely
suppressed, blocked entirely by another signal. Sometimes both signals cancel
each other out. Sometimes both get a boost. I should note an important fact
12 It can be important to get the sound right: chip companies spend a lot of effort getting the crackle
of their crisps just right, since a crisp sound translates into an experience of freshness and quality
(Spence and Shankar 2010).
here: such interactions don’t require separate modalities (or a theory of sen-
sory individuation) to get off the ground. These interactions can be easily
duplicated entirely within modalities (recall my earlier note concerning
interactions effects like the thermal-weight illusion that occur entirely within
haptic touch).
Let’s turn now to cases of complimentary overlap. This form of overlap
avoids direct conflict on the same feature, but, crucially, does not preclude
other forms of sensory interaction.13
In addition, there are cases of sensory incomparability and bi-stability.
Sometimes our sensory inputs fail entirely to resolve conflicts, leading to
unstable and distorted experiences. Think of the experiential distortions
that occur in a loud dance party filled with people and sounds and smells
and movement. In such cases there is just too much going on in our modal-
ities to bring them together into a stable, coherent state. In addition, there
are cases when streams within a modality compete, leading to unstable
representations with a relatively fragile nature. These representations easily
switch back and forth between the most likely alternatives, never firmly set-
tling on one. Our visual awareness of Necker cubes are paradigm instances of
such bi-stability, with two incompatible shape orientations made available by
the same distribution of lines and points. In such cases, where our perceptual
systems cannot determine a single most likely representation, they simply go
with one for a time, then switch as needed.
I want to end this section by making clear how pervasive these interactions
are, both within and between modalities. This isn’t really necessary, but I want
to forestall the suggestion that these interactions can be easily ignored or stip-
ulated away. I can most easily demonstrate how common these interactions
are by pointing to the continuous influence of proprioception and vestibular
awareness on our other senses.14 Most of us have little idea how active and
influential these basic bodily systems are in our perceptual awareness. What
kind of pervasive influence do I have in mind? Consider the dramatic case
of Ian Waterman, who lost his proprioception as an adult due to a sudden
deafferentation event. This loss was completely debilitating and disruptive to
13 Overall experiences will often include some combination of redundant and complimentary
overlap. Seeing and hearing a dog bark, for instance, will provide some redundant awareness (about
the location of the dog) and some complimentary awareness (about it being brown and very loud).
This complication should not be ignored.
14 This is an arbitrary choice; these interactions occur in all modalities and in nearly every interac-
tion between them.

nearly every one of his other perceptual experiences.15 Consider this survey
of evidence:
A few recent studies have shown that information from one sense can affect
perception in another sensory modality, often resulting in recognition perfor-
mance that is robust to incidental changes in the environment. For example,
Simons, Wang, and Roddenberry (2002) reported that vestibular, propriocep-
tive, and optic flow information together can be used to update the represen-
tation of an object in visual memory resulting in recognition performance that
is independent of viewpoint changes. Simons et al. (2002) found that the rec-
ognition of a novel view of an unfamiliar object was better when the observer
moved around the object to the new view of the object than when the new
view was presented to a passive observer. Similarly, Pasqualotto, Finucane, and
Newell (2005) found that observer movement can also update an object’s rep-
resentation in tactile memory thereby eliminating effects of changes in view-
point that are otherwise present in a passive viewing condition. (Ernst et al.
2007, 243)
These are not effects at the edges that occur only in highly constrained environ-
ments. They are effects that have direct influence on the operation of our other
sensory modalities.
What do these reflections reveal about perceptual experience? When one
sensory modality operates in isolation, it makes sense to think of it pro-
cessing incoming sensory signals and producing something like an output,
and we can judge its epistemic merit by how well it performs this task. We
can think of this output as a contentful representation, carrying something
like assertive information about the distal environment. We can think of
the senses as capturing environmental information through the trans-
duction of incipient energy, and then (along with some inferential pro-
cessing) making that information available to other systems downstream.
In basic outline, this is the modular, input systems picture of perception
defended by.
Problems with this model arise because sensory interactions—like beliefs,
wishes, moods, evil demons, and so on—can also make alterations to our
15 For a vivid account of Waterman’s case, see Cole 1995.

perceptual experiences.16 As we’ve seen, activity in one modality or sensory

stream can influence the operations of another in such a way that it system-
atically undermines its epistemic quality. These influences would be deeply
problematic if caused by something outside perception, and I can’t see any
good argument for ignoring them now simply because they occur within
perception.
I won’t delve into a long evolutionary story here (and doing so isn’t nec-
essary for my ultimate goals). Still, I think it is useful to say a little bit about
our unique perceptual circumstances. Here is one way to tell this story. The
first putative sensory systems were relatively simple. Depending on your fa-
vored account, the first senses to develop were either rudimentary chemical
signaling systems, used to orient basic organisms toward nutrients and away
from harms, or basic pressure sensors used by single-celled creatures to dif-
ferentiate themselves from the world around them. On either model, these
rudimentary sensory systems carried inputs from the environment and fed
into practical, life-sustaining activity (cf. Thompson 2007).
Greatly simplifying, we can see the ensuing evolutionary record as a sort of
sensory arms race. Creatures with more senses, or better senses, were more
likely to find food, avoid predators, and pass on their genes to their young.
According to one popular account, the emergence of vision might be the
key to the so-called Cambrian explosion, an episode in which the number of
species increased dramatically (see Parker 2009 for a version of this story).
According to this theory, the sudden emergence of vision extended the
sensory reach of oceanic creatures (the only creatures around at the time),
forcing extensive biological adaptation to the new predatory dangers. There
is a cost to adding senses too, of course, or else we’d see creatures with dozens
or even hundreds of distinct senses. While adding resources is in general
good, it is also a difficult physiological and computational challenge to in-
tegrate and coordinate these many different signals in a single organism.
Sensory systems operate using different sensors and integrate information in
distinct ways. Recognizing this challenge, and understanding how different
biological systems attempt to solve it, is critical for understanding the epi-
stemic role of perceptual experience.
16 Whether there is one experience influencing another or simply a unified single experience
influenced by the operations of its constituent parts, there is still an epistemic difference made by the
operation of that constituent or constituent part. This occurs, as I’ve suggested, in the generation of
representational contents. Such influence happens both within and between modalities.
This evolutionary background reminds us of some essential facts con-

cerning sensory systems that are relevant for the discussion to come: they
are evolved, messy, unevenly connected, and operate using a wide range of
different processes. This history has epistemic consequences.
2.3. Emotional Effects
Emina comes home from a long trip hungry and tired. She opens her nearly
empty fridge and pulls out a carton of milk. Worried that the milk might have
turned, she takes a sniff, and comes to have a strongly negative olfactory expe-
rience. The smell of the spoiled milk is one she finds repulsive, and she imme-
diate throws the carton into the garbage (and sets aside completely any thought
of drinking it). She thinks: What an awful smell! Her experience seemingly
grounds her evaluative judgment and provides strong motivation for her be-
havior. Examples like this are common and unexceptional, occurring in all
individual modalities and in their interactions. Call the felt pleasant and un-
pleasant qualities of these motivating perceptual experiences their affect.17
Perceptual affect is a central feature of many (if not most) of our percep-
tual experiences, yet it remains difficult to explain within current models. On
most views, perceptual experience is essentially informative; it provides us
with descriptive information about the world around us, and how we act or
react to that information is determined by cognitive and emotional processes
downstream and distinct from the perceptual experience itself. But it seems
wrong to suggest that Emina is merely acting on neutral information pro-
vided by her sensory systems. Her olfactory experience seems to be directly
presenting the smell to her as something awful, in a sense pushing her to
avoid it. In other words, she hasn’t merely judged that the smell is unpleasant
and best avoided; its extreme intensity and putrid quality completely over-
whelm her voluntary judgments. It quite literally feels awful. The experien-
tial quality provides her with an additional reason, from the nature of the
experience itself, to avoid the smell. Simply being told that there is an awful
smell—that is, merely being informed about the noxious chemicals in the
air—would not motivate in the same way or explain why the olfactory expe-
rience feels so unpleasant.
17 Many other terms are used in the literature for these and related features of experience, including
valence and hedonic tone.

On the other hand, our affective judgments and the phenomenology of

affective perceptual experience strongly suggest that affect is not simply an
emotional or cognitive reaction: the felt awfulness is a quality of the percep-
tual episode and not a co-occurring emotion or judgment (though defending
this line is quite difficult, and left here for another time). When we have
an affect-laden perceptual experience, we naturally and correctly attribute
pleasant and unpleasant qualities to the sensory features of the object. The
smell of the spoiled milk, out there in the room, really seems awful; the taste
of fresh chocolate milk, out there in the glass, really is delicious. But this isn’t
the whole story. There are no objective sensible features of the milk that are
identical with its awfulness or deliciousness. Standard arguments from inter-
and intrapersonal differences and the extreme context sensitivity of affect,
combined with their direct motivating influence on judgments and beha-
vior, strongly suggest that affective qualities are not merely objective features
of the milk itself. The first glass of chocolate milk is most delicious; by the
second or third glass it is less so, perhaps even unpleasant. This despite the
fact that the sensible features of the milk remain constant (and it seems clear
that both judgments, about the first and nth glass, are correct).
This is the central puzzle of sensory affect. Almost everyone agrees that
there is a sensory-discriminatory dimension to affective experiences; Emina
experiences the smell as extremely sour, musty, and rotten, and as emanating
from the milk carton. It is a unique smell that she can clearly distinguish
from the other smells in the room. Such a discriminatory capacity is a cen-
tral feature of nearly all current theories of perceptual experience. But what
about the awfulness? Where does that fit in our best account of perceptual
experience? It seems unlikely to be just another sensible feature of the spoiled
milk, no different from its sourness or mustiness.18 But it also seems equally
unlikely to be a non-perceptual emotional, conative, or cognitive reaction
merely caused by her perceptual experience.19
Consider the following case:
Bomb: While on a flight, Josue unexpectedly sees a wired device that looks
like a bomb in another passenger’s bag, an experience that he would de-
scribe as highly unpleasant.
18 Tye 2006 defends a view like this.

19 Heathwood 2006, for instance, defends a view according to which an associated desire for or
against a sensory experience explains its pleasant (or unpleasant) character.
Bomb is a visual experience with a strong negative affective character. It has

all of the emotional and motivational characteristics of an affective response
(negative judgments and motor inclinations with respect to the visual ob-
ject). It also results directly from a visual experience.20 Nevertheless, there
is nothing essentially visual about the affective character in Bomb. What
was unpleasant about the experience wasn’t the way the bomb looked to
Josue. None of the visually salient properties of the bomb were unpleasant.
Consider what happens if Josue knows and expects his friend to have harm-
less electronics in her bag (perhaps because she is an engineer). Finding the
equipment in the bag would no longer be unpleasant. It would completely
lose its negative affect. And yet it seems the equipment would not look any
different. While unexpectedly seeing a bomb can be an unpleasant experi-
ence, there is nothing essentially visual about the unpleasantness. What was
unpleasant was the fact that there (potentially) was a bomb in the bag. This
information could have been attained in a variety of ways. Feeling the wires
in the bag or hearing an ominous ticking sound or being told that a bomb
was in the bag would have generated the same affective response. This is a
perceptual experience whose affective character is not essentially tied to the
manner in which it was experienced.21
Compare Bomb with the following.
Lutefisk: While traveling through Minnesota, Callie tries the local delicacy
lutefisk for the first time. Her sample is made from cod (not the usual white-
fish) and is a gelatinous, lye-brined flesh with a surprisingly intense, rotten
fish odor and flavor. Callie finds its flavor extremely unpleasant.
The experience in Lutefisk also has an unpleasant character, but this nega-
tive affect is essentially tied to the experiential qualities involved in eating it.
Here it is especially the smell and the texture that seems to be doing all the
work. Merely seeing the food does not feel unpleasant in the same way, nor
does being told that there is lutefisk on the menu. The experience of eating
it is not a result merely of the information provided by the experience, but a
result of the way the fish is presented in experience, by the way its smell, tex-
ture, and flavor are integrated and combined, then connected with associated
20 For a real-life example, see the 2015 case of student Ahmed Mohamed, whose homemade clock
was mistaken by his teacher for a bomb.

21 This example and the idea of separating cases of perceptual affect in this way are adapted from
Fulkerson 2014b.
emotional and motivational activations.22 The affective character of Bomb

seems to be a byproduct of our background knowledge and other conditions
(bombs are scary and dangerous, especially on airplanes). This leads to a
kind of reactionary negative affect, which involves diffuse judgments about
the situation, and an associated emotional reaction, involving among other
things an increase in a subject’s overall state of arousal, anxiety, and negative
assessments. Lutefisk, on the other hand, involves an affective character that
is primarily sensory (it is a particular gustatory, olfactory, and tactual signal
that tastes and smells and feels awful). In the latter case, what is at issue is the
way we perceive the fish.
Unlike with bombs, fake smells and tastes don’t lose their affective qualities
when we learn they are fakes. A synthetic rotting odor indistinguishable from
the smell of lutefisk will still smell unpleasant. Further, it does not lose its un-
pleasant aspect even when one learns that it is synthetic. An unpleasant “fake”
smell is just as unpleasant as the real thing, provided all of the modality-salient
qualitative features are the same. Mohan Matthen (2012) makes a related
point when he argues that while there can be fake oranges-the-fruit (objects
that look and smell just like real oranges, but that are not oranges), there is no
such thing as fake orange-the-color. Anything phenomenally indistinguish-
able in all contexts from the color orange is orange. Similarly, anything that
shares all of the modality-salient features with a foul odor also smells foul.
Knowing it is a fake does little to lessen its affective character.23
What is interesting about this case is that it highlights something impor-
tant about perceptual affect: it does sometimes seem to involve something
non- perceptual, something distinctly emotional and motivational. The
natural move then is to suggest that all cases of perceptual affect involve
such secondary reactions. But this move is too hasty. Some cases of percep-
tual affect involve experiences that are more unified and connected than a
completely separate emotional state would allow. These reactions are more
automatic and fused to the perceptual episode, in a way that can hide their
22 Of course, flavor is essentially multisensory. But so is perception in general. Nothing in my view
assumes that modalities are discrete informational channels or requires that they have their own
proprietary affective components. In this case, the features of Callie’s flavor experience involve the
sensory-discriminative contributions of her gustatory and retronasal olfactory system along with the
texture and thermal features contributed by her oral-tactual system, and further, the contributions of
distinct affective-motivational processing areas.
23 This is not to say that it does nothing. There is considerable evidence that affective-motivational
processing is highly context sensitive and easily influenced by top-down processes. These changes are
not the same as those involved in Bomb, but changes in the very initial stages of sensory processing.
influence from subjective awareness. Just as we do with pain, then, we can

think of some perceptual experiences as being constructed from a discrimi-
natory perceptual component matched with a functionally associated affec-
tive reaction (cf. Aydede 2000; Matthen 2005). While all cases of perceptual
affect seem to involve the activation of pleasure (or pain) centers in the brain
(albeit in diverse ways), there seem at the level of perceptual experience to
be two very different ways in which this can occur. One of these creates a
domain-general emotional response that is independent of the particular
modality through which we perceive; the other involves an affective char-
acter that cannot be separated from a specific modality. It is the latter cases
that I am attempting to explain here.
The view I’ve outlined is not only theoretically advantageous; it is also
strongly supported by the best empirical evidence. Much of this evidence
comes from recent work on smell and pleasant touch. As noted in a recent
study by Lapid et al. (2011): “A large body of research has suggested that
odorant pleasantness is the primary perceptual axis of human olfaction”
(1455). In their study, Lapid et al. found that there was a correlation between
locations of receptors in the olfactory bulb and perceived pleasantness. For
them, “pleasantness” refers to “an axis ranging from very unpleasant to very
pleasant, an axis that is also referred to as odor valence, or odor hedonic
tone” (1455). Their subject, then, is precisely what I am calling the affective
character of perceptual experience.
The correlations they found between receptor location and perceived
pleasantness lead them to support the idea that olfactory experience is hard-
wired to detect pleasantness. A better interpretation, I would suggest, is that
the olfactory system comes with some hardwired biases against certain odor
qualities. The notion of bias makes better sense of the fact that the affec-
tive character of odors is malleable and open to influence from other non-
perceptual sources. As Lapid et al. summarize:
Although our results imply a hard-wired aspect of odorant pleasantness,

they do not imply that odorant pleasantness is unmalleable, as a portion
of perception that is hardwired at the receptive surface does not rule out
later remapping. . . . Here we suggest that odorant pleasantness is mapped
onto the olfactory epithelium, yet this mapping is malleable by context and
experience, which influence possibly epithelial, and without question later
processing stages. (Lapid et al. 2011, 1460)
Their studies indicate that one of the primary functions of olfactory experi-
ence is the coding of pleasantness, but while there are some initial settings
(some smells we’re immediately inclined to dislike), these settings are mal-
leable. They can be overridden by experience and context. Some odors, like
vanillin, are almost universally experienced as inherently pleasant; others,
like sulfur, are almost universally experienced as inherently unpleasant. But
these can change: there are doubtless some people who can come to dislike
the smell of vanilla and some who can come to like the smell of sulfur. The
context sensitivity and phenomenological variety of affective olfactory expe-
rience means that what is hardwired isn’t our reaction to particular smells.
Instead, we seem hardwired to have a certain biased presentation of certain
basic olfactory qualities.
Further, and more telling, evidence for the nature of affective qualities can
be found in the recent literature on pleasant touch. Francis McGlone and
colleagues recently discovered a specialized channel in touch called the C-
tactile channel, that seems partially responsible for the feelings of pleasant-
ness in hairy (but not glabrous) skin. This channel is not found in any of the
other sensory modalities (Löken et al. 2009; McGlone et al. 2007). When we
have an especially pleasant touch experience, there is a dedicated channel
that helps generate a pleasant feeling, one that cannot be recreated in any of
the other sensory modalities:
A subclass of unmyelinated afferents (C-tactile) provided us with a candi-

date for such a specific role in mediating pleasant touch. They respond vig-
orously to slow and light stroking and are found only in hairy skin. C-tactile
afferents follow ascending pathways that are distinct from those of myeli-
nated tactile fibers. Selective C-tactile stimulation activates the left anterior
insular cortex, an area that has been implicated in the processing of positive
emotional feelings. (Löken et al. 2009, 547)
Because of these findings, it is now common in touch research to distinguish

these two components.
A distinction is made throughout this manuscript between “affective” and

“discriminative” touch. In short, affective touch refers to the emotional re-
sponse to tactile stimulation, with particular emphasis on the pleasantness
of such contact. Discriminative touch refers to perceptual attributes apart
from the hedonic aspects of touch, attributes that might, in principle be
linked rather closely to quantifiable physical aspects of the stimulus. (Essick

et al. 2010, 93)
What matters here for epistemic purposes is that touch, like the other senses,
seems to involve more than simply reporting on what’s going on in the imme-
diate environment. It doesn’t simply pick up objective information and pass
it along. One of its other primary functions is essentially evaluative; touch
informs us about the emotional qualities of objects. Silk and feathers feel
pleasant; sticky slime feels unpleasant. These feels are highly malleable, con-
text sensitive, driven by learning and exposure, and sensitive to activations
in other sensory modalities. This results in highly individualized reactions
to certain sensible qualities. While I might find wool intolerably itchy and
uncomfortable, someone else might find the material pleasingly warm and
comforting. In both cases our experiences are accurately reporting on the
evaluations made about the material, but our judgments differ immensely.
Any full accounting of the basic epistemic features of perceptual experi-
ence (especially in touch) will necessarily need to account for and assess this
evaluative role.
The more general upshot for my purposes is that, among the many integra-
tion effects that occur in perception, some are between sensory systems and
subsystems. These make important differences to the accuracy and reliability
of our sensory systems. Other interactions, however, are between discrim-
inatory and evaluative (or emotional) elements of our perceptual systems.
These interactions also make an epistemic difference, for they fundamentally
change what counts as accuracy and reliability for these forms of awareness.
Rather than an objective standard, they provide context-sensitive infor-
mation that primarily concerns the comfort and survival of the organism.
Like the more traditional interaction trade-offs we discussed earlier, these
interactions often sacrifice strict, objective accuracy in order to secure coher-
ence, stability, and comfort.
2.4. Summary
Here is where we are. Senses operating in isolation are a rarity, perhaps even
an idealization. Our senses do not always simply pass along environmental
information; instead they gather up information, and then, like judges or
stock traders, they jostle, influence, dominate, suppress, assist, compete,
boost, and disrupt the operations of the other active sensory systems. Our
perceptual experiences are the result of complex interactions that seek
overall accuracy through deep, competitive exchange. A consequence of this
exchange is sometimes the rejection or alteration of the deliverances of an
individual modality. These same concerns apply within modalities. The epi-
stemic quality of a given haptic perception will involve the operation of many
distinct elements. In addition, the senses are often not simply descriptive.
They also often engage in evaluation and motivation (cf. Akins 1996).
The key point of the preceding is this: adding sensory resources isn’t (al-
ways) like adding together separate forms of evidence for consumption and
evaluation by a discerning and rational subject. More often, it’s like receiving
a detailed, univocal report from a committee whose individual members may
have vehemently disagreed with one another, and yet were forced to find con-
sensus. More worrisome: sometimes the output we receive is the consensus
report from multiple committees, some of which have primary concerns
well beyond strict, objective accuracy. What the subject is presented with,
then, are not anything like separate, independent channels of information.
In fact, these channels have been interacting, coordinating, and connecting
both with each other and with other aspects of the perceptual context. One
immediate consequence of this, even if we want to maintain our traditional
views of perceptual epistemology, is that our account needs to be holistic;
there should be, strictly speaking, no such thing as visually justified beliefs,
or haptically justified beliefs. The evidence strongly suggests that we need to
assess perceptual justification for all the senses together at a time. Perceptual
justification is multisensory justification.
The resulting situation appears to be that our empirical beliefs are justified
by all of our perceptual modalities working together, by the totality of our
perceptual faculties working in concert. This, to loosely paraphrase Quine,
treats perception as a genuine tribunal that judges incoming information all
at once, as a corporate body. If true, this is a substantive, interesting conclu-
sion stemming from some rather low-level facts about haptic and sensory
processing.
If this story is on the right track, then a more holistic approach to percep-
tion, with more sensitivity to the role of sensory interactions in generating
perceptual experience, would be welcome. Still, even if this holistic view is
just an extension of something like standard process reliabilism, it is a deep
and complex extension. Instead of talking about reliability full stop we now
need to think about the competing elements of a far more complex system.
Testimony, for instance, seems to force similar changes in standard reliabilist

positions (Goldberg 2010). There are different sources with their own re-
liability, and there need to be complex forms of assessment for their many
varied interactions. As I’ve tried to argue, when there are competing eviden-
tial sources involved, we get something more than mere summation of relia-
bility. There are many ways the senses aim for accuracy, and indeed different
conceptions of what accuracy even means. Sometimes it can be best, from
an epistemic point of view, to sacrifice the accuracy of one modality in order
to generate overall coherence or to boost the signal of another, more salient,
modality. This is especially so when one sense is more reliable concerning
some feature or domain. This suggests that we need epistemic constraints
that can decide, for instance, between discounting and facilitation, and re-
liability alone can’t do this work. In addition, of course, in our current en-
vironment many sensory interactions will be systematically unreliable (in
terms of their strict accuracy), but nevertheless epistemically meritorious
because they promote and secure experiential coherence, unity, and useful-
ness, which have their own positive epistemic features (again, consider the
ubiquity and usefulness of the ventriloquist effect in providing the inaccurate
apparent locations of sound sources).
Given this, it’s not clear that sensory interactions always lead to more accu-
rate representations overall, or to more accurate experiences than we would
get with just a single modality. For this reason, a focus on reliability strictly
in terms of accuracy will miss out on the most important elements of the
interactions I’ve discussed. While each individual stream or sensory system
carries a limited range of information about the distal environment (and so
seems entirely concerned with accuracy), the processes of integration aren’t
concerned with accuracy alone. Or at least, even if they are, there are other
important constraints that must be factored into our analysis. Just as it’s
not always better to add senses, it’s also not always better epistemically to
have one simple, highly accurate modality. When there are multiple systems
interacting, other concerns become important. We need a more nuanced,
fine-grained story to pull these possibilities apart.
We can see now why adding extra modalities can sometimes lead us astray.
The processes of integration do not seem to operate in a way that always
ensures accuracy—at least, not accuracy for each individual modality taken
in isolation. An important element of how the senses operate to ensure accu-
racy depends on balancing a large set of competing considerations. It is im-
portant for our continued survival that our overall experiences be coherent,
coordinated, unified, fast, and consistent. It is not clear that our senses can
sacrifice all of these other considerations in favor of accuracy alone. Only by
looking at the details of how these systems operate and connect their outputs
can we start to build an epistemic theory that takes these features properly
into account.
Acknowledgments
I would like to thank participants in both the UC San Diego Philosophy

of Perception reading group and my recent graduate seminar for helpful
feedback on earlier versions of this material. I also received many helpful
suggestions and valuable feedback from audiences at talks I’ve given at
Illinois State, Toronto, the UC Working Group on Sensory Interaction,
and the 2015 Society for Philosophy and Psychology. Finally, I would like
to thank Jonathan Cohen, Lana Kühle, Kevin Connolly, and Kurt Sylvan for
their more extensive and extremely helpful comments on earlier drafts.
References
Akins, K. 1996. Of sensory systems and the “aboutness” of mental states. Journal of
Philosophy 93(7): 337–372.
Aydede, M. 2000. An analysis of pleasure vis- à-vis pain. Philosophical and
Bennett, D., and C. Hill. 2014. Sensory Integration and the Unity of Consciousness.
Cambridge, MA: MIT Press.
Bergmann Tiest, W. M., and A. M. L. Kappers. 2007. Haptic and visual perception of
roughness. Acta Psychologica 124(2): 177–189.
Bertelson, P. 1999. Ventriloquism: A case of crossmodal perceptual grouping. Advances in
Psychology 129: 347–362.
Calvert, G., and T. Thesen. 2004. Multisensory integration: Methodological
approaches and emerging principles in the human brain. Journal of Physiology Paris
98(1–3): 191–205.
Campos, J. L., J. S. Butler, and H. H. Bülthoff. 2012. Multisensory integration in the esti-
mation of walked distances. Experimental Brain Research 218(4): 551–565.
Cole, J. 1995. Pride and a Daily Marathon. Cambridge, MA: MIT Press.
Craig, A. D. 2002. How do you feel? Interoception: The sense of the physiological condi-
tion of the body. Nature Reviews Neuroscience 3(8): 655–666.
Deroy, O., Y. C. Chen, and C. Spence. 2014. Multisensory constraints on aware-
ness. Philosophical Transactions of the Royal Society B: Biological Sciences
369(1641): 20130207.
Driver, J., and C. Spence. 2000. Multisensory perception: Beyond modularity and conver-
gence. Current Biology 10(20): 731–735.
Ernst, M. O., and M. S. Banks. 2002. Humans integrate visual and haptic information in a
statistically optimal fashion. Nature 415(6870): 429–433.
Ernst, M. O., C. Lange, and F. N. Newell. 2007. Multisensory recognition of actively
explored objects. Canadian Journal of Experimental Psychology /Revue Canadienne de
Psychologie Expérimentale 61(3): 242–253.
Essick, G. K., F. McGlone, C. Dancer, D. Fabricant, Y. Ragin, N. Phillips, T. Jones, and S.
Guest. 2010. Quantitative assessment of pleasant touch. Neuroscience and Biobehavioral
Reviews 34(2): 192–203.
Evans, G. 1985. Molyneux’s question. In Collected Papers. Oxford University Press.
Fodor, J. 1983. The Modularity of Mind. Cambridge, MA: MIT Press.
Frissen, I., J. L. Campos, J. L. Souman, and M. O. Ernst. 2011. Integration of vestib-
ular and proprioceptive signals for spatial updating. Experimental Brain Research
212(2): 163–176.
Fulkerson, M. 2014a. Rethinking the senses and their interactions: the case for sensory
pluralism. Frontiers in Psychology 5: 1426. https://doi.org/10.3389/fpsyg.2014.01426.
Fulkerson, M. 2014b. The First Sense. A Philosophical Study of Human Touch. Cambridge,
MA: MIT Press.
Goldberg, S. 2010. Relying on Others: An Essay in Epistemology. New York: Oxford
University Press.
Graziano, M. S. 1999. Where is my arm? The relative role of vision and proprioception in
the neuronal representation of limb position. Proceedings of the National Academy of
Sciences of the United States of America 96(18): 10418.
Grice, H. P. 1962/2002. Some remarks about the senses. In A. Noë and E. Thompson,
eds., Vision and Mind: Selected Readings in the Philosophy of Perception. Cambridge,
MA: MIT Press. 35.
Helbig, H. B., and M. O. Ernst. 2007. Optimal integration of shape information from vi-
sion and touch. Experimental Brain Research 179(4): 595–606.
Jones, L. A., and S. J. Lederman. 2006. Human Hand Function. New York: Oxford
University Press.
Klatzky, R., and S. J. Lederman. 2010. Multisensory texture perception. In M. J. Naumer
and J. Kaiser, eds., Multisensory Object Perception in the Primate Brain. New York:
Springer. 211–230.
Knox, J., P. Cordo, R. Skoss, S. Durrant, and P. Hodges. 2006. Illusory changes in head
position induced by neck muscle vibration can alter the perception of elbow position.
Behavioral Neuroscience 120(6): 1211–1217.
Lapid, H., S. Shushan, A. Plotkin, H. Voet, Y. Roth, T. Hummel, E. Schneidman, and N.
Sobel. 2011. Neural activity at the human olfactory epithelium reflects olfactory per-
ception. Nature Neuroscience 14(11): 1455–1463.
Lederman, S. J., and R. L. Klatzky. 1987. Hand movements: A window into haptic object
recognition. Cognitive Psychology 19(3): 342–368.
Löken, L. S., J. Wessberg, I. Morrison, F. McGlone, and H. Olausson. 2009. Coding
of pleasant touch by unmyelinated afferents in humans. Nature Neuroscience
12(5): 547–548.
Lyons, J. 2009. Perception and Basic Beliefs: Zombies, Modules, and the Problem of the
External World. New York: Oxford University Press.
Macpherson, F. 2011. Cross-modal experiences. Proceedings of the Aristotelian Society

111(3pt3): 429–468.
Maravita, A., C. Spence, and J. Driver. 2003. Multisensory integration and the body
schema: Close to hand and within reach. Current Biology 13(13): R531–R539.
Matthen, M. 2014. How to be sure: Sensory exploration and empirical certainty.
Philosophy and Phenomenological Research 88(1): 38–69.
McGlone, F., A. B. Vallbo, H. Olausson, L. Löken, and J. Wessberg. 2007. Discriminative
touch and emotional touch. Canadian Journal of Experimental Psychology /Revue
Canadienne de Psychologie Expérimentale 61(3): 173–183.
Naumer,M.J.,andJ.Kaiser.2010.MultisensoryObjectPerceptioninthePrimateBrain.NewYork:
Springer.
O’Callaghan, C. 2008. Seeing what you hear: Cross-modal illusions and perception.
Philosophical Issues 18(1): 316–338.
O’Callaghan, C. 2011. Against hearing meanings. Philosophical Quarterly 61(245):
783–807.
O’Callaghan, C. 2012. Perception and multimodality. In E. Margolis, R. Samuels, and S.,
Stich, eds., Oxford Handbook of Philosophy of Cognitive Science. New York: Oxford
Parker, A. 2009. In the Blink of an Eye: How Vision Sparked the Big Bang of Evolution. New York:
Basic Books.
Spence, C., and M. Shankar. 2010. The influence of auditory cues on the perception of, and
responses to, food and drink. Journal of Sensory Studies 25(3): 406–430.
Stein, B. E., ed. 2012. The New Handbook of Multisensory Processes. Cambridge, MA:
MIT Press.
Thompson, E. 2007. Mind in Life. Cambridge, MA: Harvard University Press.
Tye, M. 2006. The thesis of nonconceptual content. European Review of Philosophy 6: 7–30.
3
Multimodal Mental Imagery and
Perceptual Justification1
Bence Nanay
3.1. Two Empirical Problems

with Perceptual Justification
Some perceptual states can justify beliefs. If I look out of the window and
see that it is raining outside, this may (and as a default, would) justify my
belief that it is raining outside. But this simple picture of the relation be-
tween perception and belief has been questioned recently. A vast number
of experiments in perceptual psychology purport to show that perception
is not an encapsulated process: it can be influenced by our beliefs. And these
findings about top-down influences on perception are said to have important
implications for the potential epistemic role perception may play. The ge-
neral line of argument is this: if perception is cognitively penetrated, then it is
not an unbiased way of learning about the world, as our preexisting thoughts,
beliefs, and expectations will influence how and what we perceive.
So we get a vicious circularity: our beliefs, thoughts and expectations are
supposed to be based on and justified by our perceptual states, but these per-
ceptual states themselves are influenced by our beliefs, thoughts, and ex-
pectations (because of cognitive penetration). What we get is some kind of
“wishful seeing” (see Pylyshyn 1999, who argues that seeing is not wishful in
this sense). As Roberto Bolano says in the novel 2666, “People see what they
want to see and what people want to see never has anything to do with the
truth.”2
1 This work was supported by the ERC Consolidator grant [726251], the FWO Odysseus grant
[G.0020.12N] and the FWO research grant [G0C7416N].

2 Roberto Bolano: 2666. London: Picador, 2009, 219.
Bence Nanay, Multimodal Mental Imagery and Perceptual Justification In: The Epistemology of
Non-Visual Perception. Edited by: Berit Brogaard and Dimitria Electra Gatzia, Oxford University Press
(2020). © Oxford University Press.
DOI: 10.1093/oso/9780190648916.003.0004
The challenge from cognitive penetration was originally focusing on one

specific account of perceptual justification, namely, dogmatism (Siegel 2011;
see also Lyons 2011): the view that “whenever you have an experience as of p,
you thereby have immediate prima facie justification for believing p” (Pryor
2000, 536). The argument was that if perception is cognitively penetrated,
dogmatism is not an option because the perceptual states that our beliefs are
supposed to be justified by are themselves influenced by our existing beliefs
and expectations. This argument has been generalized to apply to other the-
ories of justification (not just dogmatism; see Siegel 2011; Tucker 2014; see
also Lyons 2015; Ghijsen 2016; Silins 2016).
The main aim of this chapter is to argue that even if the cognitive pene-
trability of perception is not something theories of perceptual justification
need to worry about, a different set of findings in perceptual psychology pose
a much more serious challenge to the very idea of perceptual justification.
These findings are about the importance of perceptual processing that is not
driven by corresponding sensory stimulation in the relevant sense modality
(such as amodal completion and multimodal completion).
I argue that these findings show that everyday perception is in fact a mix-
ture of sensory-stimulation-driven perceptual processing and perceptual
processing that is not driven by corresponding sensory stimulation in the
relevant sense modality, and that we have strong reasons to doubt the epi-
stemic pedigree of the latter process. The implication of this is not that we
should become skeptics or deny the possibility of perceptual justification. It
is, rather, that the only way in which we can understand when and whether
a perceptual state justifies beliefs is by paying close attention to empirical
facts about the reliability of perceptual processing that is not driven by corre-
sponding sensory stimulation in the relevant sense modality. In this sense (a
very narrow sense) epistemology needs to be naturalized.
3.2. The First Problem: Top-Down Influences

on Perception
One influential debate about perception is about whether it is an encapsu-

lated process that is protected from any kind of top-down influences or it is
influenced and modified by top-down information. What complicates this
debate, often referred to as the cognitive penetrability debate, is that it is not
at all clear what kind of mental state is supposed to be doing the penetrating
Multimodal Mental Imagery and Justification 79
and what kind of mental state is supposed to be penetrated. In other words,

it is not clear what is “top” and what is “below” in the debate about top-down
influences on perception.
Once we clarify these conceptual issues, it seems that there is a wealth of
empirical evidence in favor of the claim that there are indeed some top-down
influences on perception. But then the question becomes: Is this something
epistemologists would have to worry about? I will argue that they wouldn’t.
The main conceptual confusion concerning debates about top-down
influences on perception is that it is not clear what is meant by “perception”
in this context. Some (especially philosophers [Siegel 2011; Macpherson
2012; Stokes 2012], but also some psychologists [see, e.g., Firestone and
Scholl 2014, 2016]) take “perception” in this context to be perceptual expe-
rience: something we are consciously aware of. If we work with this concept
of perception, then the question is whether top-down influences can alter
the way we experience a scene—the phenomenal character of our experi-
ence: what it is like to perceive this scene.
Another way of understanding what is meant by “perception” when we
talk about top-down influences on perception is perceptual processing—
something neuroscientists (and also some psychologists) worry about.
Here the question is whether processing in, say, the primary visual cortex is
influenced in a top-down manner.
These two questions are clearly very different—one of them is about phe-
nomenology and the other is about early perceptual processing. And as
changes in early perceptual processing are neither necessary nor sufficient
for changes in perceptual phenomenology, there is no easy traffic between
these two different sub-debates.
The methodology of addressing these questions is also different. When it
comes to figuring out what influences our perceptual phenomenology, we
need to rely on self-reports and introspection (together with some form of
inference to the best explanation). When it comes to early perceptual pro-
cessing, we need to rely on neuro-imaging data (and perhaps also some be-
havioral data, like reaction-time). The problem is that it is very rarely made
clear which of these very different questions is being asked.
I argued elsewhere (Teufel and Nanay 2017) that it is doubtful that the
first of these debates could ever be resolved in a satisfactory manner. One
reason for this is the well-documented unreliability of introspection (see
Schwitzgebel 2008; Spener and Bayne 2010)— whereas introspection is
something this debate needs to heavily rely on. But an even more important
worry is that any way of resolving this debate would need to appeal to a clear
and unproblematic way of keeping apart perceptual and non-perceptual
phenomenology—something that we do not have. Given that it is not clear
what perceptual phenomenology is and how to keep it apart, introspectively,
from non-perceptual phenomenology, the question about perceptual phe-
nomenology (of whether perceptual phenomenology depends on top-down
influences) inherits all these problems. It is unclear then how we can make
any progress in answering this question—if we take the question about top-
down influences to be about perceptual phenomenology.
So, as a result, I take it that a more promising (or more productive) debate
about top-down influences on perception is about whether early perceptual
processing is influenced in a top-down manner. And here we find a lot of ex-
tremely convincing neuro-imaging and behavioral evidence for the existence
of top-down influences (see, for example, Murray et al. 2002; Gandhi et al.
1999; O’Connor et al. 2002). But then the question becomes: Are these results
relevant from the point of view of epistemology?
Observant readers may have noticed that I talked about “top-down
influences on perception” and not about cognitive penetration so far. There is
a reason for this. The term “cognitive penetration” suggests that whatever is
doing the penetration is a cognitive state, and this is not something I want to
be built into the very notion I am analyzing.
When I talk about “top-down” influences on perception, I want to allow for
any “top-down” influences—not just those that are labeled “cognitive.” And it
is not very clear why the label “cognitive” is singled out. “Cognitive” can mean
many things. It is sometimes contrasted with “affective,” but this is clearly not
something we want to do if we are interested in top-down influences on per-
ception, as there may be affective influences on perception and they may be
as important as (or more important than) nonaffective cognitive influences
(Schupp et al. 2004; Schmitz et al. 2009; Pessoa and Ungerleider 2005). The
term “cognitive” is also often contrasted with “conative,” but this also a du-
bious usage in the present context, as there may be very good reasons to posit
top-down influences on perception where it is a desire or an intention that
influences our perceptual processing (Nanay 2006; Stokes 2012).
Of course, the most straightforward use of “cognitive” may just be one
where it is contrasted with “perceptual,” but this simplifies things consider-
ably. In fact, one reason why it is better to focus on the debate about whether
there are top-down influences on early perceptual processing than on the one
about whether there are top-down influences on perceptual phenomenology
is that if we focus on the latter debate, the only kind of top-down influence
we can talk about is from mental states (normally with non-perceptual phe-
nomenology) to mental states with perceptual phenomenology (i.e., percep-
tual experiences). But we have seen that addressing any questions about the
presence or absence of such top-down influences then requires a very clear
distinction between perceptual and non-perceptual phenomenology, and we
don’t have any distinction, let alone a clear one.
If, on the other hand, we consider the debate about whether there are
top-down influences on early perceptual processing, we get a more detailed
picture. We have a very clear idea of the sequence in which “bottom-up” per-
ceptual processing proceeds. To take the visual sense modality as an example
(Katzner and Weigelt 2013; Grill-Spector and Malach 2004; Van Essen 2004;
Bullier 2004), in humans and nonhuman primates, the main visual pathway
connects neural networks in the retina to the primary visual cortex (V1) via
the lateral geniculate nucleus (LGN) in the thalamus; outputs from V1 ac-
tivate other parts of the visual cortex and are also fed forward to a range of
extrastriate areas (V2, V3, V4/V8, V3a, V5/MT). So visual processing goes
from the retina via the LGN to the primary visual cortex (V1), then to the
secondary visual cortex (V2), then (in the case of color vision) to the V4, and
so on. So we have a natural ordering of mental processes in terms of what is
bottom and what is top. If there are influences from V4 to the primary visual
cortex, it is a top-down influence because it is from a mental process (V4)
that comes later in visual processing than the primary visual cortex (V1),
which is the mental process that is being influenced.
If we raise the question of top-down influences on perceptual processing,
we get a more nuanced picture. The question of top-down influences is no
longer a yes-or-no question as in the case of the phenomenology interpreta-
tion (either there is cognitive penetration or there isn’t), but a multifaceted
one. Maybe the primary visual cortex is influenced in a top-down manner
by V2 and V4, but not by our expectations and beliefs. Or maybe it is only
influenced by V2. Or maybe also by our expectations and beliefs. All of these
claims would assert top-down influences on early perceptual processing
(of which, as we have seen, there is very strong evidence), but it matters a
lot what kind of top-down influences they are (see Nanay and Teufel 2016
for a detailed analysis of various kinds of top-down influences on early cor-
tical perceptual processing and the differences between those top-down
influences that come from within the visual system and those that come from
post-perceptual processing).
While some of these findings show very rigorously that there are top-
down influences on perception, it is important to see that many of these
findings are irrelevant in the present context. Epistemologists are worried
about whether our perception is influenced by our beliefs or other cognitive
states. So top-down influences on perceptual processing would be relevant if
the “top” in these influences were beliefs or other cognitive states. That is why
epistemologists are interested in cognitive penetration as opposed to any
form of top-down influences (where I take cognitive penetration to be the
penetration of perception by beliefs or other cognitive states—thus the label
“cognitive”). Epistemologists are not at all interested in whether the primary
visual cortex is influenced by the V4/V8 or the STS or the MT.
But don’t we have empirical evidence that perceptual processing is
influenced by non-perceptual or cognitive states? While some (including
myself; see Teufel and Nanay 2016) would say so, this premise would be easy
to dispute. Any such argument would need to presuppose some form of firm
distinction between perceptual and non-perceptual processing (not between
perceptual and non-perceptual phenomenology as before, but perceptual
and non-perceptual processing)—so that we can zero in on those top-down
influences that go across this divide. The problem is that, depending on how
we draw this divide between perceptual and non-perceptual, any finding
about top-down influences on perception could be framed as merely showing
an intra-perceptual top-down effect.
To sum up, there are two questions of top-down influences on perception.
One (concerning perceptual processing) is by and large irrelevant for episte-
mology. The other (about perceptual phenomenology) would be relevant for
epistemology, but it is difficult to see how this debate could be resolved in a
satisfactory manner. As a result, top-down influences on perception (or cog-
nitive penetration) should not give a significant cause for concern for anyone
who is interested in perceptual justification. Empirical findings about top-
down influences on perception do not seem to jeopardize any philosophical
account of perceptual justification.3
But there is another set of empirical findings about perception that should
be much more worrisome to anyone who is interested in perceptual justifica-
tion: the findings about the importance of perceptual processes that are not
triggered by corresponding sensory stimulation in the given sense modality.
3 I should note that some dogmatists take a different route to argue that the findings about cogni-
tive penetration do not threaten their account of perceptual justification (see esp. Pryor 2000; see also
Huemer 2006). The focus of this chapter is not dogmatism.
3.3. The Second Problem:

Imagery-Infused Perception
Some perceptual processing starts with sensory stimulation. The light hits
our retina, and vision is the complex visual processing of this sensory stim-
ulation. This perceptual processing may include, depending on whom you
ask, the interpretation or the elaboration or the embellishment of the sensory
stimulation, but it is the sensory stimulation that is processed /interpreted /
elaborated on.
But some other cases of perceptual processing are not the processing of
sensory stimulation because there is no sensory stimulation to be processed.
These perceptual processes are not triggered by corresponding sensory stim-
ulation in the relevant sense modality.
In psychology and neuroscience, perceptual processes that are not trig-
gered by corresponding sensory stimulation in the relevant sense mo-
dality are called “mental imagery.” Here is a representative definition
from a recent review paper: “We use the term ‘mental imagery’ to refer to
representations . . . of sensory information without a direct external stim-
ulus” (Pearson et al. 2015, 590). I realize that this way of talking about mental
imagery may be somewhat controversial for philosophers, but I will use the
concept of mental imagery as a technical concept in this chapter and use
the term “mental imagery” as a shorthand for “perceptual processing that
is not triggered by corresponding sensory stimulation in the relevant sense
modality.” Those readers who really dislike this way of using the concept of
mental imagery could just substitute “mental imagery*” (or “phantom per-
ception” or “offline perception” or whatever they like) for “mental imagery.”
I will argue that we have substantial empirical evidence that the vast
majority of what we take to be perception is really a mixture of sensory-
stimulation-driven perceptual processing and perceptual processing that
is not triggered by corresponding sensory stimulation in the relevant sense
modality. In other words, we have substantial empirical evidence that the
vast majority of what we take to be perception is really a mixture of sensory-
stimulation-driven perceptual processing and mental imagery. And these
findings pose a serious challenge to the very idea of perceptual justification.
Before we go on, I should emphasize that “perceptual processing” in this
definition means early cortical perceptual processing (Katzner and Weigelt
2013; Grill-Spector and Malach 2004; Van Essen 2004; Bullier 2004). In the
case of sensory-stimulation-driven perception, we have a correspondence
between this perceptual processing and the sensory stimulation. In the visual
sense modality, for example, early perceptual processing is retinotopic. The
primary visual cortex (and also many other parts of the visual cortex; see
Grill-Spector and Malach 2004 for a summary) is organized spatially in a way
that is very similar to the retina—it is retinotopic. While this retinotopy of
the early visual cortices (and their equivalent in the other sense modalities;
see, e.g., Talavage et al. 2004) is an extremely convenient way of gaining evi-
dence about the correspondence or lack thereof of sensory stimulation and
perceptual processing, this is just one way in which the two can correspond.
There are others. In other words, the correspondence between sensory stim-
ulation and perceptual processing does not have to be retinotopic.
But things are more complicated when it comes to other aspects of percep-
tion, for example, color perception. We can’t just read off the correspondence
or lack thereof off the relation between the retina and V4. Colorblind people
never have the kind of correspondence between the retina and V4 that we
would expect—yet, they, presumably, should not be described as always en-
tertaining mental imagery.
In order to bypass these worries, here is a more neutral way of charac-
terizing correspondence. A certain sensory stimulation reliably triggers an
early perceptual processing of a certain kind. If this early perceptual pro-
cessing happens without sensory stimulation that would reliably trigger it,
this means that there is no correspondence—we have mental imagery. It is
important that this reliability is understood restricted to the subject in ques-
tion. In a colorblind person, for example, a certain sensory stimulus would
reliably trigger certain early perceptual processing in a way that would not be
reliably triggered in other, not colorblind subjects.
Finally, visual mental imagery is visual perceptual processing that is not
triggered by visual sensory stimulation. As we shall see later, visual mental
imagery can be (and is often) triggered by non-visual (for example, auditory)
sensory stimulation. I call this form of mental imagery multimodal mental
imagery.
3.4. Varieties of Mental Imagery
I’ve defined mental imagery as perceptual processing that is not triggered

by corresponding sensory stimulation in the relevant sense modality. It
is important that this definition remains silent on a number of features of
mental imagery. For example, it allows for voluntary and involuntary mental
imagery (as perceptual processing that is not triggered by corresponding
sensory stimulation in the relevant sense modality may or may not be volun-
tary). It allows for conscious and unconscious mental imagery (as there is no
restriction that perceptual processing that is not triggered by corresponding
sensory stimulation in the relevant sense modality would need to be con-
scious). And it is also neutral about whether mental imagery is accompanied
by the feeling of presence.
As I have indicated, if the reader has a strong conviction that mental im-
agery is necessarily conscious (or necessarily voluntary or necessarily not
accompanied by the feeling of presence), she can take the rest of the discus-
sion to be about a very different technical concept, “mental imagery*.” I de-
fend the plausibility of this way of thinking about mental imagery (and, as a
result, the continuity between the psychological/neuroscientific and the phil-
osophical/everyday conception of mental imagery) elsewhere (Nanay 2016,
2018a, forthcoming), but I do not need it for the argument in this chapter.
Many different kinds of perceptual processes will count as mental imagery
according to this definition, some of which may not be normally categorized
as such. Here are a couple of examples:
The blind spot of the retina cannot be stimulated—there are no receptors
there. If the light hits this part of the retina it gives rise to no sensory pro-
cessing. So we receive no sensory information from that region of the retina.
Nonetheless, our perceptual system “fills in” the sensory input of the blind
spot on the basis of the sensory input of the surrounding parts of the retina.
The perceptual processing of information at the blind spot region of the
visual field happens already in early visual cortices (Komatsu et al. 2000;
Ramachandran 1992; Awater et al. 2005; Spillman et al. 2006; Fiorani et al.
1992), but it is not triggered by corresponding sensory stimulation because
there is no sensory stimulation at the blind spot, let alone corresponding sen-
sory stimulation.4 Processing at the blind spot counts as mental imagery.
4 One may object: hasn’t Daniel Dennett’s repeated skepticism about “filling in” the blind spot
(Dennett 1991, 335ff.) demonstrated that this story is incorrect? I don’t think so. First, there is
plenty of empirical evidence that the early cortices do actively “fill in” the missing part of the visual
scene (see, e.g., Komatsu et al. 2000 and also Churchland and Ramachandran 1993—and also Akins
and Winger 1996 for a very good overview of this debate). Second, I’m not even sure that Dennett
would disagree with anything I say here—his concern in Dennett 1991 was about phenomenology—
whether there is conscious filling in. And I’m certainly not arguing that there is—there is cortical
filling in. The mental imagery involved in the filling in of the blind spot is almost always unconscious.
As does peripheral vision. Peripheral regions of the retina are much less
sensitive than focal ones. And this focal preference is even stronger in early
cortical processing. As a result, the properties of the peripheral regions of
the visual field that our perceptual system processes are much less determi-
nate than the properties of the focal regions. This asymmetry is especially
striking when it comes to color vision, as there are very few retinal cells in
the periphery that are sensitive to color information (Hansen et al. 2009).
But the same is true for all other perceptually processed properties, like size
or shape. Peripheral vision can also “fill in” some regions of the periphery.
“Artificial scotoma” is a region of the visual field where different sensory
stimulation is induced from what surrounds it (and this can be no sensory
stimulation surrounded by a pattern). If this is presented in the periphery,
the visual system fills in the scotoma, making it blend in. This filling in pro-
cess starts very early in the visual processing (De Weerd et al. 1995, 1998,
2006; Ramachandran and Gregory 1991; Weil et al 2007, 2008; Troncoso
et al. 2008; Welchman and Harris 2001). Again, we have perceptual pro-
cessing that is not triggered by corresponding sensory stimulation (as the
sensory stimulation at the artificial scotoma is very different from what is
perceptually processed).
Amodal completion also counts as mental imagery according to my defini-
tion. Amodal completion is the representation of those parts of the perceived
object that we get no sensory stimulation from. In the case of vision, it is the
representation of occluded parts of objects we see: when we see a cat behind
a picket fence, our perceptual system represents those parts of the cat that
are occluded by the picket fence. In tactile perception, it is the completion
of those parts of the objects we touch that are not in direct contact with our
hand, for example. We complete those parts amodally.5 Amodal completion
is, by definition, perceptual processing that is not triggered by corresponding
sensory stimulation (Nanay 2010, 2018b). The mental imagery involved in
amodal completion may bring about very similar issues for the very idea of
perceptual justification as the ones I raise about multimodal mental imagery
in this chapter—see Helton and Nanay 2019).
But this chapter is about a form of mental imagery where perceptual pro-
cessing in one sense modality is triggered by sensory stimulation in another
sense modality: this chapter is about multimodal mental imagery.
5 Note that the term “amodal” is a bit of a misnomer here: the completion by any account happens
visually—the term “amodal,” traditionally, was supposed to indicate that this process is not triggered
by sensory stimulation.
3.5. Multimodality and Mental Imagery
Multimodal perception is the norm and not the exception—our sense modal-
ities interact in a variety of ways (see Spence and Driver 2004; Vroomen et al.
2001; Bertelson and de Gelder 2004 for summaries; and O’Callaghan 2008a,
2011, as well as Macpherson 2011, for philosophical overviews). Information
in one sense modality can influence and even initiate information processing
in another sense modality at a very early stage of perceptual processing (even
in the primary visual cortex in the case of vision, for example; see Watkins
et al. 2006).
A simple example is ventriloquism (Bertelson 1999; O’Callaghan 2008b).
The auditory sense modality identifies the ventriloquist as the source of the
voices, while the visual sense modality identifies the dummy. And as a result
of the influence of vision, we auditorily experience the voices as coming from
the dummy.
What I am interested in here is not multimodal perception, but multi-
modal mental imagery: cases where there is perceptual processing in one
sense modality that is not triggered by corresponding sensory simulation in
that sense modality, but rather by corresponding sensory stimulation in a
different sense modality.
Here is an example from Nanay 2018a: When I am looking at my coffee
machine that makes funny noises, this is an instance of multisensory
perception—I perceive this event by means of both vision and audition. But
very often we only receive sensory stimulation from a multisensory event by
means of one sense modality. If I hear the noisy coffee machine in the next
room, that is, without seeing it, then the question arises: how do I represent
the visual aspects of this multisensory event?
We know that our visual system in these circumstances does get acti-
vated (and even the very early visual cortical areas can; see Hertrich et al.
2011; Pekkola et al. 2005; Zangaladze et al. 1999; Ghazanfar and Schroeder
2006; Martuzzi et al. 2007; Calvert et al. 1997; James et al. 2002; Iurilli et al.
2012; Vetter et al. 2014). In other words, there is early cortical activation in
the visual sense modality without corresponding sensory stimulation in this
sense modality. That is, we represent these features by means of mental im-
agery. I call this form of mental imagery multimodal mental imagery.
Remember the last phrase in the definition of mental imagery: perceptual
processing that is not triggered by corresponding sensory stimulation in the
relevant sense modality. This phrase is crucial in the present context. Mental
imagery can be triggered by sensory stimulation as long as it is not in the rel-

evant sense modality.6
In other words, depending on what perceptual processing is triggered by,
we get different perceptual phenomena. If perceptual processing is triggered
by corresponding sensory stimulation in the relevant sense modality, we get
sensory stimulation-driven perception. If it is triggered by sensory stimu-
lation in another sense modality, we get multimodal mental imagery. If it is
triggered by something else, we get some other kind of (non-multimodal)
mental imagery. In short, multimodal mental imagery is mental imagery in
one sense modality induced by sensory stimulation in another sense mo-
dality. And, as we have seen, we have strong empirical evidence that non-
sensorily driven perceptual processing in any sense modality can be induced
by sensory stimulation in any other sense modality.
But the vast majority of the entities we encounter are multisensory entities
(that is, entities that we could encounter by means of more than one sense
modalities). And our perceptual access to these multisensory entities is very
rarely absolute (that is, encompassing all relevant sense modalities). If we put
together these two claims, what we get is that multimodal mental imagery is
the norm, not the exception.
Here is a nice experimental illustration of this point: the double-flash
illusion. Subjects are presented with one flash and two beeps simultane-
ously (Shams et al. 2000). So the sensory stimulation in the visual sense
modality is one flash. But they experience two flashes, and already in the
primary visual cortex, two flashes are processed (Watkins et al. 2006). This
means that the double-flash illusion is really about multimodal mental im-
agery: we have perceptual processing in the visual sense modality (again,
already in V1) that is not triggered by corresponding sensory stimulation
in the visual sense modality (but by sensory stimulation in the auditory
sense modality).
The picture we ended up with is one where perceptual processes consist of a
sensory-stimulation-driven and a non-sensory-stimulation-driven compo-
nent (where by sensory-stimulation-driven, I mean driven by corresponding
6 Much of this section is about the intricate connections between different sense modalities.
Nonetheless, in the definition of mental imagery in general and of multimodal mental imagery in
particular, I am relying on the difference between perceptual processing in different sense modali-
ties. It is important to emphasize that there is no tension between these two claims—in spite of all the
intricate links between the perceptual processing in different sense modalities, we can nonetheless
identify what distinctively visual perceptual processing amounts to. Multimodality does not imply
that there are no distinct sense modalities.
sensory stimulation of the relevant sense modality). In the vast majority of

perceptual scenarios, sensory- stimulus-driven perceptual processing is
combined with mental imagery. And in these cases, much of what we take
ourselves to perceive we really (at least partly) represent by means of mental
imagery.
3.6. Back to Perceptual Justification

(and Top-Down Influences)
I argued in section 3.2 that top-down influences on perception should not be

considered to be a very strong reason to re-evaluate how we think about per-
ceptual justification. But the empirical results I introduced in sections 3.5–7
should. The primary reason why mental imagery is important in the discus-
sion of perceptual justification is not that it can be influenced in a top-down
manner. In this section, I argue that mental imagery can be influenced in a
top-down manner—but, again, this is not the primary reason why it should
pose a problem for any account of perceptual justification.
Some cases of mental imagery are fully bottom-up, but some others are
influenced or even initiated in a top-down manner. There are many forms
of mental imagery that do not depend on any top-down information. The
kind of mental imagery our perceptual system uses to fill in the blind spot
is one clear example. The perceptual processing of information that would
correspond to the blind spot is not triggered by corresponding sensory stim-
ulation in the given sense modality because there is no corresponding sen-
sory stimulation in the given sense modality: the blind spot has no receptors.
But this perceptual processing is determined fully by bottom-up informa-
tion about the sensory stimulation of the parts of the retina that surround
the blind spot. We have no evidence that there are any top-down influences
on this form of mental imagery. Another fully bottom-up example of multi-
modal mental imagery is the double-flash illusion.
But some other cases of multimodal mental imagery are influenced in a
top-down manner. One widely used and researched example of what I call
multimodal mental imagery is seeing someone talking on television with the
sound muted. The visual perception of the talking head in the visual sense
modality leads to auditory mental imagery in the auditory sense modality
(e.g., Calvert et al. 1997; Hertrich et al. 2011; Pekkola et al., 2005; see also
Spence and Deroy 2013 for a philosophical summary).
The auditory mental imagery in this case clearly depends on bottom-up

factors like the lip movements of the person on the screen. But not only
these. If this person is someone you know or have heard speak, your auditory
mental imagery will be influenced by this information. If it is Barack Obama
(someone you have, presumably, heard before), you will “hear” him speaking
with his distinctive tone of voice or intonation, for example. This is a case
(and cases like this are not at all rare) where multimodal mental imagery is
influenced by top-down information.
If we put together these findings with the claim about the impor-
tance of mental multimodal imagery in everyday perception, we get a
straightforward argument for top-down influences on perception: mul-
timodal mental imagery is influenced by top-down information, and
perception is very much influenced by multimodal mental imagery. By
transitivity, it would follow that perception is influenced by top-down
information.
Note that while this argument may establish top-down influences on
perception, it provides no arguments for or against the dependence of per-
ception (either perceptual phenomenology or perceptual processing) on
cognitive states—which would be the sense of top-down influences on per-
ception that is relevant in the context of perceptual justification. From the
point of view of epistemology and of the question of perceptual justification,
the only debate about top-down influences on perception that matters is
about cognitive influences on perception. Whether there are some forms of
top-down influences on early perceptual processes is strictly irrelevant. The
reason why the importance of multimodal mental imagery in everyday per-
ception is important from the point of view of epistemology lies beyond the
focus on top-down influences.
3.7. A Deeper Problem about

Perceptual Justification
Findings about mental imagery show that much of what we take ourselves to
perceive we really represent by means of mental imagery. Perceptual states
are almost always of multisensory entities, and unacquainted parts of these
multisensory entities are represented by means of mental imagery. In the
world of multimodal perception, what we take ourselves to perceive is partly
represented by multimodal mental imagery. But then we have much stronger
reasons to think that perception might not be the right kind of mental state to
base our beliefs on and to justify our beliefs by.
Mental imagery may or may not be influenced by top-down information.
And even when it is, it is not clear how far up this top-down information
comes from. But this is cold comfort for those who expect perception to
serve as some unbiased basis for perceptual justification. Even in those cases
where mental imagery is not at all influenced by top-down information, it
fails to be caused by what it represents.
Even in the somewhat trivial case of the blind spot, where, supposedly,
no top-down information is being used, the blind spot is filled in by mental
imagery—by perceptual processes not triggered by corresponding sensory
stimulation. So no matter what way the blind spot is filled in, that has no
causal connection with whatever is in front of that part of the retina. Mental
imagery, by definition, fails to be caused by what it is about.
Sensory-stimulation-driven perception is caused by what it is of. Light
from the perceived object hits our retina and that is the sensory stimulation
that gets processed. But mental imagery is by definition not sensory stimula-
tion driven. So it is cut off from the object it is about. One link from the causal
chain is missing. Light from the perceived object hits our retina, the percep-
tual processing in the visual cortices is not triggered by this retinal stimula-
tion. So the perceptual state that is supposed to do the epistemic heavy-lifting
partly depends on (or maybe even partly constituted by, see Nanay 2016,
2018a, forthcoming) mental imagery— something not particularly well
suited at all for any epistemic role.
Why not? Because at least on the face of it, it violates both the safety and
the sensitivity conditions of justification (see Helton and Nanay 2019 for a
parallel argument concerning amodal completion). The sensitivity condition
is clearly violated: the beliefs we form on the basis of multimodal mental im-
agery are not sensitive to possible changes in the multimodally completed
bits: no change in the multimodally completed part of the multisensory event
would show up in your beliefs.
The safety condition is a bit more complicated (Sosa 1999). Here is one
version of the safety condition: in those close possible worlds where one
believes that p on the basis of multimodal mental imagery, p is the case.
A lot will depend on how one construes the scope of these possible worlds
(all the closest possible worlds or most of the closest possible worlds). But
even if we go with the least restrictive characterization, there will be cases,
for example, the double-flash illusion previously described, where it will
not be the case that in most close possible worlds where I believe p on the
basis of my multimodal mental imagery, p is true. I will believe in most,
but maybe even all close possible worlds that there were two flashes. But
there is only one flash. This is an extreme case, but the point is that we can’t
exclude the possibility of multimodal mental imagery violating the safety
condition in any scenario where multimodal mental imagery plays a role
(which means, as we have seen, in the vast majority of everyday perceptual
scenarios). In short, almost all instances of perception are mixtures of a
state that is supposed to track the truth (perception) and a state that, on the
face of it, isn’t (mental imagery).
Perception is supposed to be a good way of justifying our beliefs because per-
ception tracks truth. But mental imagery is, by definition, a step removed from
the truth it is supposed to track. Of course it can track truth albeit in a fallible
manner. The mental imagery used for filling in the blind spot, for example, is
really very reliable. It can be fooled, but in the vast majority of cases it isn’t. So the
mental imagery that is used to fill in the blind spot does track truth—not 100%
reliably, but nonetheless reliably enough. And the reason we know this is that
we know the exact mechanisms of how the visual system uses the sensory stim-
ulation around the blind spot as an input when filling in the blind spot. If this
mechanism were less reliable, this mental imagery would fail to track the truth.
But then the same question needs to be asked about those forms of mental
imagery that play a more important role in everyday perception: about
whether the mechanisms that construct these forms of mental imagery are
reliable enough. Whether perception can justify beliefs depends on empir-
ical facts about the reliability of the mechanisms of mental imagery involved
in perception.7
Again, we can make the default assumption that our perceptual system
built pretty good mechanisms for mental imagery on the basis of contextual
or cross-modal information that does co-vary with the scene in front of us.
The blind spot is a good example. We can fool the filling in of the blind spot,
but this happens very rarely and only in exceptional circumstances (and only
in monocular vision, for a start).
But if what we take to be perception is really a mixture between sensory-
stimulation-driven perception and mental imagery, then we cannot take
it for granted that perceptual justification is unproblematic. We need to
7 This argument clearly does not apply to dogmatist theories of perceptual justification—I am not
discussing dogmatism about perceptual justification in this chapter.

examine the mechanisms of mental imagery to see how reliable they are and
what role they can play in perceptual justification.
3.8. Conclusion: Epistemology Naturalized
A lot more work needs to be done in order to show that we are justified to
move from (imagery-infused) perception to belief. Again, this is not to say
that we can’t eventually do so, we surely can. But any such move would need
to involve a close empirical examination of the reliability of the processes
that constitute mental imagery.
It is also important to stress that when I say that almost all of our visual
perceptual states are in fact mixed sensory-stimulation-driven /mental
imagery states, I do not mean to suggest that the contribution of sensory-
stimulation-driven processes and not-sensory-stimulation-driven processes
(that is, mental imagery) is approximately equal. In fact, it happens very
rarely that they are equal. Normally, the mental imagery component is neg-
ligible. But the very fact that it is always lurking in the background should
prevent us from taking perception at face value when it comes to perceptual
justification.
The conclusion is that the question of perceptual justification is at least in
part an empirical question—it requires the examination of the reliability of the
forms of mental imagery that play a role in perception per se. This is a sense (a
fairly narrow sense, to be sure) in which epistemology needs to be naturalized.
References
Akins, K. A., and S. Winger. 1996. Ships in the night: Churchland and Ramachandran
on Dennett’s theory of consciousness. In K. A. Akins, ed., Perception. Oxford: Oxford
Allefeld, C., et al. 2011. Flicker-light induced visual phenomena: Frequency dependence
and specificity of whole percepts and percept features. Consciousness and Cognition
20: 1344–1362.
Awater, H., J. R. Kerlin, K. K. Evans, and F. Tong. 2005. Cortical representation of space
around the blind spot. Journal of Neurophysiology 94: 3314–3324.
Bakin, J., K. Nakayama, and C. Gilbert. 2000. Visual responses in monkey areas V1 and V2
to three-dimensional surface configurations. Journal of Neuroscience 20: 8188–8198.
Ban, H., H. Yamamoto, T. Hanakawa, S. Urayama, T. Aso, H. Fukuyama, and Y. Ejima.
2013. Topographic representation of an occluded object and the effects of spatiotem-
poral context in human early visual areas. Journal of Neuroscience 33: 16992–17007.
Bertelson, P. 1999. Ventriloquism: A case of cross-modal perceptual grouping. In G.

Aschersleben, T. Bachmann, and J. Müsseler, eds., Cognitive Contributions to the
Perception of Spatial and Temporal Events. Amsterdam: Elsevier. 347–362.
Bertelson, P., and B. de Gelder. 2004. The psychology of multimodal perception. In C.
Spence and J. Driver, eds., Crossmodal Space and Crossmodal Attention. Oxford: Oxford
Bolano, R. 2009. 2666. Trans. N. Wimmer. London: Picador.
Bullier, J. 2004. Communications between cortical areas of the visual system. In L. M.
Chalupa and J. S. Werner, eds., The Visual Neurosciences. Cambridge, MA: MIT Press.
522–540.
Bushnell, B. N., P. J. Harding, Y. Kosai, and A. Pasupathy. 2011. Partial occlusion
modulates contour-based shape encoding in primate area V4. Journal of Neuroscience
31: 4012–4024.
Calvert, G. A., E. T. Bullmore, M. J. Brammer, R. Campbell, S. C. R. Williams, P. K.
McGuire, P. W. R. Woodruff, S. D. Iversen, and A. S. David. 1997. Activation of auditory
cortex during silent lipreading. Science 276: 593–596.
Chen, J., B. Liu, B. Chen, and F. Fang. 2009. Time course of amodal completion in face per-
ception. Vision Research 49: 752–758.
Churchland, P. S., and V. S. Ramachandran. 1993. Filling in: Why Dennett is wrong. In B.
Dahlbom, ed., Dennett and His Critics. Oxford: Blackwell. 28–52.
De Weerd, P., R. Desimone, and L. G. Ungerleider. 1998. Perceptual filling-in: A para-
metric study. Vision Research 38: 2721–2734.
De Weerd, P., R. Gattass, R. Desimone, and L. G. Ungerleider. 1995. Responses of cells
in monkey visual cortex during perceptual filling-in of an artificial scotoma. Nature
377: 731–734.
De Weerd, P., E. Smith, and P. Greenberg. 2006. Effects of Selective attention on percep-
tual filling-in. Journal of Cognitive Neuroscience 18: 335–347.
Dennett, D. C. 1991. Consciousness Explained. Boston: Little, Brown.
Fiorani, J. M., M. G. Rosa, R. Gattass, C. E. Rocha-Miranda. 1992. Dynamic surrounds of
receptive fields in primate striate cortex: A physiological basis for perceptual comple-
tion? Proceedings of the National Academy of Sciences USA 89: 8547–8551.
Firestone, C., and B. J. Scholl. 2014. “Top-down” effects where none should be found: The
El Greco fallacy in perception research. Psychological Science 25(1): 38–46.
Firestone, C., and B. J. Scholl. 2016. There are no top-down influences on perception.
Behavioral and Brain Sciences 39: 1–72.
Gandhi, S. P., et al. 1999. Spatial attention affects brain activity in human primary visual
cortex. Proceedings of the National Academy of Sciences 96: 3314–3319.
Ghazanfar, A. A., and C. E. Schroeder. 2006. Is neocortex essentially multisensory? Trends
in Cognitive Sciences 10: 278–285.
Ghijsen, H. 2016. The Real Epistemic Problem of Cognitive Penetration. Philosophical
Studies 173: 1457–1475.
Gregory, R. L., ed. 1987. The Oxford Companion to the Mind. Oxford: Oxford
University Press.
Grill-Spector, K., and R. Malach. 2004. The human visual cortex. Annual Review of
Neuroscience 27: 649–677.
Grossberg, S., and E. Mignolla. 1985. Dynamics of form perception: Boundary comple-
tion, illusory figures, and neon color spreading. Psychological Review 92: 173–211.
Hansen, T., L. Pracejus, and K. R. Gegenfurtner. 2009. Color perception in the interme-
diate periphery of the visual field. Journal of Vision 9(4): 26, 1–12.
Hardin, C. L. 1988. Color for Philosophers. Indianapolis: Hackett.
Hegdé, J., F. Fang, S. O. Murray, and D. Kersten. 2008. Preferential responses to occluded
objects in the human visual cortex. Journal of Vision 8: 16–35.
Helton, G., and B. Nanay. 2019. Amodal completion and knowledge. Analysis.
Hertrich, I., S. Dietrich, and H. Ackermann. 2011. Cross-modal interactions during
perception of audiovisual speech and nonspeech signals: An fMRI study. Journal of
Cognitive Neuroscience 23: 221–237.
Huemer, M. 2001. Scepticism and the Veil of Perception. Lanham, MD: Rowman and
Littlefield.
Iurilli, G., D. Ghezzi, U. Olcese, G. Lassi, C. Nazzaro, R. Tonini, V. Tucci, F. Benfenati, and
P. Medini. 2012. Sound-driven synaptic inhibition in primary visual cortex. Neuron
73: 814–828.
James, T. W., G. K. Humphrey, J. S. Gati, P. Servos, R. S. Menon, and M. A. Goodale.
2002. Haptic study of three-dimensional objects activates extrastriate visual areas.
Neuropsychologia 40: 1706–1714.
Katzner, S., and S. Weigelt. 2013. Visual cortical networks: Of mice and men. Current
Opinion in Neurobiology 23: 202–206.
Kentridge, R. W., C. A. Heywood, and L. Weiskrantz. 1999. Attention without awareness
in blindsight. Proceedings of the Royal Society of London B 266: 1805–1811.
Komatsu, H. 2006. The neural mechanisms of perceptual filling-in. Nature Review
Neuroscience 7: 220–231.
Komatsu, H., M. Kinoshita, and I. Murakami. 2000. Neural responses in the retinotopic
representation of the blind spot in the macaque V1 to stimuli for perceptual filling-in.
Journal of Neuroscience 20: 9310–9319.
Kosslyn, S. M., M. Behrmann, and M. Jeannerod. 1995a. The cognitive neuroscience of
mental imagery. Neuropsychologia 33: 1335–1344.
Kovacs, G., R. Vogels, and G. A. Orban. 1995. Selectivity of macaque inferior temporal
neurons for partially occluded shapes. Journal of Neuroscience 15: 1984–1997.
Lamme, V. A., and P. R. Roelfsema. 2000. The distinct modes of vision offered by feedfor-
ward and recurrent processing. Trends in Neuroscience 23: 571–579.
Lee, T. S., and M. Nguyen. 2001. Dynamics of subjective contour formation in the early
visual cortex. Proceedings of the National Academy of Sciences 98: 1907–1911.
Lerner, Y., M. Harel, and R. Malach. 2004. Rapid completion effects in human high-order
visual areas. NeuroImage 21: 516–526.
Li, W., V. Piëch, and C. D. Gilbert. 2004. Perceptual learning and top-down influences in
primary visual cortex. Nature Neuroscience 7: 651–657.
Lommertzen, J., R. Van Lier, and R. Meulenbroek. 2009. Grasping partly occluded
objects: Effects of global stimulus information on Action. Perception 38: 200–214.
Lupyan, G. 2015. Cognitive penetrability of perception in the age of prediction: Predictive
systems are penetrable systems. Review of Philosophy and Psychology 6: 547–569.
Lupyan, G., S. L. Thompson-Schill, and D. Swingley. 2010. Conceptual penetration of
visual processing. Psychological Science 21(5): 682–691.
Lyons, J. C. 2011. Circularity, reliability, and the cognitive penetrability of perception.
Lyons, J. C. 2015. Unencapsulated modules and perceptual judgment. In J. Zeimbekis

and A. Raftopoulos, eds., The Cognitive Penetrability of Perception: New Philosophical
Perspectives. New York: Oxford University Press. 103–122.
Macpherson, F. 2012. Cognitive penetration of colour experience. Philosophy and
Phenomenological Research 84: 24–62.
McGrath, M. 2013. Phenomenal conservatism and cognitive penetration: The bad basis
counterexamples. In Chris Tucker, ed., Seemings and Justification. Oxford: Oxford
Martuzzi, R., M. M. Murray, C. M. Michel, J.-P. Thiran, P. P. Maeder, S. Clarke, and R.
A. Meuli. 2007. Multisensory interactions within human primary cortices revealed by
BOLD dynamics. Cerebral Cortex 17: 1672–1679.
Murray, S. O., D. Kersten, B. A. Olshausen, P. Schrater, and D. L. Woods. 2002. Shape per-
ception reduces activity in human primary visual cortex. PNAS 99: 15164–15169.
Nanay, B. 2006. Does what we want influence what we see? In R. Sun, ed., Proceedings of
the 28th Annual Conference of the Cognitive Science Society (CogSci 2006). Hillsdale,
NJ: Lawrence Erlbaum. 615–621.
Nanay, B. 2010. Perception and imagination: Amodal perception as mental imagery.
Philosophical Studies 150: 239–254.
Nanay, B. 2011. Do we see apples as edible? Pacific Philosophical Quarterly 92: 305–322.
Nanay, B. 2013. Between Perception and Action. Oxford: Oxford University Press.
Nanay, B. 2016. Hallucination as mental imagery. Journal of Consciousness Studies
23(7–8): 65–81.
Nanay, B. 2018a. Multimodal mental imagery. Cortex 105: 125–134.
Nanay, B. 2018b. The importance of amodal completion in everyday perception. i-
Perception 9(4): 1–16. https://doi.org/10.1177/204166951878887.
Nanay, B. Forthcoming. Seeing Things You Don’t See. Oxford: Oxford University Press.
Nieder, A. 2002. Seeing more than meets the eye: Processing of illusory contours in ani-
mals. Journal of Comparative Physiology A 188: 260–294.
O’Callaghan, C. 2008a. Object perception: Vision and audition. Philosophy Compass
3: 803–829.
O’Callaghan, C. 2008b. Seeing what you hear: Cross-modal illusions and perception.
O’Callaghan, C. 2011. Lessons from beyond vision. Philosophical Studies 153: 143–160.
O’Callaghan, C. 2014. Intermodal binding awareness. In D. Bennett and C. Hill, eds.,
Sensory Integration and the Unity of Consciousness. Cambridge, MA: MIT Press. 73–96.
O’Connor, D. H., M. M. Fukui, M. A. Pinsk, and S. Kastner. 2002. Attention
modulates responses in the human lateral geniculate nucleus. Nature Neuroscience
5(11): 1203–1209.
Pearson, J., and F. Westbrook. 2015. Phantom perception: Voluntary and involuntary
nonretinal vision. Trends in Cognitive Sciences 19: 278–284.
Pearson, J., T. Naselaris, E. A. Holmes, and S. M. Kosslyn. 2015. Mental imagery: Functional
mechanisms and clinical applications. Trends in Cognitive Sciences 19: 590–602.
Pekkola, J., V. Ojanen, T. Autti, I. P. Jaaskelainen, R. Mottonen, A. Tarkainen, and M.
Sams. 2005. Primary auditory cortex activation by visual speech: an fMRI study at 3 T.
NeuroReport 16: 125–128.
Pessoa, L., and L. G. Ungerleider. 2005. Visual attention and emotional perception. In L.
Itti, G. Rees, and J. K. Tsotsos, eds., Neurobiology of Attention. San Diego, CA: Elsevier.
160–166.
Phillips, I. 2013. Afterimages and sensation. Philosophy and Phenomenological Research

87: 417–453.
Pryor, J. 2000. The skeptic and the dogmatist. Noûs 34(4): 517–549.
Pylyshyn, Z. 1999. Is vision continuous with cognition? The case for cognitive impenetra-
bility of visual perception. Behavioral and Brain Sciences 22: 341–365.
Ramachandran, V. S. 1992. Filling in the blind spot. Nature 356: 115.
Ramachandran, V. S., and R. Gregory. 1991. Perceptual filling-in of artificially induced
scotomas in human vision. Nature 350: 699–702.
Rauschenberger, R., T. Liu, S. D. Slotnick, and S. Yantis. 2006. Temporally unfolding
neural representation of pictorial occlusion. Psychological Science 17: 358–364.
Rauschenberger, R., and S. Yantis. 2001. Masking unveils pre-amodal completion repre-
sentation in visual search. Nature 410: 369–372.
Rolls, E. T., and M. J. Tovee. 1994. Processing speed in the cerebral cortex and the neu-
rophysiology of visual masking. Proceedings of the Royal Society: Biological Sciences
257: 9–15.
Roseboom, W., T. Kawabe, and S. Nishida. 2013. The cross-modal double flash illu-
sion depends on featural similarity between cross-modal inducers. Scientific Reports
3: 3437–4359.
Scherzer, T. R., and V. Ekroll. 2015. Partial modal completion under occlusion: What do
modal and amodal percepts represent. Journal of Vision 15: 1–20.
Schmitz, T. W., E. De Rosa, and A.K. Anderson. 2009. Opposing influences of affective
state valence on visual cortical encoding Journal of Neuroscience 29: 7199–7207.
Schupp, H. T., B. N. Cuthbert, M. M. Bradley, C. H. Hillman, A. O. Hamm, and P. J. Lang.
2004. Brain processes in emotional perception: Motivated attention. Cognition and
Emotion 18: 593–611.
Schwitzgebel, E. 2008. The unreliability of naive introspection. Philosophical Review
117: 245–273.
Sekuler, A. B., and S. E. Palmer. 1992. Perception of partly occluded objects: A
microgenetic analysis. Journal of Experimental Psychology General 121: 95–111.
Sekuler, R., A. B. Sekuler, and R. Lau. 1997. Sound alters visual motion perception. Nature
285: 308.
Shams, L., Y. Kamitani, and S. Shimojo. 2000. What you see is what you hear. Nature
408: 788.
Siegel, S. 2011. Cognitive penetrability and perceptual justification. Nous 46: 201–222.
Siegel, S. 2013. The epistemic impact of the etiology of experience. Philosophical Studies
162: 697–722.
Silins, N. 2016. Cognitive penetration and the epistemology of perception. Philosophy
Compass 11: 24–42.
Smith, F. W., and L. Muckli. 2010. Nonstimulated early visual areas carry information
about surrounding context. PNAS 107: 20099–20103.
Sosa, E. 1999. How to defeat opposition to Moore. Philosophical Perspectives 33: 141–153.
Spence, C., and J. Driver, eds. 2004. Crossmodal Space and Crossmodal Attention. Oxford:
Spener, M., and T. Bayne. 2010. Introspective humility. Philosophical Issues 20: 1–22.
Sperandio, I., P. A. Chouinard, and M. A. Goodale. 2012. Retinotopic activity in V1
reflects the perceived and not the retinal size of an afterimage. Nature Neuroscience
15: 540–542.
Spillmann, L., T. Otte, K. Hamburger, and S. Magnussen. 2006. Perceptual filling-in from
the edge of the blind spot. Vision Research 46: 4252–4257.
Stokes, D. 2012. Perceiving and desiring: A new look at the cognitive penetrability of ex-
perience. Philosophical Studies 158: 479–492.
Sugita, Y. 1999. Grouping of image fragments in primary visual cortex. Nature
401: 269–272.
Talavage, T. M., M. I. Sereno, J. R. Melcher, P. J. Ledden, B. R. Rosen, and A. M. Dale.
2004. Tonotopic organization in human auditory cortex revealed by progressions of
frequency sensitivity. Journal of Neurophysiology 91(3): 1282–1296.
Teufel, C., and B. Nanay. 2017. How to (and how not to) think about top-down influences
on perception. Consciousness and Cognition 47: 17–25.
Thorpe, S., D. Fize, and C. Marlot. 1996. Speed of processing in the human visual system.
Nature 381: 520–522.
Troncoso, X. G., S. L. Macknik, and S. Martinez-Conde. 2008. Microsaccades counteract
perceptual filling-in. Journal of Vision 8: 15–19.
Tucker, C. 2014. If dogmatists have a problem with cognitive penetration, you do too.
Dialectica 68: 35–62.
van Essen, D. C. 2004. Organization of visual areas in macaque and human cerebral
cortex. In L. Chalupa and J. S. Werner, eds., The Visual Neurosciences. Cambridge,
MA: MIT Press. 507–521.
Vance, J. 2015. Cognitive penetration and the tribunal of experience. Review of Philosophy
and Psychology 7: 1–23.
Vetter P., F. W. Smith, and L. Muckl. 2014. Decoding sound and imagery content in early
visual cortex. Current Biology 24: 1256–1262.
Vrins, S., T. C. J. De Wit, and R. Van Lier. 2009. Bricks, butter, and slices of cu-
cumber: Investigating semantic influences in amodal completion. Perception 38: 17–29.
Vroomen, J., P. Bertelson, and B. de Gelder. 2001. Auditory- visual spatial
interactions: Automatic versus intentional components. In B. de Gelder, E. de Haan,
and C. Heywood, eds., Out of Mind. Oxford: Oxford University Press. 140–150.
Vul, E., et al. 2008. The McCollough effect reflects permanent and transient adaptation in
early visual cortex. Journal of Vision 8: 4.
Watanabe, T., and T. Sato. 1989. Effects of luminance contrast on color spreading and
illusory contour in the neon color spreading effect. Perception and Psychophysics
45: 427–430.
Watkins, S., L. Shams, S. Tanaka, J. D. Haynes, and G. Rees. 2006. Sound alters activity in
human V1 in association with illusory visual perception. NeuroImage 31: 1247–1256.
Weil, R. S., J. M. Kilner, J. D. Haynes, and G. Rees. 2007. Neural correlates of perceptual
filling-in of an artificial scotoma in humans. Proceedings of the National Academy of
Sciences USA 104: 5211–5216.
Weil, R. S., S. Watkins, and G. Rees. 2008. Neural correlates of perceptual completion of an
artificial scotoma in human visual cortex measured using functional MRI. Neuroimage
42: 1519–1528.
Welchman, A. E., and J. M. Harris. 2001. Filling-in the details on perceptual fading. Vision
Research 41: 2107–2117.
Zangaladze, A., V. D. Weisser, E. Stilla, S. C. Prather, and K. Sathian. 1999. Involvement of
visual cortex in tactile discrimination of orientation. Nature 401: 587–590.
4
How Reliably Misrepresenting Olfactory
Experiences Justify True Beliefs
Angela Mendelovici
4.1. Introduction
Olfactory experiences are the experiences we have when we do what we com-

monly call “smelling.” For example, when we smell a burning piece of toast,
a wet dog, the smoke outside our window, or sour milk, we have olfactory
experiences. This chapter is about olfactory experiences, their contents, and
how they justify beliefs, such as beliefs about burning toast, wet dogs, and
sour milk.
I will argue that olfactory experiences have contents (section 4.2). They
represent everyday objects or ad hoc “bare olfactory objects” as having prim-
itive olfactory properties (subsections 4.3.1–4.3.2). Olfactory properties
happen never to be instantiated, so olfactory experiences misrepresent the
world around us. However, I will argue, since olfactory experiences misrep-
resent in the same way on distinct occasions, they reliably misrepresent. This
provides the basis for an explanation of how, despite misrepresenting, olfac-
tory experiences can not only be useful, but also justify true beliefs about pu-
tatively smelly objects (section 4.4).
Before beginning, a brief note about terminology is in order: it is useful to
have a fairly neutral and everyday way of describing olfactory experiences
without presupposing any particular view of whether they have contents or
of what they represent. I will reserve the term “smell” for this use, using it to
stand for an act, a property, or an object, depending on context. For instance,
we can say that we smell the roses, that the roses smell nice, that the roses
have a nice smell, or that we notice a nice smell. “Smell” talk is neutral on the
contents of the states described. For example, saying that Eleni smells the
roses does not imply any particular view of the content of her state or even
that she has a corresponding contentful state at all.
Angela Mendelovici, How Reliably Misrepresenting Olfactory Experiences Justify True Beliefs
In: The Epistemology of Non-Visual Perception. Edited by: Berit Brogaard and Dimitria Electra Gatzia,
Oxford University Press (2020). © Oxford University Press.
DOI: 10.1093/oso/9780190648916.003.0005
4.2. Olfactory Experiences Have Contents
Before considering the question of what particular contents olfactory

experiences represent, let us first motivate the view that they do represent.
This view has received much explicit endorsement (see especially Perkins
1983; Lycan 1996, 2000; Matthen 2005; Batty 2009, 2010a, 2010b, 2010c,
2010d; Richardson 2013; Budek and Farkas 2014; Mizrahi 2014; Roberts
2015; Cavedon-Taylor 2018), but it has at least a few detractors (Reid 1785/
2002; Peacocke 1983) and Lycan (1996) argues it is far from obvious.
There are various ways of elucidating the notion of content at play here.
My favored way is ostensive, relying on introspectively accessible para-
digm examples of contentful states, such as thoughts and certain percep-
tual experiences. For example, thoughts and perceptual experiences, such as
visual experiences, seem to “say,” or be “of ” or “about,” certain things. For
example, a thought that the cat is on the mat might “say” that the cat is on the
mat or be “about” the proposition that the cat is on the mat, and a visual ex-
perience of a red cup might be “of ” or “about” a (putative) red cup or the (pu-
tative) fact that there is a red cup. The foregoing mental states are examples
of intentional mental states, and what they “say,” are “of ” or “about” or, more
generally, represent, are their contents.1
I will now motivate the claim that olfactory experiences have contents. I will
do so by outlining two key reasons for thinking that visual and other experiences
have contents and arguing that they also apply to olfactory experiences.
That visual experiences, and many other kinds of experiences, are inten-
tional has been argued on phenomenological grounds. Many visual and
other experiences seem to present, represent, or “tell us” something. For
example, a visual experience of a blue cup seems to represent a cup and a
blueness that qualifies the cup, which suggests that it in fact represents a cup
or cup-shaped object as being blue.2 In the same way, olfactory experiences
1 Note that although our paradigm cases of intentionality are introspectively accessible mental
states, this does not mean that all cases of intentionality must be introspectively accessible or that
they must be mental states. There might be introspectively inaccessible or non-mental states with the
same feature as our paradigm cases that are not introspectively accessible. Notice also that my defi-
nition does not make reference to conditions of truth or reference. This is because that there are such
things as conditions of truth or reference is a substantive further claim about the nature of intention-
ality (see Mendelovici 2018a, chap. 1). However, for the purposes of my discussion, I will assume that
intentional states have conditions of truth or reference.
For a more precise ostensive definition of intentionality, see Mendelovici 2018a, 2010, forth-
coming, and Kriegel 2011.
2 Such arguments are often made by representationalists, who take phenomenal consciousness to
be a species of intentionality. See Harman 1990; Lycan 1996; Dretske 1995; Tye 2000; Mendelovici
2013a, 2014, 2018a; Bourget and Mendelovici 2014; and Bourget 2015, 2017a, 2017b.
Reliably Misrepresenting Olfactory Experiences 101
seem to represent objects, such as putatively smelly objects, and qualities that
seem to qualify them. For example, an olfactory experience of a basil leaf
seems to represent the basil leaf and a basil-y smell that qualifies the basil
leaf, which suggests that the olfactory experience represents the basil leaf as
having certain olfactory properties. Even olfactory experiences that do not
seem to present any everyday objects, such as an experience of a foul smell
with an unknown source, seem to tell us something about the world, even
if it is simply that there is a smell present. So even these kinds of olfactory
experiences seem to have contents.
That visual and other experiences have contents has also been argued on
epistemological grounds: McDowell (1994) roughly argues that experiences
have contents because they justify beliefs and, in order for them to do so, they
must have contents. For instance, a perceptual experience of a blue cup be-
fore you might all by itself justify the belief that there is a blue cup before you.
It is unclear how the perceptual experience can by itself justify the belief if it
were a mere sensation or other non-intentional state. But if the perceptual
experience has a content like <there is a blue cup before me>, it can justify
a belief with this or related contents in the same way that a belief can justify
another belief with a related content. Similarly, an olfactory experience of a
basil leaf by itself justifies the belief that the basil leaf has a certain smell. This
suggests that the experience represents the basil leaf as having particular ol-
factory properties.
4.3. The Contents of Olfactory Experiences
In the previous section, I motivated the idea that olfactory experiences have
contents. But what exactly do they represent? Since olfactory experiences
generally represent (putative) objects as (putatively) having various prop-
erties, the question can be divided into two more specific questions, which
I will address in the following two subsections:
1. What are the objects represented by olfactory experiences?

(Section 4.3.1)
2. What are the properties represented by olfactory experiences?
(Section 4.3.2)
Although much of the literature on the contents of olfactory experiences

focuses on the first question, I will focus on the second question, since it is
the answer to that question that compels us to say that olfactory experiences
generally misrepresent. However, it is still useful to briefly address the first
question in order to attain a more complete picture of the contents of olfac-
tory experiences.
Before we begin, it is helpful to clarify the notions of represented objects
and represented properties at play: I will take a represented property to be a
content that can qualify other contents, and a represented object to be a con-
tent that can be qualified by other contents. For example, a visual experience
of a blue ball represents a represented object, such as <the ball>, and various
represented properties, such as <blueness> and <roundness>, which qualify
<the ball>. Importantly, we can accept that visual experiences represent both
objects and properties without requiring that represented objects are, con-
tain, or otherwise involve externally existing objects, such as concrete exter-
nally existing balls or that represented properties are, contain, or otherwise
involve universals, or other items that might be identified with properties.
The distinction between represented objects and represented properties is a
superficial distinction between the roles that contents play, not a metaphys-
ical one concerning their deep natures.3
4.3.1. The Objects Represented by Olfactory Experiences
There are various views of the objects represented by olfactory experiences.

One option is to say that olfactory experiences represent everyday objects,
like cakes, roses, and burning pieces of toast. Another option is to say that ol-
factory experiences represent odors, which might be taken to be collections
of airborne molecules (Richardson 2013; Cavedon-Tayler 2018). Another
option is that an olfactory experience represents some object as having ol-
factory properties, but no particular object (Batty 2010a). Yet another option
is that olfactory objects are the “stuffs” that everyday objects are composed
of, rather than the everyday objects themselves (Mizrahi 2014). Sometimes
3 Mendelovici 2018b and 2018a, chap. 1, make a distinction between the deep natures and the su-
perficial characters of contents. Mendelovici 2018b argues that the debate over propositionalism is
best construed as a debate over the superficial characters of contents. The notions of represented
objects and represented properties operative in this chapter pertain to the superficial character of
contents, not their deep natures.
these options are combined: Lycan (1996) claims that every olfactory ex-
perience has two olfactory objects, an odor and an everyday object, while
Budek and Farkas (2014) take some olfactory experiences to represent odors
and others to represent odors as well as the everyday objects that are their
sources. My favored view is a different combined view on which some ol-
factory experiences represent everyday objects, while others represent “bare
olfactory objects,” which are ad hoc objects whose sole purpose is to be the
bearers of olfactory properties.
Any of these views of olfactory objects is compatible with the main points
of this chapter. However, in order to develop a full view of the contents of
olfactory experience, I will briefly present and overview the motivations for
my favored view of olfactory objects, which is defended in greater detail in
Mendelovici n.d.
My favored view, on which there are two different types of olfactory
experiences that differ in their represented objects, can be motivated by a
pair of everyday phenomenological observations: First, sometimes we expe-
rience objects as having smells, as when a basil leaf smells basil-y. Second,
sometimes smells seem to take on a life of their own, leaving their sources far
behind, as when the smell of a baking cake travels through the house and out
the window.
The first phenomenological observation is that we at least sometimes
experience objects as having smells. When we sniff a basil leaf, it seems
that it is the basil leaf itself that smells basil-y, and when we smell a rose, it
seems that the rose itself has a certain sweet, flowery smell. Our olfactory
experiences seem to tell us about features of the basil leaf and rose them-
selves. Taken at face value, such cases suggest that the objects represented by
olfactory experiences are at least sometimes everyday objects like basil leaves
and roses.
The case of multimodal experiences involving olfactory components also
supports the view that the represented objects of olfactory experiences are
at least sometimes everyday objects. A multimodal experience of a halved
orange might involve visual, tactile, and olfactory experiences representing
the halved orange as having various properties, such as being orange, having
a particular size, shape, and texture, and having an orangey, citrusy smell.
These properties are all represented as being properties of the same object.
Since the properties of the visual and tactile components of the experience
are represented as properties of the halved orange, the object of the olfactory
component of the experience is also the halved orange.
One objection to the view that some olfactory experiences represent

everyday objects is that we cannot always discriminate between different
scenarios in which different everyday objects are causing our olfactory
experiences. Batty (2010d, 2011) raises such an objection, arguing that we
cannot tell from our olfactory experience alone whether a kitchen is com-
pletely filled with air freshener, or whether a spot above the sink was missed.
However, that we cannot discriminate between two scenarios does not mean
that we do not represent one or the other as obtaining. In the case of vision,
stimuli reaching our retinas are massively ambiguous in that there are mul-
tiple visual scenes that can cause them (see Marr 1982). The visual system
does not simply fail to see in cases where it cannot discriminate between dif-
ferent possibilities, but instead makes various “assumptions” about what the
visible world is like and uses them to resolve the ambiguity. This shows that,
in general, not being able to discriminate between a set of possibilities does
not prevent us from representing one of them.
Our second introspective observation is that in some cases it seems that
smells can take on a life of their own, travel through space, and leave their
sources far behind. The baking cake’s smell wafts around the room, fills
the whole house, and goes out the window. The wet dog’s smell lingers in the
foyer, long after the wet dog has gone. Even though one knows what is the
source of the smell, one’s olfactory experience seems to represent the smell
as existing distinctly, independently of its source. The intuition is clearer in
cases where the source of a smell is unknown, as in an example offered by
Budek and Farkas (2014) in which you enter a room and are greeted by an
unexpected smell. In this case, you do not represent a particular everyday
object as having the smell. You just smell the smell. We can call such olfactory
experiences lone smell experiences.
Taken at face value, lone smell experiences suggest that at least some ol-
factory experiences represent something that can exist independently from
everyday objects like cakes and roses and that is the bearer of olfactory prop-
erties. Whatever these things are, they may be experienced as causally origi-
nating from particular everyday objects, as in the case of the baking cake, or
they may not, as in the case of the unexpected smell.
Some views of the objects of lone smell experiences take them to be odors
(collections of airborne molecules), spatial locations, or portions of air. The
problem with these views is that they receive little phenomenological sup-
port. Olfactory experiences do not seem to represent the objects of lone
smell experiences as having a molecular constitution, so their objects are not
odors. Lone smell experiences do not represent portions of space as being
smelly—it is not as if space itself seems to bear the smell—so their objects
are not spatial locations. And lone smell experiences can represent smells
as moving through the air without representing the air itself as moving, so
their objects are not portions of air. (These arguments are fleshed out in
Mendelovici n.d.)
My suggestion, instead, is that the objects of lone smell experiences are
bare olfactory objects, objects that are represented as having no properties
other than olfactory properties and locations. When we represent a lone
smell, we create an ad hoc representation of an object that plays the role of
being the bearer of the relevant olfactory properties. These bare olfactory
objects are the objects that we commonly call “smells,” as in when we say that
there is a smell in the foyer, or that the cake’s smell is wafting out the window.
Before continuing, it is worth considering another alternative view of lone
smell experiences, some versions of which are not compatible with my claim
in section 4.4 that olfactory experiences are systematically false: the view
that lone smell experiences do not represent any objects at all but instead
represent mere olfactory properties (and hence cannot be true or false). In
Mendelovici 2013a, 2013b, I argue for a similar view of mood experiences
on which they represent mere affective properties, such as elatedness and
sadness. However, there is a key difference between the two cases that make
moods, but not lone smell experiences, amenable to such treatment. Moods
do not seem to tell us about how the world is. In contrast, even though lone
smell experiences do not tell us about the sources of lone smells, they do tell
us something about how the world is. At the very least, they tell us that there is
a smell, whatever this amounts to. This suggests that lone smell experiences,
but not moods, have propositional contents.
One might suggest an alternative version of the proprietal content view on
which what olfactory experiences represent is not mere olfactory proprieties
but rather olfactory properties as being instantiated. Now, in perceptual ex-
perience, the way that we generally represent the instantiation of properties
is by binding them to represented objects. So, we represent contents of the
form <o is F>, rather than <F is instantiated>, the latter of which does not
involve the representation of an object. This makes sense, since it is argu-
ably less demanding to perceptually represent <o is F> than it is to perceptual
represent <F is instantiated>; the latter content, but not the former, requires
a special representation of instantiation, which requires either a concept of
instantiation or a special non-conceptual representation of instantiation. So,
we might expect that lone smell experiences, which can be had by relatively
unsophisticated thinkers such as young children, have the form <o is F and at
L> rather than <F is instantiated at L>. But then we end up with my favored
view, the view that lone smell experiences represent ad hoc objects as having
olfactory properties, rather than a view on which they represent the mere in-
stantiation of olfactory properties.
In summary, olfactory experiences represent either everyday objects or
bare olfactory objects as having olfactory properties.
4.3.2. What Properties Do Olfactory Experiences

Represent?
Let us now turn to the question of what olfactory properties are, that is, of
what are the properties represented by olfactory experience.
As Batty (2010b) notes, the options here correspond to the options in the
debate on the contents of perceptual color experiences. Here are three main
options:
Physicalism: Olfactory properties are mind-independent physical prop-

erties of olfactory objects, such as molecular properties.
Relationalism: Olfactory properties are relations between olfactory
objects and minds, such as dispositions to cause certain kinds of mental
states in subjects.
Primitivism: Olfactory properties are primitive properties.
I will argue in favor of primitivism about olfactory properties. The problem

with the alternative physicalist and relationalist views is that there is a mis-
match between the contents they ascribe to olfactory experiences and the
contents supported by pre-theoretic considerations, considerations that
are independent of any theories of intentionality or views of the partic-
ular contents of particular states. This problem is the mismatch problem.
Primitivism, in contrast, ascribes contents that are supported by pre-
theoretic considerations. The argument presented here is a special case of a
general form of argument from the mismatch problem, which I’ve developed
elsewhere in greater detail.4
4 See Mendelovici 2018a for arguments against tracking theories of mental representation based
on actual mismatch cases, Mendelovici 2013c and 2016 for arguments against tracking theories
based on the possibility of mismatch cases, Mendelovici and Bourget 2014 and forthcoming-a
for overviews of the mismatch problem, and Mendelovici 2013b for an application of the mis-
match problem to the case of moods and emotions and. See also Pautz 2006, 2013 for the structural
Let us consider physicalism. According to physicalism, olfactory proper-

ties are mind-independent physical properties that are represented as being
had by olfactory objects. One natural physicalist view is that these properties
are the molecular properties that are the categorical bases of smelly objects’
dispositions to cause olfactory experiences in us.
The problem with this view is that pre-theoretic introspective and episte-
mological considerations tell against it. What we introspectively notice when
we smell a halved orange is not a particular configuration of molecules or
some other related physical property. Such properties form no part of our
phenomenology. This is evidenced by the fact that we cannot tell through
introspection alone which particular molecular or other physical proper-
ties particular smells are, or even that smells are molecular or other kinds of
physical properties. Instead, what we introspectively notice when we intro-
spect on our olfactory experience of the halved orange is a sweet, tangy, and
orangey smell. This intangible and difficult-to-describe smell is nowhere to
be found in the molecular properties or other physical properties that the
physicalist might identify with olfactory properties. In short, the contents
ascribed by physicalism do not match the contents supported by introspec-
tive considerations.
Our epistemological situation provides further considerations against
physicalism: from olfactory experiences alone, we are not justified in infer-
ring that olfactory objects have or involve certain molecular properties or
other kinds of physical properties. Such inferences would involve a leap, going
beyond the evidence provided by our olfactory experiences. But if olfactory
experiences did represent the relevant physical properties, then we should
be justified in making such inferences on the basis of olfactory experience
alone. In contrast, it seems we are justified in believing that olfactory objects
represent difficult-to-describe smells like the sweet, tangy, and orangey smell
previously mentioned. In sum, the contents ascribed by physicalism do not
match the contents supported by epistemological considerations.
There are two parts to the mismatch problem: First, the physicalist view
attributes contents to olfactory experiences that they do not have, namely
contents having to do with molecular properties or other physical proper-
ties. This is an error of commission, since the problem is that physicalism
mismatch problem for tracking representationalism, which is the problem of there being a mismatch
in the relations of similarity and difference between what a set of representations represent and the
relations of similarity and difference between the items that they track. See Mendelovici and Bourget
2019 and forthcoming-b for discussion of the application of both kinds of mismatch problem to
Karen Neander’s (2017) theory of mental representation.
includes material in the content of olfactory experiences that it is inappro-

priate to include. Second, the physicalist view fails to attribute contents to
olfactory experiences that they do in fact have, such as the sweet, tangy, and
orangey smell involved in the olfactory experience of an orange. This is an
error of omission, since the problem is that the physicalist’s description of
the content of olfactory experiences inappropriately leaves something out.
Both problems show that the physicalist view delivers the wrong answers: the
properties physicalism takes olfactory experiences to represent are not iden-
tical to the properties they in fact represent, which means that the physicalist
view is false.
Similar arguments can be used to argue against the relationalist view of the
properties represented by olfactory experiences. According to relationalism,
olfactory properties are relational properties, and minds, subjects,
organisms, or states of minds, subjects, or organisms are one of the relata.
The problem is that introspective and epistemological considerations sug-
gest against taking the properties represented by olfactory experiences to be
such relational properties. When we introspect upon olfactory experiences,
we seem to experience everyday objects or bare olfactory objects (“smells”)
themselves as having olfactory properties independently of us. We do not
experience olfactory properties as somehow attaching to both us or some-
thing to do with us and external olfactory objects, which might be everyday
objects or bare olfactory objects. Indeed, on the relational view, olfactory
experiences represent at least two objects: us (or something to do with us)
and external olfactory objects. But there is little introspective reason to think
that we or something to do with us are objects of olfactory experience.
Epistemological considerations also suggest against the view: from ol-
factory experience alone, we are not justified in drawing conclusions about
our relations to external olfactory objects. Of course, from your olfactory
experiences, you might infer that you have olfactory experiences, and
from this and the belief that external olfactory objects cause your olfac-
tory experiences, you can infer that your olfactory experiences are caus-
ally related to external olfactory objects. So perhaps olfactory experiences
do play a role in justifying beliefs about our relations to olfactory objects.
However, your belief about your relation to external olfactory objects is
not justified by your olfactory experience alone in the way that would be
predicted by the relational view. It is justified by a combination of (per-
haps your knowledge of) the fact that you have an olfactory experience, re-
gardless of its content, together with an auxiliary belief about the causes of
olfactory experiences. But if the relational view were true, such conclusions
should (also) be justified by the contents of olfactory experiences. Olfactory
experiences should “say” that you are related to external olfactory objects
in such-and-such ways, and this should justify your conclusion that you are
thus related to them.
The preceding discussion shows that the relational view clearly makes an
error of commission: it includes in the content of olfactory experiences ma-
terial that should not be included. It is less clear, however, that it makes an
error of omission. Whether it does depends on whether, say, the sweet, tangy,
and orangey smell that we experience when we smell an orange can be found
in the relational contents attributed by the relationalist. This will depend on
the particular relationalist view. On a relationalist view on which the rela-
tional contents are, say, dispositions to cause behavioral reactions in us, it is
doubtful that anything in the picture can capture the sweet, tangy, and or-
angey smell we experience, and the view will make an error of omission. But
a relationalist view taking the relevant contents to be dispositions to cause
phenomenal experiences in us might plausibly capture the sweet, tangy, and
orangey smell in the relevant phenomenal experiences, so it avoids an error
of omission. Such a view still faces the problem of commission, however,
since the extra relational material it includes in its account of olfactory prop-
erties does not match the contents we have introspective and epistemological
reasons to ascribe. It is also not clear what would motivate such a view over
the primitivist alternative described in what follows other than the desire to
avoid the error theory that (as we will see) primitivism leads to. As we will see
in the next section, it is not clear that there is good reason to reject such an
error theory.
Unlike physicalism and relationalism, primitivism does not face the mis-
match problem. Primitivism takes olfactory experience at face value, taking
the properties represented by olfactory experiences to be just what they
appear to be and not something else: properties like the sweet, tangy, and
orangey smell we experience when we sniff the halved orange. These prim-
itive olfactory properties are sui generis (not reducible to non-olfactory
properties), categorical (not dispositional), non-relational, and non-mental
properties. These are the contents revealed by introspection of olfactory
experiences. These are also the contents that justify inferences based on ol-
factory experiences. From an experience of a halved orange as having a sweet,
tangy, and orangey smell, we justifiably infer that the halved orange has a
sweet, tangy, and orangey smell. The property the inferred belief ascribes
to the halved orange is a sui generis, categorical, non-relational, and non-

mental property.5
One key objection to these arguments is that olfactory experiences rep-
resent their contents in special ways that make these contents introspec-
tively inaccessible and block our using them to draw content-appropriate
inferences.6 For instance, perhaps olfactory experiences represent their
contents under special guises or special modes of presentation. When we
introspect upon them, what we notice are these modes of presentation, not
their contents. And when we draw an inference from them, what we base
our inference on is the modes of presentation, not the contents. For example,
one might say that an olfactory experience of a halved orange represents the
halved orange as having a physical property but that it does so under a special
“sweet, tangy, and orangey” mode of presentation, which is what we intro-
spect upon and reason upon.
One problem with this proposal is that it risks making it unmotivated to
attribute contents to olfactory experiences. Recall that the key reasons to
take olfactory experiences to have contents were introspective and episte-
mological reasons (see section 4.2). We introspectively notice that olfactory
experiences “say” something, and olfactory experiences play certain roles in
justifying inferences. If modes of presentation do all the work in accounting
for what is introspectively accessible and in licensing inferences, then there
is little motivation for taking olfactory experiences to also have contents.
Indeed, it seems that on this picture, modes of presentation play the role of
contents (see our definition of “intentionality” in section 4.2), so they would
simply be contents (and, for the reasons previously discussed, they would
have to be primitivist contents), and there would be little reason to ascribe
any additional contents that physicalism or relationalism are true of to olfac-
tory experiences.7
5 Perkins (1983) defends a related projectivist view of olfactory experiences, on which they repre-
sent properties of experiences that we mistakenly project onto the world. While such a view avoids the
mismatch problem, it is unnecessarily stronger than the alternative primitivist view I am defending
in that it requires that olfactory properties be instantiated in the mind, whereas the primitivist view
I am defending allows that they are represented whether or not they are instantiated anywhere.
6 I consider these and other objections to the mismatch problem in greater detail in Mendelovici
2018a, chap. 3.
7 This does not mean, of course, that there would not be reasons to ascribe distinct referents to
olfactory experiences, where referents are the external items that contents pick out. However, it is
not clear that the referents would be physical or relational properties either, unless we allow an inten-
tional state’s referents to be determined wholly independently of its contents.
We are considering the objection that olfactory experiences represent

their contents in special ways that make them introspectively inaccessible
and that block our using them to draw content-appropriate inferences.
Another version of this objection takes olfactory experiences to repre-
sent olfactory properties nonconceptually, that is, in a way that does not re-
quire having any concepts. Perhaps nonconceptually represented contents
cannot be discerned by introspection and cannot license inferences to beliefs
with conceptually represented contents, even when the conceptually and
nonconceptually represented contents are the same.
One problem with this suggestion is that it is not generally true that non-
conceptual contents cannot be discerned by introspection or that they cannot
justify inferences to conceptually represented contents. Visual experiences of
shapes represent particular shapes nonconceptually, since we needn’t have
shape concepts in order to have them. But visual shape experiences are acces-
sible to introspection. We can tell from introspection that a particular visual
shape experience represents a circular shape rather than a square shape. And
visual shape experiences can justify beliefs about shapes, which presumably
involve conceptual representations of shapes. For example, from a visual
shape experience of a square before you, you might justifiably infer that there
is a square before you. In order for the nonconceptual content reply to help
the physicalist or relationalist about olfactory properties, there would need
to be a special reason to think that nonconceptual contents prevent olfac-
tory properties from being discernable to introspection and from licensing
inferences to related beliefs.
There’s another problem with the suggestion relying on nonconcep-
tual representation: It is arguably a datum that olfactory experiences have
contents that are introspectable and that play a role in justifying conceptual
states, such as beliefs (see section 4.2). Even if we accept some “hidden” non-
conceptual physicalist or relationalist contents, we should also accept primi-
tivist contents. But then it is unclear why we should also accept the “hidden”
physicalist or relationalist contents at all.
4.4. False Olfactory Experiences, True Beliefs
I have argued that olfactory experiences represent everyday objects or bare

olfactory objects as having primitive olfactory properties. In this section,
I will argue that olfactory experiences reliably misrepresent: they get things
wrong in the same way all the time.8 This might be thought to prevent them
from justifying true beliefs about the properties of apparently smelly objects,
but I will argue that it does not.
Presumably, the primitive olfactory properties represented by olfactory
experiences are never instantiated in the actual world. The halved orange
doesn’t really have a sweet, tangy, orangey smell. Instead, it has certain phys-
ical properties and certain dispositions to affect us in certain ways. So our ol-
factory experiences generally misrepresent. The result is an error theory, one
on which olfactory experiences generally misrepresent the world around us.
One might worry that an error theory is incompatible with the role that
olfactory experiences play in justifying beliefs about the sources of smells,
which often turn out to be true. For example, suppose I take a whiff of a
halved orange and experience a sweet, tangy, and orangey smell. From this
olfactory experience, I might truly and justifiably conclude that the orange
is ripe. Similarly, I might smell a smoky, cigarette-y smell. From this experi-
ence, I might truly and justifiably conclude that someone is smoking nearby.
I want to suggest that an error theory about olfactory experience is perfectly
compatible with such cases in which we use our olfactory experiences to
make true and justified inferences.
The key point is that although olfactory experiences misrepresent, they
reliably misrepresent, in that they misrepresent in the same way in similar
circumstances. When presented with a halved orange with particular molec-
ular properties on multiple occasions, we misrepresent it as having the same
sweet, tangy, orangey smell each time. The fact that olfactory experiences re-
liably misrepresent explains why they are so useful despite being false.
As I’ve argued elsewhere (Mendelovici 2013b), mental states that reli-
ably misrepresent can nonetheless help us perform various cognitive tasks.
This is also true specifically of olfactory experiences. For instance, olfactory
experiences that reliably misrepresent can help us discriminate between dif-
ferent kinds of items. We can tell the difference between fresh and sour milk
on the basis of our olfactory experiences alone, even though we misrepresent
both fresh and sour milk samples as having olfactory properties they do not
have. Olfactory experiences that reliably misrepresent can also help us iden-
tify particular objects or kinds of objects. For example, on the basis of our
olfactory experiences alone, we might be able to tell that there’s butter chicken
cooking in the kitchen, even though we misrepresent the butter chicken as
8 See Mendelovici 2013b, 2016, and 2018a for related discussion of reliable misrepresentation.
having olfactory properties it does not have. Similarly, a father might correctly
identify his newborn baby on the basis of smell alone, even though he reliably
misrepresents her as having olfactory properties she does not in fact have.
Mere reliability, even in the absence of veridicality, is useful for many tasks.9
The preceding examples also show that olfactory experiences that reliably
misrepresent can lead us to true conclusions. Your false olfactory experi-
ence of a glass of milk might nonetheless cause true beliefs about whether
the milk is fresh or sour. Your false experience of that characteristic butter
chicken smell allows you to correctly infer that butter chicken is cooking in
the kitchen.
This does not yet show that inferences involving mental states that reliably
misrepresent are ever justified. Even if your olfactory experience of an or-
ange as sweet, tangy, and orangey causes you to truly believe that the orange
is ripe, this is not enough to conclude that the inference is justified. After all,
not every true inference we make or are disposed to make is justified. And in
the case of olfactory experiences, there is special reason to think that the rele-
vant inferences might not be justified: the inference’s premises do not seem to
support its conclusion. It seems we are reasoning as follows:
P1. The orange is sweet, tangy, and orangey. (Content of olfactory

experience, false)
C. The orange is ripe. (Content of belief, true)
The problem is that an orange’s having a particular primitive olfactory pro-

perty does not entail or make it more likely to be true that the orange is
ripe. Now, I want to acknowledge that in some cases, we might be making
inferences like these, and that in such cases, our inferences are not justified,
even though our conclusions might be true. But I also want to suggest that
in other cases, there is an implicit bridge premise in our reasoning that, to-
gether with the content of the olfactory experience, justifies our conclusion.
This implicit bridge premise links the olfactory properties that our olfactory
experience ascribes to its object (like that of having a sweet, tangy, and or-
angey smell) to the non-olfactory properties our conclusion ascribes to the
9 This provides a response to arguments against error theories on the basis of the general well-
functioning of our olfactory systems. Batty writes: “Widespread error on the part of the olfactory
system does not accord with its being a functioning olfactory system and this is a consequence we
ought to avoid both for intra-and inter-species considerations” (Batty 2010b, 1153). However, as
we’ve seen, when accompanied by reliability, widespread error does accord with the general well-
functioning of the olfactory system.
object (like that of being ripe). In the case of our inference based on the olfac-
tory experience of the halved orange, our reasoning might go like this:
P1. The orange is sweet, tangy, and orangey. (Content of olfactory

experience, false)
P2. Oranges that are sweet, tangy, and orangey are likely to be ripe.
(Content of implicit bridge premise)
C. Therefore, the orange is ripe. (Content of belief, true)
P2 is our implicit bridge premise. From P1 and P2, we can justifiably infer
C, since P1 and P2 together make C more likely to be true. If we do in fact at
least implicitly accept such bridge premises, false olfactory experiences can
indeed help justify true related beliefs.10
It is not implausible that we do accept such premises in many cases.
Someone who infers C from P1 is likely to at least implicitly believe P2. She
likely takes the sweet, tangy, and orangey smell to be a sign of an orange’s
being ripe, and this is likely to be because she takes it to be a general truth
that oranges that are sweet, tangy, and orangey are ripe. It is not hard to see
how we might come to have beliefs like P2 and how these beliefs might come
to be justified: since olfactory experiences reliably misrepresent, we are likely
to be confronted with strong inductive evidence that sweet-, tangy-, and
orangey-smelling oranges are ripe and little disconfirming evidence. Every
sweet, tangy, and orangey orange we seem to encounter is ripe, and no unripe
oranges we encounter are sweet, tangy, and orangey. And, so, we might justi-
fiably come to believe that oranges that are sweet, tangy, and orangey are ripe.
If we understand P2 as stating that all or most oranges that are sweet, tangy,
and orangey are likely to be ripe, it is trivially true, since there are no oranges
that have the relevant olfactory properties. But if we understand P2 as stating
a counterfactual-supporting law, such that its truth would require that any
actual or merely possible oranges that are sweet, tangy, and orangey are likely
to be ripe, then it is most likely false. There is no reason to think that possible
oranges with the relevant primitive olfactory properties would be more likely
to be ripe than oranges without these properties. For our purposes, it doesn’t
10 In Mendelovici 2018a, chap. 7, I argue that thoughts generally originally represent different
contents than experiences, though they can derivatively represent many of the same contents thanks
to their relations to experiences. If this is right, then the contents <sweet, tangy, and orangey> and
<smoky and cigarette-y> are originally represented in P1, but derivatively represented in P2. See also
Mendelovici 2019 for a condensed discussion of the view.
matter which implicit belief we are most likely to have in such cases, since ei-
ther way of understanding P2 would allow it to help P1 justify C.
Olfactory experiences representing bare olfactory objects can similarly
justify related beliefs. Suppose you notice a smoky, cigarette-y smell, but do
not experience it as attaching to any everyday object. We might reason to the
conclusion that someone is smoking nearby as follows:
S1. There is a smoky and cigarette-y smell located around here. (Content
of olfactory experience, false)
S2. When there is a smoky and cigarette-y smell at a location, it is likely
that there is someone smoking near that location. (Content of im-
plicit bridge premise)
SC. Therefore, someone is smoking near here. (Content of belief, true)
It is likely that someone who infers SC from S1 at least implicitly believes

S2. Again, S2 is either trivially true or false, since, strictly speaking, there
are no smoky, cigarette-y smells. But S2 nonetheless helps S1 justify the true
belief, SC.
I have argued that even though olfactory experiences misrepresent, they
can lead us to true and justified beliefs about the sources of smells. Since they
reliably misrepresent, they can lead us to true beliefs about the sources of
smells, and since we sometimes at least implicitly accept bridge premises like
P2 and S2, many of these beliefs are justified.
4.5. Conclusion
I have argued that olfactory experiences represent everyday objects or bare

olfactory objects as having primitive olfactory properties, which they do not
in fact have. Olfactory experiences misrepresent, but since they tend to mis-
represent in the same way on multiple occasions, they reliably misrepresent,
which explains how they can be useful and lead to true and justified beliefs.
It is sometimes thought that since perception is useful in helping us get by
in the world and acquire true and justified beliefs, it is generally veridical. This
kind of reasoning can lead us to reject error theories of various kinds of per-
ceptual experiences, which might otherwise be well motivated. But we’ve seen
that experiences that don’t just misrepresent, but instead reliably misrepresent,
can be useful and help us acquire related true and justified beliefs. If what is
true of olfactory experience is true of many other types of experiences, as I’ve

argued elsewhere (Mendelovici 2018a, chap. 3, 2013c, 2013a), this suggests a
picture of perception on which perception provides us a model of the world
that is largely non-veridical but that corresponds in a systematic way to how
things actually are and that can be used to successfully guide our behavior
and draw true and justified inferences about the world. For most practical
purposes, being reliably wrong is just as good as being right.
References
Batty, C. 2009. What’s that smell? Southern Journal of Philosophy 47(4): 321–348.
Batty, C. 2010a. Olfactory experience I: The content of olfactory experience. Philosophy
Compass 5(12): 1137–1146.
Batty, C. 2010b. Olfactory experience II: Objects and properties. Philosophy Compass
5(12): 1147–1156.
Batty, C. 2010c. A representational account of olfactory experience. Canadian Journal of
Philosophy 40(4): 511–538.
Batty, C. 2010d. What the nose doesn’t know: Non-veridicality and olfactory experience.
Journal of Consciousness Studies 17(3–4): 10–17.
Batty, C. 2011. Smelling lessons. Philosophical Studies 153: 161–174.
Bourget, D. 2015. Representationalism, perceptual distortion and the limits of phenom-
enal concepts. Canadian Journal of Philosophy 45(1): 16–36.
Bourget, D. 2017a. Representationalism and sensory modalities: An argument for inter-
modal representationalism. American Philosophical Quarterly 54(3): 251–268.
Bourget, D. 2017b. Why are some phenomenal experiences “vivid” and others “faint”?
Representationalism, imagery, and cognitive phenomenology. Australasian Journal of
Philosophy 95(4): 673–687.
Bourget, D., and A. Mendelovici. 2014. Tracking representationalism. In A. Bailey, ed.,
Philosophy of Mind: The Key Thinkers. New York: Bloomsbury Academic. 209–235.
Budek, T., and K. Farkas. 2014. Which causes of an experience are also objects of the
experience? In B. Brogaard, ed., Does Perception Have Content? New York: Oxford
Cavedon-Taylor, D. 2018. Odors, objects and olfaction. American Philosophical Quarterly
55(1): 81–94.
Dretske, F. 1995. Naturalizing the Mind. Cambridge, MA: MIT Press.
Harman, G. 1990. The intrinsic quality of experience. Philosophical Perspectives 4: 31–52.
Kriegel, U. 2011. The Sources of Intentionality. New York: Oxford University Press.
Lycan, W. 1996. Consciousness and Experience. Cambridge, MA: MIT Press.
Lycan, W. 2000. The slighting of smell. In N. Bhushan and S. Rosenfeld, ed., Of Minds and
Molecules: New Philosophical Perspectives on Chemistry. New York: Oxford University
Press. 273–289.
Marr, D. 1982. Vision. San Francisco: Freeman.
Matthen, M. 2005. Seeing, Doing, and Knowing: A Philosophical Theory of Sense
Perception. New York: Oxford University Press.
McDowell, J. 1994. Mind and World. Cambridge, MA: Harvard University Press.
Mendelovici, A. 2010. Mental representation and closely conflated topics. PhD diss.,
Princeton University.
Mendelovici, A. 2013a. Intentionalism about moods. Thought: A Journal of Philosophy
2(1): 126–136.
Mendelovici, A. 2013b. Pure intentionalism about moods and emotions. In U. Kriegel,
ed., Current Controversies in Philosophy of Mind. New York: Routledge. 135–157.
Mendelovici, A. 2013c. Reliable misrepresentation and tracking theories of mental repre-
sentation. Philosophical Studies 165(2): 421–443.
Mendelovici, A. 2016. Why tracking theories should allow for clean cases of reliable mis-
representation. Disputatio 8(42): 57–92.
Mendelovici, A. 2018a. The Phenomenal Basis of Intentionality. New York: Oxford
University Press.
Mendelovici,A.2018b.Propositionalismwithoutpropositions,objectualismwithoutobjects.
In A. Grzankowski and M. Montague, eds., Non-propositional Intentionality. New York:
Mendelovici, A. 2019. Immediate and reflective senses. In D. Shottenkirk, M. Curado, and
S. Gouveia, eds., Perception, Cognition, and Aesthetics. New York: Routledge. 187–209.
Mendelovici, A. n.d. The objects of olfactory experience.
Mendelovici, A., and D. Bourget. 2014. Naturalizing intentionality: Tracking theories
versus phenomenal intentionality theories. Philosophy Compass.
Mendelovici, A., and D. Bourget. 2019. Review of Karen Neander’s A Mark of the Mental.
Philosophical Review 128(3): 378–385.
Mendelovici, A., and D. Bourget. Forthcoming-a. Consciousness and intentionality. In U.
Kriegel, ed., Oxford Handbook of the Philosophy of Consciousness. New York: Oxford
University Press.
Mendelovici, A., and D. Bourget. Forthcoming-b. Kolors without colors, representation
without intentionality. Philosophy and Phenomenological Research.
Mizrahi, V. 2014. Sniff, smell, and stuff. Philosophical Studies 171(2): 233–250.
Pautz, A. 2006. Sensory awareness is not a wide physical relation: An empirical argument
against externalist intentionalism. Noûs 40(2): 205–240.
Pautz, A. 2013. The real trouble for phenomenal externalists: New empirical evidence
for a brain-based theory of consciousness. In R. Brown, ed., Consciousness Inside and
Out: Phenomenology, Neuroscience, and the Nature of Experience. New York: Springer.
237–298.
Peacocke, C. 1983. Sense and Content: Experience, Thought and Their Relations. New York:
Perkins, M. 1983. Sensing the World. Indianapolis: Hackett.
Reid, T. 1785/2002. Essays on the Intellectual Powers of Man. University Park: Pennsylvania
State University Press.
Richardson, L. 2013. Sniffing and smelling. Philosophical Studies 162(2): 401–419.
Roberts, T. 2015. A breath of fresh air: Absence and the structure of olfactory perception.
Pacific Philosophical Quarterly 96(3). https://doi.org/10.1111/papq.12058.
Tye, M. 2000. Consciousness, Color, and Content. Cambridge, MA: MIT Press.
5
Hearing As
William G. Lycan
The expression of a change of aspect is the expression of a new per-

ception and at the same time of the perception’s being unchanged.
—Wittgenstein, Philosophical Investigations
Ambiguity— rabbit or duck?— is clearly the key to the whole

problem of image reading. . . . [S]‌uch interpretation involves a ten-
tative projection, a trial shot which transforms the image if it turns
out to be a hit.
—E. H. Gombrich, Art and Illusion
It continues to puzzle me, and I’m sure others, that so little is written by
philosophers on aspect perception, a phenomenon both ubiquitous and fas-
cinating. It doesn’t surprise me at all that every bit of what has been written
is on vision.
One of my purposes in this chapter is to review some views and lessons
about “seeing as,” and then to see which of those may carry over to aspect
perception in hearing. The latter may actually be of more general interest,
because, while visual aspect perception is fun, it is best known through
gimmicks such as ambiguous figures, while hearing-as plays at least two
more central roles in human life.
But first a little groundwork is needed. (1) What is perceived-as is per-
ceived, and more specifically an object of perception; to see X as such and
such one must see X. (2) There is a contrast between merely seeing and
seeing-as, in that two people may see X and see X qua X, but be seeing X
under different aspects. (3) Aspect perception is commonly contrasted with
representation; seeing-as is thought to outrun what is visually represented.
(4) It may be held that aspect perception is necessarily or at least normally a
William G. Lycan, Hearing As In: The Epistemology of Non-Visual Perception. Edited by: Berit
Brogaard and Dimitria Electra Gatzia, Oxford University Press (2020). © Oxford University Press.
DOI: 10.1093/oso/9780190648916.003.0006
Hearing As 119
conscious phenomenon (in whichever sense of the “c”-word). So: What are
the objects of hearing? What are the typical contrasts between merely hearing
X and hearing X as such and such, and between auditorily representing X
and hearing X under an aspect? And what is the relation between hearing-as
and consciousness?
5.1. Background Assumptions
1. I shall simply assume that perceiving is a matter of representation.1 I shall

further assume that we are talking of representations hosted and employed
by whole persons, not of the merely subpersonal representational contents
assigned to elements of our perceptual modules or other specialized brain
agencies.
And we must ask, what does hearing characteristically or proprietarily
represent? Here I have little of my own to contribute. There is already a
prominent contender: O’Callaghan’s (2007) carefully defended view that
(a) hearing represents sounds, which are worldly events rather than objects
or properties; and (b) a particular sound is an “event . . . of oscillating or
interacting bodies disturbing or setting a surrounding medium into wave
motion” (60).
Nowadays nearly every philosopher agrees that sounds are particulars,2
and that audible qualities like volume, pitch, and timbre are properties of
sounds. There are two immediate contrasts with vision and its objects. First,
though it is generally agreed that vision represents properties or property-
instances such as color and shape, it is controversial whether vision represents
individual things or events in the world. Second, as Kania (2010) observes, it
is nearly tautologous that what we hear are sounds and sounds are what we
hear; but there is no parallel tautology for seeing, except (I note) the slightly
strange “What we see are sights.”3 One might take this to suggest that sounds
1 With one small qualification: There is a puzzle about what taste (proper, unaided by smell)
represents at the whole-person level (Lycan 2018); one possible solution to the puzzle is to fall back
and say that taste represents only in the debased, primitive way set aside by Ramsey (2007, chap. 4).
A literature that burgeoned in the 2000s rejected the whole idea that sense modalities represent,
e.g.: Campbell 2002; Travis 2004; Fish 2009. But for critique, see Siegel 2010a and Schellenberg 2011.
2 An exception was Pasnau (1999), who took sounds to be qualities.
3 “Sights” in the ordinary sense are special things, as in “sightseeing,” “seeing the sights.” But no-
where more special than at Macbeth’s banquet: “You make me strange /Even to the disposition that
I owe, /When now I think you can behold such sights, /And keep the natural ruby of your cheeks, /
When mine is blanch’d with fear.” The dumbfounded Ross naturally asks, “What sights, my lord?”
are in some way more subjective or more closely mind-related than are the
objects of seeing, but if so, that does not show in the current array of theories,
for each of them takes sounds to be real phenomena in the environment that
affect us physically. (Which is hardly tautologous, though I take it to be fairly
obvious.)4
Some of the competing ontologies: Casati and Dokic (1994) had contended
that sounds are vibration events, with no mention of a medium (notice the
implication that there would be sounds even in a perfect vacuum). Before
that, Perkins (1983) had identified sounds with “trains of airwaves stretching
from the listener’s ear back to the sound-making object throughout the dura-
tion of the sound” (172)5; Roy Sorensen (2008) also defends a wave theory, as
do O’Shaughnessy (2009) and Nudds (2009). Roger Scruton (1997) puts for-
ward a less forthrightly physicalist view: that sounds are “pure” (subjectless)
events and, even though they are real and external to perceivers, they are not
reducible; their intrinsic properties are limited to ways in which they appear
or would appear. (Scruton thinks this is important for the understanding of
music, because, he says, it is important to hear music as independent of its
source.)
These theorists duke it out with each other on grounds that include the
perceived durations and locations of sounds, and intuitions regarding iden-
tity and individuation. I shall not take sides, except to assume physicalism
about sounds. But I shall insist, following Sorensen (2009) and Farennikova
(2013), that absences can be perceived as such,6 for in listening to music it is
sometimes important to hear the silence of a dramatic rest (think of that last
rest in the “Hallelujah Chorus”), or the absence of the third from an unex-
pected open fifth.
4 I would, being temperamentally an arch-realist about everything. But I gather the traditional
Berkeleyan question about the tree in the forest persists among the folk, even though its counterpart
about the tree in the quad does not. (No one contributes to popular culture by asking, “If a tree falls
in a forest and no one is there to see it, is it really on the ground?”) So perhaps, without accepting
Berkeleyan idealism generally, some speakers are tempted to withhold the word “sound” if there were
no sentient creature available to hear it. In 1884 (April 5 issue), Scientific American declared as fact
that “sound is vibration, transmitted to our senses through the mechanism of the ear, and recognized
as sound only at our nerve centers. The falling of the tree or any other disturbance will produce vibra-
tion of the air. If there be no ears to hear, there will be no sound” (218, quoted in Wikipedia).
5 It is astonishing that neither O’Callaghan nor any other recent writer on sounds cites Perkins.
Sensing the World won the American Philosophical Association’s Johnsonian Prize in 1983. It was
perhaps the most significant contribution since Reid’s to the dethroning of vision. (“Vision” comes
only as c hapter 6, following “Hearing,” which follows “Feeling Heat and Cold.”)
6 This creates the obvious problem for perceptual psychosemantics, for perceiving an absence
cannot be a state caused by the object perceived, nor is it a case of perceiving an object that is not re-
ally there. The most straightforward approach here is to hold that absences can themselves be causes,
though that will constrain theories of causation.
Hearing As 121
2. We hear sound-events, then, and we do so in virtue of our auditory sys-

tems’ representing them. We hear some of their properties, ditto. Very likely,
as part of their total inputs and detailed computational processing, our au-
ditory systems also represent sounds or bits of sound and properties that we
as whole persons do not hear; following usual practice I will continue to call
those representations “subpersonal.”
There is an important question about the level of conceptual sophistica-
tion at or under which events and properties are perceived. A conservative
view would be that we strictly perceive only the environmental properties
that our sensory systems are built to detect. For vision, this might be just
colors, shapes, and locations, or a bit less conservatively, also depth, mo-
tion, and change as such. As noted previously, it would be slightly more
liberal (and already controversial) to contend that we visually represent
individuals, or objects such as blocks. Still more liberally, it may be argued,
as does Siegel (2010b), that we represent objects under Rosch’s “basic cat-
egories” (Rosch et al. 1976) such as “dog,” “shoe,” “chair,” “red,” and “lady,”
and even members of specific natural kinds such as pine trees. At a Kuhnian
extreme (Hanson 1958), which preceded and in part inspired The Structure
of Scientific Revolutions (1962),7 but especially Churchland (1988), we di-
rectly see things described in terms of any theory you like: a woman suf-
fering from Bright’s disease, the advancement of global warming, an
economic recession, or a relation in Riemannian space.8 For audition, a
conservative-to-liberal spectrum of examples might be a sound of such-
and-such volume in a certain timbre range; a tone at this or that pitch; a
repeated banging of two objects together; splashing liquid; a waterfall; an
explosion; a bell ringing; a melody; a melody played on an oboe; a human
voice; a voice saying words; a voice beginning to recite “Tintern Abbey”;
the sound of a Rolls-Royce’s door closing; the opening instrumental bars of
Handel’s “Lift Up Your Heads”; Martin Luther King concluding his “I Have
a Dream” speech; the Chicago Symphony Chorus under Margaret Hillis;
the bowel sounds of one suffering from Baird’s syndrome; the launch of a
Zircon missile; the end of the world.
Elsewhere (Lycan 2014) I have contended that in the case of vision, the lib-
eralism issue is hard if not impossible to settle. It is not an entirely empirical
7 I wonder if you have to be as old as I am to remember Hanson’s extraordinary life and
work: https://en.wikipedia.org/wiki/Norwood_Russell_Hanson.
8 Though Churchland denies that perception is propositional representation, so he is not saddled
with the claim that the visual system “has” concepts like “economic recession.”
question, but no ordinary appeal to common sense or to introspection will

do, and psychosemantic ideas cannot help. Austen Clark (2000, 2004) puts
forward a conservative view, but does not rule out more liberal possibilities.
I know of only one actual detailed argument for a robust liberalism: Siegel’s
aforementioned, based on her “method of phenomenal contrast,” but I main-
tain (Lycan 2014) that it breaks down on each of several points.9 Most perti-
nently for our purposes here, the argument fails to distinguish actual visual
representation from merely “seeing as”; does the subject see, by visually
representing, a pine tree per se, or does she merely see a tree-shaped object
and see it as a pine? As we shall learn, this distinction is even more vexed in
the case of hearing.
3. On questions of consciousness: To be blunt, I think they have nothing
to do with the present issue. What sort of “consciousness” might be thought
to be required for aspect perception? The idea may be that in a paradigm
case, that of deliberate aspect-flipping as with the duck-rabbit, one cannot
do that unawares, and so it may be held that aspect perception is neces-
sarily conscious in that the perceiver must be aware of the relevant mental
aspect over and above what is more strictly seen. But that is not so. To stay
with the example, it’s widely observed that were you to see what is in fact
a duck-rabbit figure amid a whole scene involving ducks, you would auto-
matically and quite unconsciously see it as a duck, but were you to see the
very same figure in a rabbit-filled picture, you would just as automatically
and unthinkingly see it as a rabbit. Aspect perception is ubiquitous—or so
I shall argue, contra Wittgenstein himself—and we are rarely more aware
of doing it than we are of perceiving at all. (More generally, we are seldom
aware of our own mental state of perceiving, though of course we can be and
sometimes are. To be introspectively aware of it, we must at least to some
small degree attend to it rather than to the external objects and properties
it represents [Lycan 1996, 2008]. But for a mental state to be a “conscious”
one in this sense, as opposed to, say, Dretske’s [1993] or Block’s [1995] quite
different respective senses, is a very superficial phenomenon of little onto-
logical significance.)10
9 And see the debate between Siegel and Alex Byrne in Siegel and Byrne 2016.
10 I here ignore the view that goes under the heading of “phenomenal intentionality,” according
to which intentionality and on some versions all genuine representation is a phenomenological,
hence conscious, phenomenon (Horgan and Tienson 2002); I have rebutted that idea at length in
Lycan 2008.
Hearing As 123
4. I’ve been assuming that there is a distinction between what is strictly seen,
which I identify with being represented by vision alone, and what is represented
only by dint of “cognitive penetration” or other influence of background infor-
mation on the visual system’s output that would result in perceptual belief. But
the idea of “vision alone” depends on a modular picture of a sense modality
(Fodor 1983), or at least a clear distinction between perception and cognition,
that may or may not be supported by the most current neuroscience.
I shall continue to assume that wide-open Kuhnian-Churchlandian rel-
ativism is false, that perception is not “theory-laden” through and through,
and that there is a distinction between what is more or less strictly perceived
and what appears a certain way but can do so only because of the subject’s
knowledge or beliefs.11 The distinction may be weakened in one way or an-
other; Lycan (2014) argued that for vision, it comes in grades or stages cor-
responding to layers of processing. And recent literature on “cognitive
penetrability” shows that even for vision, which neuroscientifically is fairly
well understood, the influence of empirical beliefs remains contentious; in-
deed, there are different kinds of “penetration” that have only just begun to
be distinguished.12 Unfortunately this muddies our question of what the dif-
ference is between merely seeing X by seeing X’s visible properties and seeing
X as having this property or that.
Finally, we cannot assume that a single perceptual representation or per-
ceptual experience has just one intentional object or content. Peacocke
(1992), Lycan (1996), Noë (2004), and Schellenberg (2008) have each argued
that a representation can have more than one content, systematically related
to each other by some asymmetrical priority relation. Principally, the percep-
tual state may represent one object or property by representing a more primi-
tive or less ambitious one. For example, Schellenberg argues that we perceive
“situation-dependent” properties of external objects, and thereby the higher-
level properties of the same objects. I shall call views of this kind “layering”
views. My own model for layering of this sort is that of deferred linguistic
referring: uncontroversially, we refer to a thing by referring to a distinct but
in some way more accessible thing that is saliently associated with it.13
11 For resistance to the radical view on the basis of vision science, see Gilman 1992, 1994.
12 See the essays collected in Zeimbekis and Raftopoulos 2015.
13 A number by a numeral, a novel by a copy of it or by the author’s name, a color by a sample, a
military unit by its commander, hospital patient by her complaint (“That pancreas was discharged
this morning and went home”), or a restaurant patron by what he ordered (“The ham sandwich wants
his check”). Multiple layering: In the same gesture, we can refer to a numeral, thereby referring to a
number, thereby referring to an office in the building, thereby referring to its occupant.
Layering may be of some aid to liberals such as Siegel, since we might rep-
resent something like a pine tree by representing something more primitive.
And there are obvious auditory candidates (I say only “candidates” because
I do not want to beg any questions about what higher-level properties may be
represented in audition): by hearing a sound of such-and-such timbre and
duration, you hear a bang, and by hearing the bang you hear a collision; you
hear a major triad by hearing three tones sounded simultaneously or in close
succession, and perhaps you hear a D♭ minor triad by hearing the triad; by
hearing noises as of a human voice, you hear words, and (some contend) by
hearing those you hear meanings.
5.2. Aspect Perception
5. It is more than time to set out some facts and standard claims about as-
pect perception. Naturally, they have always been put in terms of vision.
It is even more unfortunate that for the public and for most professionals,
our entrée to aspect perception has been through ambiguous figures, para-
digmatically the duck-rabbit (figure 5.1).
We say the figure can be seen as a duck or as a rabbit, but that is inac-
curate. The figure is a picture, not an animal or other environmental object.
Except in unusual circumstances such as a very rare trompe l’oeil, we cannot
see the picture either as a duck or as a rabbit, but only as picturing a duck or a
rabbit. Or to put it better, following Wollheim (1996), we can distinguish be-
tween seeing-as and “seeing in.” It is less that we see the duck-rabbit figure as
Figure 5.1 Duck/rabbit picture

Hearing As 125
a picture of a duck (though of course we can do that) than that we see a duck
and alternately a rabbit in the picture, so to speak not noticing the picture.
Since aspect perception is of crucial importance in understanding visual art
(Gombrich 1960), seeing-in has been more important to aesthetics and art
criticism than has everyday seeing-as.
There are auditory pictures too, in the sense of real-world sounds being de-
liberately used as representations of events actual or nonactual. A raconteur
may interpolate oral sound effects in his funny story. A radio play without
words (the opposite of a silent movie) might depict an occurrence in nature,
or for that matter a sequence of human interactions. Occasionally this is done
in music, typically when there is a battle theme. Isaac wrote “A La Battaglia”
in 1487. Apparently Byrd wrote a keyboard suite called “The Battell” (ca.
1588); I don’t know that piece. Janequin produced a madrigal, “La Bataille,”
in which choral voices imitate battle sounds (much as he portrayed flocks of
birds in the much better known “Chant des Oiseaux”). Famously, the “1812
Overture” is actually scored for, and with sufficient funds is played using, real
cannons rather than percussion instruments to represent the cannon fire that
had actually occurred sixty-odd years before.14
In this chapter I should try to focus on perceiving-as, and set aside per-
ceiving-in; so, a bit ironically, the duck-rabbit should not be a paradigm.
For the same reason, neither should the Necker cube (figure 5.2).
There is no actual cube, but only the representation of one, and the aspect-
flipping it invites is as between the view from the top (with front facing left-
ward) and the view from the bottom (front rightward), in three dimensions.
And likewise for most of the ambiguous figures displayed on pages or seen
in paintings. A seeing-as proper in visual art will ordinarily be of a piece of
sculpture. For example, Henry Moore’s Nuclear Energy can in part be seen
as a mushroom cloud, or as a skull—and yet, isn’t that phenomenology still
more like seeing-in? That’s why I said “in part.” Of course, Moore could have
constructed a purer case, a sculpture that itself really could be aspect-flipped
as between a mushroom cloud and a skull. For that matter, someone could
make a real three-dimensional but transparent Necker cube and hang it
14 But cp. Peter Schickele’s hilarious parody, in which the cannon shots are represented by a tri-
angle going “ting.”

Apparently Tchaikovsky had been preceded in this use of actual cannon by Beethoven, in a short
1813 piece called “Wellington’s Victory.”
What should we say of leitmotifs? Do they represent, strictly speaking, or is there merely an in-
tended association? Think of a version of “Peter and the Wolf ” that is souped up in this regard.
Figure 5.2 Necker cube
somewhere. For all I know someone has, and ditto the duck-rabbit; in fact,
I’m sure that somewhere there is a three-dimensional duck-rabbit figure.
And the preceding point exhibits a third difference between seeing-as
and hearing-as: I doubt anyone could ever have introduced the whole topic
of aspect perception by playing examples of representational music. For one
thing, in drawing and in painting, ordinary three-dimensional objects are
represented by (not counting impasto) two-dimensional ones, while in rep-
resentational music events are represented by sounds, which are already real-
world events. Also, the parallel question would be, not whether the musical
sound itself could be aspect-flipped, but whether the represented sound could
be, and I find it hard to think of a good example of doing that.15 At least histor-
ically, the paradigm cases of “seeing as” are really seeings-in; but hearings-in
are rare at best, and the paradigm of “hearing as” is hearing-as proper.
6. Truisms coming from Wittgenstein (1953):
(1) “The expression of a change of aspect is the expression of a new percep-

tion and at the same time of the perception’s being unchanged” (196).
Indeed, we are tempted to say that the object seen has “changed,” even
15 What about sounds heard on recordings? A musical note or chord heard on a CD can certainly be
heard-as and aspect-flipped; and a muffled sound heard on a cassette tape can be heard as a cough or
a distant pistol shot. Perhaps these are straightforward cases of hearing-in. Note that there is a close
analogue for vision: people and their actions seen on television. But my own view is that those are not
cases of pictorial representation. Rather, they are cases of technologically enhanced hearing and seeing
per se. Thanks to audio and video recordings, you hear the orchestra itself, and you see the president. It
lacks only the cachet and the prestige of being there in person. Granted, this is a large subject.
Hearing As 127
though we “see that it has not changed” (193) and know that it in itself
has not.
(2) We can speak both of the “continuous seeing” of aspects and of the
“dawning” of aspects (194). Such a dawning may be very dramatic.
Think of those pointillist screens that appear to be just mélanges of var-
icolored dots; handed one, we are told to try to see a human figure or
rudimentary scene in it. The task is hard, and when or if we succeed,
there is a rush of revelation.
(3) “The flashing of an aspect on us seems half visual experience, half
thought” (197). But:
(4) We “describe the alteration like a perception; quite as if the object had
altered before my eyes” (195). Accordingly:
(5) The perception/cognition distinction does not apply straightforwardly.
In particular, it is inaccurate to divide the experience into what we
strictly see and how we interpret it in thought. Seeing-as is “like seeing
and again not like” (197).
(6) Seeing-as is part of visual phenomenology (though Wittgenstein would
never have put it that way).
(7) Seeing-as can be relational. E.g., we can “see a likeness” between two
faces (193); this seems to amount to more than just seeing the first as
like the second and the second as like the first).
(8) Perceiving-as is profoundly influenced by context. Two perfectly con-
gruent duck-rabbits will be seen as being quite dissimilar if one is
surrounded by ducks and the other by rabbits.
(9) In the striking cases that called our attention to the topic, the alternate
aspects compete. We cannot see both the duck and the rabbit simulta-
neously, but must flip between them.
(10) Aspects can be consciously and deliberately flipped, at will. (Doing that
is fun.)16
(11) Perceiving under an aspect is irreducibly conceptual. Even if strictly
visual content is nonconceptual, perceiving-as is or at least requires the
application of a concept. (That is strongly suggested by the grammar
alone: To see X as . . . is to see X as F or as bearing R to Y.)
16 For most anyone, not just for philosophers and psychologists. Why?
7. At least two more theoretical claims have since been made, and each is
plausible for the case of seeing:
Expectation: Perceiving X under a particular aspect is at least in part a

matter of expectation, of how one would tentatively expect X to behave or
its parts to be perceived should circumstances alter. Gombrich speaks of
how we “project” possibilities from what we perceive.
Attention: Aspect perception is an attentional phenomenon. Empirical

studies by Ricci and Blundo (1990) and Kleinschmidt et al. (1998) suggest
that the neural correlates of ordinary attention are activated during their
subjects’ aspect-flipping activities. Earlier, Chastain and Burnham (1975)
had shown that the aspect seen in an ambiguous figure depends on which
part of the figure the subject focuses on.
Attention in particular is a very important idea. For one thing, it

opens the aspect-perception issue to empirical findings from a different
and burgeoning area of research. More importantly, it would reduce
Wittgenstein’s mystery to a more tractable question about attending: granted
that visual attending in particular is a visual phenomenon, in that at least it
affects visual phenomenology, but what is the sense in which what is seen
does not change? Here the latter question has a nonrhetorical answer: By
every current psychological theory,17 attending is a matter of “heightening”
or “strengthening” or “enhancing” (or something) an existing representation,
as opposed to creating a new representation. What has not changed when an
aspect dawns is the content of the existing representation—which is a good
thing to mean by “What is strictly seen does not change.”
Orlandi (2011, 2012) uses attention to argue both that (11) is false, i.e.,
that aspect perception does not require the application of concepts, and that
aspect perception does not require interpretation of any sort, if “interpreta-
tion” means an inference-like subpersonal process.18
17 See the references in Mole 2010.

18 Gauker (2017) defends an even stronger anti-concept view of aspect perception. He holds that
all perceptual reference is merely “marking” in similarity spaces, and argues that aspect perception is
just a special case of that.
Even if we are unconvinced by Orlandi and Gauker and insist on the truth of (11), there is a se-
rious question of which concepts are mobilized in aspect-perceiving. DeBellis (1995) points out that
although music listening and appreciation are shot through with aspect perception and the details
are given extensive analysis by music theorists both traditional and psychology-informed, even
Hearing As 129
Figure 5.3 Dot array
8. If we do restrict consideration to aspect perception in propria persona

as opposed to perceiving-in, it still comes in several subcases. First, before
we get to ambiguous pictures where it seems obvious that everyday concepts
are involved, there are simpler organizational examples such as dot arrays
(figure 5.3), nonpictorial figures in the sense that there is no issue of a three-
dimensional interpretation.
The array in figure 5.3 can be seen either as four horizontal rows of three
dots each or as three columns of four each.
Figure 5.4 can be seen as a diamond shape standing on one of its points or
as a square tipped up onto one of its corners. (Notice that this case doubles as
an illusion, for seen as a diamond it looks bigger.)19
And Nickel (2007) offers a richer example (figure 5.5).
Nickel says we can see an arbitrarily chosen set of constituent squares “as
prominent.” For example, in figure 5.5 we can see the top row as prominent,
fairly sophisticated listeners do not often have the precise technical concepts applied by the music
theorists (“second inversion,” “hemiola,” “Neapolitan sixth”). DeBellis tends to identify the concep-
tual with the doxastic, so is disinclined to grant even that the listeners hear things under concepts
corresponding crudely to those technical terms, but we could insist on the latter without attributing
particular beliefs to the listeners.
Notice that a similar issue infects even the duck- rabbit paradigm. Even sophisticated
philosophers and psychologists do not aspect-see ducks or rabbits in the strict zoological senses of
those terms. I suppose there are ordinary concepts corresponding to Putnamian stereotypes as op-
posed to Putnamian natural kinds. But a person need not have heard of ducks or rabbits even in those
ordinary senses in order to aspect-flip on the duck-rabbit figure. It may be that the person must have
some rudimentary idea of “ear” or “beak.”
19 This is for the obvious reason that in diamond mode, both its top-to-bottom measurement and
its side-to-side measurement are longer by √2 times than the square’s sides.
Figure 5.4 Square /regular diamond
Figure 5.5 Nickel’s array
or alternately the left column, or alternating squares starting with the top left,
or the complement of that—or most any smaller or larger combinations of
squares—while representing just the same figure and its elements.20
20 Peacocke (1983), Macpherson (2006), and Nickel (2007) mobilize these and other examples in
objecting to the representation theory of sensory qualities; Block (2010) uses Gabor patches simi-
larly. But that issue is not our concern here.
Hearing As 131
There are auditory analogues. Think of a regular succession of beeps at

half-second intervals. You can hear them as grouped in twos, or in threes, or
in fours. You can hear them in different meters—3/4 versus 4/4, for example.
If you’re a musician you may be able to hear the difference between 3/4 and
6/8. Or imagine that the “piece” is notated in 3/4, but it is coming to a big ca-
dence and the penultimate two measures are a hemiola. A classic hemiola is
precisely a dramatic aspect-switch to duple meter within triple.21
Moreover, there are Nickelian differences of prominence that can be flipped
at will. Imagine, or actually listen to, a major triad that is being sustained in-
definitely. Hear the tonic as being merely harmonized by the third and fifth.
Now switch to hearing the third as the melody note, with the tonic as the bass
and the fifth as harmony. Now hear the fifth as the melody and imagine that
the third and tonic are about to drop a tone each to make V in first inversion;
better, imagine that the tonic’s drop to the leading tone will be slightly delayed,
making a suspension; that will reinforce hearing the fifth as melody.
For these kinds of cases, “organizational” as we may call them, atten-
tion looks pretty good. In the visual ones it does seem that we are doing
small shifts of attention, even though Nickel says that to get his effect we need
not change the focus of our vision. Attention is less obvious in the musical
cases, but still perhaps tenable.
Second subtype: The standard cases of seeing ordinary objects under aspects.
My favorite example (please pardon my using it yet again): in chapter 17 of
Booth Tarkington’s Penrod and Sam, Penrod is sitting reluctantly in church and
comes out of a doze to find himself looking at a coconut, sitting on an inverted
dinner plate “miraculously balanced” on the back of a pew a few rows in front
of him. He pulls himself together and realizes that the “coconut” is Georgie
Bassett’s head. Had he wanted to, he could then have relieved the continuing
tedium by aspect-flipping between Georgie, coconut, and perhaps other kinds
of object. Less amusingly, we can see a dog as a wolf or vice versa, a spot on the
wall as a bug, a rock formation as a human face, a convex surface as concave.
(Wittgenstein himself held, on the basis of ordinary-language consid-
erations since discredited by Gricean pragmatics, that except in abnormal
circumstances, an ordinary object is not perceived as the kind of object it is,
or rather that to say it is commits a linguistic error. That’s just wrong; concave
21 And how wonderful that is. Your whole body will move to emphasize the effect, if you don’t stop
yourself. My choral conductor wife reminds me that Handel’s “And the Glory of the Lord” is a joyous
riot of hemiolas.
surfaces can be and normally are seen as concave, a spot on the wall is seen as
a spot on the wall, and in the novel, Georgie’s head was nearly always seen as
a head if not as Georgie’s.)
There are auditory analogues for these ordinary cases too. A gunshot can
be heard as a popping like that of popcorn, or as a gunshot. Wind whistling
through an aperture can be heard as a human voice, as can a human voice.
A voice can be heard as speaking English, or just as making noises. A tune
can be heard as being played on a wind instrument, or as on a more specific
type of wind instrument.
Third subcase: Perceiving ordinary objects under very high-level aspects.
A cloud can be seen as a human head or as a particular type of animal. A dis-
play of colors and dots on a map can be seen as a Republican triumph. Naked
peaks in New Zealand can be seen as the advancement of global warming.22
Here too there are auditory cases; merely consider the more high-level
possibilities considered previously as candidates for representata on a very
liberal view: a distant rustling as Margaret Hillis conducting, or a distant
rumble as the launch of a Zircon missile.
Is aspect perception itself a kind of representation? Some authors assume
that it is,23 but, supposing that vision “itself ”/“alone” does not represent the
cloud we’re looking at as a Tibetan antelope having stomach trouble, is there
really a visual state of mine that does represent a Tibetan antelope having
stomach trouble? (Obviously some mental state does, viz., the thought that
occurs to the viewer, but that’s not the present question.) I suppose this just
reproduces Wittgenstein’s original mystery.
5.3. Music
9. In beginning this chapter I said that hearing-as plays at least two central
roles in human life. I have in mind its function in listening to music and, even
more central, its function in understanding speech.
22 I here ignore a fourth and still more diffuse sort of case mentioned in Lycan 2014: The nonce use
of everyday objects to represent the movements of ships, or of military units in a battle. The phenom-
enon is visual, because through the isomorphism intended by the players, a viewer can learn how the
real units did move in relation to each other, and possibly “see” when one destroyed another. But it
seems a stretch to say that a particular spoon or coffee cup can be seen-as a brigade of infantry.
23 And some just assume that it isn’t. Hence the use of seeing-as examples as objections to the rep-
resentation theory of sensory qualities, noted in note 20. (But if seeing-as is a specifically visual and
not merely cognitive phenomenon, then how is seeing a human head as a coconut not, at least in part,
visually representing the head as a coconut?)
Hearing As 133
But care is needed here, because if hearing-as per se is to have a func-

tion, or even play a distinctive role, in those activities, it must somehow be
contrasted with representation, or (not to beg the question raised a moment
ago) at least with some more basic kind of representation. To start, we may
insist that aspect perception is a whole-person-level achievement and con-
trast it as before with the subpersonal representations that figure in the me-
chanics of visual, musical, or language processing. Interestingly, there are a
number of person-level musical illusions.24
But if the content of a perceptual state is multiple and layered, as I be-
lieve, matters are complicated further. Layered perceptual representation
is not per se aspect perception, but it is whole-person rather than merely
subpersonal, and its more distal intentional contents may coincide with
aspects. Suppose Siegel is right and we can visually represent pine trees as
such. I would insist that we do that by representing objects having shapes.
And suppose I do see a pine tree by visually representing a pine tree. I may
be seeing the tree as a pine. But I might not be; I might even know it is a
pine but for some reason be seeing it as some other variety or as something
else entirely. So, under our suppositions, there are those two incompatible
aspect-perception possibilities. (And if Siegel is wrong, the third possibility
is that I cannot visually represent a pine tree as such, but can see what I do
represent as a pine tree.)
Is there an auditory analogue? Suppose I can represent a high-level mu-
sical entity, such as a closing cadence in E♭ major. Normally I would hear it as
such, but might I know it to be one and still hear it as something else? That is
hard to imagine—what could a cadential formula be but a cadence?—though
we cannot rule out the possibility a priori. But, obviously, if hearing itself
cannot represent anything so abstract, I can still hear the sound as a closing
cadence in E♭ major.
Music commentators and theorists use the “as” locution all the time.
A selection culled from Mark DeBellis’s (1995) book on conceptualization
of music:
Harmonically, one can hear m. 3 as revolving around F, and m. 4 around

G. In that case, the entire progression of mm. 1–5 can be heard as an expan-
sion of the I–IV–V–I of the opening measure.
24 See Diana Deutsch’s collection: http://deutsch.ucsd.edu/psychology/pages.php?i=201.

Example 5.1 The three diminished 7th chords
It is . . . possible to hear the establishment of the D-flat triad in m. 49 as

the arrival of tonic harmony.
This rhythmic arrival is heard also as the beginning of a new
hypermeasure.25
But such usage leaves it open whether a trained listener’s auditory system
actually represents things like harmonic expansions and hypermeasures. In
their monumental mingling of traditional music theory with cognitive and
perceptual psychology, Lerdahl and Jackendoff (1983) are well on the opti-
mistic side of that question, thereby diminishing our scope for musical as-
pect perception that outruns representation.
Perhaps we should approach the matter by considering ambiguity (un-
equivocal ambiguity, so to speak). Tonality in music results from unequal
division of the octave, in Western music into half steps versus whole steps;
a scale, regardless of key or absolute pitch as notated by letter name, is pre-
cisely a division of the octave into half steps and whole steps, and different
scales will sound quite different in character even to one who knows nothing
about music. For ambiguity, then, we look to equal division of the octave. We
can divide it into tritones (on a keyboard, diminished fifths or augmented
fourths),26 into minor thirds, into whole tones, or into half tones. The classic
case of tonal ambiguity is the (“full”) diminished 7th chord, of which there
will be three in any key, along with the diminished triads that are their proper
parts. Example 5.1 shows them in C.
(You can get a remarkable effect by rippling up and down these chords
for a while without resolving any of them.)27 In tonal music a full dimin-
ished 7th can be resolved directly to a 6th chord by lowering two of its notes
25 See DeBellis’s references (1995, 10). The writers are music theorists, not philosophers or
psychologists.
26 On a keyboard instrument, different tritones will not be truly the same interval, because of equal
temperament. And a true tritone can be heard either as an augmented fourth or as a diminished
fifth. Mediaeval music theorists, who had a Platonic/Christian way of parsing the cosmos in musical
terms, did not like the tritone, and reportedly called it “diabolus in musica.” I’m pretty sure some of
them would have meant that literally.
27 My composer friend Arthur Wenk once wrote a lied that had such rippling as its piano
accompaniment.
Hearing As 135
Example 5.2 From Mozart’s Sonata No. 6 in D Major, K. 284
a semitone each (if a 6th is considered sufficiently restful in the context),

but more commonly it will resolve to a straight 7th chord, which can serve
as V7 of a new key. The breadth of the ambiguity is appreciated when we see
that lowering any one of its four tones will produce a different 7th. For ex-
ample, in the first of the three chords in example 5.1, we can get either B7,
D7, F7, or A♭7.
Just listening to a diminished 7th being sounded out of context, it takes
real sophistication to hear that four-way ambiguity, but it is there to hear.28
And that experience should count as hearing-as, for what is represented does
not determine any resolution at all.
Contrast example 5.2: a whopping cadential formula, especially when it’s
tricked out with a conventional device such as the trill on the supertonic.
Not much ambiguity there! Now consider it minus its final I: If you were to
play all but the last chord and stop abruptly, even hearers who know nothing
about music would yelp with frustration, or at least feel frustrated. (But
there is still a potential ambiguity: that cadence, however final-sounding,
could be the closing cadence of a movement within a longer work, where the
movement itself is in V relative to the key of the whole. So a true disciple
of Schenker adept at “structural hearing” would try to hear the cadence as,
rather, a half-cadence, and might succeed.)
28 Perhaps this is a musical counterpart of an ambiguous figure. Casey O’Callaghan has suggested
to me that Deutsch’s surprising “tritone paradox” is one also: http://deutsch.ucsd.edu/psychology/

pages.php?i=206.
Example 5.3 From Mozart’s Sonata No. 6 in D Major, K. 284
Example 5.4 From Schubert’s “Der Wegweiser,” in “Winterreise”
But the really ubiquitous case of deliberate and indispensable ambiguity is

the pivot chord in a modulation,29 ubiquitous because modulation (temporary
change of key) is an important feature of any very interesting piece. Another
Mozart example, very mild, is example 5.3, where the pivot chord is the vi6 in
measure 5 that doubles as ii6 in the new key (A), setting up the new dominant
(E7) in measure 6. In analysis, pivot chords are sometimes notated with a hy-
phen: vi6-ii6. (I called this example “mild” because the new key is closely related
to the old, and the pivot chord is not highly ambiguous. The use of a diminished
chord as pivot is of course less conservative because of its degrees of freedom
aforementioned, and may introduce a more striking modulation.)
A more daring one, from a minor key to the minor one whole step higher,
is shown in example 5.4.
29 I’m speaking of a classic modulation that does work by pivot chord or “common chord”; not
all do, and some modulations, for one reason or another, are entirely unprepared. (Some are semi-
conventional, as when in pop music the last verse of a song is moved up a semitone.)
Hearing As 137
Here the pivot occurs by an intervening step of chromatic alteration, me-

lodically lowering the third of old (Fm) V to make new (Gm) iv.30
I’ve suggested that hearing-as is crucial because modulation of this type
requires tonal ambiguity in the pivot. But it is tricky to say when the pivot
chord aspect-dawns. It does not dawn the instant the pivot is sounded, be-
cause at that point the progression is still within the old key and could simply
continue in that key. It does not seem to dawn at any time until the next, alien
chord sounds; so far as the pivot chord was being aspect-heard in its split
second of existence, it was being heard as in the old key. But by the time that
next chord sounds, the pivot chord is past and no longer being heard at all.
The best we can say is that the alien chord makes us hear the pivot ambigu-
ously in retrospect; it is partly a memory phenomenon. (What if the pivot
had been the opening though pickup chord of a piece, and the next chord
established the piece’s key? But that merely reveals the fact that the opening
chord of any musical work is ambiguous and cannot by itself establish any
key at all; indeed, even the first and second chords together do not suffice.)
We might think this is a reductio of the idea that aspect switch is doing the
work in modulation. But (a) that the auditory ambiguity is doing the work
is undeniable, (b) “ambiguity” in this sense is defined in terms of hearing-
as, and (c) the advent of the new key certainly presents as the dawning of an
aspect. Another puzzle. Perhaps I am looking for too sudden a dawn; what’s
certain is only that by the time a couple of measures have gone by, I am now
hearing certain chords as I that I had been hearing as V, and the like.
30I will not resist footnoting a favorite of mine from Lennon (example 5.5).
This qualifies as a modulation because of the suddenness and prominence of the E major chord
in what was C. Actually E is just V of the relative minor (Am) and was set up by Am as pivot, but the
piece could have continued in Am or for that matter in A (major). As is, re-pivoted by F major that is
VI of Am, it modulates beautifully and restfully back to C.
Example 5.5 From the Beatles’ “I Want to Hold Your Hand”

What of the two theoretical claims formulated in section 7? Expectation

is nearly a byword in music theory. Indeed, much more so than Gombrich
would have imagined in thinking about visual art, because music is a tem-
poral and a performance art, and we are literally expecting events that have
not yet occurred. The great Leonard Meyer (1956) went so far as to explicate
the value and interest of a melody or of a harmonic progression entirely in
terms of expectation, formalized in turn in terms of Shannon-Weaver infor-
mation theory. More to the point, the aspect under which we hear an ambig-
uous tone or chord is indicated by what we expect to follow it—in the case of
a dissonance, how we expect it to resolve.
But now for an unexpected moral from our attention to aspect percep-
tion in music: hearing-as does not seem to have much to do with attention;
and so it casts doubt on attention. The psychologists cited previously were
thinking of our ways of attending to pictures: focusing on one or more parts
of a picture, or scanning the picture in this order or that. But we do not do
that with chords, not even when a chord is sustained for an appreciable pe-
riod of time (as in the ordinary case of modulation, it is not). This is sad news,
because attention offered real hope for the understanding of aspect per-
ception generally.
10. Music also affords an entirely different category of hearing-as: the
emotional. We hear a passage, or a melody, or a scale or (some say) even a
key as having affective qualities, the paradigm being happiness or sadness
and specifications of those. I personally think the understanding of music
in terms of emotion is absurdly overemphasized, but there’s no denying that
many go in for it, or that it has a solid psychological basis. If there were no
natural association between musical properties and affect, there would be no
movie soundtracks, to say nothing of opera.
The literature on the expressing (in some sense) of affect by music is huge,
and I cannot enter into any of those debates; I shall only comment on emo-
tional properties as they may be aspect-heard. It is sometimes said that a
piece or passage “is sad,” “expresses sadness,” or “is heard as sad.” Given the
all too obvious fact that a musical sound-event is not itself a sentient crea-
ture, what could such statements mean?—but now merely as concerning as-
pect perception.
I shall just follow Peter Kivy (2002) in supposing that when we say such
things and we are talking about the music itself, they do not mean things
about real emotional effects the music will have, or would have, or tends
to have, and so on, upon listeners. Listening to music obviously can affect
Hearing As 139
listeners emotionally, for any of a variety of reasons, but there is no evi-

dence for the claim made by “arousal” theorists that merely listening to
well-performed “sad” music has per se any tendency to make a concertgoer
actually feel sad; indeed the better the performance, the more exhilarated the
audience members are likely to be.31 The “sadness” is a property of the music,
even though a sound-event cannot be sad in the same sense a person can
be. I sympathize with the “contour” theory defended by Kivy (1980) (though
I acknowledge the objections he himself went on to make in Kivy 2002, I be-
lieve they can be overcome): a piece of music or passage “is sad” in that its
literal or near-literal properties are analogous to salient properties of a sad
person, much as (Kivy’s central example) a St. Bernard’s face looks like a car-
icature of a sad human being.
Sad music being in slow, halting tempo, subdued in dynamics, with

drooping, faltering melodies, and sad people walking in slow, halting gait,
with drooping bodies, speaking in subdued, halting voice, cannot be alto-
gether coincidental. . . .
Melancholy people’s voices tend to “sink,” and tend to remain in the low
vocal register. . . .
[T]‌he “contour” of music, its sonic “shape,” bears a structural analogy to
the heard and seen manifestations of human emotive expression.
(2002, 37–40)
How, then, can such music be heard as expressing sadness? Simple an-
thropomorphism. For obvious evolutionary reasons, we have at least a slight
propensity for seeing things as animate, and (again following Kivy) I suggest
this carries over to hearing as well. At least, we can hear a sound-event as
expressing sadness (even when we know it is actually doing no such thing)
in much the way we see the St. Bernard’s face, or a natural configuration, as
doing so.
And this is a clear case of hearing-as that is not musical representation. So
far as it is representation at all, it defies all existing psychosemantics. And if it
qualifies as cognitive penetration, it seems to be of a special kind.
31 A personal example: I once attended a performance of Lyell Cresswell’s orchestral piece “Ylur,”
whose title is Gaelic for “sorrow” or “deep grief ” and which according to the New Zealand website
SOUNZ (http://sounz.org.nz/) “is concerned with different aspects of sorrow.” I loved the piece, espe-
cially for its novel orchestration effects. At a party following the concert I exclaimed to Cresswell him-
self that I got nothing from it about sorrow or grief but had expressed my listening pleasure in excited
mental exclamations of “That is so cool!” I think he was disappointed; perhaps he holds an arousal view.
5.4. Speech
11. Moore famously noted the huge phenomenological difference between

hearing someone utter a sentence in a language we do not understand and
hearing someone utter a sentence in our native language. O’Callaghan (2011)
points out that this difference can be made fairly precise and controlled for,
in experiments making use of “sinewave speech.”32 The difference offers a
prime candidate for hearing-as, because the sense in which what is “strictly”
heard is obvious to common sense.
But open to correction by science. Psychologists speak of the “language
module,”33 and it is held by some that we do not first hear sounds that are
speech sounds and then use background information to infer or construct
a linguistic entity; rather, we are built to just hear meaningful utterances as
such. According to a strong version of this view, we simply hear meanings.
According to an even stronger version, we hear strong implicatures, and
in such cases do not even hear literal meanings.34 If some such modularist
views are correct, then we actually represent semantic sound events, and are
not merely hearing-as. But O’Callaghan (2011, 2015) argues that the empir-
ical evidence for the modularist claims is scanty at best. So, as we did in the
case of music perception, let’s turn to known ambiguities.
Of course there are oronyms and mondegreens, phonological ambiguities
that depend in part on dialect.35 My old colleague Stan Munsat tells a story
in which “I buried it” is misheard as “A bear ate it” or vice versa. My personal
favorite is the possibility of hearing the Beatles’ line “The girl with kaleido-
scope eyes” (from “Lucy in the Sky with Diamonds”) as “The girl with colitis
goes by.” And my wife remembers from 60 years ago an entire genre of pun
jokes that started with “She was only . . . ,” an instance of which is “She was
only a stablehand’s daughter, but all the horsemen knew her” (“horse ma-
nure”). But I am more interested in ambiguities of meaning.
32 Remez et al. 1981. To hear examples: http://www.lifesci.sussex.ac.uk/home/Chris_Darwin/
SWS/. (Thanks to Casey O’Callaghan for the reference.)

33 See, e.g., Pinker 1994.
34 Actually most of those who claim we hear meanings are philosophers, e.g., Strawson (1994).
I know of no one who has defended the implicature claim in print, but I have heard more than one
linguist say it. It is supported both by phenomenology and by appeal to memory, but those consider-
ations are not strong.
35 The term “mondegreen” is attributed to the columnist Sylvia Wright (http://wordinfo.info/unit/
3347), who reportedly misheard the Scottish folk song line “They hae slain the Earl of Murray, and
laid him on the Green” as “They hae slain the Earl of Murray and Lady Mondegreen.”
Hearing As 141
As Paul Ziff used to emphasize, nearly every English sentence has permis-
sible readings other than the obvious one(s), readings that would never have
occurred to anyone. How these are ruled out in normal disambiguation is
entirely unknown; even to speak of “disambiguation” is to suggest a process
of some sort, but Gricean reasoning or even relevance-theoretic adjustment
would be far too cumbersome. Still, we may suppose for the sake of argu-
ment that the intended reading is linguistically represented by the hearer. But
sometimes the sentence uttered is more conventionally and obviously am-
biguous, and the hearer genuinely can’t tell from intonation which reading
is intended. Moreover, suppose the utterance is recorded, and we can play it
over and over. Then we could deliberately aspect-flip it.36
For syntactic ambiguities perhaps attention has some purchase; when
we aspect-flip Groucho Marx’s “This morning I shot an elephant in my
pajamas,” we may be relying on tacit emphasis and focusing via slightly
different-sized gaps between words. But semantic ambiguity is less ame-
nable; I don’t see how attentional differences would apply to uniformly pro-
nounced lexemes.
12. There is a more interesting and I think indisputable sort of aspect
perception in listening to speech: interpreting illocutionary force. As hearer
of a particular utterance I may be genuinely undecided as between a mere
statement of fact, a warning, an actual threat, or possibly even a reassurance.
As children say, “Is that a threat or a promise?” Assuming there is no ambi-
guity in the sentence uttered, syntactic and semantic representations will be
fixed, and my uncertainty will concern only the speaker’s intentions. Or, al-
ternately, the context may leave no uncertainty even when the speaker’s tone
is flat and does not contribute. I hear the threat very definitely as a threat.
An even more striking example is the hearing-as we do when attending a
play. The illocutionary forces conveyed by the actors are only simulated; the
human beings on stage are not really making statements, asking questions, or
issuing requests or commands, but only simulating those actions. But in any-
thing like the normal case of watching a play, we vividly hear their utterances
36 It is now forgotten, and certainly was by me until I happened to see Davies 2011, that
Wittgenstein explicitly applied his discussion of aspect perception to linguistic understanding. He

held that sometimes we “experience” the meanings of words (210). He even considers a hypothetical
condition of “aspect-blindness” applying to pictures (213–214) in which the sufferer could see only
one aspect and not the other(s). That would translate, he goes on to say, into inability to hear alternate
meanings of ambiguous words.
And Wittgenstein actually mentions both the appreciation of puns and the phenomenon of “sec-
ondary” lexical meanings that depend asymmetrically on normal ones.
as having those forces, and react emotionally, even when we know that the
utterances do not actually have those forces.
Even in theater there can be uncertainty and genuine ambiguity on the
latter point. Some avant-garde drama is interactive, well past “A-effect”; au-
dience members may contribute genuine speech acts, to which an actor may
respond but while still in character. Or it may be deliberately left open which
part of the venue is the stage and which human beings are actors, stagehands,
or audience members.37 Or—never mind the avant-garde—the real world
might intrude: The stage manager might step out, distraught, to announce
that (backstage) one of the actors has been shot.38
Thus, illocutionary aspect ambiguity can run riot. Here again it is not
explained by linguistic representation. And here again, although differences
in attending may play a role, attention cannot be the whole story or even
much of it.
5.5. Epilogue
13. Aspect perception is not literally paradoxical, but it remains mysterious.

My tentative diagnosis: To sort out exactly what aspect perception is and how
it works, we will have to gain a much fuller empirical story than we have now,
not just about vision or just about vision and audition but about all the ex-
teroceptive sense modalities. But even complete empirical information will
not settle the matter, because the mystery involves philosophically conten-
tious questions such as the nature of representation and how to understand
matters of cognitive penetration generally. Moreover, it will have to be de-
cided how to lay folk categories and concepts over the neuropsychological
facts (and for that matter how much of the relevant folk psychology must be
respected and preserved). Some of that decision may turn out to be purely
verbal, but some will not.
I’m quite sure that none of that will happen in my lifetime, and probably
not in yours. It may happen before the heat death of the universe.
37 Bruce Wilshire (1982, ix–xiv) vividly and appreciatively describes attending a performance of
Robert Whitman’s Light Touch, in which this conceit is employed.

38 I owe this sort of example to Mayank Bora. An actor friend of mine once appeared in a perfor-
mance during which one of the other actors died, suddenly, actually, on the stage.
Hearing As 143
Acknowledgments
I am grateful to Casey O’Callaghan, Nico Orlandi, Jenny Judge, and Erica

Skoe for bibliographic help, and to John Brackett for technical discussion of
diminished chords and others.
References
Block, N. J. 1995. On a confusion about a function of consciousness. Behavioral and Brain
Sciences 18: 227–287.
Block, N. J. 2010. Attention and mental paint. In E. Sosa and E. Villanueva, eds.,
Philosophical Issues, 20: Philosophy of Mind. Oxford: Wiley-Blackwell. 23–63.
Campbell, J. 2002. Reference and Consciousness. Oxford: Oxford University Press.
Casati, R., and J. Dokic. 1994. La philosophie du son. Nimes: Chambon.
Chastain, G., and C. A. Burnham. 1975. The first glimpse determines the perception of an
ambiguous figure. Perception and Psychophysics 17: 221–224.
Churchland, P. 1988. Perceptual plasticity and theoretical neutrality: A reply to Jerry
Fodor. Philosophy of Science 55: 167–187.
Clark, A. 2000. A Theory of Sentience. Oxford: Oxford University Press.
Clark, A. 2004. Feature-placing and proto-objects. Philosophical Psychology 17: 451–477.
Davies, L. 2011. Aspect perception and understanding the meanings of words in
Wittgenstein’s Philosophical Investigations. Sophia 12.
DeBellis, M. 1995. Music and Conceptualization. Cambridge: Cambridge University Press.
Dretske, F. 1993. Conscious experience. Mind 102: 263–283.
Farennikova, A. 2013. Seeing absence. Philosophical Studies 166: 429–454.
Fish, W. 2009. Perception, Hallucination, and Illusion. Oxford: Oxford University Press.
Fodor, J. A. 1983. The Modularity of Mind. Cambridge, MA: MIT Press.
Gauker, C. 2017. Three kinds of nonconceptual “seeing-as.” Review of Philosophy and
Psychology: 1–17. DOI: 10.1007/s13164-017-0339-2.
Gilman, D. 1992. A new perspective on pictorial representation. Australasian Journal of
Philosophy 70: 174–186.
Gilman, D. 1994. Pictures in cognition. Erkenntnis 41: 87–102.
Gombrich, E. H. 1960. Art and Illusion. London: Phaidon.
Hanson, N. R. 1958. Patterns of Discovery. Cambridge: Cambridge University Press.
Horgan, T., and J. Tienson. 2002. The intentionality of phenomenology and the phe-
nomenology of intentionality. In D. Chalmers, ed., Philosophy of Mind: Classical and
Contemporary Readings. Oxford: Oxford University Press. 520–533.
Kania, A. 2010. Review of Nudds and O’Callaghan 2009. Notre Dame Philosophical
Reviews, August 17.
Kivy, P. 1980. The Corded Shell: Reflections on Musical Expression. Princeton, NJ: Princeton
University Press.
Kivy, P. 2002. Introduction to a Philosophy of Music. Oxford: Oxford University Press.
Kleinschmidt, A., C. Buchel, S. Zeki, and R. S. Frackowiak. 1998. Human brain activity
during spontaneously reversing perception of ambiguous figures. Proceedings of the
Royal Society of London, Series B: Biological Sciences 265: 2427–2433.
Kuhn, T. 1962. The Structure of Scientific Revolutions. Chicago: University of

Chicago Press.
Lerdahl, F., and R. Jackendoff. 1983. A Generative Theory of Tonal Music. Cambridge,
MA: MIT Press.
Lycan, W. G. 1971/ 1985. Gombrich, Wittgenstein and the duck- rabbit. Journal of
Aesthetics and Art Criticism 30: 229–237. Reprinted in J. V. Canfield, ed., The Philosophy
of Wittgenstein: Aesthetics, Ethics and Religion (New York: Garland Publishing).
Lycan, W. G. 2008. Phenomenal intentionalities. American Philosophical Quarterly
45: 233–252.
Lycan, W. G. 2014. What does vision represent? In B. Brogaard, ed., Does Perception Have
Content? Oxford: Oxford University Press. 311–328.
Lycan, W. G. 2018. What does taste represent? Australasian Journal of Philosophy
96(1): 28–37. http://www.tandfonline.com/eprint/P7RszBA2tMZfhF3WqfKA/full.
Macpherson, F. 2006. Ambiguous figures and the content of experience. Noûs 40: 82–117.
Meyer, L. 1956. Emotion and Meaning in Music. Chicago: University of Chicago Press.
Mole, C. 2010. Attention Is Cognitive Unison: An Essay in Philosophical Psychology. New York:
Nickel, B. 2007. Against intentionalism. Philosophical Studies 136: 279–304.
Noë, A. 2004. Action in Perception. Cambridge, MA: MIT Press.
Nudds, M. 2009. Sounds and space. In M. Nudds and C. O’Callaghan, eds., Sounds and
Perception: New Philosophical Essays. Oxford: Oxford University Press. 69–96.
Nudds, M., and C. O’Callaghan, eds. 2009. Sounds and Perception: New Philosophical
Essays. Oxford: Oxford University Press.
O’Callaghan, C. 2007. Sounds: A Philosophical Theory. Oxford: Oxford University Press.
O’Callaghan, C. 2011. Against hearing meanings. Philosophical Quarterly 61: 783–807.
O’Callaghan, C. 2015. Speech perception. In M. Matthen, ed., The Oxford Handbook of the
O’Shaughnessy, B. 2009. The location of a perceived sound. In M. Nudds and C.
O’Callaghan, eds., Sounds and Perception: New Philosophical Essays. Oxford: Oxford
Orlandi, N. 2011. The innocent eye: Seeing-as without concepts. American Philosophical
Quarterly 48: 17–32.
Orlandi, N. 2012. Embedded seeing-as: Multi-stable visual perception without interpre-
tation. Philosophical Psychology 25: 1–19.
Pasnau, R. 1999. What is sound? Philosophical Quarterly 49: 309–324.
Peacocke, C. 1983. Sense and Content. Oxford: Oxford University Press.
Peacocke, C. 1992. A Study of Concepts. Cambridge, MA: MIT Press.
Pinker, S. 1994. The Language Instinct: How the Mind Creates Language. New York: W.
Morrow.
Ramsey, W. 2007. Representation Reconsidered. Cambridge: Cambridge University Press.
Remez, R. E., P. E. Rubin, D. B. Pisoni, and T. D. Carell. 1981. Speech perception without
traditional speech cues. Science 212: 947–950.
Ricci, C., and C. Blundo. 1990. Perception of ambiguous figures after focal brain lesions.
Neuropsychologia 28: 1163–1173.
Rosch, E., C. Mervis, W. Gray, D. Johnson, and P. Boyes-Braem. 1976. Basic objects in nat-
ural categories. Cognitive Psychology 8: 382–439.
Hearing As 145
Schellenberg, S. 2008. The situation-dependency of perception. Journal of Philosophy

105: 55–85.
Schellenberg, S. 2011. Perceptual content defended. Noûs 45: 1–37.
Scruton, R. 1997. The Aesthetics of Music. Oxford: Clarendon Press.
Siegel, S. 2010a. Do visual experiences have contents? In B. Nanay, ed., Perceiving the
World. Oxford: Oxford University Press. 333–368.
Siegel, S. 2010b. The Contents of Visual Experience. Oxford: Oxford University Press.
Siegel, S., and A. Byrne. 2016. Rich or thin? Debate in B. Nanay, ed., Current Controversies
in Philosophy of Perception. London: Routledge. 59–80.
Sorensen, R. 2008. Seeing dark things: The philosophy of shadows. Oxford: Oxford
University Press.
Sorensen, R. 2009. Hearing silence: The perception and introspection of absences. In M. Nudds
and C. O’Callaghan, eds., Sounds and Perception: New Philosophical Essays. Oxford:
Strawson, G. 1994. Mental Reality. Cambridge, MA: MIT Press.
Travis, C. 2004. The silence of the senses. Mind 113: 57–94.
Wilshire, B. 1982. Role Playing and Identity: The Limits of Theatre as Metaphor.
Bloomington: Indiana University Press.
Wittgenstein, L. 1953. Philosophical Investigations. Trans. G. E. M. Anscombe. Oxford:
Basil Blackwell.
Wollheim, R. 1996. Seeing-as, seeing-in, and pictorial representation. In Art and Its
Objects. Cambridge: Cambridge University Press.
Zeimbekis, J., and A. Raftopoulos, eds. 2015. The Cognitive Penetrability of Perception: New
Philosophical Perspectives. Oxford: Oxford University Press.
6
Is Tactual Knowledge of Space Grounded
in Tactual Sensation?
John Campbell
6.1. Objective and Subjective Poles in Touch
Katz’s The World of Touch, as reported by Krueger, drew a basic distinction

between two aspects of tactual knowledge. First, there is our knowledge of
the external reality we’re touching; second, there is our knowledge of our
own bodies. These are the “objective” and the “subjective” poles of touch.
There are, then,
two distinct poles or aspects of the touch experience: (l) the subjective pole
(e.g., “I feel a prickling sensation”) . . . and (2) the objective pole (e.g., “I feel
a pointed object out there”)
In vision, by contrast, only the objective pole exists, with all impressions
being projected and felt “out there,” even when one views an afterimage
with the eyes shut or experiences a subjective gray which really depends
on the state of the visual system rather than on an external object.
(Krueger 1970, 337–338)
I want to connect this point with an issue that comes up about the role of our
sensory experience in getting knowledge. Everyone agrees that perception
gives knowledge of our surroundings. There’s a bifurcation in how people
think about the role of consciousness in this process. Epistemologists, and
common sense generally, take it as the most obvious point that it is percep-
tual experience that gives knowledge of our surroundings. Cognitive science,
on the other hand, is the discipline that seems most obviously responsible for
explaining how we know about our surroundings. And in its models of how
knowledge is achieved, it generally makes no use of concepts of experience
or consciousness. Cognitive science uses a concept of representation, but it
John Campbell, Is Tactual Knowledge of Space Grounded in Tactual Sensation? In: The Epistemology
of Non-Visual Perception. Edited by: Berit Brogaard and Dimitria Electra Gatzia, Oxford University
Press (2020). © Oxford University Press.
DOI: 10.1093/oso/9780190648916.003.0007
What grounds tactual knowledge of space? 147
doesn’t have any evident relation to experience. Having identified firings in a

cell assembly, scientists will ask (a) To what external phenomenon are these
cell-firings a response? And having provisionally identified external phe-
nomenon X as the thing to which these cell-firings are a response, scientists
will ask, (b) Is it plausible that the adaptive point of these cell-firings is to
identify the presence of X? In that case, the scientists will say that these cell-
firings represent the presence of X. Philosophers have done a lot of work to
make this idea fully explicit (cf., e.g., Millikan 1996). But for present purposes,
the point is that there’s no presumption that anything so represented is re-
flected in consciousness. The whole story about how computations over such
representations generate our representations of our surroundings seems to
leave consciousness without any work to do. But if there is any work for con-
sciousness to do, it is not explicitly part of the information-processing story
supplied by cognitive science.
What, if anything, could in principle be added to a story about the role
of experience in generating knowledge than its having representational
content? The classical, and still popular answer is that in addition to repre-
sentational contents, experiences have sensational contents, distinctively
experiential aspects. Perhaps, as Burge (2010) says, any representational
aspects of experience could be had by non-conscious states. But if our know-
ledge of the world is grounded, at least partly, in perceptual sensations, then
we can see how perceptual experience, and not merely perceptual represen-
tation, could play a role in the generating knowledge.
This approach does not, however, seem very promising for the case of vi-
sion. The trouble is the transparency of vision, which makes it difficult to
understand how we can make sense of the idea of visual sensations. Talk
about visual sensations cannot be grounded ostensively, in our use of “inner
demonstratives” to point out aspects of visual experience; any demonstrative
we might use in connection with vision seems bound to refer only to some-
thing seen. As Katz put it, vision does not have a “subjective pole.” Of course,
in vision, there are phenomena such as “one’s eyes feeling tired,” or “feeling
overwhelmed by the light and color and variety” of the scene in the bazaar.
But what we do not have is the possibility of, for various particular visual
stimuli, oscillating back and forth between noticing the thing in the envi-
ronment, and noticing the idiosyncrasies of one’s visual response specifically
to that thing. Unlike vision, though, touch does have, as Katz also pointed
out, a subjective pole. You can oscillate back and forth between noticing the
shape of the kettle you have lifted up during a power outage in the kitchen,
and noticing the way in which it specifically is impacting on your body. Of

course, oscillating back and forth may not always be possible: as Katz notes,
in the case of a hand exploring a surface, attention to one’s fleeting sensations
may actually be obstructive, preventing one from finding whether the sur-
face is level and hard, for example, whereas in the case of areas of the body
that are not often used for tactual exploration of the environment, such as the
inside of the ear or nose, if there is an alien intruder, it is usually only possible
to attend to the subjective pole of touch.
How to demonstrate the existence of such things as “the visual sensation
of squareness” or “the visual sensation of redness,” as opposed to just seeing
something square or something red, is something of a puzzle. For example,
Ned Block, in a much-discussed series of papers, has argued that various
experiments in vision science demonstrate the existence of something he
calls “mental paint,” which is supposed to be a version of the classical idea
that in vision we have not just representation, but this other aspect, sensa-
tion. When we step back and view what is going on here synoptically, the
methodology is puzzling. We are supposed to be considering an aspect of
personal-level perceptual experience. Why it should take an experiment to
demonstrate the existence of such a thing is hard to understand. If it’s right
there in the dazzle of ordinary visual experience, how can its existence be
an arcane matter requiring subtle experiments for its demonstration? And
if it does require experiment to demonstrate its existence, presumably it is
to be thought of as a construct, like the Higgs boson, postulated to explain
observation. But then we should expect some elaboration of the construct.
For example, is all mental paint intrinsically the same? Is it homogenous? Or
are there different varieties of mental paint? If so, what are the dimensions of
variation? Are there “principal types”? Does mental paint come in different
quantities? It seems quite evident that no one has the slightest idea how to
answer these questions; certainly Block’s papers provide no clue at all. It is
natural to wonder whether this is a valid area of inquiry.
In sharp contrast, in the case of touch, there really are tactual sensations: it
does not require experiment to establish that there is the bodily sensation one
gets when a needle is injected into one, or a dentist’s drill touches a nerve, for
example. Tactual sensations are a ubiquitous feature of ordinary tactual ex-
perience, distinct from our tactual perception of the external environment;
we notice them at will, and ordinary language provides an extensive vocab-
ulary for describing and classifying the variety of our tactual sensations. So
in this case we do have a sensible question: we can legitimately ask whether
these tactual sensations play an essential role in grounding our knowledge of

the world. Could touch provide us with knowledge of the environment even
if, like vision, it lacked a “subjective pole”?
When we consider the role of consciousness in our knowledge of our
environment, it is natural to begin by reflecting on what cognitive science
seems to tell us about how we gain knowledge, and then to reflect on the
fact that the notions of representation and information processing that it
uses make no mention at all of phenomena of consciousness. Then one won-
ders if there might be a role for sensation in the workings of perception, and
this suggests that maybe the role of consciousness comes in when we see
how the processing of representations has to be supplemented by an appeal
to sensation. But, of course, we do not have to think of consciousness in
terms of sensation. One alternative model would be to think of conscious-
ness of one’s environment as a relation that one stands to it, rather than as
a matter of the internal generation of sensations. So a belief that conscious-
ness matters for knowledge of one’s surroundings does not commit one to
thinking that this knowledge is grounded in sensation. So even if we believe
consciousness matters for tactual knowledge of the world, that does not of
itself settle, either way, the question whether tactual knowledge of our sur-
roundings is grounded in tactual sensation.
If you miss the possibility that “consciousness” might fundamentally be
a relation between the perceiver and the environment, then one natural,
perhaps inescapable option is to think of consciousness in terms of sensa-
tion: a conscious tactual perception is one in which sensation figures. As
I say, if you think that “tactual experience of x” is a relation between the
perceiver and x that can hold whether or not the experiencer is having any
tactual sensations, then you don’t have to suppose that conscious tactual
perception is always accompanied by tactual sensations. But if you think
that the consciousness of a tactual perception simply is a matter of there
being a tactual sensation in the perceiver, then you are forced to conclude
that there must be tactual sensations present whenever there is conscious
experience. Moreover, the consciousness of a tactual perception seems to
be key to its playing a role in grasp of concepts of the external world. So
that means, on this model, that the sensation must be doing some work
in our grasp of concepts. What could it be? The situation we are in at this
point is extremely puzzling. We seem somehow to have been forced to a
position on which there “must” be all these tactual sensations of which we
do not ordinarily have any knowledge, and these hypothesized sensations
are doing some important cognitive work for us in providing our concep-
tion of the external world. But what on earth could that work be? When you
employ the “subjective pole” of tactual awareness ordinarily, noticing your
own bodily sensations, you do not seem to be getting information that could
provide you with any significant understanding of what is going on in your
external environment.
Consider, for example, sweeping one’s hand across a table. It seems possible
that you could use this kind of exercise to explain to someone the meaning of
the concept “flat surface.” And it seems that there would have to be conscious
tactual perception of the surface of the table for this exercise to provide you
with knowledge of what a “flat surface” is. Having your body pick up some
information from a hand-sweep, remote from consciousness, couldn’t of it-
self help you to understand the concept of a flat surface. If we think of con-
sciousness in terms of sensation, this means that the sensation must be doing
some work in your understanding of the concept of a flat surface. But how
could that be? The flatness of a surface is one of the elements that explains
the course of your tactual sensations (other elements including, for example,
that you’ve moved your hand in relation to the table). But how could merely
having a tactual sensation generate knowledge of the factor explaining why
your tactual sensations are as they are? Of course, on this picture, merely
having the sensation would provide you with some data—the occurrence of
the sensations—that call out for explanation, so having the sensations could
provide you with knowledge of the existence of a hypothesized factor “F” to
explain the course of one’s sensations. But tactual experience seems to pro-
vide one with knowledge of what a flat surface is, not merely knowledge of
the functioning of some unknown hypothesized cause.
Incidentally, it seems to me that within Katz’s notion of the “subjective
pole” of touch we can make a distinction between (a) noticing and thus
knowing about the tactual sensation, and (b) the tactual sensation itself. But
the distinction should not be overdrawn. The characteristics of the tactual
sensation are not independent of our ways of knowing about it. For example,
consider cases in which attention is distracted from bodily pain, as in the case
of soldiers on a battlefield who do not notice their injuries, football players
who sustain damage of which they’re unaware, or even people who watch
an engrossing movie and forget their headache. In such cases we naturally
feel some skepticism as to whether the sensation is there at all, even though
unnoticed; the metaphysics of sensation here cannot be sharply separated
from the epistemology of sensation. Otherwise we would have no protection
against such hypotheses as that, for example, all humans are constantly
experiencing very high levels of bodily pain, that the world is an ocean of suf-
fering, but that because of inherited biases in our attention, we are in practice
incapable of exhibiting or even noticing this pain.
It seems to me that the idea that our tactual knowledge of our surround-
ings is grounded in tactual sensation gets things exactly the wrong way
round. It is easiest to see this, I think, in the case of spatial aspects of sensa-
tion, and we will look at them in more detail later. The situation is, I think,
that our knowledge of our tactual sensations is grounded in our knowledge
of our external environment, and that consequently, the characteristics of
our tactual sensations are grounded in our knowledge of our external en-
vironment. Whether consciousness is essential to our tactual knowledge
of our surroundings is a separate matter. We can keep them separate by
acknowledging that perceptual consciousness should not be analyzed in
terms of sensation.
6.2. Reid’s Framing of the Problem:

“Fugitive Sensations”
The classic discussion of tactual perception and tactual sensation is Thomas

Reid’s, in An Inquiry into the Human Mind on the Principles of Common
Sense. Reid centers the discussion around his experimentum crucis, which we
will come to in a moment. But there is a lot of interest in the way he frames
the discussion, so let’s begin by looking at that.
Reid’s examples of properties known by touch are extension, and hard-
ness or softness. The key move in his discussion comes almost parentheti-
cally, and right at the start, when he says: “There is, no doubt, a sensation by
which we perceive a body to be hard or soft” (V.ii.37). Now although I have
begun by emphasizing that there really are tactual sensations, it certainly
doesn’t follow from this that it is by having tactual sensations that we have
tactual perception of the characteristics of external objects. In Katz’s terms,
when we are attending to the objective pole of touch, we are tactually aware
of characteristics of external objects. When we are attending to the subjec-
tive pole of touch, we are aware of bodily sensations. So far, there are here
simply two different domains of tactual knowledge, the external properties
and the sensations. It doesn’t follow at all that one achieves tactual percep-
tion of external properties by having tactual sensations. Why should one
think that it does, that there is, “no doubt” a sensation by which we per-
ceive a body to be hard or soft? I think there are two complementary lines of
thought one might have here:
(1) How is it, in general, that one tactually perceives an external stimulus?
Presumably the thing must have some impact on the perceiver. There
must be something about the perceiver that is changed by the external
stimulus. And that internal change must be sufficiently specific to allow
recovery of information about the external stimulus.
If you think of the “internal change” that must be made in the per-
ceiver as a sensation, then it will immediately follow that the sensation
must be sufficiently specific to allow recovery of information about the
external stimulus. But, of course, there is no particular reason to think
that the internal change made must be the production of a sensation.
It may be a disturbance in the nerves and joints and brain of the per-
ceiver, rather than a sensation.
There may, indeed, be the consequential generation of a sensation
in the perceiver, but there may not be. And there is no reason to think
that it is by having the sensation, if one is produced, that one perceives
the hardness of the external object. The sensation may anyhow not be
sufficiently specific to recover information about the characteristics of
the external thing.
(2) You might point out that it is conscious tactual perception in particular
that gives one knowledge of the properties of the thing—its hardness or
softness, for example. One certainly might unconsciously adjust one’s
clothing or posture in response to changes in the room temperature,
or in the orientation of the floor one is walking across, that one wasn’t
aware of, for example. But that exercise wouldn’t mean that one’s expe-
rience could be used to explain the meaning of concepts of temperature
or of orientation. Tactual information-processing without tactual ex-
perience would not of itself provide for conceptual understanding or
propositional knowledge of one’s surroundings.
What does this point have to do with the idea that one perceives
the hardness or softness of a body by having tactual sensations? Well,
suppose you think that tactual experience just is a matter of having
tactual sensations. Then when you acknowledge that it is tactual ex-
perience specifically that provides propositional knowledge of one’s
surroundings, you seem to be forced to say that the sensations are

what provides the knowledge; that you have the knowledge by having
sensations.
But, of course, we do not have to think of tactual experience as a
matter of having tactual sensations. We can think of perceptual con-
sciousness as a relation between the perceiver and the environment. It
may be, after all, that the experience one reports when one says, “I feel
the hardness of the metal,” can’t be analyzed into an inner component,
the sensation, plus some external causation, but is rather a matter of a
relation, “x tactually experiences y,” holding between the perceiver and
the thing.
To put these two points together, the “disturbance in the nerves and joints
and brain of the perceiver” that is caused in tactual perception may be the
ground of conscious experience of some external phenomenon, such as
hardness or sharpness—a relation to the external phenomenon—but the re-
lational “conscious experience of the external” that one is having may not
require sensation. If sensations are generated, that is a further fact.
I’ve worked through a couple of possible motivations for Reid’s casual
statement that there is “no doubt, a sensation by which we perceive a body
to be hard or soft,” partly because it seems intrinsically interesting, but also
because Reid’s subsequent discussion is an attempt to pinpoint the exact role
of the tactual sensation in tactual perception. We have to be aware from the
outset that the problem may be ill-posed. Maybe tactual sensation is simply
a phenomenon independent of conscious tactual perception, though sharing
some underlying causal networks.
Whatever the reason, Reid brings to his discussion a conviction that
tactual sensation must be playing a role in all cases of tactual perception.
Suppose Sam bangs his head into a pillar; then, Reid says, he will not per-
ceive the characteristics of the stone, but only feel the violent pain in his head.
However,
It is quite otherwise when he leans his head gently against the pillar; for
then he will tell you that that he feels nothing in his head, but feels hard-
ness in the stone. Hath he not a sensation in this case as well as in the other?
Undoubtedly he hath: but it is a sensation which nature intended only as
a sign of something in the stone; and accordingly, he instantly fixes his
attention upon the thing signified; and cannot, without great difficulty, at-
tend so much to the sensation, as to be persuaded that there is any such
thing, distinct from the hardness it signifies. (V.ii.56)
This line of thought postulates sensations of which the subject is not

aware, pervading and making possible conscious tactual perception.
Reid elaborates at some length the doctrine of these “fugitive sensations,”
explaining our difficulty in recognizing them as a result of our focus on
the aspects of the external environment of which the sensations are signs.
Then the question is what exactly the role is of these sensations in tactual
perception. Just how do these sensations play a role in providing us with
perception of the external environment? Reid considers the idea that these
sensations provide us with knowledge of, for example, hardness or softness,
because of some intrinsic resemblance between the sensation and the ex-
ternal property. But there evidently is no such resemblance: “No man can
give a reason why the sensations of smell, or taste, or sound, might not have
indicated hardness, as well as that sensation, which by our constitution,
does indicate it” (V.ii.57). So we have a problem. It is only through tactual
sensation that we have any tactual perception of the environment. But if
it is not by means of intrinsic resemblance between the sensation and the
surroundings, how can the sensation be providing knowledge of the envi-
ronment? Here is Reid stating the problem:
Hardness of bodies is a thing which we conceive as distinctly, and believe as

firmly, as any thing in nature. We have no way of coming at this conception
and belief, but by means of a certain sensation of touch, to which hardness
hath not the least similitude; nor can we, by any rules of reasoning, infer
one from the other. The question is, How we come by this conception and
belief? (V.ii.57–58)
This is, of course, reminiscent of the kind of argument that we saw currently
occupies many philosophers studying vision. Just as Reid argues that there
are “fugitive sensations” of touch, Block, as we saw, argues that there are “fu-
gitive sensations” in vision, the “mental paints” whose detection requires the
analysis of experiments in vision science, and which ordinarily go unno-
ticed by the perceiver. The puzzle then is to discover exactly what role these
fugitives are playing in perception of one’s surroundings. Of course, the alter-
native possibility is that there are no such sensations; that tactual awareness
of one’s surroundings does not require tactual sensations, even though there
clearly are cases in which one has and can attend to one’s tactual sensations.
At any rate, Reid’s solution to his puzzle is this:
I see nothing left, but to conclude, that, by an original principle of our con-
stitution, a certain sensation of touch both suggests to the mind the concep-
tion of hardness, and creates the belief of it. (V.ii.58)
Or, as we’d say nowadays, the connection between tactual sensation and tac-
tual perception is innate, hardwired into us. So in tactual perception the tac-
tual sensation is always present, and we tactually perceive the external world
because the sensations cannot but causally generate concepts of these ex-
ternal properties, and belief that the objects have them. That is what goes on
in tactual perception, even though we are often so focused on the external
that we do not notice the postulated sensations or the causal work that they
are doing.
6.3. Reid’s Experimentum Crucis
Reid works through three different roles that tactual sensations might be
playing in our knowledge of the environment: (1) there might be empirical
correlations between our sensations and external phenomena, empirical
correlations we discover through scientific inquiry. Here there is no role for
sensation in providing us with our conceptions of the particular factors with
which the sensations are correlated, except perhaps as unknown hypothe-
sized causes of sensation. (2) The connection between a sensation and an ex-
ternal phenomenon is one that does not require scientific investigation by us,
but is known by us by “a natural principle”; of this kind is the relation between
the experience of seeing an angry expression on someone’s face, and knowing
that the person is angry. Here the knowledge of what anger is is not provided
by the visual experience; but given a subject who already does know, by some
other route, what anger is, the “natural principle” connects that visual expe-
rience with the person’s anger. Reid thinks this kind of connection is widely
and distinctively used in the arts; it’s these “natural principles” that allow us
to interpret works of art. And finally, (3) there are sensations that “though we
never before had any notion or conception of the things signified, do suggest
it, or conjure it up, as it were, by a natural kind of magic, and at once give us
a conception, and create a belief of it” (V.iii.60). The possibility that tactual
sensations do not play a significant role in grounding our tactual knowledge
of our surroundings is not considered by Reid.
He gives an argument, his experimentum crucis, that knowledge of exten-
sion, figure, and motion can’t be grounded in tactual sensation by means of
resemblance or by means of some rationally discernible connection between
the sensation and the external property. Therefore, the only possibility that is
left is that it’s an “original principle of our constitution”—a hardwired, innate
connection—that links the tactual sensation to the external property. The
experimentum crucis aims to consider someone who has only tactual sensa-
tion, but no tactual perception of external properties and relations. The ques-
tion is whether there are relations of resemblance, or some other rationally
penetrable link, between the sensations and external characteristics such as
hardness or extension. If there are, then it ought to be possible, in principle,
for the person who has only tactual sensations to exploit the resemblance,
or the rationally penetrable connections, between the sensations and the ex-
ternal phenomena, to arrive at the concepts of, for example, hardness or ex-
tension, and beliefs about the hardness or softness of bodies around them, or
about the spatial relations among the things around them.
Reid’s strategy is to give a sufficiently vivid picture of the situation of the
individual who has only tactual sensation that it’s compellingly evident that
the mere possession of those sensations, together with a general ability to
exploit relations of resemblance, or rationally penetrable connections, to ex-
ternal phenomena, will be of no use to the subject in deriving the concepts
of external characteristics or beliefs about the external world. But our tac-
tual sensations are the drivers of our tactual concepts of the external, and
beliefs about the external world. How then can our sensations be playing this
driving role? Only, the conclusion is, because there are “original principles
of our constitution” that mean that possession of those sensations will cause
one to have the concepts of the external, and various particular beliefs about
one’s surroundings. Let’s go.
Suppose a blind man, by some strange distemper, to have lost all the ex-
perience and habits and motions he had got by touch; not to have the least
conception of the existence, figure, dimensions or extension, either of his
own body, or of any other; but to have all his knowledge of external things
to acquire anew, by means of sensation, and the power of reason, which we
suppose to remain entire. . . .
Suppose him first to be pricked with a pin; this will, no doubt, give a
smart sensation: he feels pain, but what can he infer from it? Nothing surely
with regard to the existence or figure of a pin. . . . Common sense may lead
him to think that this pain has a cause; but whether this cause is body or
spirit, extended or unextended, figured or not figured, he cannot pos-
sibly, from any principles he is supposed to have, form the least conjecture.
Having had formerly no notion of body or of extension, the prick of a pin
can give him none. (V.vi.65)
Of course it’s possible, if one already has some concept of what a pin is, to
guess that this might have been the culprit when one feels a pinprick for the
first time; that would be like the case of recognizing the presence of anger
from someone’s expression, even if one hasn’t seen an expression of anger
before. But we are considering someone who doesn’t have any conception
of pins, or of extension or hardness, at all. And it is, as Reid says, hard to see
how such an individual could generate the concepts of extension and shape
merely from the sensation. Reid then elaborates the example. So far, we
have considered only the tactual sensation produced by a single pinprick.
It does, indeed, seem compelling that this will not have an “intrinsic re-
semblance” to any characteristic of physical objects that would be of a sort
that would allow one to form the concept of that characteristic of objects,
or beliefs about that characteristic. Of course, there is a notion of “resem-
blance” on which there are endless resemblances between any two things
(there are endlessly many sets of which both are members), but we are here
only talking about resemblances that could be used by us in concept for-
mation. And the mere having of a sharp pain would not of itself be enough
to allow one to form the concept of a sharp pin. Moreover, it is not as if the
exercise of reason would help. At best, it might prompt one to form hypoth-
eses about the cause of one’s sensation. Now the having of the pain might
play an essential role in one’s forming the description “the cause of this
particular sensation.” But it would not thereby provide one with knowledge
of the form “the cause of this particular sensation is F.” Reid wants to know
the role of the sensation in providing the concept of F-ness—that is, the
hardness and pointedness of the pin. It must play a role—Reid’s framing of
the problem guarantees that—but we cannot yet see how it is playing that
role, given that his subject cannot get the concept from the sensation by
exploiting either a “resemblance” of the sensation to the external property,
or the application of reason.
Perhaps the problem is that in considering a single pinprick, we are not

taking into account the variety of tactual sensations that one can have.
Suppose, Reid says, warming to his task, “a body not pointed but blunt, is
applied to his body with a force gradually increased until it bruises him.”
Even this ruthless experiment will evidently not allow the distempered blind
man to derive concepts of the external. Suppose that the body applied to
him touches a larger or a lesser part of his body; that still won’t provide a
conception of extension, if the subject doesn’t have it already. So far, we’ve
considered only passive touch, rather than active manipulation or grasping,
for example, and we’ve only considered touch as a sensitivity to pressure,
rather than, for example, sensitivity to vibration or to temperature. But Reid
thinks it’s evident that no such amplification of the resources available to the
distempered blind man will help:
Suppose, again, that a body is drawn along his hands or face, while they
are at rest. . . . Let us next suppose, that he makes some instinctive effort to
move his head or his hand, but that no motion follows, either on account of
external resistance, or of palsy. . . . Last of all, let us suppose, that he moves
a limb by instinct. . . . He has here a new sensation, which accompanies the
flexure of joints, and the swelling of muscles. But how this sensation can
convey into his mind the idea of space and motion, is still altogether myste-
rious and unintelligible. (V.vi.66)
No amount of variety in the tactual sensations will allow the subject to de-
rive concepts of the external. That is Reid’s main point. And since tactual
sensations, he thinks, do play a role in our forming concepts of the external,
the only way this can be happening is that the “original principles of our con-
stitution” mean that the sensation will brutely cause one to form concepts of
the external and to have knowledge whose content uses those concepts.
It is a striking picture, the blind subject peppered with various types of
tactual sensation, unable even to form the conception of the external en-
vironment. But the case of location has often given commentators pause.
Don’t tactual sensations involve location? Couldn’t we give our subject two
spatially separated pinpricks, for example, a few inches apart on the arm?
Wouldn’t they be experienced as spatially separated? According to van Cleve:
Reid’s reply would be that when we strip away all hardwiring (as we are
supposed to do in the experimentum crucis), we also strip away beliefs
about location. So the subject of our experiment would not feel the pricks
as separated. If they were qualitatively alike, perhaps he would feel them as
one; if they were qualitatively different, perhaps he would feel them as two
in the way that one discerns two notes within a chord as two. He would not
experience them as spatially separated. (2015, 43–44)
Now this is not obvious. The experimentum crucis did indeed have to strip
out beliefs about external location. But the locations of sensations seem, as
it were, to be internal to them. What we had to strip out were the hardwired
connections between sensations and concepts of the external world. That
does not, on the face of it, mean that the spatial relations among sensations,
and their bodily locations, have to be stripped out. In fact, if the spatial re-
lations were stripped out, we would be left with a collection of tactual
sensations quite unlike those we ordinarily experience. It would be no sur-
prise to be told that someone with tactual sensations quite unlike those we
usually have could not exploit resemblance or reason to generate concepts of
the external.
Or is the locatedness of bodily sensation essential to it? In the case of pain,
there is a familiar puzzle that goes like this: pains are mental states; they are
“in the mind” if anything is. Pain is often felt as being in one body part or
another; but how can the pain be “in your mind” if it’s in your finger? One
resolution of the puzzle is to say that the apparent locations of pains are
illusions; pains don’t really have the bodily locations they seem to. So perhaps
on this view we could make sense of a subject not subject to illusion, who
apprehends tactual sensations as not located.
That said, it’s not obvious that the locatedness of sensation makes a differ-
ence to Reid’s experimentum. Suppose we have an animal with a tough and
leathery exterior, which does not have and does not need perception of its
environment. Perhaps, for example, it lives deep in a trench in the ocean, and
encounters no predators and effortlessly absorbs nutrients from the local sea-
water. Perhaps, though, it has a complex internal organization, with a number
of internal organs and a constant transport of various complex chemicals be-
tween them, which it has to regulate. So let’s suppose it has an extensive range
of bodily sensations, and is very sensitive to the locations of these sensations,
which it uses in organizing the internal flow of chemicals. The idea of locations
“outside its body” might seem simply incoherent to this animal.
This animal stands to its own internal space somewhat as some thinkers
have suggested God stands to the space we occupy (as Newton said, “Space
is God’s sensorium”). There is no coherent concept for God to form of “the

space outside this space,” but though he does pervade this space, and knows
everything about the things located in it and their spatial relations to one an-
other, he himself is not an occupant of this space.
Or again, some medieval thinkers took the universe to be “closed”: bounded
by the sky, with there being no coherent speculation to be had as to what there
might be “on the other side of the sky.” Such a thinker might be confronted
with a puzzle: can’t the spatial relations that hold between ordinary physical
things be iterated indefinitely, past the sky itself, so that we can coherently
ask what, if anything, is to be found at locations defined by such iterations?
But the answer will be that pressing this kind of iteration is simply the way in
which philosophers often do generate utter nonsense. Suppose, for example,
we consider Wang’s classic argument that all numbers are small: 0 is a small
number; if n is a small number then n + 1 is a small number; so all numbers
are small. The conclusion is preposterous, because there are lots of numbers
that are quite big. And a natural reaction to the argument is that it presses
too hard the idea that the principle, “If n is a small number, then n + 1 is a
small number” can be iteratively applied to endlessly many cases. At a certain
point, you might suggest, the iteration ceases to be legitimate. Similarly, the
spatial relations among sublunary things cannot be iterated indefinitely, to
generate the concept of what’s on “the other side of the sky.” And similarly,
our blind man might have tactual sensations and know the spatial relations
among them, but resist as spurious the idea that these spatial relations can be
iterated indefinitely to generate a concept of external space.
On the face of it, the localization of a sensation typically amounts to no
more than its being placed in a particular part of the body. This localiza-
tion relates to what O’Shaughnessy called the “long term” body image. You
know about the structure of your body: what the parts of your body are,
how they are related to one another, and what the possibilities of move-
ment are. The long-term body image does not hold information about how
your limbs are currently configured, whether they are moving, and so on.
You may have a pain in your finger, and your finger may be moving, but the
pain is not thereby moving. The baby may grip your finger, enfolding it in
its fist, but the baby does not thereby have a pain in its fist. Our localization
of tactual sensation typically assigns them to body parts, as identified in the
long-term body image, but stops there. The body part may have a location
in some broader frame of reference, but we typically do not assign the tac-
tual sensations themselves locations in such broader frames of reference.
So even if we accept that there is a sense in which tactual sensation is often

localized, the nature of the localization does not allow it to be used as the
basis for a frame of reference that could specify the places of ordinary ex-
ternal objects and landmarks.
Of course, the body part in which a tactual sensation is located can itself,
at any moment, be placed in a broader frame of reference. But this broader
frame of reference may itself depend on one’s ordinary tactual and visual per-
ception of one’s surroundings. We can indeed say that the “pain was in the
finger,” and that “the finger was gripped in a vice,” and we could modify our
usual way of talking and go on to say that the “pain was in the vice,” but if we
do that we are using a frame of reference that depends on ordinary percep-
tion of the environment, not purely tactual sensation. And if we do not do
that, we are using a way of specifying the locations of sensations that relates
only to the long-term body image, and thus cannot be the sole basis for one’s
conception of spatial location and extension generally.
So even if we grant the essential spatiality of tactual sensations, the point
of Reid’s experimentum crucis still stands. If we begin with tactual sensations,
there may be no way in which concepts of the external can be derived by
exploiting resemblances or rational connections. Therefore, by Reid’s argu-
ment, the connection between sensations and concepts of the external can
only be a hardwired, brutely causal connection.
6.4. Sensations and Relations
The general problem Reid is addressing is to find the epistemic role of tactual
sensations in tactual perception. As we saw, he takes it for granted that tac-
tual sensations must have an epistemic role in our conceptualization of and
knowledge of the external environment. He isn’t explicit about why they must
have such a role. But I think that even today there are many philosophers who
would sympathize with his framework here. The datum is that conscious tac-
tual experience has a role to play in our conceptualization and knowledge of
the external. As Reid puts it,
I take it for granted, that the notion of hardness, and the belief of it, is first
got by means of that particular sensation, which, as far back as we can re-
member, does invariably suggest it; and that if we had never had such a
feeling, we should never have had any notion of hardness. (V.iii.61)
The datum is that conscious tactual experience has an epistemic role

in our understanding of and knowledge of the external; and the further,
apparently innocuous, step that Reid further takes is to interpret “con-
scious” tactual perception as tactual perception in which tactual sensations
are implicated. As we saw, this interpretation is one with which many
philosophers would sympathize today. If consciousness cannot be found
in the information processing postulated by cognitive science, perhaps it
is sensations that charge up the system to make the tactual perception con-
scious. But then, since it is specifically conscious tactual experience that
plays a role in providing for concept-formation and knowledge of the ex-
ternal, those sensations must be playing an epistemic role in providing for
that concept-formation and knowledge.
As I said earlier, the natural alternative to this view is to resist the step
of interpreting the consciousness of ordinary tactual perception in terms of
sensation. We should take our ordinary talk of conscious tactual perception
at face value, as describing a relation between the perceiver and the envi-
ronment. “Consciousness” is a characteristic of that relation. Think of it in
these terms, and the problem of finding the epistemic role of the sensation
disappears.
To see why we should drop the idea of analyzing tactual consciousness
in terms of sensations, and move to thinking of it as fundamentally a rela-
tion between the perceiver and the external phenomenon perceived, we have
only to look a little more closely at Reid’s response to his problem. He has to
explain (a) why tactual sensations are essential to forming concepts such as
hardness or extension (“If we had never had such a feeling, we should never
have had any notion of hardness”) and (b) how sensations can be playing any
epistemic role at all in the formation of concepts of the external. His answer,
as we saw, is that there is a brute causal connection between having the sen-
sation, on the one hand, and forming the concept of the external, and beliefs
formed using that concept, on the other. But notice (a) this proposal does not
explain why tactual sensation should be essential to the formation of these
concepts—why couldn’t there be some other stimulus which also, as a matter
of brute causal fact, brings about the formation of concepts of the external
and beliefs involving them? and (b) this proposal leaves the intrinsic char-
acter of tactual experience with no role in grounding the rightness of using
the concepts as one does; the intrinsic character of tactual experience, on this
proposal, is merely a matter of which sensations are involved, and it could be
any arbitrary collection of sensations which serves as the causal prod to use
of our ordinary concepts of the external. In contrast, if we think of tactual ex-

perience of the external as a relation between the perceiver and the environ-
ment, then the experience is partly constituted by the external phenomenon
of which one is forming a concept; the experience brings the external phe-
nomenon itself into one’s subjective life. This might reasonably be thought to
be an essential element in concept formation, and it explains why there is a
normative relation between the experience and the concept; the experience,
by bringing the external phenomenon itself into the subjective life of the per-
ceiver, in effect explains to perceivers what is going on when they are using
the concept, what it is that they are thinking or talking about.
Let me contrast this assessment of Reid’s argument with van Cleve’s. After
some preamble, van Cleve turns to what he suggests might be a really funda-
mental difficulty with Reid’s argument:
the doctrine of the transparency of sensations, as espoused by Harman

(1990) and others. Proponents of this doctrine typically make two
points: first, they say that when we attend to our sensations, we never no-
tice any of their intrinsic features; second, they say that what our sensations
reveal to us is not their own properties, but those of the objects they repre-
sent. Both of these points are anti-Reidian, the first because Reid thinks that
we can know our own sensations perfectly and the second because Reid
holds that sensations are nonrepresentational. If transparency is the expla-
nation of the difficulty of Reid’s thought experiment, it is a subversive one.
(2015, 53)
In fact, though, as we have seen, properly understood, the talk of tactual

sensations that we attend to and fully know is perfectly legitimate. There
really are tactual sensations; the feeling on your shoulder when someone
touches it, the disturbing sensation when an insect crawls into your ear, or in
Reid’s example, the pain you feel when you bang your head into a pillar. You
can and do attend to these sensations and notice their intrinsic features. It
is quite different with vision. There are some scattered phenomena that you
might call “visual sensations”—tired eyes after a day repairing jewelry, for
example—but in general, proponents of transparency are surely right about
vision. But transparency really is something special about vision.
Even though there are tactual sensations, however, it is a mistake to sup-
pose that they can be the ground of our understanding of concepts of exten-
sion and location. Even if we think of the sensations as located, the sense in
which they are located does not allow us to extend the concept of location to
places in the external environment. And we should not even give the tactual
sensations the minimal place Reid finds for them, as brute causal prods to
our formation of concepts of and beliefs about the external. Our conscious
tactual experience may be part of the ground of our understanding of the ex-
ternal without there being any role for sensation at all.
References
Block, N. 2010. Attention and mental paint. Philosophical Issues 20: 23–63.
Burge, T. 2010. Origins of Objectivity. Oxford: Oxford University Press.
Krueger, L. E. 1970. David Katz’s Der Aufbau der Tastwelt (The world of touch): A syn-
opsis. Perception and Psychophysics 7: 337–341.
Millikan, R. 1984. Language, Thought, and Other Biological Categories. Cambridge,
MA: MIT Press.
Millikan, R. 1996. On swarpkinds. Mind and Language 11(1): 103–117.
O’Shaughnessy, B. 1980. The Will: A Dual Aspect Theory. Cambridge: Cambridge
University Press.
Reid, T. 1764/1997. An Inquiry into the Human Mind on the Principles of Common Sense.
Ed. Derek R. Brookes. University Park: Pennsylvania University Press.
Van Cleve, J. 2015. Problems from Reid. New York: Oxford University Press.
7
“Unless I Put My Hand into His Side,
I Will Not Believe”
The Epistemic Privilege of Touch
Olivier Massin and Frédérique de Vignemont
Except I shall see in his hands the print of the nails, and put my finger
into the print of the nails, and thrust my hand into his side, I will not
believe. (John 20:25)
Thomas’s emphasis on touch echoes the long-standing idea that touch

enjoys some epistemic advantage over the other senses when it comes
to attesting to the reality of external objects. This tactile preference can
be found as early as in the atomists, as Plato depicted them, “who think
nothing is except what they can grasp firmly with their hands” (Theaetetus,
155e), and “maintain stoutly that alone exists which can be touched and
handled” (Sophist, 247c). The epistemic privilege of touch has been con-
stantly recurring in philosophical writings since then1 and appears to be
constitutive of our ordinary thinking as well, as reflected in our use of “tan-
gible” as a synonym for “real.”
In this chapter we first clarify how to interpret exactly the rather vague
idea of an epistemic advantage of touch over the other senses. We then de-
fend a “muscular thesis,” to the effect that only the experience of resistance
to our motor efforts, as it arises in effortful touch, presents us with the inde-
pendent existence of some causally empowered object. We finally consider
whether this muscular thesis applies to the perception of our own body.
1 See Massin 2010, 13–15, for references.
Olivier Massin and Frédérique de Vignemont, “Unless I Put My Hand into His Side, I Will Not Believe”
In: The Epistemology of Non-Visual Perception. Edited by: Berit Brogaard and Dimitria Electra Gatzia,
Oxford University Press (2020). © Oxford University Press.
DOI: 10.1093/oso/9780190648916.003.0008
7.1. What Privilege?
7.1.1. A More Reliable Sense?
A first way to interpret the epistemic privilege of touch is in terms of its re-
liability. This is the view adopted by Descartes, among others (see also Kant
2006, §17; Schopenhauer 1958, bk. 2, chap. 3): “Of all our senses, touch is
the one considered least deceptive and the most secure” (Descartes 1998, 5).
Why should it be so? A common hypothesis is that unlike auditory or visual
objects, tactual objects are in contact with our body, and thus at least one kind
of misperception is excluded: those that arise from perturbations in the per-
ceiving medium:
There are fewer ways of going wrong about what we touch than there are
about what we see. Our eyes may deceive us, and our eyes may be deceived.
Mirrors, sleight of hand, queer conditions of light or atmosphere, mirages
and visions, optical illusions, even ordinary perspective—in all sorts of
cases, for various reasons and in various ways, we may be led into mistakes.
It is much less easy, though not of course impossible, to play tricks on the
sense of touch. After all we are always very close to what we touch, and we
are not dependent upon a variable intervening “medium.” (Warnock 1953,
54; see also Heider 1959, 19–20)
However, one may question whether touch is a contact sense (Fulkerson

2014), and correlatively one may consider the body as a tactile medium
(Massin and Monnoyer 2003; Vignemont and Massin 2015). The spatial con-
tent of tactile experiences are indeed structured by bodily awareness: one
feels pressure on one’s right hand thanks to the awareness of the structural
organization of one’s body (O’Shaughnessy 1980; Martin 1993). The fact is
that one can play many tricks on bodily awareness, and thus on the sense
of touch. For instance, if you are touched on two spots on your finger after
observing the image of a smaller version of your hand, you experience the
distance between the two tactile stimuli as relatively smaller (Taylor-Clarke
et al. 2004). This is just an example of the many ways that the complexity of
bodily awareness can pave the way for a variety of tactile illusions. We do
not believe that such criticisms are fatal to the view of a higher reliability
of touch, but we think that they have together sufficient weight to shift the
burden of proof to the upholder of this approach.
7.1.2. A More Materialist Sense?
A second way of interpreting the hypothesis of an epistemic advantage of

touch is to argue that only touch gives us direct access to material objects
or properties and to assume that material objects or properties are in some
sense prior or more fundamental that nonmaterial ones: “Touch is the sense
that gives us ‘access to reality,’ because even in immediate tactual perception
the object of perception is something material” (Armstrong 1962, 31).2 We
agree that touch, more than the other senses, gives access to material objects
or properties (see Vignemont and Massin 2015 for a proposal along these
lines). Bracketing complex issues about fundamentality, we also agree that
material objects or properties are in some sense more fundamental that
nonmaterial ones (such as phenomenal colors or sounds). One question,
however, is why material entities are ordinarily conceived of as being more
fundamental than nonmaterial ones. Perhaps the reason of our naive materi-
alism is that material entities are accessed through touch, which enjoys some
epistemic advantage over other senses. If so, rather than explaining the epi-
stemic privilege of touch by appealing to ordinary preference for matter, one
may want to explain our ordinary preference for matter by appealing to our
ordinary preference for touch. Bennett makes a close suggestion:
If one could explain the differences between primary and secondary qual-
ities by adducing facts about their respective sensory bases or correlates,
I suspect that the crux of the explanation would turn out to be the fact that
the sense of touch—or rather of touchandmovement—is involved in all the
primary qualities in a way in which it isn’t with any of the secondary. But
that is only a suspicion. Someone should write a book on the epistemology
of the sense of touch. (Bennett 1971, 102)
Although we cannot deliver a book in defense of Bennett’s suspicion, we want

at least to make room for it. Accordingly, the view that the epistemic priority
of touch explains common-sense materialism (or the distinction between
primary and secondary qualities) should not be precluded by one’s interpre-
tation of the epistemic privilege of touch. But starting from the assumption
2 See also Sanford 1967, 333; Strawson (Hampshire and Strawson 1961, 107); Perkins 1983,
250–251—who, however, rejects the view. This may have been Aristotle’s view as well, as argued by
Freeland 1992.
that material properties are prior to nonmaterial ones does seem to preclude
such an explanation. We thus turn to another account of the tactual privilege.
7.1.3. A More Realist Sense?
On our view, the epistemic privilege of touch is neither due to its higher re-
liability, nor to the materiality of its objects, but to the fact that only tactile
perception presents us—under certain conditions—with its objects as being
real, in a sense to be specified later.
Let us readily dismiss a possible misunderstanding. The proposal is
not that only what appears in tactile experiences is real. One thing is to
ask whether the objects of perceptual experience are real; another is to
wonder whether their reality is experienced. Call these questions respec-
tively the question of the reality of appearances and the question of the
appearance of reality:
1. Reality of appearances: are appearances real? I.e., do the objects of our

perceptions exist independently from us?
2. Appearances of reality: do appearances seem real? I.e., are the objects of
our perceptions presented to us as existing independently from us?
The question of the reality of appearances has historically drawn most of the
philosophical attention. It is at stake in the debate between direct realists,
indirect realists, and phenomenalists about perception. The question of the
phenomenal character of reality has received comparatively less attention
(see, however, Siegel 2006; Campbell and Cassam 2014, chap. 3), maybe be-
cause its answer seems obvious to many, for whom it is a basic phenome-
nological fact that the perceptual world appears to be independent from us.
Whatever the reason of the relative neglect of the question of the appearances
of reality, the question matters for the epistemology of realism. For if some
perceptual objects indeed seem real, then there is an internalist answer to the
question “On what grounds do you believe that what you perceive is real?”
(whereby “internalism” we mean the view that such beliefs have justifiers
of which one can become aware). The answer is the same as to the ques-
tion “On what grounds do you believe that what you perceive is round/red/
hard . . . ,” namely: “Because it seems so.” All these cases are instances of the
schema “S believes that x is F for the reason that x seems F to S.” Such an
internalism is a version of phenomenal conservatism (Huemer 2006, 2007),

according to which we are justified in believing that p if it seems to us that
p (and if no defeaters appear to us). In the present case, we are justified in
believing that x is real if it seems to us that x is real. As an internalist pro-
posal, it may not readily address skeptical objections of the evil demon sort
(see, however, Huemer 2007). But it may at least help us to refute other kind
of objections against realism, such as the challenge—associated with “qui-
etism”—according to which the notion of reality is not meaningful. If some
things appear to be real, then one may answer to this challenge by arguing
that predicate in “to be real” is as meaningful as the predicate “to be red,” in-
sofar as the corresponding properties are accessible through experience. Our
aim now is to specify in what sense touch can give rise to a specific appear-
ance of reality.
7.2. The Muscular Thesis
Let us start with a description of a specific type of touch that has played a
central role in the discussion of the epistemic advantage of touch, namely,
effortful touch. All the examples provided in support of an epistemic priv-
ilege of touch involve some voluntary effort: Plato’s atomists believe only in
what they “can grasp firmly”; Thomas “thrusts” his hand into Jesus’s side; and
Samuel Johnson kicks a stone with “mighty force” so as to refute Berkeley. As
it happens, the feeling of effort has often been ascribed an epistemic privilege
akin to the one ascribed to touch: “There is no commoner remark than this,
that resistance to our muscular effort is the only sense which makes us aware
of a reality independent from ourselves,” writes James (1880, 29). The pro-
posal has indeed a very rich history.3 France appears to be the country where
effortful touch has been granted most privileges. Malebranche (1687/1991,
40–43) already argues that if resistance to our physical effort gives us reason
to believe in the reality of solid bodies, then a fortiori, resistance to our will
should give us reason to believe in the reality of ideas.4 Condillac (1754/1997,
3 See Massin 2017 for detailed references.

4 Arguments of the sort, we think, pass too quickly from the premise that x cannot be modified
at will to the conclusion that x resists to our will. Another possibility is that x is a kind of thing that
we cannot want or try to modify: with respect to idea, and other phenomena that are beyond our
agentive powers, we might be in a state of abulia, rather than in a state of being impeded (see Massin
2010, 652). Note that Descartes (1983, pt. 2, sec. 1) had already pointed that we experience the fact
that what we perceive comes from something else than ourselves, since cannot modify what we per-
ceive at will.
pt. 2, chap. 5) argues that it is only when equipped with effortful touch and
the capacity of motion that his statue becomes aware of a world distinct from
itself. But it arguably with Destutt de Tracy (1801) and Maine de Biran (1812/
2001) who gave the first careful examinations of the idea that only effortful
touch presents us with the mind-independency of the external world.
That touch and efforts are granted similar epistemic privilege is no accident,
we take it. There are not two different kinds of experiences—the feeling of ef-
fort on the one hand, and tactile experiences on the other hand—which present
us independently with the reality of external objects. Instead, there is one com-
plex form of tactile perception, effortful touch, that provides us with an exclusive
entry point to the reality of external objects.
7.2.1. Effortful Touch
What is effortful touch? It involves four ingredients: cutaneous pressure sensi-

tivity, muscular sensitivity, effortfulness, and the sensation of effortfulness.
(1) Cutaneous pressure-sensitivity.
By touch in the strict sense we mean here the sense of pressure (Vignemont
and Massin 2015). Pressures and tensions are states of material entities (solids,
liquids, gazes) that arise as a result of being acted on by several opposite forces.
Exerting two opposite inward forces on both ends of a stick, as in pushing,
amounts to putting the stick under pressure. Exerting two opposite outward
forces on these ends, as in pulling, amounts to putting the stick under tension.
(2) Muscular sensitivity.
Tactile sensations are felt to be located in the parts of the body that are in con-
tact with the objects, such as the feet and the hands. By contrast, the muscle
sense gives rise to the awareness of tension in our muscles and in some
larger parts of our body. The muscular sense, which is only a part of what
Sherrington (1906) calls proprioception, involves muscle spindles, which are
sensitive to muscle stretch and Golgi tendon organs.5
5 Joints receptors are involved in another species of proprioception, and are not part of the muscle
sense. The reason why we exclude these receptors from the muscle sense is that they are responsive
Body Body
Agent
exert
Body
F1 F2
Figure 7.1 Effort: a force-based account
(3) Effortfulness.
Cutaneous pressures and muscular tensions can occur in absence of any in-
tentional action of the subject. For touch to be effortful, the muscular tensions
and continuous pressures must result from some activity of the subject. In
short this means that one of the forces constitutive of pressures or tensions
must be intentionally exerted by the subject. More precisely, making a motor
effort, we submit, amounts to
a. exerting a force (F1) on an object in order to make it move (or stay

at rest)
b. this intentionally exerted force being at least partly counterbalanced by
an external opposite resistive force (F2) (figure 7.1).
Opposite forces act on a same body; they have opposite direction and they
may have different magnitudes. To say that they “counterbalance” each other
is to say that each prevents (part of) the other from causing the acceleration
of the body that it would have caused, had it acted on it alone. For instance,
to kinematic (and trigonometric) properties such as positions or motions, by contrast to muscle

spindles and tendon organs, which are sensitive to dynamic properties such as forces or tensions
(Massin 2010, 468).
to make an effort to lift a bag is to exert some upward muscular force on it,
which is at least partly counteracted by the weight of the bag—a downward
gravitational force. That effort essentially involves some external resistive
force allows one to understand the connection between effort and resist-
ance: necessarily, in virtue of the nature of effort, if a subject makes an effort
on some object, then that object opposes some resistance to the subject’s ef-
fort. It is beyond the scope of this chapter to defend such a force-based ac-
count of efforts (see however Massin 2010, 2017).6
What unifies these three features (pressure sensitivity, muscular sensi-
tivity, effort) in a single perceptual system is that at the heart of each of them
stands a relation of opposition between forces: cutaneous pressures consist
in antagonist inward forces being exerted on the dermis or epidermis; mus-
cular tensions consist in antagonist outward forces exerted on muscles; effort
consists in exerting a force against an opposite one.
(4) The feeling of effortfulness.
Effortful touch, as we use the term, does not only involve the exertion of
a force against some resistive force, but also the feeling that one’s inten-
tionally exerted force is resisted. Cutaneous and muscular sensitivity to
pressures and tensions by itself only presents us with pressures in a passive
way: the pressure between one’s back and the backrest of one’s chair, and
the pressure we feel when a cat jumps on our knee, are not presented as
constituted by a force we intentionally exert and another force that resists
it. In such passively felt pressures, the opposite forces felt are on an equal
footing: none of them is intentional or resistive. In feeling efforts, by con-
trast, we experience one of the forces constitutive of the felt pressure as one
force that we exert, and we experience the other force as a force that we do
not exert and that resists the force we exert. An effort consists in one inten-
tionally exerting forces counteracted by a resistive force. Correspondingly,
on top of the feeling of pressure or tension the feeling of effort involves
agentive feelings, namely, the feeling of intentionally exerting some force
that meets opposition.
6 Three main aalternative accounts of efforts are (1) efforts are primitive feelings; (2) efforts consist
in, or arise from, a comparison between efference copy of the motor command, and the subsequent
afferent signals received from the muscles; (3) efforts consist in expending some energy or limited
resource in order to reach one’s goal.
We can now better understand the epistemic privilege of effortful touch. The
thesis we propose, which we call the muscular thesis, is that effortful touch, and
only it, provides us with an experiential access to the reality of external objects
in the following sense:
Muscular thesis: Only the experience of resistance to our motor efforts, as it

arises in effortful touch, presents us with the independent existence of some
causally empowered object.7
Accordingly, when we say that effortful touch is the only sense that pre-
sent us its objects as being real, our sense of being real encompasses two main
components: mind-independency and causal efficiency. Such a sense of being
real is not pulled out of the hat but combines two common views. A first in-
fluential understanding of being real is indeed to equate it with some relational
property: that of being subject-or mind-independent. Something is real, on this
understanding, if and only if it exists independently of us, by which is generally
meant independently of mental episodes directed towards it. Another common
way of understanding being real, tracing back to Plato’s Sophist, is to equate
it with having causal power (see, e.g., Armstrong 1997, 41–43; Berto 2012).
Something is real if and only if it can act, or have an effect, on other entities. The
sense of “real” that we adopt is a combination of both proposals: to be real is to
have both mind-independent and efficient existence. Key to the reconciliation
of these two understanding of “being real” is the idea that mind-independency,
instead of being understood in terms of independency from perception, should
be understood in terms of independency from the will—where willing is under-
stood as a kind of striving.8
The privilege of effortful touch is thus the following: only effortful touch
presents us with the contrast between ourselves as striving agents and an
independent causally empowered being that resists our effort. We identify
with the powerful being that exerts the first force—by contrast to the external
being that exerts the resistive force. Thus, while passive touch is mute with
respect to the self-world distinction, effortful touch, which brings agency in,
7 The contrapositive thesis has also been considered in some details: in the same way that effortful-
ness has been claimed to be the chief source of our knowledge of the distinction between ourselves
and the world, effortlessness has been advanced as explaining cases in which the subject-object dis-
tinction vanishes.
8 On the idea that willing is a kind of trying or striving, see, e.g., Anscombe 1957, §36;
O’Shaughnessy 1980, 100; Lowe 1996, 157ff.; 2000, 246ff.; McCann 1998, 89; Massin 2014.
discloses that distinction, and thereby, the reality of its objects.9 Only a cer-
tain kind of touch presents us with there being something that has causal
power independent from us. That species of touch does not only present
us with there being something beyond us; neither does it only present with
there being something causally empowered. It presents us with both, at once.
A tour de force.
7.2.2. Feeling of Reality in Others Senses
The immediate corollary of our view is that in sight or hearing we are not
presented with the reality of visual or auditory objects. This may seem obvi-
ously false: “all (or almost all) serious theories of perception agree that our
perceptual experience seems as if it were an awareness of a mind-independent
world” (Crane 2005; see also Strawson 1979, 97; Cassam and Campbell 2014,
chap. 3; Allais 2015, 53). Let us focus here on sight: is it the case that visual
objects do not appear to be real in our sense? A first possible answer on be-
half of the muscular thesis is to bluntly reject Crane’s description of our visual
experiences and maintain that vision does not present us its object as existing
independently from us. Such a view was indeed standard among modern
philosophers, who regarded the experience of mind-independency as clearly
impossible: “As to the independency of our perceptions on ourselves, this
can never be an object of the senses” (Hume, Treatise, 1.4.2; see also Berkeley,
Principles, §18; Dialogues, 201; Reid, Inquiry, 687).
Indeed, claiming that visual objects do not appear to exist independently
of our perception of them does not necessarily commit one to an absurd ac-
count of the phenomenology of experience.
This is so first because denying that existential mind-independency is
presented in sight is compatible with other forms of subject-independency
being visually presented. Siegel (2006) usefully distinguishes the subject-
independence of the existence of a thing from the subject-independence of
the properties of that thing. For instance, changing one’s perspective relative
to a tree does not change the perceived location of that tree. The subject-
independence that is visually available here, however, does not concern the
existence of the tree itself, but only its location. Second, not experiencing
9 See, e.g., Scheler: “Reality is not given to us in perceptual acts, but in our instinctive and conative
conduct vis-à-vis the world” (1927/1973, 318).

visual objects as mind-independent does not entail experiencing them as

mind-dependent. The present proposal merely assumes that visual phe-
nomenology is mute with respect to the mind-(in)dependency of its objects.
Finally, the muscular thesis is compatible with the claim that the background
beliefs of every perception may include a belief in the present existence of its
objects, as Reid claimed. Reid endorses epistemic foundationalism with re-
spect to such a belief:
The sceptic asks me, Why do you believe the existence of the external ob-
ject which you perceive? This belief, sir, is none of my manufacture; it came
from the mint of Nature; it bears her image and superscription; and, if it is
not right, the fault is not mine: I even took it upon trust, and without suspi-
cion. (Reid, Inquiry, chap. 6, sec. 20)
The upholder of the muscular thesis may thus grant that in vision the ex-
istential belief arises irresistibly, although without being justified by the
visual experience. By contrast, in the case of effortful touch there is another
answer to the skeptic’s question: “Why do you believe the existence of the
external object which you perceive?” Namely, because it seems to exist
apart from ourselves. Thus, one plausible proposal, we surmise, is this: all
sensory perceptions come with an instinctive belief in the reality of their
object. In all ordinary perceptual experiences, that belief in the reality of per-
ceived objects is not justified by the content of perception—call this Reidian
foundationalism. By contrast, effortful touch provides us, on top of this, with
some experiential justification for this belief—in accordance with phenom-
enal conservatism.
Although we confess being tempted by this bold line of answer—denying
that mind-independency is present in visual experiences—a second more
cautious line of answer is available to us. Recall that our sense of “real” is
quite demanding: mind-independency and causal efficacy. Hence our pro-
posal is compatible with sight presenting its object as mind-independent,
as long as it does not also present its objects as causally empowered. This
will still sound implausible to those who think that sight presents us with
causal relations. But they may be convinced by the following comment: in
effortful touch, we become aware not just of two causally empowered objects
interacting in front of us, but of an interaction between ourselves and another
external causally empowered object. This causal contrast corresponds to the
spatial contrast between the perception of the distance between two objects
and the perception of the visual depth between some objects and oneself.
Thus we need only to deny that vision presents us with “causal depth,” that is,
with our causal encounter with an external object.
7.3. The Body as One Object among Others?
We argued in the previous section that it is in virtue of the feeling of resistance

that tactile experiences provide the appearance of external reality. But do
they also provide the appearance of bodily reality? In other words, does touch
also have an epistemic advantage when it comes to one’s own body? One of
the peculiarities of touch is its duality. So far we have focused on its exter-
oceptive dimension, but tactile experiences are also bodily experiences: we
feel the pressure of the ball on our hand. How is the hand presented to us in
our tactile sensations?10
7.3.1. The Resisting Body
Let us consider the following bodily version of the muscular thesis.
Muscular bodily thesis: The experience of resistance to our motor efforts

presents us with the independent existence of our causally empowered
body.11
At first sight, this bodily thesis seems counterintuitive: efforts do not appear
to break the normal unity of the body and the self, as famously described by
Descartes in Meditation VI:
I am not only lodged in my body as a pilot in a vessel, but . . . I am besides so

intimately conjoined, and as it were intermixed with it, that my mind and
body compose a certain unity.
10 To be clear, the question here is not about self-touch. To some extent, touching one’s body is like
touching a book or a chair. One might then claim that in self-touch I can become aware of the exist-
ence of my body in the same way as I am aware of the existence of other objects. What we are inter-
ested in here is whether the appearance of bodily reality can be provided by the bodily component of
tactile sensations.
11 What is at stake here is not whether you experience this body as being your own (sense of bodily
ownership), but whether you experience it as being you.

But should this normal unity between the subject and her body be conceived
of as a counterexample to the bodily thesis? Let us analyze in detail four dis-
tinct cases, in which our body seems to resist us.
The Heavy Box: Your friend asks you to help him move out from his house.
You must carry boxes full of books. One of them, however, is too heavy for
you. You manage to lift it a little bit but not for long and it falls down on the
floor. You are simply not strong enough.
There is no doubt that one experiences effort while trying to lift a heavy box.
There is also little doubt that one becomes vividly aware of one’s body and
of its incapacity to lift the box. Yet in such a case, instead of describing one’s
body as something that is causally empowered independently of the subject,
one takes one’s bodily incapacity to be one’s own incapacity: “I cannot lift the
box” is the natural description of the situation. If anything, then, one might
say, pace the muscular thesis, that such a situation reinforces the subject’s
identification with her resisting body, rather than presenting it as distinct
from her. However, it is actually unclear that the subject experiences the re-
sistance of her body. Instead, what she feels is the incapacity of her body to
lift the box. Feeling of resistance and feeling of incapacity are two distinct,
although tightly interrelated, experiences. In particular, the feeling of re-
sistance is largely sensory and instantaneous, whereas the feeling of bodily
incapacity involves monitoring one’s performance over time, and can thus
be qualified as being metacognitive along other noetic feelings (Vignemont,
forthcoming). One can become aware of what one can and cannot do by
exploiting two cues: the ratio of success to failure and the ease or difficulty
associated with performing the movement. While the latter consists in one’s
feeling of effort, the former does not; the intensity of an effort is not essen-
tially tied to its chance of success or failure: intense efforts can be vain, and
slight efforts can be pointless. The feeling of incapacity is thus clearly dis-
tinct from the feeling of resistance insofar as it involves a more enduring
awareness of agency (e.g., a repeated experience of failure). Furthermore, the
sensations of resistance and of effort normally lead to success, whereas one
can repeatedly fail independently of any sensation of resistance. Suppose, for
instance, that you try to move your arm by uttering “Move!” in your mind.
Such an attempt to move your arm will fail, and repeatedly so, but it is unclear
that it involves significant effort on your part. One may reply that repeating
such a mental order requires some effort. But the resistance involved in such
an effort is not that of your arm, but rather of your mind. The next case will
help making that distinction clearer.
The Trek: Your friend invites you for a relatively challenging trek in the
Swiss Alps. Although you rarely do any kind of sport, you decide to go, but
you soon realize that it is more difficult than expected. At the end of each
day, you feel that you will never be able to make one more step, and still you
do it. Every morning when you wake up, you feel that you will never even be
able to get up, and still you do it. Until the last day, you struggle.
Again, the subject’s body is not presented to her as being causally empowered
independently of herself. Yet there is little doubt that she makes efforts.
However, we need to distinguish between two kinds of effort. First, there is the
physical effort to raise one’s legs. But there is also the mental effort to keep on
walking instead of giving up. One should not reduce the sensation of making
an effort on the body to the sensation of trying to move the body with effort—the
latter not being sufficient for the former. For instance, Delboeuf (1881) argues
that mountaineering efforts and physical efforts usually require making
second-order mental efforts: efforts to overcome aversion to first-order phys-
ical efforts: the mountaineer struggles against the nagging temptation to stop
walking. On his view, the mountaineer’s effort is more mental than physical.
Mental efforts are ubiquitous. One may, for instance, talk of effortful
hearing—for example, listening. But if effort could be combined with any
modality, would touch then lose all epistemic privileges? This is not so.
Despite similar surface grammar, effortful hearing and effortful touch cap-
ture distinct categories of phenomena. Effort, we have here assumed quite
standardly, is a goal-directed phenomenon: we make effort in order to achieve
some goal. Now the goal of the effort involved in effortful hearing is typically
to focus one’s attention on and to understand what is being heard: these are
mental episodes. By contrast, the goal of the effort involved in effortful touch
is to move things around: that is, a physical, kinematic episode. If anything,
mental effort presents us with the resistance of our mental inclinations—of
our distracted attention—not of physical bodies. Whether mental efforts pre-
sent us thereby with the reality of our psychological tendencies is a question
we leave here open. We are only concerned with the muscular thesis, which is
literally muscular, and it is not challenged by the Trek example insofar as the
mountaineer’s effort is mental.
Still even Delboeuf grants that the mountaineer also experiences a first-
order physical effort, such as struggling against the weight of her body and
making herself step over a high rock. This muscular effort requires the
mountaineer to exert an intense physical effort with her limbs, and she may
use her hands to get over the rock, for instance. And yet it is unlikely that
the feeling of resistance it prompts gives to the mountaineer the impression
that her body is distinct from her. Why not? Maine de Biran suggested that
the body displays only a relative resistance because it obeys one’s voluntary
efforts, whereas the rest of the world can display an absolute resistance that
can be invincible. But this is hardly convincing. It is true that the external
world can oppose a resistance so important that no effort can overcome it: no
matter how hard we push the mountain with our hands, it will not move.
However, the resistance of the external world does not need to be invincible
to yield the impression of its reality: the swimmer who feels the resistance of
the water to the motion of her hands has a clear sense of the mind-indepen-
dence of the causal power of the water. On the other hand, the resistance of
our own body might be invincible: no matter how much we try, we cannot fly.
Yet we do not experience the resistance of our body in the way we experience
the resistance of the mountain.
A this point the upholder of the bodily thesis might simply bite the bullet and
insist that the tired mountaineer’s body really seems to her to exist apart from
her. One should, however, first consider two other options open to the bodily
thesis, which allow explaining the contrast between the resistance of our body
and the resistance of objects external to our body.
The bodily thesis concerns effortful touch, which, we submitted, includes
cutaneous pressure sensitivity. One may then suggest that what is missing in
the Trek example is this latter component. It is only if one has both a mus-
cular and a cutaneous feeling of resistance that one can have the impression
that there is a mind-independent object at the origin of the force exerted
against one’s body. When hiking, one can feel the contact of the ground on
one’s feet, and when lifting a box, one feels its pressure on one’s hands. But the
sense of bodily resistance is not part of the passive tactile phenomenology;
it is part of the kinesthetic, effortful phenomenology. One may then suggest
that the muscular thesis is equally a tactual thesis: it involves the sense of re-
sistance, but it needs to be experienced through touch for it to provide the
appearance of reality. This would explain why one does not experience the
body in the same way one experiences the mountain.12
12 This account, however, is not satisfying if one assumes that proprioception and touch, which
both consist in perceiving forces, are of a same kind (Massin and Monnoyer 2003).
A second strategy consists in looking for defeaters to the mind-

independent causal power of the body. There may be nothing missing in the
case of bodily resistance, but there may be something extra that prevents us
from experiencing our body as being distinct from us. The fact is that in the
cases described earlier, the sense of resistance is associated with pain: each of
your muscles feels painful while hiking, and you feel that you are breaking
your back while trying to carry the box. Unless in very extreme circumstances,
pain does not separate the self from the body but instead anchors it there: one
does not feel distinct from one’s suffering body (Vignemont 2017). If this
analysis is correct, then one should experience more the feeling of the inde-
pendence of the body when there is bodily resistance with no pain. Consider
the two following examples:
The Anarchic Hand: After a brain lesion you suffer from the anarchic hand
syndrome. You feel that you have no control over one of your hands,
which is often undoing what your other hand is doing or displaying
aggressive behavior toward you (including trying to strangle you). You
try to stop your anarchic hand, but it resists.
The Safety Strap: You wake up and you see that you are in hospital. You
feel that your legs are restrained by safety straps. You try to move them
a bit, but you completely fail. You try again, but the pressure exerted by
the straps does not increase and your legs remain still. You then realize
that your legs are paralyzed.
In both cases, patients can experience their body as being distinct from
them. In the anarchic hand syndrome, patients do not deny that this is their
own hand, but they still deny that the movements performed by the hand are
their own. The anarchic hand seems to have a will of its own (Pacherie 2007).
In other words, its causal power is experienced as being mind-independent.
Paralyzed patients can also treat their legs as if they were alien or foreign to
them, as if they were mere objects (Scandola et al. 2017). This seems to be in
line with the bodily hypothesis. The problem, however, is that it is not clear
whether what they experience is due specifically to their sense of bodily re-
sistance. Indeed the paralyzed patients may try to move their limbs at the
beginning, but soon give up because they have updated what they can and
cannot do and do not attempt to perform actions that are impossible. They
then have a feeling of bodily incapacity, but as said earlier, this is different
from a feeling of resistance. The question is: do they experience a feeling of
alienation toward their body exclusively at the beginning when they still try
to move their limbs?
The fundamental problem that this question raises is that the sense of
bodily resistance is at most occasional and there are good reasons for the
normal lack of bodily resistance: actions are planned on the basis of the
knowledge of one’s bodily capacities, and to guarantee the success of one’s
actions, one generally plans only movements that are physically possible:
“The self-awareness of a self-consciously competent bodily agent includes
a familiarity with the possibilities for bodily acting that come with having the
kind of body she has: for instance, a familiarity with the different movements
that are feasible at different joints” (McDowell 2011, 142).
It is thanks to the awareness of bodily possibilities that one does not
over-or under-reach when trying to get an object. It is also thanks to it that
one does not attempt to move in biologically impossible or painful ways.
In short, the body does not resist because it does not have to: one usually
asks only what it is possible for it to do. Hence, the sense of bodily resist-
ance is the opposite of the way one normally experiences the control that
one has over one’s body. One does not have to force one’s body to do things.
Instead, one’s body typically obeys one’s will. The difference between con-
trolling one’s body versus controlling another object is the transparency of
the bodily medium and the fluidity of control. Therefore, experiencing the
mind-independency of the causal power of one’s body is at most occasional
and anecdotal, revealing the breaking down of the unity between the body
and the self.
7.3.2. Derealization and Depersonalization:

Feeling of Reality or of Presence?
To finish, we would like to consider a last potential objection to the account

we have presented so far by analyzing in detail the two related psychiatric
conditions of depersonalization and derealization. Depersonalized patient
report that they feel detached from the world and from their body, as if they
were external observers of their mental and bodily processes or as if their
body had disappeared:
I do not feel I have a body. When I look down I see my legs and body but it
feels as if it was not there. When I move I see the movements as I move, but
I am not there with the movements. I am walking up the stairs, I see my legs
and hear footsteps and feel the muscles but it feels as if I have no body; I am
not there. (Dugas and Moutier 1911, 28; translated by Billon 2017, 194)
Interestingly, these patients often feel the urge to touch their body or to pour
hot water on it to reassure themselves of its existence.
Even if I touch my face I feel or sense something but my face is not there. As
I sense it I have the need to make sure and I rub, touch, and hurt myself to
feel something. (Sierra 2009, 29)
Self-touch is thus used for them as a mean to try to re-establish their feeling
of bodily objectivity. In the same way, we might pinch ourselves to make sure
that we are not dreaming.
Their behavior may at first sight appear as in line with the epistemic advan-
tage of touch that we have defended so far. However, it is not clear how the
syndrome of depersonalization fits with the muscular thesis. The difficulty
is twofold. On the one hand, depersonalized patients experience the lack of
reality of the external world, and yet they can still experience the resistance of
objects on their skin, and their sense of effort is not disturbed (Billon 2017).
On the other hand, they no longer experience the reality of their body and
feel alienated from it, and yet they have no motor deficit and their body does
not appear to resist them more than healthy individuals.
These objections, however, are not fatal. They simply invite us to clarify
further the view that we defend. One way to answer is to suggest that these
patients suffer from a more fundamental disruption, namely, a disruption
of the sense of the self. Let us return to our origin proposal. We argued that
self-world dualism finds its experiential origin in the experience of effortful
touch. The main issue we were concerned with was how one was aware of ex-
ternal objects as being independent from the subject. In other words, we were
interested in how to differentiate the world from the self. But this question
makes sense only if one has already a sense of self. However, in deperson-
alization this may not be the case. Billon (2017) argues that depersonalized
patients have a fundamental distortion of subjectivity. They may feel resist-
ance but they no longer experience their sensations as being their own. For
effortful touch to provide the appearance of reality, one needs some level of
self-awareness. In short, if there is no sense of the self, then there cannot be
self-world dualism. On this view, these patients lack not only the feeling of
presence, but also the appearance of reality normally provided by effortful
touch despite their preserved tactile and motor abilities. The case of deper-
sonalization thus does not invalidate the muscular thesis; it merely shows
that effortful touch does not suffice for the appearance of reality: one also
needs subjectivity.
An alternative reply is that what these patients have lost is not what we
target with our muscular thesis. As we said earlier in our discussion of vi-
sion, we propose a highly specific definition of the sense of reality, according
to which external objects are presented to us as having mind-independent
causal efficacy. In other words, in effortful touch one experiences that there
is something that exerts the force that opposes us. One might then argue that
patients who suffer from depersonalization/derealization do not lack this
specific sense of reality. What they lack is something different. One might,
for instance, suggest that they lack the feeling of presence only (Vignemont,
forthcoming).
The notion of feeling of presence has originally been proposed to charac-
terize the distinctive visual phenomenology associated with actual scenes,
which is lacking in visual experiences of depicted scenes (Noë 2005; Matthen
2005; Dokic 2010). When you see a picture of your children, your experi-
ence of them feels different from the experience that you have when you
see them in front of you: they do not experience them as being here. Seeing
an object as present involves being aware of it as a whole object located in
three-dimensional space, as an object that one can explore from different
perspectives and that one can grasp, while seeing a picture of the same ob-
ject only involves being aware of its material surface with certain con-
figurational properties. In the same way that there is a feeling of presence
associated with visual experiences of actual objects, we suggest that although
there is no feeling of bodily reality as we defined it, there is nonetheless a
feeling of bodily presence normally associated with bodily sensations: we nor-
mally experience our body as being “here” in a space larger than our body
(Vignemont, forthcoming). As Martin notes: “In being aware of one’s skin as
a boundary of one’s body, one has some sense of space extending beyond that
boundary” (1993, 213). The awareness of bodily boundaries and the aware-
ness of a larger space are two facets of the same coin. Insofar as we are aware
of the boundaries of our body, we are also aware that there is something be-
yond these boundaries. Being aware of bodily boundaries indeed involves
being able to contrast what is inside from what is outside. Martin concludes
that we are aware not only that there is a space larger than our body but also
that our body is part of this larger space:“We have a sense of ourselves as
being bounded and limited objects within a larger space, which can contain
other objects” (Martin 1993, 211).
When we feel touch on our hand, we experience the pressure in a spe-
cific location within the map of our body (i.e., bodily location), but we also
experience this part of our body in a specific location in the external world
(i.e., egocentric location). As O’Shaughnessy (1980, vol. 1, p. 262) claims,
bodily sensations are “sensations-at-a-part-of-body-at-a-point-in-body-
relative-space.” Thanks to the egocentric locations of bodily sensations,
the body that one feels is experienced as being here in three-dimensional
external space.13
Now experiencing something as being here or as being present is not the
same as experiencing it as being real. Although the feeling of presence and
the feeling of reality are often confused, we believe that it is important to keep
them apart. Unlike the feeling of reality, the feeling of presence indeed is not
specific to tactile phenomenology but can be found in all sensory phenom-
enology. It might possibly be found even in imagination and in dream, and
we clearly do not experience a feeling of reality there. The feeling of pres-
ence only expresses the awareness of the spatial relationship between the per-
ceived object and the perceiver. It may not be as fundamental as the feeling
of reality as we defined it, but it is important, and its absence can be cruelly
experienced, as in depersonalization.
13 Smith (2002) claims that bodily sensations are devoid of phenomenal three-dimensional spa-
tiality because there cannot be any felt distance between the sense organ and the object since they
are one and the same, namely, the skin. Consequently, one cannot occlude one’s bodily sensations
by changing one’s spatial perspective on one’s body. This is true, but this does not prevent bodily
sensations from being experienced within an egocentric frame of reference (Vignemont 2018). This
is well illustrated by the Japanese illusion. Cross your wrists, your hands clasped with thumbs down.
Then turn your hands in toward you until your fingers point upward. If now I touch one of your
fingers, you will have difficulty not only in moving the finger that is touched, but also in reporting
which finger it is. This difficulty shows that the relative location of body parts matters. Further evi-
dence in favor of the egocentricity of tactile sensations can be found in the actions that we perform.
As Evans (1985) defines it, the egocentric space is a “behavioural space,” the space within which one
acts. Imagine that you feel an intense itching sensation on your back. It is true that you cannot oc-
clude it by changing your perspective on it, but you can still try to relieve it by scratching your back.
In order for you to appropriately guide your actions toward your body, you need information about
the itching location in external space. Furthermore, you need to be aware of your body as a three-
dimensional object in this space, an object that you have to get around to reach your goal.
References
Allais, L. 2015. Manifest Reality: Kant’s Idealism and His Realism. Oxford: Oxford
University Press.
Anscombe, G. E. M. 1957. Intention. Cambridge, MA: Harvard University Press.
Armstrong, D. M. 1962. Bodily Sensations. London: Routledge and Kegan Paul.
Armstrong, D. M. 1997. A World of States of Affairs. Cambridge: Cambridge
University Press.
Baldwin, T. 1995. Objectivity, causality, and agency. In J. L. Bermùdez, A. Marcel, and N.
Eilan, eds., The Body and the Self. Cambridge, MA: MIT Press. 107–125.
Baldwin, T. 2003. Perception and agency. In J. Roessler and N. Eilan, eds., Agency and Self-
Awareness. Oxford: Oxford University Press. 188–200.
Bennett, J. 1971. Locke, Berkeley, Hume: Central Themes. Oxford: Oxford University Press.
Berkeley, G. 1998a. Three Dialogues between Hylas and Philonous. Ed. J. Dancy. Oxford:
Berkeley, G. 1998b. A Treatise Concerning the Principles of Human Knowledge. Ed. J.
Dancy. Oxford: Oxford University Press.
Berto, F. 2012. “To be is to have causal powers”: Existence and nature in analytic meta-
physics. In M. Favaretti and M. Plebani, eds., Existence and Nature: New Perspectives.
Frankfurt am Main: Ontos. 33–63.
Billon, A. 2017. Mineness first. In F. de Vignemont and A. Alsmith, eds., The Subject’s
Matter: Self-Consciousness and the Body. Cambridge, MA: MIT Press. 189–216.
BonJour, L. 1992. Externalism/ internalism. In J. Dancy and Ernest Sosa, eds., A
Companion to Epistemology. Oxford: Blackwell.
Boswell, J. 1973. The Life of Samuel Johnson. Vol. 2. London: Everyman.
Brewer, B. 1995. Bodily awareness and the self. In J. Bermudez, A. J. Marcel, and N. Eilan,
eds., The Body and the Self. Cambridge, MA: MIT Press. 291–309.
Brown, T. 1827. Lectures on the Philosophy of the Human Mind. Edinburgh: William Tait.
Cabanis, P. J. G. 1802. Rapport du physique et du moral de l’homme. Paris: Crapart, Caille
and Ravier.
Campbell, J., and Q. Cassam. 2014. Berkeley’s Puzzle: What Does Experience Teach Us? Oxford:
Cassam, Q. 2005. Self and World. Oxford: Oxford University Press.
Condillac. 1997. Traité des sensations, traité sur les animaux. Paris: Fayard.
Crane, T. 2005. The problem of perception. In E. N. Zalta, ed., The Stanford Encyclopedia
of Philosophy, Spring 2008 ed. http://plato.stanford.edu/archives/spr2008/entries/
perception-problem/.
Delboeuf, J. 1881. Le sentiment de l’effort. Revue Philosophique de la France et de l’Étranger
12: 513–527.
Descartes, R. 1724/1979. Les méditations métaphysiques. Paris: Garnier Flammarion.
Descartes, R. 1983. Principles of Philosophy. Trans., Miller, Valentine Rodger and Miller,
Reese P. Dordrecht, Reidel Distributed by Kluwer Boston.
Descartes, R. 1998. The World and Other Writings. Ed. and trans. S. Gaukroger. Cambridge:
Destutt de Tracy, A. 1801. Projet d’éléments d’idéologie à l’usage des écoles centrales de la
République française. Paris: Chez Pierre Didot l’aîné, Firmin Didot et Debray.
Dokic, J. 2010. Perceptual recognition and the feeling of presence. In B. Nanay, ed.,
Perceiving the World. Oxford: Oxford University Press.
Dugas, L., and F. Moutier. 1911. La dépersonnalisation. Paris: Félix Alcan.

Fine, K. 1994a. Essence and modality. In J. Tomberlin, ed., Philosophical Perspectives
8: 1–16.
Fine, K. 1994b. Ontological dependence. Proceedings of the Aristotelian Society
95(1): 269–290.
Freeland, C. 1992. Aristotle on the sense of touch. In M. Nussbaum, ed., Essays on
Aristotle’s “De anima”. Oxford: Oxford University Press. 227–248.
Fricke, M. F., and P. Snowdon. 2003. Solidity and impediment. Analysis 63(3): 173–178.
Fulkerson, M. 2014. The First Sense: A Philosophical Study of Human Touch. Cambridge,
MA: MIT Press.
Garnett, A. C. 1965. The Perceptual Process. London: George Allen & Unwin.
Hamlyn, D. W. 1990. In & Out of the Black Box. Oxford: Basil Blackwell.
Hampshire, S. 1959. Thought and Action. London: Chatto and Windus.
Hampshire, S., and P. F. Strawson. 1961. Symposium: Perception and identification.
Proceedings of the Aristotelian Society 35: 81–120.
Heider, F. 1959. Thing and medium. Psychological Issues 1(3): 1–34.
Huemer, M. 2006. Phenomenal conservatism and the internalist intuition. American
Philosophical Quarterly 43(2): 147–158.
Huemer, M. 2007. Compassionate phenomenal conservatism. Philosophy and
Hume, D. 2000. A Treatise of Human Nature. Ed. D. Norton and M. Norton. Oxford:
James, W. 1880. The Feeling of Effort. Anniversary Memoirs of the Boston Society of
Natural History. Boston: Published by the Society.
James, W. 1950. The Principles of Psychology. Vol. 2. New York: Dover.
Kant, I. 2006. Anthropology from a Pragmatic Point of View. Ed. and trans. Robert B.
Louden. Cambridge: Cambridge University Press.
Lowe, E. J. 1996. Subjects of Experience. Cambridge: Cambridge University Press.
Lowe, E. J. 2000. An Introduction to the Philosophy of Mind. Cambridge: Cambridge
University Press.
Maine de Biran, F. P. G. 2001. Essai sur les fondements de la psychologie. Paris: Vrin.
Malebranche, N. 1687/1991. Entretiens sur la métaphysique et sur la religion. In A. Robinet,
ed., Oeuvres complètes. Paris: Vrin.
Martin, M. G. F. 1993. Sense modalities and spatial properties. In N. Eilan, R. McCarty,
and B. Brewer, eds., Spatial Representations. Oxford: Oxford University Press. 206–218.
Massin, O. 2010. L’objectivité du toucher, métaphysique et perception. PhD dissertation,
Aix-Marseille University.
Massin, O. 2014. Quand vouloir, c’est faire. In R. Clot-Goudard, ed., L’explication de
l’action: Analyses contemporaines Paris: Vrin. 71–114.
Massin, O. 2017. The composition of forces. British Journal for the Philosophy of Science
68(3): 805–846.
Massin, O. 2017. Towards a definition of efforts. Motivational Science 3(3): 230–259.
Massin, O., and J.-M. Monnoyer. 2003. Toucher et proprioception. Voir (Barré) 26: 48–73.
Matthen, M. P. 2005. Seeing, Doing, and Knowing: A Philosophical Theory of Sense
Perception. Oxford: Oxford University Press.
McCann, H. 1998. The Works of Agency: On Human Action, Will, and Freedom, Ithaca,
NY: Cornell University Press.
McDowell, J. 2011. Anscombe on bodily self-knowledge. In A. Ford, J. Hornsby, and F.

Stoutland, eds., Essays on Anscombe’s Intention. Cambridge, MA: MIT Press. 128–146.
Noë, A. 2005. Real presence. Philosophical Topics 33: 235–64.
O’Shaughnessy, B. 1980. The Will. Vol. 1. Cambridge: Cambridge University Press.
Pacherie, E. 2007. The anarchic hand syndrome and utilization behavior: A window onto
agentive self-awareness. Functional Neurology 22(4): 211–217.
Reid, T. 1969. Essays on the Active Powers of the Human Mind. Ed. B. Brody. Cambridge,
MA: MIT Press.
Reid, T. 2000. An Inquiry into the Human Mind on the Principles of Common Sense.
University Park: Pennsylvania State University Press.
Russell, J. 1995. At two with nature: Agency and the development of self-world dualism.
In J. Bermudez, A. J. Marcel, and N. Eilan, eds., The Body and the Self. Cambridge,
Russell, J. 1996. Agency: Its Role in Mental Development. Hove, East Sussex, UK: Erlbaum
(UK), Taylor & Francis.
Sanford, D. 1967. Volume and solidity. Australasian Journal of Philosophy 45: 329–340.
Scandola, M., E. Tidoni, R. Avesani, G. Brunelli, S. M. Aglioti, and V. Moro. 2014. Rubber
hand illusion induced by touching the face ipsilaterally to a deprived hand: Evidence
for plastic “somatotopic” remapping in tetraplegics. Frontiers in Human Neuroscience
8: 404.
Scheler, M. 1927/ 1973. Idealismus- Realismus. In Philosophischer Anzeiger, vol. 2.
Bonn: Verlag Friedrich Cohen. Trans. D. R. Lachterman in M. Scheler, Selected
Philosophical Essays. Evanston, IL: Northwestern University Press.
Scheler, M. 1977. Zu “Idealismus—Realismus,” Aus Teil V. Das Emotionale realitäts
problem. Trans. in part by T. J. Sheehan as Reality and resistance: On Being and Time,
section 43, Listening 12(3): 61–73.
Schopenhauer, A. 1958. The World as Will and Representation. Trans. E. F. J. Payne. New York:
Dover.
Sherrington, C. S. 1906. The Integrative Action of the Nervous System. New Haven: Yale
University Press.
Siegel, S. 2006. Subject and object in the contents of visual experience.” Philosophical
Review 115: 355–388.
Sierra, M. 2009. Depersonalization: A New Look at a Neglected Syndrome. Cambridge:
Smith, A. D. 2002. The Problem of Perception. Cambridge, MA: Harvard University Press.
Strawson, P. F. 1979/2002. Perception and its objects. In A. Noë and E. Thompson,
eds., Vision and Mind: Selected Readings in the Philosophy of Perception. Cambridge,
Taylor-Clarke, M., P. Jacobsen, and P. Haggard. 2004. Keeping the world a constant
size: Object constancy in human touch. Nature Neuroscience 7(3): 219–220.
Vignemont, F. de. 2017. Pain and touch. The Monist 100(4): 465–477.
Vignemont, F. de. 2018. Mind the Body. Oxford: Oxford University Press.
Vignemont, F. de. Forthcoming. Varieties of bodily feelings. In U. Kriegel, ed., The Oxford
Handbook of the Philosophy of Consciousness. Oxford: Oxford University Press.
Vignemont, F. de., and O. Massin. 2015. Touch. In M. Matthen, ed., Oxford Handbook of
8
Experiences of Duration and
Cognitive Penetrability
Carrie Figdor
8.1. Introduction
While “everyday life consists of picnics and meetings just as it consists of

chairs and birds” (Zacks et al. 2007, 23), events have been given relatively
short shrift in perception research. Similarly, problems in the epistemology
of perception are framed almost exclusively in terms of objects and their
properties, in particular those perceived visually. This chapter extends the
philosophical debates to include the epistemology of event perception. The
main phenomenologically accessible difference between object and event
perception is the experience of temporal information in the latter. Although
temporal experience includes temporal order, simultaneity, and duration,
I focus on our experiences of duration, and consider mainly their relation
to the problem of cognitive penetrability.1 Is the perception of the temporal
unfolding of events cognitively penetrated? If so, is this cognitive penetration
epistemically pernicious?
In section 8.1, I present the problem of the experience of duration from a
philosophical perspective. I then offer an empirically based account of how
we perceive durations. I focus on perceptions of everyday events occurring
within timescales of seconds to minutes to hours. This is the range of interval
timing, which lies between millisecond timing (involved in speech percep-
tion and motor control) and circadian timing (regulating the 24-hour sleep-
wake cycle). On this account, such experiences are a consequence of event
segmentation and interval timing processes. This account provides insight
into the philosophical problem of duration, and shows how experiences of
the now are derived from experiences of durations.
1 I will use “perception” throughout to mean “conscious perception,” which is standard usage.
Carrie Figdor, Experiences of Duration and Cognitive Penetrability In: The Epistemology of
Non-Visual Perception. Edited by: Berit Brogaard and Dimitria Electra Gatzia, Oxford University
Press (2020). © Oxford University Press.
DOI: 10.1093/oso/9780190648916.003.0009
Experiences of Duration and Cognitive Penetrability 189
In section 8.2, I consider the problem of cognitive penetrability and its

implications for the justification of perceptual beliefs about duration. I argue
that duration experiences and beliefs can’t be divided into the stages required
to formulate the problem. In interval processing, stored temporal informa-
tion is combined with nonstored nontemporal information to produce both
duration experiences and duration beliefs. The problem of cognitive pene-
trability requires belief to influence perception, whereas interval processing
yields both belief and perception.
An independent philosophical debate regarding temporal experience,
which I will set aside, concerns the metaphysics of time. This is the debate be-
tween the A-and B-theories of time. They agree that time exists, but differ on
the nature of tense. An A-theorist holds that the present (the “now” or “spe-
cious present”) is ontologically special, and that the past and future do not
exist. The B-theorist holds that past, present, and future are all equally real,
and that the present is not ontologically special (although it may be special
for psychological reasons; of course, it can also be psychologically special for
A-theorists). Temporal experiences can be veridical if there is a past, present,
and future that we track, whether the A-theory or B-theory of time is cor-
rect, although not if time itself is an illusion. So aside from global skepticism
about the reality of time, this debate does not affect the question of whether
event percepts justify event beliefs.2
For the same reason, I also set aside background issues in the metaphysics
of events. While events seem to be perennially suspect from an ontological
perspective, arguments from parsimony for denying ontological commit-
ment to events cut both ways. Four-dimensionalists hold that objects are
very slow events, and process metaphysicians hold that processes are fun-
damental. Everyday perception is consistent with a four-dimensionalist
metaphysics as well as one that downgrades events to properties of or rela-
tions between objects. The prevailing contemporary view is that events are
particulars, not universals (Casati and Varzi 2008). Even so, there are sharply
different ways of individuating them (Quine 1985; Davidson 1969; Kim
1973). I will assume the coarse-grained view of individuation adopted ex-
plicitly or implicitly in the empirical literature on event perception. A coarse-
grained view entails that the same token event might have changed in certain
2 Callender (2010) summarizes the contemporary debate in physics about the existence of time as a
fundamental physical magnitude. This debate leaves open its existence as an emergent magnitude, in
relation to which temporal experiences can be veridical or illusory. I will assume time is (at least) an
emergent property.
ways, including in its duration, and remained the same event. This is plau-
sible because there are modality-specific limits below which we cannot
distinguish distinct events in that sense modality, and because we tend to
individuate longer events by goals and intentions rather than durations.
8.2. The Philosophy, Psychology, and Neuroscience

of Experiences of Duration
8.2.1. The Experience of Duration in Philosophy
Kelly (2005, 210) isolates the experience of concern here in terms of what
he calls the puzzle of temporal experience: “How is it possible for us to have
experiences as of continuous, dynamic, temporally structured, unified events
given that we start with (what at least seems to be) a sequence of independent
and static snapshots of the world at a time?” I call this the puzzle of duration,
given that the phrase “temporal experience” also includes experiences of si-
multaneity and temporal order.3 While it is not obvious phenomenologically
that we do start with what seems to be a sequence of snapshots, the funda-
mental question is how we come to experience events as occurring through
time at all.
A vivid illustration of this type of experience involves hearing a soprano
at the opera. The soprano hits a high C, and holds it, and holds it, and holds
it, and holds it . . . and at some point, as Kelly (2005, 208) puts it, what we
hear “no longer seems to be limited to the pitch, timbre, loudness, and other
strictly audible qualities of the note. You seem in addition to experience, even
to hear, something about its temporal extent. . . . one is tempted to say . . . that
the note now sounds as though it has been going on for a very long time.”
A minimally apt description of this experience is that we do experience it
as having an unusually long duration, which entails that we have an expe-
rience of its duration. This type of experience is also present when we con-
sider ordinary events, such as brewing coffee or checking out items at the
grocery store. Unlike a passive acoustic experience of hearing a single held
3 These experiences generate distinct puzzles and are also targets of empirical research. The time-
stamp problem is about when in temporal order we experience something as having occurred, and
the simultaneity problem is about which events we experience as simultaneous (Kelly 2005). He also
distinguishes the puzzle of duration from the hard problem of consciousness, which he sets aside.
In contrast, Merino-Rajme (2014) considers the hard problem of consciousness extended to certain
temporal experiences.
note, many of these events have temporal subparts, involve multiple objects,
include our active participation, and integrate multimodal sensory inputs,
each with proprietary temporal-perceptual properties. Simple or complex,
the experience of duration marks a psychologically important difference be-
tween perceptual experiences of ordinary events and ordinary objects. How
do we explain it?
Kelly considers two philosophical theories of how we experience du-
ration, and argues that both are inadequate. These are what he calls the
specious present theory and the retention theory. On the specious present
theory, which he associates with William James, the present of which we
are directly aware may be considered to be somewhat extended in time, like
a saddleback. We do not perceive a static snapshot, but instead a tempo-
rally extended duration that includes what has recently occurred and what
is about to occur.4 This proposal requires either that we have direct per-
ceptual access to the recent past and future moments close to the present,
or else that we are directly aware of the duration. Neither choice makes
sense, Kelly argues: we don’t have direct perceptual awareness of what has
recently occurred or what is about to occur, while the “default position” on
direct awareness seems to be that we are directly aware of what is present
now, not of duration.
The retention theory, which Kelly associates with Husserl, holds that per-
ception presents us with snapshots, but that these momentary snapshots of
which we are directly aware in perception are augmented by states of reten-
tion and protension. These are not memories and anticipations, but sui ge-
neris intentional acts. Kelly argues that this theory merely provides labels for
what must be explained. For example, retention is defined as a way of being
directed toward objects and events as just-having-been. It is neither instan-
taneous remembering (such as when one suddenly remembers having left
one’s keys in the kitchen), nor entertaining a memory (in the way one can
relive now a past event). What we need is an explanation of what it is to ex-
perience something as just-having-been that goes beyond saying it is a phe-
nomenon involved in the experience of duration. Kelly does not elaborate on
the nature of protension, but presumably a similar problem would arise for
that state as well.
4 Poppel (2004) has argued that our neurocognitive machinery is such that the sensed now is in the
range of three seconds. Kelly also takes note of Poppel’s claim.

Kelly does not try to solve the puzzle of duration, but hopes to have shown
that there is an interesting problem to be addressed. In what follows, I will
elaborate an account of duration based in contemporary science and artic-
ulate its solution to the puzzle. With the account’s prima facie empirical and
philosophical validity thus established, our experiences of duration so un-
derstood will be examined in the discussion of the cognitive penetrability of
duration perception in section 8.3.
8.2.2. The Psychology and Neuroscience of Event Perception
Event segmentation is analogous to, and as basic as, object individua-

tion: even infants are able to parse evolving scenes into discrete segments
(Wynn 1996). This is a fruitful perspective from which to understand dura-
tion perception because durations are a matter of interval timing, and event
segmentation yields the intervals perceived as durations. Interval timing is
the midrange of timing involved in conscious time estimation, cognitive pro-
cessing (such as decision-making), and behavior (such as foraging) (Buhusi
and Meck 2005, fig. 1). Many ordinary events, including those targeted in
event segmentation research, fall in this range.
A leading contemporary theory of how we segment continuous activity to
generate perceptions of discrete events is event segmentation theory (EST)
(Zacks and Tversky 2001; Zacks et al. 2007; Kurby and Zacks 2007; Zacks
et al., 2001; Reynolds et al. 2007; Tversky et al. 2008; Shipley and Zacks 2008;
for a predecessor, see Newtson et al. 1977). The targets of EST are percep-
tual experiences of Quinean events, whether dynamic (events) or concrete
(objects) (Zacks and Tversky 2001, 5).5 An event is defined as a segment of
time at a given location that is perceived by an observer to have a beginning
and an end. This definition, while not exhaustive, includes events that in-
volve goal-directed human activity and have durations from seconds to tens
of minutes (Zacks and Tversky 2001, 5).
In EST, event segmentation is a side effect of our use of prediction for
perception. It is thus a specific case of the predictive processing or predic-
tive coding framework of overall brain function (Kurby and Zacks 2007,
5 This perspective coheres with Kelly’s position that adequate answers to the puzzle of duration
should not fundamentally distinguish between object and event perception, even though their rela-
tion to interval timing and perceived duration differs phenomenally. I set aside here the question of
how interval timing is related to object perception.
73; Friston and Stephan 2007; Clark 2013; Hohwy 2014).6 It is considered a
spontaneous, ongoing process that does not require conscious attention and
occurs at various timescales simultaneously. The model posits event models
and event schemata. Event models are multimodal, actively maintained
representations in working memory of “what is happening now” (Zacks et al.
2007, 7). Their content is influenced by event schemata, which are semantic
memory representations of information from previously encountered events
and are encoded in permanent synaptic changes.
The mechanism of EST is the detection of variations in the incoming sen-
sory stream and comparison of them to values predicted by event models.
Transient increases in prediction error based on comparison between ac-
tual and predicted sensory input at a given timescale triggers updating of
the relevant event model, after which another period of stability begins. This
increased processing is perceived as the subjective experience that a new
event has begun, while periods of stability are perceived as ongoing events
(Kurby and Zacks 2007, 72). We can selectively attend to particular timescales
in response to instruction, but also spontaneously segment at finer grains
when there is less predictability and we seek more information to understand
what is going on. The account implicitly distinguishes experiences of dura-
tion from those of succession: succession implies at least two events, while
durations are of one. Philosophical accounts of duration sometimes conflate
succession and duration as one phenomenon (Phillips 2014, 140).7
Zacks et al. (2007, 4) illustrate EST with the example of a man scraping
plates in the course of washing dishes. The whole plate-scraping segment
of the dishwashing event is predictable until the last plate is scraped, when
the goal of scraping the plates would no longer have predictive value and
updating mechanisms would kick in. At a more fine-grained timescale, each
plate-scraping activity will generate a small predictive error when that plate
has been scraped, and this error will correspond to a boundary between
each individual plate-scraping. It follows that prediction errors will be rel-
ative to timescales, such that variations in input that count as prediction
errors at one timescale (the start of a particular plate-scraping) fall within
expected or predicted limits at another (the whole dishwashing activity). The
6 While publications elaborating EST do not cite major advocates of the predictive processing
framework, the conceptual relationship is obvious and sometimes explicit (e.g., Radvansky and
Zacks 2017, 133).
7 It also may conflict with Phillips’s (2014) “naive” account of the relation between objective
durations and subjective experiences of duration, in which the latter “inherit” the “temporal struc-
ture” of the former. It depends on how “inheritance” is elaborated.
boundaries of each dish-scraping are prediction errors relative to the time-

scale of the event model for plate-scraping, but are within predicted values
at the timescale of the dishwashing event model in which the dish-scrapings
occur. Many everyday events will exhibit this complexity: they have temporal
segments (entailing part-whole relations) and temporal abstraction relations
across timescales and levels of generality (entailing hierarchical relations).8
While event schemata are explicitly thought to include such items as goals
and statistical information about paths and motions of objects, they also
presumably contain statistical information about how long types of events
generally take. In classical conditioning, after learning, the subject is able to
predict the duration between stimulus onset and reward presentation and
will begin to exhibit anticipatory behavior when the reward is nigh. But top-
down expectations of duration can also influence the quality of ongoing
experiences of duration without inducing prediction errors and thus seg-
mentation. For example, we may know from experience that a typical epi-
sode of standing in line to buy groceries takes a few minutes. An event model
for a token of this event may reliably predict a temporal range within which
the event should end. With this range and goal set, there is no need to allo-
cate attention to its duration if it is taking about as long as expected. When
the duration extends to the far end or beyond the approximate range, our
experience may qualitatively change: we suddenly become aware that we are
in a very slow checkout line. The qualitative difference may be felt as a change
of affect (e.g., impatience) that is assigned to the duration rather than some
other aspect. But there is no segmentation: the duration that comes to seem
too long is the duration of the same event.9
These changes in our experiences of ongoing durations make vivid the fact
that while EST tells us when and why event boundaries are perceived, it says
little about how event durations are perceived. It says only that they may be
perceived during periods of processing stability. Nevertheless, since event
segmentation yields intervals, EST is ipso facto a theory of what activates
or triggers interval timing and thus of what makes experiences of durations
possible. We can therefore expect to gain further insight into experiences of
duration by augmenting EST with what we know about our internal interval-
timing mechanism, sometimes known as the stopwatch. While the circadian
8 Thanks to Andy Clark for raising the need for this clarification.
9 Animals likely have similar qualitative differences: your dog experiences a duration between
when you walk in the door and when he is fed, but this experience (and not just his increasingly
frantic behavior) can change in quality if you are slower than usual in feeding him.
clock has been localized in the suprachiasmatic nucleus and millisecond

timing in the cerebellum, the stopwatch involves thalamo-cortico-striatal
circuitry, including the basal ganglia, supplementary motor area, prefrontal
cortex, and posterior parietal cortex (Allman 2014, 746 and table 1; Buhusi
and Meck 2005; Grondin 2010).10 After two preliminary remarks, in the rest
of this subsection I briefly summarize main themes from the literature on
interval timing, combine EST and the leading theory of interval timing, and
indicate how the combined account makes contact with epistemic concerns.
The first, minor, remark is that much psychological and neurobiolog-
ical research on interval timing has been directed at unimodal events with
timescales of a few hundred milliseconds to a few seconds (Matthews and
Meck 2016; Grondin 2010). Scaling-up assumptions are implied when
results are extended to longer intervals and more complex events within the
midrange.11
The second, major, remark is related to the distinction in research
protocols between those eliciting prospective timing judgments and those
eliciting retrospective timing judgments (Wearden 2005). In prospective
timing experiments, human or animal subjects are told (or trained) in ad-
vance that an interval estimate (provided verbally or behaviorally) will be
sought (e.g., “Hold down the button for one second”). In retrospective timing
experiments, human subjects are asked unexpected questions about intervals
(e.g., “how much time has passed since you started reading this paragraph?”).
The targets of these estimates are considered distinct: prospective judgments
are reports of episodes of interval timing, and retrospective judgments are
outputs of general cognitive mechanisms not specifically related to time. For
illustration, in Kelly’s case of the soprano presumably we could have made a
prospective judgment of the note’s duration had we been asked to do so. In
contrast, cases where (e.g.) after 25 minutes of a boring movie you say, “It
feels like the movie has been going on for an hour” (Merino-Rajme 2014), we
are giving retrospective judgements (and arguably in this case we are really
reporting an affective response). The internal clock theory discussed below
is about the mechanisms of interval timing whose experienced outputs are
10 These timing mechanisms interact: for example, older adults make more accurate duration
estimates in the morning, young adults in the evening (Allman et al. 2013, 758).
11 For example, Hommel et al. 2001 propose a theory of common coding for event perception
and action planning. While their focus is simple, brief events like key-pressing tasks, Zacks (2001,
910) and Hommel et al. (2001, 914) agree that a modified version of their basic framework should
apply to longer events (e.g., making coffee).
reported in prospective estimates. The importance of this point will become

clear in section 8.2.
The stopwatch temporally integrates multimodal sensory signals that
arrive at different speeds and are processed at different speeds by sensory
organs, compensating for these differences so that we can track discrete
events and objects. The dominant information-processing hypothesis of in-
terval timing is the pacemaker-accumulator or PA model (Treisman 1963;
Wearden 2005; Allman et al. 2014; Buhusi and Meck 2005; Eagleman et al.
2005; Grondin 2010; Block and Grondin 2014). The model implements the
scalar expectancy theory or SET (Gibbon 1977; Gibbon et al. 1984; Buhusi
and Meck 2005) in that it was formulated in part to explain the scalar pro-
perty of interval estimates. The scalar property is the fact that these estimates
are more variable relative to the mean as the length of the interval grows, by
a fixed proportion that follows Weber’s law. The errors are larger for larger
intervals, although not relatively larger than errors for shorter intervals
(Wearden 2005). Gibbon et al. (1997, 170) liken this feature of subjective
timing to a rubber ruler that can be stretched to measure any arbitrary target
interval, but where interval estimation error increases proportionally as the
ruler is stretched.
Note that the scalar property also applies to neural responses (Buhusi and
Meck 2005, 756, fig. 2). So while interval estimates by human and animal
subjects are reliably correlated with and interpreted as reports of experienced
intervals, the psychological concept of subjective interval timing includes
more than just experienced intervals. This coheres with EST’s view that
event segmentation proceeds nonconsciously as well as consciously. Also,
Block and Grondin (2014, 1) question whether the scalar property should
be taken for granted as an explanandum of an adequate theory of interval
timing, given that it is directly a feature of time estimation rather than of time
perception (which in their usage refers to what interval timing mechanisms
do, whether subjects perceive the intervals or not). However, by seeking to
explain both the PA model affirms a tight relationship between experienced
and estimated intervals in prospective timing.12
12 A minority view of interval timing (e.g., Staddon and Higa 2006) is that it is not performed
by a specialized mechanism but by basic memory and learning processes. Buhusi and Meck (2005,
763) hold that interval timing is a specialized mechanism, but that it shares neural circuits with non-
temporal processing. (For comparison: the circadian clock is considered a specialized mechanism,
and it has dedicated neural wetware.)
The PA model posits a pulse counter (a clock stage), a reference memory (a

memory stage), and a comparator (a decision stage), each with its own forms
of variance that can contribute to inaccurate perceptions of duration. At the
onset of an event, a pacemaker emits pulses (neural spikes) that are gated
into an accumulator by attention. The pulse tally is transferred to working
memory for comparison to an interval value for that type of event stored in
reference memory. (In prospective timing experiments, this is often an in-
terval just presented to the subject.) This step is also described in terms of the
transfer of the pulse count directly to reference memory for comparison to
the stored reference interval (Allman et al. 2014, 750). Either way, when these
values are close enough, a decision rule determines an interval estimate and
appropriate response.13
EST and the PA model can be integrated in part in the following way.
If EST is correct, working memory contains an active event model whose
predictive processing of incoming stimuli triggers the first stage of the PA
model; presumably non-attentional mechanisms can also do the gating in
this stage. Working memory also comes to contain the pulse tally received
from the accumulator in the second stage of interval processing. Importantly,
there must be coordination between the event model that generates predic-
tion errors (event boundaries) and the reference interval to which the accu-
mulated pulse count that begins at a given boundary is compared. A simple
hypothesis is that the reference interval is part of the event model, although
other relations are possible. However coordination occurs, event schemata
presumably influence duration perception as well as boundary perception
by helping determine which reference interval is relevant. This may be in-
cluded in determining which event model is active. For example, the event
model for dish-scraping includes information about how long this type of
event normally takes, and the interval that is triggered at the start of the
dish-scraping is compared to this reference interval. Given that perceived
boundaries and subsequent perceived intervals together constitute perceived
durations, it follows that both constituents of duration perception are subject
to top-down influences. I return to this point when discussing cognitive pen-
etration in section 8.3.
13 The coincidence-detection (or striatal beat frequency, SBF) model of interval timing holds that
patterns of spikes of cortical neurons are continuously compared by striatal spiny neurons with a
reference pattern. It may be that coincidence-detection and pulse counting are just two ways of com-
paring a stimulus-dependent quantity to a reference interval (Buhusi and Meck 2005, 763). If so, the
SBF model is the neurobiological implementation of the first step of the PA model, not an alternative
theory.
In addition, the fact that we make proportionally larger errors when esti-
mating larger intervals may help explain why, as events increase in temporal
length, they are more characterized by goals, plans, or intentions rather
than physical features such as motion (Zacks and Tversky 2001), and why
beginnings loom larger than endings (Teigen et al. 2017). We compensate
for the stopwatch’s loss of precision at longer intervals by using more reliable
alternative methods to fix when longer events are over. A non-elite marathon
runner may have a vague idea of her time when she crosses the finish line, but
she knows exactly when she has finished. We can of course get better at time
estimation. Elite marathon runners are duration experts analogous to the
way sommeliers are wine experts: they can accurately distinguish durations
of two hours from two hours and five minutes. In any case, there is no reason
to think estimates of durations of everyday events must be precise to be accu-
rate. Accuracy often only requires falling within a range.14
The neurobiology of interval timing is being actively investigated. A bio-
logical basis for the stopwatch was initially inferred from the fact that higher
bodily temperatures in fever resulted in altered subjective judgments of
passing time (Wearden 2005, 9, fig. 1). Within the PA model, heating implies
more clock ticks per objective interval and more quickly accumulating the
number of pulses in the reference interval associated with the event being
timed. The result is experiencing and estimating intervals as longer than they
really are: when instructed to count out a minute, we may count out (what
feels like) 60 seconds in less than 45 seconds. Perceptions of durations are
also affected by a variety of sensory, psychological, and physiological factors,
including attention, arousal, memory, affect, psychiatric disorders, and
drugs affecting neural and neurotransmitter activity (Matthews and Meck
2016; Allman et al. 2014; Droit-Volet and Meck 2007; Buhusi and Meck 2005;
Grondin 2010; Cheng et al. 2006; Terhune et al. 2014).
In terms of the PA model, these factors can change the pulse rate (clock
speed), the working memory representation of the reference interval, or the
baud rate at which accumulated pulses are transferred into working memory.
For example, dopaminergic agonists (e.g., methamphetamine) increase
clock speed, while antagonists (e.g., haloperidol) decrease it (Allman et al.
14 Our baseline accuracy for judging intervals differs across modalities (e.g., audition is generally
more accurate than vision [Allman et al. 2013, 746]). We also tend to perceive events as occurring
closer in time than they are when the events are an action we did and its effect (Ebert and Wegner
2010; Andersen 2013), although measured effects are in the millisecond range. Interval timing ability
also varies across subjects.
2014, 749; Buhusi and Meck 2005, 757), and cholinergic antagonists (e.g.,
atropine) affect reference memory in that the criterion interval a subject uses
is longer than the intervals that were presented in training. Given the PA
model, at least some of these influences can explain why car accidents may
be experienced as occurring in slow motion. Due to sharp, rapid increases in
attention, affect, and stress, external time is experienced as abruptly slowing
down because the internal clock rate has rapidly accelerated without a cor-
responding shift in the reference interval for normal driving that was active.
Yet while we may perceive everything in slow motion, our millisecond timing
mechanisms can still control our motor responses automatically.15
8.2.3. The Philosophical Adequacy of the Account
The grounding of duration perception in perception of the now appears to

be a common background assumption of the specious present and retention
views. Perception of now is the basic item of temporal perception, whether
it is instantaneous or somewhat stretched. However, if the above account
is on the right track, we do not perceive duration as a number of stitched-
together nows (or overlapping brief successive experiences: Dainton 2014).
Perceiving durations is perceiving boundaries plus intervals— that is,
maxima of processing changes that trigger the stopwatch and periods of pro-
cessing stability between these boundaries while the stopwatch ticks away.
The experienced now is just an ordinary, smallish experienced duration—
ordinary and smallish because we are easily able to perceive ordinary events
with durations far briefer than three seconds. “Now” is a vague term that
picks out these intervals. There is nothing left to explain of the now once we
explain duration in terms of event segmentation and interval timing. Note
that this is not an account of simultaneity or temporal order of successive
events, which are distinct explananda.
An imperfect metaphor for our experiences of duration is many reels
of film with irregular frames (events at varying timescales with varying
durations) running through a movie projector at once. The now is what is
15 Arstila (2012) suggests that norepinephrine is likely to play an important role as well, in partic-
ular for attention shifts. On a higher-level account (Tse et al. 2004), the added attention may result in
fewer temporal cues being missed (“missed temporal cues” interpretation) or the counting of more
units or pulses by the stopwatch (“attentional boost” interpretation), or both. The account in the text
is not intended to be complete.
in the projector’s light, whether this includes at least one consciously acces-
sible boundary or not. In these terms, the traditional debate of extensional
versus retentional views of experiences of duration concerns whether what
is in the projector’s light is extended or momentary (see, e.g., Dainton 2014,
103, fig. 6.2). The empirically based account does not ignore philosophical
concerns to explain how conscious experience seems (continuous, unified,
etc.); it interprets phenomenological descriptions of experiences of duration
and the now in the light of what we know about interval processing.
But Kelly might insist: how is it that we both perceive the soprano as
singing the note now and yet perceive that this acoustic event has been going
on for a long time? I suggest that this description picks out temporal and non-
temporal elements of our phenomenologically accessible experience. The
nowness of the event’s duration is the accessible segment of an ongoing dura-
tion. A second aspect of our experience is generated by enhanced attention to
this duration, which is prolonged beyond expectation on the assumption that
the aria is not familiar. (If we are opera experts, we will not experience the
note as too long—not unless it violates the composer’s actual notation.) This
is worthy of additional attentional resources in the light of our limited know-
ledge. Kelly describes this sudden change in our experience of the note as
perceiving that it has been going on for a long time. But we are not reporting
experiencing the soprano’s singing as slower or faster than it actually is (as
we might if we are feverish). Those changes would be reported as changes in
experienced duration. Our stopwatch is clicking away at its normal speed.
Instead, we are reporting another aspect of our experience—an experience
of astonishment directed at its duration, rather than, say, its pitch. This is a
phenomenological change in our experience of its duration, but it is not due
to a change in timing, and so a report of it is not a report of our experience
of timing. The account of duration helps us distinguish features of duration
experiences that phenomenological description does not.
Kelly also mentions that the received view of our experience of the now
is that it is direct. For some, perception is direct in those cases when it can
be appropriately contrasted with cases where inference (or some other cog-
nitive operation) is added to perception. For example, I perceive my car by
looking at it, but may use inference to recognize it as my car if it has been
totaled (Gallagher 2008). By the account given above, experience of the now
is direct by this definition. On the other hand, for others perception is direct
if percepts contain exclusively sensory information, independently of any
other conscious source of information, such as belief or memory (Chuard
2011). By the account given here, experience of the now is not direct by this
definition. Chuard also asks (2011, 4) whether temporal properties and rela-
tions are among the “strictly speaking perceptually accessible” properties—
that is, if they are perceptible “in the same way, that is, as shapes, colours,
and spatial relations” are perceptible. They aren’t, but so what? Appeals to
directness or strict perceptual accessibility do not seem to isolate important
features of temporal experiences.
8.3. The Epistemology of Event Perception
8.3.1. Perception, Justification, and Cognitive Penetration
The main question in the epistemology of perception is whether beliefs

about the external world are justified by the perceptual experiences they are
occasioned by (Siegel and Silins 2015).16 Three main sources of doubt that
they are so justified include the underdetermination or ambiguity of percep-
tual content by sensory stimuli (underdetermination), the biasing influence of
cognitive states (cognitive penetrability), and the possibility of global illusion
(skepticism). Underdetermination and cognitive penetrability can overlap
when cognitive states (or sources of information considered cognitive) dis-
ambiguate the sensory input. Philosophers have discussed these concerns pri-
marily in relation to visual perception of and beliefs about objects and their
features. Here I extend the discussion of cognitive penetrability to events and
their durations, setting aside the problems of skepticism (or illusion) and
underdetermination.17 Also, in contemporary discussions, experiences are
taken to have externally directed contents, and perceivers need not believe
that things are as experiences represent them to be (Siegel and Silins 2015,
782). Just so, our experiences of duration are about external events, and we
16 Siegel and Silins distinguish between propositional justification, when an experience provides
reason to believe something whether one comes to believe it or not, and doxastic justification, in
which a belief is based on experience—more specifically, a belief is doxastically justified iff it is ra-
tionally formed, adjusted, or maintained on the basis of experience (Siegel and Silins 2015, 784). In
another use of “doxastic,” doxastic states are those accessible to consciousness and inferentially inte-
grated (Macpherson 2017, 11).
17 The ambiguity of event perception is revealed by the McGurk effect and the sliding/bouncing
effect (Sekuler et al. 1997; Watanabe and Shimojo 2005). Experiments inducing illusory duration
judgments suggest that the stopwatch is constantly being calibrated to the external world (Eagleman
et al. 2005). Moreover, there must be coordination between the timing and durations of our own
movements and the movements or motions of other entities for the purpose of adaptive action
(Hecht 2000, 18; Kurby and Zacks 2007, 78; Hommel et al. 2001, 877).
can have experiences of durations that we do not believe are veridical (e.g., the
illusion of time passing in slow motion during a car accident).
To a first approximation, cognitive penetration of perception occurs when
beliefs, expectations, desires, hopes, goals, and other cognitive states inter-
vene in the perceptual process so as to affect the nature of the perceptual ex-
perience (Stokes 2013; Silins 2016, 24; Vance 2015; Siegel 2012, 205–206).
This difference in the nature of the experience is usually spelled out at least
in part in terms of a difference in the experience’s content. Macpherson
(2012, 29) holds that a perception is not cognitively penetrable if it is not
possible for any two perceivers (or the same perceiver at different times) to
have experiences with distinct content or character when one holds fixed the
object or event of perception, the perceptual conditions (e.g., lighting), the
spatial attention of the subject, and the conditions of the sensory organs(s).
We may add temporal attention to the list (e.g., attention to the soprano’s
singing). Macpherson (2017) further specifies that in cognitive penetration
there must be a semantic and causal link at each step from the belief (or other
cognitive state) to the perception.18
Cognitive penetrability of perception is an epistemic problem in the light
of traditional views of how perception works and how it is supposed to pro-
vide epistemic justification. The metaphysical presupposition is the existence
of a theoretically important perception/cognition distinction. The episte-
mological presupposition is the idea that percepts must be linked to the ex-
ternal world without mediation by cognitive resources to have justificatory
power. They should be “untainted” by prior assumptions made by the subject
or a subpersonal part of her perceptual system (Vance 2015, 643). As Silins
(2016) puts it, your experience “reflects what is before you and does not re-
flect your own mind. Given that your experience is not influenced in any way
by your theories or expectations, it is thereby in an optimal position to con-
firm or disconfirm hypotheses about the world.” These two presuppositions
ground the idea “that the content of the perception underwrites that of the
belief it justifies, in the sense that it logically implies it, or makes it probable,
or maybe just in the sense that an inference from the perception to the be-
lief would, in the present context, be reliable” (Heck 2000, 499). Cognitively
penetrated perceptions lack the content independence required for them to
provide rational justification for belief.
18 Macpherson (2017) also distinguishes between cognitive penetration of early vision (e.g.,
Pylyshyn 1999, 343) and of perceptual experience; the latter is of concern here.
Siegel (2012, 2017) provides a compelling example of the worrisome cir-

cularity: Jill believes (without good grounds) that Jack is angry at her, she
sees his face as being angry because of this belief, and she takes this percep-
tion at face value to justify her belief that he is angry. He does, after all, appear
angry to her. In this way, the penetrating cognitive states are “stacking the tri-
bunal of experience” in their favor: the experience does not provide an inde-
pendent reason for your belief (but see Lyons 2011). Of course, we also know
that learning can improve perceptual experiences and lead to better beliefs
than one might otherwise have. A tree expert knows what an elm looks like
and will perceive elms more quickly and accurately, and these perceptions
will redound to the credit of her prior elm beliefs. If these are cases of cogni-
tively penetration, then cognitive penetration is not necessarily pernicious.
Regardless, it seems clear that a problem can arise in some cases.
8.3.2. Experiences of Duration and Cognitive Penetrability
On the EST-based account, experiences of duration are made possible by the

triggering of the stopwatch by prediction errors generated in the course of
processing nontemporal sensory inputs. EST builds top-down influence into
event segmentation, and thus potentially into experiences of event bound-
aries, via the influence on event models of prior knowledge contained in
event schemata. I elaborated this account in plausible ways. Schemata pre-
sumably include statistical information about previously experienced event
durations. The influence of this information can be expected to extend to
experiences of event duration, given the need to coordinate segmentation
and interval timing processes. Coordination would be achieved if event
models simply contained previously learned reference intervals. In short, the
proposed account of event perception appears to raise worries of cognitive
penetrability about experiences of duration.
One reason not to worry stems from the fact that the presupposed divide
between perception and cognition is undermined by scientific advances.
Without that distinction, cognitive penetrability may be impossible given
how it is defined (Macpherson 2017). Even if it is not ruled out by defini-
tion, in predictive processing or broadly Bayesian accounts of overall neural
function, the outputs are our best models of the world. What is fed forward
from the sensory stream is a prediction error (if any) generated by a differ-
ence between the expected sensory stimulation and what is detected. As Shea
(2015, 76–77) puts it, in predictive processing there is no “cascade of illusory

justification” bubbling up from below, just the nudging of preexisting models
toward better predictions using prediction errors. The account of duration
perception defended here does not quite escape the worry for this reason,
because while EST is a predictive processing model, the PA model is not. But
it could easily be revised into one—for example, by interpreting the refer-
ence interval as a predicted value of an active event model. In this case, the
worry would be alleviated for duration perception as it is for any predictive
coding model.
However, Shea suggests that an updated reformulation of the perception/
cognition distinction that might enable us to pose some of the questions
raised in the cognitive penetrability debate. It has turned out that many input-
driven processes important in cognition turn out to straddle the traditional
divide. They have features that usually belong to systems in the cognitive cat-
egory. For example, while they are input driven, they have an “amodal” phe-
nomenology in that they “take as input whichever kinds of sensory inputs are
relevant in the circumstances” (Shea 2015, 85). Among these hybrid systems
are the system for representing one’s spatial position, language processing,
and Carey’s systems of core cognition, such as numerosity for representing
quantities of items or perception of agency.
Our interval timekeeping mechanisms also belong in the hybrid class.
The stopwatch is amodal in that it is driven by nontemporal inputs in the
sensory modalities that are relevant in the circumstances. It also operates at
many timescales, some fast and mandatory, others slower and under cog-
nitive control. Note that our timing mechanisms are not on a par in this
respect. Millisecond and circadian timing are automatic and largely not con-
sciously controlled. Circadian timing in particular satisfies many if not all of
Fodor’s (1983) original criteria for modularity—it is automatic, has dedi-
cated neural wetware, and is to some degree encapsulated in that its opera-
tion is impervious to belief about the actual time (such as right after a long
flight). Interval timing operates like a central system by Fodor’s standards.
It is subserved by neural circuitry spanning many cortical and subcortical
regions of the brain. Influences on the activity of the stopwatch that affect
our experiences of duration can come from below (e.g., drug-induced or
externally stimulated variation in neurotransmitter activity, affecting the
pulse rate of the stopwatch) and above (e.g., memory defects and emotional
states, affecting the reference interval). These effects occur in normal and
abnormal experienced intervals.
Shea suggests that we can at least characterize how much of a mechanism’s

operation is driven by current input and how much by top-down informa-
tion. Not all preexisting information is top-down. For example, a system can
contain prior information in the form of expectations, with the latter un-
derstood as dispositions to transition from one representation or another in
certain ways. These do not count as top-down effects, but are the effect of
learning and may be beneficial. Top-down information is occurrent and ex-
plicitly represented, such that it can act as input to many different systems.
More precisely, a top-down influence is “a representationally mediated ef-
fect of an explicit representation R on a psychological process, where R is
not computed more directly than the representational influence of current
sensory inputs on the process” (Shea 2015, 81).19 This makes top-down influ-
ence a matter of degree. The surviving epistemic question would be whether
the output of one system A is suitable for belief formation (or other subse-
quent processing) in another system B if B’s representations have affected
A’s processing. On this view, independence from top-down influence is not
needed for epistemic suitability. Top-down information can influence a psy-
chological process whose outputs reinforce that information without perni-
cious circularity.
Shea’s proposal has the virtues of updating the debate to take into account
the actual complexity of much processing, and to allow that the epistemic
implications of top-down influence vary from case to case. The traditional
presuppositions are gently modified to leave some form of the traditional
worry behind. Unfortunately, once we acknowledge the complexity of ac-
tual processing, a third presupposition of the traditional debate is left unsup-
ported, and is unsupportable.
Perceptual belief is traditionally understood as being formed on the basis
of experience: the beliefs are occasioned by the perceptions. We see a red to-
mato and form the belief that it is red. Structurally, there are two processes
connected by a route for information flow. Both the original perception/cog-
nition distinction and Shea’s top-down/bottom-up distinction embed this
routing assumption in which information flows from process A to process
B, and maybe from B back to A. (There is no circularity without a routing
assumption.) This structure presupposes that A and B are (or are the outputs
of) distinguishable processes. We have percept-forming processes, and their
19 The “directness” requirement is not patently clear (as Shea recognizes) but I will not raise any
issues here about it.

outputs can be inputs to belief-forming processes. In cognitive penetration,

the routing also goes the other way: outputs of belief-forming processes in-
fluence percept-forming processes. Predictive coding and Shea’s analysis
show that these processes cannot be neatly separated. Duration processing
shows that even Shea’s reformulation does not go far enough in taking ac-
count of the complexity.
As noted previously , interval timing researchers distinguish between pro-
spective and retrospective judgments or estimates of durations. Prospective
judgments are of experienced intervals produced by interval timing
mechanisms: e.g., you are listening to a familiar aria, and the soprano should
begin holding the long note in 3-2-1 . . . . Their close relation suggests that
duration experiences and prospective duration estimates are outputs of the
same interval timing mechanism. They are correlated because they have
a common cause in the same endogenous process subject to top-down
influences, among many others. The duration estimates are not formed on
the basis of perception, but along side perception. These estimates are dis-
tinguished from retrospective judgments, which are outputs of distinct pro-
cessing mechanisms not specific to time. In short, researchers distinguish
duration beliefs that are about the experiences of duration arising from in-
terval timing mechanisms, and duration beliefs that are not occasioned di-
rectly by interval timing at all. Of course, we can describe this processing
structure in terms that fit the routing structure behind traditional episte-
mology of perception and the problem of cognitive penetration: just call the
stored or top-down information that influences interval timing a belief. But
the duration estimate that is the outcome of interval timing mechanisms is
not identical to this belief. The perceptual belief (the estimate of the experi-
enced duration) and the stored belief are not the same. So there is no circu-
larity no matter how much top-down information influences interval timing.
Duration perception and estimation do not map onto the abstract structure
of the problem.
This analysis points to a larger moral. Most philosophers agree that the be-
havior and states of agents (or persons, in a bland sense) are to be explained
in terms of the operations of subpersonal states. But belief and perception are
personal-level states when we refer to them for epistemic purposes. The visual
system does not perceive, the person does. It is a big and not entirely coherent
step to argue from the fact that subpersonal processing explains personal-
level behavior to the idea that epistemic features attach to subpersonal pro-
cessing stages. This treats perception and belief as personal and subpersonal
at the same time. It might have been harmless to continue to ascribe epi-
stemic features directly to subpersonal states if the processing details had
turned out to be as simple as tradition holds. But in a complex information-
processing system, what is the value of distinguishing particular stages by
epistemic features when it’s the overall outcome that is epistemically good or
bad? Complex systems routinely adjust to compensate for normal variation
or change, as well as defect or insult, without any observable behavioral dif-
ference. All stages might contribute a little bit to a lousy epistemic outcome.
There are simply too many influences, from too many different sources and
functional levels, that can spoil the epistemic broth.
Thus, suppose an on-duty police officer on a call, service weapon in hand,
perceives a man’s arm movement following a pulling-object-out-of-pocket
event as the start of a taking-aim event rather than the start of a raising-hands
event, and he shoots the man. A taking-aim event reliably has a much briefer
duration than a raising-hands event in these circumstances, but which type
of event has begun at the perceived event boundary is ambiguous. So we as-
sume the officer perceives the start of an event of predictably very brief rather
than predictably somewhat longer duration: the underdetermination of the
initial motion is resolved in one way rather than another. Suppose it has been
resolved by past experience stored in the officer’s event schemata for dan-
gerous policing situations, of which this context is a token. There may or may
not also have been a detected difference in motion at the millisecond time-
scale that governs his automatic response.
Something has gone morally awry if the officer shoots an innocent man.
What is not at all clear is whether anything has gone epistemically awry
that can be pinned on his subpersonal processing. If we suppose the victim
is black, one might argue that the officer had antecedent racist beliefs that
somehow influenced his perception, perhaps by making it more likely that
this context would be classified as dangerous (or more dangerous than oth-
erwise). But racist beliefs need not impinge on the officer’s event processing
for them to play a role in his behavior, nor on the explanatory interests that
lead us to blame the officer for his response (or else excuse him, if we are also
racists). Beliefs can make us insensitive to evidence (Lyons 2011, 301). Much
of the epistemic threat blamed on cognitive penetration may instead arise
from confirmation bias and other forms of motivated reasoning, not pro-
cessing details.
Of course, there are times when we can reliably pick out a particular faulty
stage in information processing: she believed that the boulder was a big dog,
but she wasn’t wearing her glasses. But we extend these simple cases far be-
yond the point of validity. Our penchant for doing so appears to answer to
pragmatic concerns for personal responsibility and reasons-giving, not to
concerns for getting the processing details right. It hardly seems to matter
how subpersonal information processes work: in the case of the officer, via a
heightening of fear that increases the probability of trigger-pulling whatever
the current perceptual input, or via top-down influence on perceptual pro-
cessing of the current sensory input from the man’s arm movement, or via
an implicit expectation built into his activated event models, or via misap-
plied reference intervals, or via all of the above. What matters epistemically
is uncorrected, uncompensated for, or evidentially problematic learning or
evidence gathering.
Perhaps we should distinguish an epistemic role we call belief from a
subpersonal psychological state of belief that might fill the epistemic
role. In any case, once we accept that we cannot divide information pro-
cessing into intuitive routes to which subpersonal assignments of epistemic
features might be made, the epistemic problem is just that of whether the
person would have behaved better if he had better beliefs or perceptions.
The accused officer may plead that he felt he was in danger; he may say, as
a retrospective judgment of the man’s action, that he perceived the man
taking aim. Others may try to pin the bad outcome on racism-penetrated
perception. Both responses reflect the custom of assigning responsibility
to individuals, extended now to their subpersonal processes or outputs, for
when something goes wrong. Such assignments do not answer to the facts
of actual information processing. But we should not think that is their pur-
pose anyway.
8.4. Conclusion
This chapter had two main aims. First, I provided an empirically based ac-
count of our experiences of duration. My account elaborates event segmenta-
tion theory with current research on interval timing mechanisms at both the
information-processing and neural levels. Along the way I used this account
to provide an explanation of changes in quality of duration experiences and
illusory durations. I thus demonstrated its prima facie adequacy in terms of
taking both empirical research and phenomenology into account. Second,
I considered the problem of cognitive penetrability for duration perception.
The EST-based account strongly suggests that experiences of duration in-

volve top-down influences. However, I also argued that the structure of in-
terval perception and belief does not map onto the structure required by
the epistemological problem. Maybe subpersonal processing stages can be
assigned personal-level epistemic roles for the purposes of assigning respon-
sibility and reasons-giving. But we should not confuse the structure of these
assignments with the structure of the underlying processing.
Acknowledgments
I would like to thank Andy Clark, Carlos Montemayor, Nicholas Shea,

and two anonymous reviewers for very helpful comments on drafts of this
chapter.
References
Allman, M., S. Teki, T. Griffiths, and W. Meck. 2014. Properties of the internal clock: First-
and second- order principles of subjective time. Annual Reviews of Psychology
65: 743–771.
Andersen, H. 2013. The representation of time in agency. In A. Bardon and H. Dyke, eds.,
A Companion to the Philosophy of Time. Malden, MA: Wiley-Blackwell. 470–488.
Anderson, M. 2015. Mining the brain for a new taxonomy of the mind. Philosophy
Compass 10(1): 68–77.
Arstila, V. 2012. Time slows down during accidents. Frontiers in Psychology 3 (196).
https://doi.org/10.3389/fpsyg.2012.00196
Block, R., and S. Grondin. 2014. Timing and time perception: A selective review and
commentary on recent reviews. Frontiers in Psychology 5: 648. http://doi.org/10.3389/
fpsyg.2014.00648
Buhusi, C., and W. Meck. 2005. What makes us tick? Functional and neural mechanisms
of interval timing. Nature Reviews Neuroscience 6: 755–765.
Callender, C. 2010. Is time an illusion? Scientific American 302(6): 58–65.
Chuard, P. 2011. Temporal experiences and their parts. Philosophers’ Imprint 11(11): 1–28.
Casati, R., and A. Varzi. 2008. Event concepts. In T. Shipley and J. Zacks, eds.,
Understanding Events. Oxford: Oxford University Press. 31–53.
Cheng, R.-K., C. MacDonald, and W. Meck. 2006. Differential effects of cocaine and keta-
mine on time estimation: Implications for neurobiological models of interval timing.
Pharmacology, Biochemistry, and Behavior 85: 114–122.
Clark, A. 2013. Whatever next? Predictive brains, situated agents, and the future of cogni-
tive science. Behavioral and Brain Sciences 36(3): 181–204.
Dainton, B. 2014. The phenomenal continuum. In D. Lloyd and V. Arstila, eds. Subjective
Time: The Philosophy, Psychology, and Neuroscience of Temporality. Cambridge,
Davidson, D. 1969. The individuation of events. In N. Rescher, ed., Essays in Honor of Carl
G. Hempel. Dordrecht: Reidel. 216–234.
Droit-Volet, S., and W. Meck. 2007. How emotions color our perception of time. Trends in
Cognitive Sciences 11(12): 504–513.
Eagleman, D., and A. Holcombe. 2002. Causality and the perception of time. Trends in
Cognitive Sciences 6(8): 323–325.
Eagleman, D., P. Tse, D. Buonomano, P. Janssen, A. Nobre, and A. Holcombe. 2005. Time
and the brain: How subjective time relates to neural time. Journal of Neuroscience
25(45): 10369–10371.
Ebert, J., and D. Wegner. 2010. Time warp: Authorship shapes the perceived time of
actions and events. Consciousness and Cognition 19(1): 481–489.
Fodor, J. 1983. The Modularity of Mind. Cambridge, MA: MIT Press.
Friston, K., and K. Stephan. 2007. Free-energy and the brain. Synthese 159: 417–458.
Gallagher, S. 2008. Direct perception in the intersubjective context. Consciousness and
Cognition 17: 525–543.
Gibbon, J. 1977. Scalar expectancy theory and Weber’s law in animal timing. Psychological
Review 84(3): 279–325.
Gibbon, J., R. Church, and W. Meck. 1984. Scalar timing in memory. Annals of the
New York Academy of Sciences 423: 52–77.
Gibbon, J., C. Malapani, C. Dale, and C. Gallistel. 1997. Towards a neurobiology of tem-
poral cognition: Advances and challenges. Current Opinion in Neurobiology 7: 170–184.
Grondin, S. 2010. Timing and time perception: A review of recent behavioral and neu-
roscience findings and theoretical directions. Attention, Perception, and Psychophysics
72(3): 561–582.
Hecht, H. 2000. The failings of three event perception theories. Journal for the Theory of
Social Behaviour 30(1): 1–25.
Heck, R. 2000. Nonconceptual content and the “space of reasons.” The Philosophical
Review 109(4): 483–523.
Hohwy, J. 2014. The Predictive Mind. Oxford: Oxford University Press.
Hommel, G., J. Muesseler, G. Aschersleben, and W. Prinz. 2001. The theory of event
coding (TEC): A framework for perception and action planning. Behavioral and Brain
Sciences 24(5): 849–937.
Kelly, S. 2005. The puzzle of temporal experience. In A. Brook and K. Akins, eds., Cognition
and the Brain: The Philosophy and Neuroscience Movement. Cambridge: Cambridge
Kim, J. 1973. Causation, nomic subsumption, and the concept of an event. Journal of
Kononowicz, T., and V. van Wassenhove. 2016. In search of oscillatory traces of the in-
ternal clock. Frontiers in Psychology 7: 224. http://doi.org/10.3389/fpsyg.2016.00224
Kurby, C., and J. Zacks. 2007. Segmentation in the perception and memory of events.
Trends in Cognitive Sciences 12(2): 72–79.
Le Poidevin, R. 2015. The experience and perception of time. In E. N. Zalta, ed., Stanford
Encyclopedia of Philosophy. Summer 2015 ed. http://plato.stanford.edu/archives/
sum2015/entries/time-experience.
Lloyd, D., and V. Arstila, eds. 2014. Subjective Time: The Philosophy, Psychology, and
Neuroscience of Temporality. Cambridge, MA: MIT Press.
Lupyan, G. 2015. Cognitive penetrability of perception in the age of prediction: Predictive
systems are penetrable systems. Review of Philosophy and Psychology 6: 547–569.
Lyons, J. 2011. Circularity, reliability, and the cognitive penetrability of perception.

Macpherson, F. 2012. Cognitive penetration of colour experience: Rethinking the
issue in light of an indirect mechanism. Philosophy and Phenomenological Research
84(1): 24–62.
Macpherson, F. 2017. The relationship between cognitive penetration and predictive
coding. Consciousness and Cognition 47: 6–16.
Matthews, W., and W. Meck. 2016. Temporal cognition: Connecting subjective time to
perception, attention, and memory. Psychological Bulletin 142(8): 865–907.
Meck, W., and C. Malapani. 2004. Neuroimaging of interval timing. Cognitive Brain
Research 21: 133–137.
Merino- Rajme, C. 2014. A quantum theory of felt duration. Analytic Philosophy
55(3): 239–275.
Michotte, A. 1946/1963. The Perception of Causality. Trans. T. Miles and E. Miles.
London: Methuen.
Montemayor, C. 2013. Minding Time: A Philosophical and Theoretical Approach to the
Psychology of Time. Boston: Brill.
Newen, A., and P. Vetter. 2017. Why cognitive penetration of our perceptual experience is
still the most plausible account. Consciousness and Cognition 47: 26–37.
Newtson, D., G. Engquist, and J. Bois. 1977. The objective basis of behavior units. Journal
of Personality and Social Psychology 35(12): 847–862.
Phillips, I. 2014. The temporal structure of experience. In D. Lloyd and V. Arstila,
eds., Subjective Time: The Philosophy, Psychology, and Neuroscience of Temporality.
Cambridge, MA: MIT Press. 139–158.
Poppel, E. 2004. Lost in time: A historical frame, elementary processing units and the 3-
second window. Acta Neurobiologiae Experimentalis 64: 295–301.
Pylyshyn, Z. 1999. Is vision continuous with cognition? The case for cognitive impenetra-
bility of visual perception. Behavioral and Brain Sciences 22: 341–423.
Quine, W. V. O. 1985. Events and reification. In E. LePore and B. McLaughlin, eds., Actions
and Events: Perspectives on the Philosophy of Donald Davidson. Oxford: Blackwell.
162–171.
Radvansky, G., and J. Zacks. 2017. Event boundaries in memory and cognition. Current
Opinion in Behavioral Sciences 17: 133–140.
Raftopoulos, A. 2001. Neural pathways and the theory- ladenness of perception.
Philosophy of Science 68(3) Supplement: Proceedings of the 2000 Biennial Meeting of
the Philosophy of Science Association: S187–S199.
Reynolds, J., J. Zacks, and T. Braver. 2007. A computational model of event segmentation
from perceptual prediction. Cognitive Science 31: 613–643.
Schwan, S., and B. Garsoffsky. 2008. The role of segmentation in perception and under-
standing of events. In T. Shipley and J. Zacks, eds. Understanding Events. Oxford: Oxford
University Press.
Sekuler, R., A. Sekuler, and R. Lau. 1997. Sound alters visual motion perception. Nature
385: 308.
Shea, N. 2015. Distinguishing top-down from bottom-up effects. In D. Stokes, M.
Matthen, and S. Biggs, eds., Perception and Its Modalities. Oxford: Oxford University
Press. 73–91.
Shipley, T., and J. Zacks. 2008. Understanding Events: From Perception to Action.
Siegel, S., and N. Silins. 2015. The epistemology of perception. In M. Matthen, ed., The
Oxford Handbook of Philosophy of Perception. Oxford: Oxford University Press.
781–811.
Staddon, J., and J. Higa. 2006. Interval timing. Nature Reviews Neuroscience 7: 678.
https://doi.org/10.1038/nrn1764-c1.
Stokes, D. 2013. Cognitive penetrability of perception. Philosophy Compass 8(7): 646–663.
Teigen, K., G. Bohm, S. Bruckmiller, P. Hegarty, and O. Luminet. 2017. Long live the
king! Beginnings loom larger than endings of past and recurrent events. Cognition
163: 26–41.
Terhune, D., S. Russo, J. Near, C. Stagg, and R. Kadosh. 2014. GABA predicts time percep-
tion. Journal of Neuroscience 34(12): 4364–4370.
Tse, P., J. Intriligator, J. Rivest, and P. Cavanagh. 2004. Attention and the subjective expan-
sion of time. Perception & Psychophysics 66(7): 1171–1189.
Treisman, M. 1963. Temporal discrimination and the indifference interval: Implications
for a model of the “internal clock.” Psychological Monographs 77(13): 1–31.
Tversky, B., J. Zacks, and B. Hard. 2008. The structure of experience. In T. Shipley and J.
Zacks, eds., Understanding Events. Oxford: Oxford University Press. 436–464.
Vance, J. 2014. Emotion and the new challenge from cognitive penetrability. Philosophical
Studies 169: 257–283.
Vance, J. 2015. Cognitive penetration and the tribunal of experience. Review of Philosophy
and Psychology 6: 641–663.
Watanabe, K., and S. Shimojo. 2005. Crossmodal attention in event perception. In L. Itti,
G. Rees, and J. Tsotsos, eds., Neurobiology of Attention. Amsterdam: Elsevier. 538–543.
Wearden, J. 2005. Origines et developpement des theories d’horloge interne du temps
psychologique. Psychologie Francaise 50: 7–25.
Wynn, K. 1996. Infants‘ individuation and enumeration of actions. Psychological Science
7(3): 164–169.
Zacks, J. 2001. Scaling up from atomic to complex events. Behavioral and Brain Sciences
24(5): 9009–9910.
Zacks, J., N. Speer, K. Swallow, T. Braver, and J. Reynolds. 2007. Event perception: A mind/
brain perspective. Psychological Bulletin 133(2): 273–293.
Zacks, J., and B. Tversky. 2001. Event structure in perception and cognition. Psychological
Bulletin 127(1): 3–21.
Zacks, J., B. Tversky, and G. Iyer. 2001. Perceiving, remembering, and communicating
structure in events. Journal of Experimental Psychology: General 130(1): 29–58.
9
Experiencing Emotions
Aesthetics, Representationalism, and Expression
Rebecca Copenhaver and Jay Odenbaugh
9.1. Introduction
In this essay, we offer an account of the basic emotions and their expression.
On this account, emotions are experiences that indicate and have the func-
tion of indicating how our body is fairing and how we are fairing in our en-
vironment. Emotions are also objects of experience: our perceptual systems
are more or less sensitive to the expression of emotion in our environment by
features that indicate and have the function of indicating emotions. We apply
our account to expression in art. What does it mean to say that an artwork
expresses sadness? Is perceiving joy in an artwork the same kind of experi-
ence as perceiving joy in a friend’s face? How may artworks express emotions
without having emotions or any other mental states?
In the next section, we provide an overview of unrestricted representa-
tionalism about experience. In section 9.3, we offer a representationalist ac-
count of the basic emotions that combines exteroception and interoception.
On our view, emotions are perceptual experiences that represent properties
of our viscera and properties in our extra-bodily environment. Exteroceptive
and interoceptive systems combine to constitute a system whose states—
emotions—indicate how we are fairing in our environment. In the fourth
section, we survey aesthetic theories of expression in art, including the re-
semblance, persona, and arousal theories, and argue that each faces signficant
problems. Building on the work of Dominic Lopes (2005) and Mitchell
Green (2007), we offer a teleosemantic account of emotional expression in
art that is impersonal and continuous with a representationalist account of
the basic emotions. Finally, in section 9.5, we apply our view to an example—
Théodore Géricault’s The Raft of the Medusa—in order to illustrate how we
experience emotions as represented properties of a painted canvas.
Rebecca Copenhaver and Jay Odenbaugh, Experiencing Emotions In: The Epistemology of Non-Visual
DOI: 10.1093/oso/9780190648916.003.0010
We experience emotions by having them: by being happy, or frightened,

or disgusted. Emotions are also expressed—in smiles and cringes, tones and
gestures. They are expressed in artworks as well—in depicted figures and
scenes, and in design, using non-depictive elements. Because emotions are
expressed, we may experience them a way other than by having them; we
may experience them as objects: surprise expressed in a friend’s face; de-
light expressed in Mattise’s Dance. In doing so, we need not be surprised
or delighted. Indeed, the expression of emotion does not require that any
person have the emotion expressed. Your friend’s face may express surprise
even though she knew about her surprise-party. Mattise’s Dance may ex-
press delight even when no one is delighted—no one depicted or imagined
in the painting and no one viewing the painting and its expression of delight.
Whether a feature expresses an emotion—be it a smile or a design element—
depends on whether it has the function of indicating the emotion, not on the
presence of the emotion indicated.
9.2. Representationalism
We hold a representationalist—or intentionalist— view of experience.

Experiences are mental states for which there is something it is like for
the subject of the experience to have or be in that state. In other words,
experiences have a phenomenology. Experiences also convey something to
the subject of the experience. Experiences present, to the subject, things in
the subject’s environment as being a certain way, as having certain features.1
Representationalism, in the broadest sense, is a view about the relationship
between these two apparent facts about experience. Put most generally, it
holds that the phenomenology of experiences is a matter of how things are
presented to the subject in her experience. How things look, the way things
feel, what things sound like—these are presented in experience as features
of things: things look to have a certain shape, things feel to be a certain
temperature, things sound to be at a certain pitch, and so on.2 We hold an
1 Some reserve the verb “to present” for describing disjunctivist views of experience while
reserving the verb “to represent” for describing representationalist views. This is a fine distinction
with some merit. We use the terms interchangeably for what experience conveys to the subject of the
experience.
2 For a selection of standard representationalist views, see Byrne 2001; Dretske 1995; Harman
1990; Jackson 2004; Lycan 1996; Rey 1991; Thau 2002; Tye 1995, 2000, 2009. Byrne (2001) credits
Armstrong (1968) with presenting the view first in the analytic tradition.
Experiencing Emotions 215
unrestricted version of representationalism that applies to perceptual expe-

rience but also to other experiential mental state kinds: bodily sensations,
memories, and emotions (Byrne 2001, 205).
The contents of experience need not be conceptual. Adult humans have
experiences, but so do infants and non-human animals, creatures that lack
the conceptual sophistication that adult humans possess. A book can look
large and red to an infant though she may yet be incapable of forming the
perceptual belief that the book is large and red. A refrigerator may sound
rhythmic to a dog, though he is unlikely to experience it as a refrigerator
or as rhythmic. Nevertheless, the infant and the dog have experiences
with content. This is evident from the fact that their experiences can be
inaccurate—the book may look large and red to the infant when it is not, and
the refrigerator may sound low-pitched and rhythmic to the dog when it is
not. The contrast between adult humans on the one hand and infants or non-
human animals on the other illustrates how the contents of experience need
not be conceptual, but beware of drawing the conclusion that the contents
of adult human experience are conceptual, while the contents of infant and
non-human animals are not. Rather, the point is that an experience may have
content—it may present things to the subject as being a certain way without
the subject whose experience it is possessing any of the concepts used when
describing the accuracy conditions.3
The contents of experience differ from the contents of mental states such
as belief in being rich and fine-grained.4 Take Rebecca’s perceptual belief that
there is a large, red book, formed on the basis of her experience of the book
looking large and red. This experience conveys more information than is
contained in her belief. It includes information about the placement of the
book relative to other features in her visual field and relative to her perspec-
tive as an observer. It contains information about features of the book other
than its relative size and color: perhaps the words on the spine, the color of
the pages, the shadow it casts on the table, and so on. And the experience
conveys information about other objects, properties, and relations in the
environment occupied by the book. One way to express this difference in
3 May adult human experience have conceptual content? The answer depends on an adequate ac-
count of concepts and concept-possession. For a survey of standard views on the nonconceptual
content of experience, see section 4.1 of Bermúdez and Cahen 2015.
4 These features are used to support experience nonconceptualism. See Bermúdez and Cahen 2015
and Siegel 2015. We offer them here merely as observations about the phenomenology of experience
and about what information experience conveys.
informational content is as a difference between digital and analog encoding.5

We can think of beliefs as encoding information digitally: Rebecca’s belief
that the book is large and red conveys that information and no more. By con-
trast, experiences are analog. Rebecca’s experience of the book conveys in-
formation in addition to what it conveys about the book. Her experience of
the book is also more fine-grained than her belief that it is large and red. She
experiences the book not only as large, but as a determinate (if relative) size,
and a determinate shade of red (Raffman 1995).
In addition to representationalism, we hold a liberal view about the
contents of experience (Bayne 2009).6 Conservatives hold that the contents
of perceptual experience are confined to low-level properties. The contents
of visual experience are confined to color, illumination, figure, motion, while
the contents of audition are confined to pitch, volume, timbre, and so on.
Conservatives hold that the features presented in the contents of experience
are best understood as features that are visible, audible, gustable, olfactable,
and so on. By contrast, liberals hold that experience presents some features
that are not strictly visible, audible, gustable, and so on. In particular, liberals
hold that some high-level features are represented in the contents of expe-
rience: “These include being an artificial kind, being a natural kind, being a
specific individual, causation, the nature of the backsides of objects, the na-
ture of the occluded parts of objects, directionality” (Macpherson 2011, 9).
Some liberals, including us, hold that the high-level features that may
figure in the contents of experience include multimodal features (what a
strawberry looks to taste, how a bell sounds to look) as well as evaluative
features (sounding dangerous, smelling rotten, looking ill) and affective
features (looking sad, sounding angry). The liberal view is intended to be
compatible with nonconceptualism about the contents of experience: some-
thing can look to be an acorn to a squirrel; something can smell rotten to an
infant, something can sound angry to a dog.
Humans are social primates, and the basic emotions—happiness, sad-
ness, fear, anger, surprise, and disgust—lie at the center of our practical
lives. These emotions are experiences. There is something it is like for us to
have them and something they convey: something it is like to fear what is
behind you, to be disgusted by a den of snakes, to be happy with the sunlight
5 See Dretske 1981; Goodman 1976.

6 For a discussion of liberalism vs. conservatism about the contents of experience, see Bayne 2009
and the other papers collected in Hawley and Macpherson 2011. See also Siegel 2006; Siewert 1998.
on your skin, to be sad at the close of day. The phenomenology of emotions

is a matter of how things are presented to the subject in her experience. As
with other experiences, the features presented in emotional experience are
represented as features of the world: threatening movements, revolting food,
comforting embraces. But emotions are also objects of experience—they are
among the features represented in our experience: the happiness expressed
in a friend’s smile, the fear in a child’s trembling hand, the surprise in an
award-winner’s voice. We experience these emotional features as features of
the world. And, we will argue, among the experiences that present emotions
are experiences of art.
Our subject is thus twofold: emotions as representations and the representa-
tion of emotion.
9.3. Emotions as Representations
The basic emotions include happiness, sadness, fear, anger, surprise, and dis-
gust (Ekman 1992). The non-basic emotions include guilt, shame, jealousy,
envy, and so on. We focus on the basic emotions. What are emotions? We
contend that emotions are experiences.7 Because we take experiences to be
representational, we hold that emotions are representational. Emotions pre-
sent various properties and features; they are not mere sensations. In this sec-
tion, we elaborate on this proposal.
There are a variety of theories of representational content, one of which
is teleosemantics.8 On a teleosemantic account of the representational con-
tent of emotions, emotions indicate certain properties and have the func-
tion of indicating those properties. What properties do emotions have the
function of indicating? One proposal is that emotions represent evaluative
properties—specifically, relational properties concerning how we are fairing
in our environment. For example, anger tracks threat, fear tracks danger, and
sadness tracks loss.9
7 Psychologists use the term “emotion” narrowly and broadly. Narrowly, they are referring to emo-
tional experiences; broadly, they are referring to emotional experiences, physiological responses,
and behavior. We are focused on emotions narrowly construed, which is compatible with broad
approaches, provided emotional experience is at least a component of such an approach.
8 Classical discussions of teleosemantics include Millikan 1984 and Dretske 1988. For applications
of teleosemantics to the emotions see Price 2006 and Prinz 2004.

9 This is roughly the view of psychologist Richard Lazarus, who calls these features core relational
themes (Smith and Lazarus 1993; see Prinz 2004 as well.

How do emotions come to have these functions? One possibility is that

the basic emotions and their expressions evolved by natural selection to do
certain jobs. Perhaps the basic emotions contribute to reproductive success
by alerting and motivating us with regard to important features of our social
and non-social environment (Millikan 1995). Emotional expressions often
signal our commitments. Perhaps they enable us to forge alliances (Trivers
1971).10 However, we are wary of casual evolutionary hypothesizing. A trait
evolves by natural selection if, and only if, there is heritable variation in fit-
ness (Sober 1993). Too often, evidence for the relevant variability and herita-
bility is lacking. Likewise, adaptationist hypothesizing ignores the possibility
that emotions and their expressions are homologous traits rather than
adaptations, and screens off other sources of “design” such as culture (e.g.,
learning and imitation).11
Following Charles Darwin (1998), Paul Ekman (1971, 1983,
1984) argues that basic emotions form distinct families each composed
of perceptions, facial expressions, physiological responses, and behav-
ioral tendencies. And William James (1884) and Antonio Damasio (1994,
1999) hold that the emotions are experiences of characteristic bodily
states. You see a bear in a clearing: your heart rate increases, you begin to
sweat, the hair on the back of your neck stands up, and you prepare to run.
According to James, emotions consist in perceiving these changes in one’s
bodily state. He asks:
What kind of an emotion of fear would be left if the feeling neither of quick-
ened heart-beats nor of shallow breathing, neither of trembling lips nor of
weakened limbs, neither of goose-flesh nor of visceral stirrings, were pre-
sent, it is quite impossible for me to think. Can one fancy the state of rage
and picture no ebullition in the chest, no flushing of the face, no dilatation
of the nostrils, no clenching of the teeth, no impulse to vigorous action, but
in their stead limp muscles, calm breathing, and a placid face? The present
writer, for one, certainly cannot. (James 1884, 193–194)
10 The most important discussion of emotions as “commitment devices” is in Frank 1988, 2001.
This is due to the fact that expressions are “hard to fake.” Evolutionary biologists have studied the ev-
olution of costly behaviors that reliably signal phenotypic features, and some have claimed that facial
expressions are “hard to fake” signals of this sort. For detailed discussions see Ekman and Friesen
2003; Green 2007; Zahavi 1997.
11 Important recent essays on the evolution of emotions and their expressions include Panksepp
1998; Plutchik 1980; Tooby and Cosmides 1990.

We deny that emotions are simply perceptions of one’s bodily states, though
we hold that such perceptions are one component of emotions. The main
reason for rejecting a purely somatic theory such as James’s is that the phe-
nomenology of emotions is at least in part directed at the world outside our
bodies. The fear you feel in the forest clearing is very much bear-directed.
Jesse Prinz has developed a sophisticated neo-Jamesian theory of emotions
(Prinz 2004). Following James, he argues that emotions are perceptions of
one’s bodily states, including one’s heart rate, breathing, sweating, facial
expressions, and so on. These perceptions have a valence: they may be pos-
itive or negative. When you see the bear, your perception of the hairs at the
back of your neck, your sweating, and your heart rate has a negative valence.
When you see the bear retreat, your perception of your muscles relaxing,
your breath returning, and your heart slowing has a positive valence. These
perceptions are also appraisals in the sense that they represent core relational
themes: they represent how you are fairing in your environment. When you
see the bear, your body responds immediately—you freeze, you sweat, your
hair bristles, your heart rate increases, and your facial expression changes.
Your body registers the danger. Additionally, you perceive those bodily
changes. Following Dretske, Prinz argues that our bodily states indicate and
have the function of indicating such properties.12
But Prinz’s perceptual theory of the emotions is subject to a worri-
some objection. On his view, an emotion is a perception of a bodily state
that represents a core relational theme.13 But perceptions themselves
are representations. Thus, Prinz’s view commits him to the position that
emotions are representations of bodily states that are representations of core
relational themes. But transitivity is not guaranteed in intensional contexts.
Emotions may represent bodily states, and bodily states may represent the
properties in the environment that comprise core relational themes without
the emotions themselves representing these external properties. Consider
an analogy. While browsing in your local bookstore, a curious book catches
your eye: The Nematode and Nasturtium: A Story of Greed, Madness, and the
12 Walter Cannon (1927) argued that the Jamesian approach was doomed to fail since states of the
autonomic nervous system are shared across different emotions. Hence, they cannot indicate dif-
ferent emotions. However, we now know that emotions are associated with distinctive states of the
autonomic nervous system (Ekman et al. 1983).
13 “If emotions are perceptions of bodily states, they are caused by changes in the body. But if
those changes in the body are reliably caused by the instantiation of core relational themes, then
our perceptions of the body may also represent those themes” (Prinz 2004b, 55). See also Prinz
2004a, 68–69.
Race to Design the Perfect Garden.14 The book is about the lives of two rival
landscape gardeners in the 19th century (though you do not know this). You
wonder whether the book is worth reading. Your thoughts are about the
book, and the book is about rival landscape gardeners, but your thoughts
are not about rival landscape gardeners—otherwise, why read the book? So
too, your emotions may be about your bodily state, and your bodily state may
be about properties in the environment, but this does not entail that your
emotions are about properties in the environment. Prinz’s account has diffi-
culty accounting for the representational content of emotions as perceptions
directed toward the world.
We hypothesize that emotions are complex experiences that represent
both external objects and their properties, and properties regarding the
viscera; emotions are complexes composed of and integrating exterocep-
tion and interoception (Zaki et al. 2012). Interoception is a sensory system
that tracks physiological states of the body used for achieving homeostasis
(Craig 2002).15 Interoceptors include mechanoreceptors, chemoreceptors,
thermoreceptors, and osmoreceptors, each indicating specific properties of
the body. Plausibly, interoceptors evolved in order to maintain bodily ho-
meostasis. But the overall content of emotional experience is directed be-
yond the body toward the world. As Antoine Bechara and Nasir Naqvi write,
“When we feel joy on seeing someone we love, information from the viscera
is passed on to a second-order map to be re-represented in relation to an
emotional stimulus in the world” (Bechara and Naqvi 2004, 103).
The content of emotions is not merely the conjunction of the contents of
interoception and exteroception, but a content unavailable to either as an iso-
lated representational system. Emotions are an instance of multimodal per-
ception (O’Callaghan 2012). Perceptual experience is not the conjunction of
visual experience, auditory experience, tactile experience, and so on. Rather,
experience is integrated across modalities such that the contents of experi-
ence includes the representation of features that are neither confined nor
available to a single modality in isolation. When we see the bear in the forest
clearing, our visual, auditory, and olfactory experiences of the bear and our
experiences of physiological disturbances in our bodies are integrated in the
emotion fear. This emotion indicates both the properties in the world that
14 The example is fictional.

15 Whether proprioception and the vestibular system are part of interoception is subject to debate
(Ritchie and Carruthers 2015).
compose core relational themes—how we are fairing in our environment—

and the states of our bodies.
It is worth noting that on our view, James’s subtraction argument still
holds. If we subtract the perception of physiological changes in our body
from the overall content of the state, the emotion disappears. However, the
emotion also disappears if we subtract the perception of properties in the
environment from the overall content of the state. Exteroceptive and intero-
ceptive representational systems combine to constitute a system whose states
(emotions) indicate how we are fairing in the environment. How these sys-
tems come to be so combined is an empirical matter. One might hypothesize
they are merely correlated (Prinz 2004, 181). One might suppose that exter-
oceptive perceptions cause interoceptive ones (Prinz 2004, 62). Finally, one
might hold that emotional experiences are composed of and integrate exter-
oceptive and interoceptive experiences (Barlassina and Newen 2014; Goldie
2002; Tye 2008b). We assume that the problem of how these modalities bind
will be resolved by the sciences, but we believe that the integrative approach
is the most plausible.16
Traditionally, philosophers have recognized two broad positions on the
nature of emotions. Cognitivists claim that necessarily emotions contain
judgments (propositional attitudes) as parts. Non-cognitivists deny this. We
find this way of viewing the options misleading. With regard to the basic
emotions, we side with non-cognitivists for several reasons (Zajonc 1980;
LeDoux 1998). First, infants and some non-humans experience emotions
though they lack the ability to form judgments. Second, some emotions—
such as fear—are directly elicited through the amygdala, bypassing the ce-
rebral cortex, making it implausible to account for such emotions in terms
of judgment. Third, in experimental (and cultural) contexts we may induce
emotions directly absent the formation of judgments by administering
drugs. Fourth, affect-inducing images shown to subjects at speeds too fast for
recognition can induce emotional reactions absent judgment.
Still, emotions are information-carrying mental states. They convey some-
thing to the subject of the emotion. Emotions present, to the subject, the state
16 Some have worried that representationalism about emotions, and more specifically moods, fails
because they possess phenomenal character that cannot be captured in terms of representational
content (Kind 2013). One may feel “down” but think that this is not directed toward any external ob-
ject. In addition to other representationalist strategies (Crane 1998; Tye 2008b; Mendelovici 2013),
we suggest that interoception is important in our understanding of moods. For example, often when
one has an emotion, the bodily disturbance continues longer than perception of the external elicitor.
Moods then would be a subset of emotions for which the bodily disturbance is especially long lived.
of her body and things in her environment as being a certain way, as having
certain features. Moreover, emotions may misrepresent. We may feel fear
when there is no danger, be angry where there is no threat, feel sadness when
nothing has been lost. Emotions are experiences with content. They present
things to the subject as being a certain way, which way is assessable for accu-
racy. Thus, in an important sense, emotions are cognitive.
9.4. Representations of Emotion
Emotions are experiences, or so we have been arguing. But emotions are also
objects of experience. Among the high-level features that may figure in the
contents of experience are emotional features: something may sound to be
angry, look to be happy, or feel to be frightened. Emotions are experiences
that are expressed in humans (and in nonhuman animals) through various
embodied states: anger expressed in tone of voice, happiness expressed in fa-
cial expressions, fear expressed in trembling, disgust expressed in wretching,
and so on. Embodied states such as facial expressions express emotions be-
cause they indicate and have the function of indicating these emotions.
Jay’s small dog Sam is frightened of paper bags. Jay knows this because he
sees her trembling before them and feels her trembling when he picks her
up. In doing so, Jay has an experience that represents Sam as frightened. Jay
experiences Sam’s fear as a feature of the world—as a property of Sam. Jay
does not thereby experience Sam’s fear as experienced by Sam. That is, Jay’s
experience of Sam’s fear is not identical with Sam’s fear, which is also an ex-
perience. This much is clear from the fact that Jay’s experience of Sam’s fear is
not directed at paper bags, while Sam’s fear is. The point is the same as the one
presented as an objection to Prinz’s perceptual theory of the emotions: tran-
sitivity is not guaranteed in intensional contexts.
Typically, artworks do not experience emotions and thus do not express
emotions in the ordinary way that humans do.17 How then do artworks ex-
press emotions? How can an artwork express fear when it cannot be afraid?
At first sight, this is puzzling. But it is no more puzzling (or perhaps just as
puzzling) as how sentences and utterances express thoughts though they
cannot think. Traditionally, there are three approaches to this problem: the
17 We say “typically” because there are instances of performance, dance for example, in which an
artwork is a human.
resemblance, persona, and arousal theories. We consider a fourth, minimal

contour, approach with which we are sympathetic.
The first approach is the resemblance theory of expression. Stephen
Davies, a defender of the view, writes the following concerning instru-
mental (“pure”) music: “What form does it take when what is experienced is
music’s expressiveness? I believe it is an experience of resemblance between
the music and the realm of human emotion” (Davies 2011, 181). Consider
Miles Davis’s Blue in Green—what resemblance is there between it and say,
sadness? For there to be a resemblance between the two, there must be at
least one property that they share. Davies supposes that the shared prop-
erties are gestural. For example, the cadence of sad people is typically slow
and their voices are in a low register. Insofar as Blue and Green and the ges-
ture of a sad person resemble one another, the former expresses the latter’s
emotion.18 Besides the gestural similarity, the listener must experience the
resemblance. Thus:
Resemblance: an artwork expresses emotion if, and only if, there is an expe-
rienced resemblance between the work and a person’s emotion.
There are several problems with this approach. First, though some
aspects of some artworks may resemble some emotional gestures, surely
there are emotions expressed by artworks that do not resemble emotional
gestures. The scope of emotional expression in artworks appears wider
than the scope of gestural resemblance. Second, resemblance is sym-
metric, while expression is asymmetric. Suppose Rothko’s Black on Maroon
(1958) resembles and (thereby) expresses sadness. It follows that sadness
resembles Rothko’s Black on Maroon, but it does not follow that sadness
expresses Black on Maroon. Davies proposes that the expressive asymmetry
arises because we animate objects with human emotions. Black on Maroon
expresses sadness, while sadness does not express Black on Maroon because
we animate Black on Maroon with the sadness we experience it as resem-
bling. However, this adjustment threatens to account for expression pri-
marily in terms of animation, since unrestricted experienced resemblance
itself is insufficient.19
18 According to Davies, emotional predicates do not apply in their primary sense to pure music but
only in a secondary sense.

19 It should be noted that Davies thinks that the notion of animation does not commit him to the
persona theory described subsequently.

Another approach—the persona theory of expression—adopts just this

strategy. Jay’s dog Sam expresses fear when around paper bags—she is fright-
ened by them and expresses her fear by whining and trembling. Her fear is ev-
ident in these expressions. Edvard Munch’s The Scream expresses fear as well.
But The Scream is not frightened. On the persona view, Munch’s The Scream
expresses fear because it makes fear evident: it is evidence of someone’s fear, of
someone’s being frightened. The painting is animated by supposing a persona
whose fear the paining expresses. Jerold Levinson writes,
As a number of philosophers have rightly underlined, expression is essen-
tially a matter of something outward giving evidence of something inward.
Otherwise put, expression is essentially the manifesting or externalizing of
mind or psychology. (Levinson 2006, 191). On this view, though a painting
itself cannot be afraid, it expresses the fear of an imagined persona, who
is frightened. Bruce Vermazen has argued that when a thing expresses an
emotion, it is evidence of that emotion (Vermazen 1986). Expressions are
evidence of emotion because expressions indicate emotion. For example,
suppose Rebecca spills coffee on Jay’s expensive art book. Jay experiences
anger, and his furrowed brow, clenched fists, and flushed face indicate his
anger. These observable embodied states are thereby evidence of Jay’s anger.
However, artworks often do not evidence a specific person’s emotion. The
artist need not feel the emotion expressed in her work, and there need not
be a specific person, fictional or real, depicted in an artwork that expresses
an emotion. Thus, we must imagine a person—a persona—who expresses
the emotion evidenced in the work. In the context of pure music, Levinson
writes, “A passage of music P is expressive of an emotion E iff P, in context,
is readily heard by a listener as the expression of E by a persona” (Levinson
1996, 192). Thus,
Persona: an artwork expresses an emotion if, and only if, the artwork can be
imagined as evidence of a persona’s emotion.
There are two problems with this approach. First, it is possible to imagine
any number of mutually incompatible personae expressing emotions in an
artwork. A single abstract painting could then express any emotion imag-
inable, quite literally. But even the most abstract paintings that express
emotions express a particular range of emotions. Second, in order to con-
strain the limits of imagination, we would have to identify expressive qual-
ities independently of the imagined personae. But this implies that it is the
expressive qualities of the artwork that are responsible for the emotions
expressed—the personae are incidental.
Finally, let us consider arousal theories of expression. Like persona theo-
ries, arousal theories locate the emotions expressed by artworks in a person.
However, the person so located is the spectator. A painting expresses sadness,
for example, in virtue of arousing sadness in the viewer. Derek Matravers has
offered the most sophisticated arousal account:
A work of art x expresses the emotion e, if, for a qualified observer p
experiencing x in normal conditions, x arouses in p a feeling which would be
an aspect of the appropriate reaction to the expression of e by a person, or to
a representation of the content of which was the expression of e by a person.
(1998, 146) We may summarize this account as follows,
Arousal: an artwork expresses an emotion if, and only if, the artwork would
arouse that emotion (or a distinct but suitable emotion) in a qualified ob-
server in appropriate circumstances.
As Dominic McIver Lopes notes, the arousal theory is not the trival claim
that some artworks do arouse emotions in observers (Lopes 2005, 66).
Rather, according to the arousal theory, what it is for an artwork to express
an emotion is for it to arouse that emotion (or a distinct but suitable emo-
tion) in the observer. Emotional expression in artwork is constituted at least
in part by the arousal of emotions in persons observing the artwork. Note as
well that the emotion aroused need not be the emotion expressed. It may be
a distinct but suitable emotion: a painting may express fear by arousing pity
in the observer.
As with each account so far, the arousal view faces some problems. First
is the problem of dry eyes: qualified viewers in ordinary circumstances may
perceive an emotion expressed by an artwork without feeling the emotion
expressed (or a distinct but suitable emotion). Perhaps this condition is to
be lamented, perhaps not. The arousal theory holds that dry eyes are impos-
sible. It is committed to the view that an artwork cannot express an emotion,
and thus that an observer cannot perceive the emotion expressed, unless that
emotion or a distinct but suitable emotion is aroused in the observer. But
dry eyes are possible. Second, recall that the requirement that the emotion
aroused be suitable allows the arousal theory to accommodate the insight that
an artwork may express fear by, for example, arousing pity. Unfortunately,
this requirement also loosens the restrictions on which aroused emotions
are sufficient for successful expression, rendering the crucial emotional con-
tribution to expression entirely generic, or, as Lopes puts it, “The emotion
aroused is phenomenological window-dressing” (2005, 69).
We suggest a different approach. Following Lopes, we recognize three ex-
pressive devices in paintings (Lopes 2005). First, figure expression is expres-
sion attributable to a figure depicted in the painting. Degas’s The Absinthe
Drinker expresses despair because the woman in the painting is depicted as
despairing. Second, scene expression is expression attributable to the depicted
scene and not any depicted figure. Turner’s Snow Storm: Steam Boat off a
Harbor’s Mouth expresses fear though there are no figures in the scene; the
whole depicted scene expresses fear. Third, design expression (i.e., color, form,
texture) is expression attributable to the picture’s designed surface, and not
to any figure or scene (Lopes 2005, 62–68).20 Picasso’s The Weeping Woman
expresses sorrow and torment through clashing colors and discordant geo-
metric shapes. A figure and scene are depicted in The Weeping Woman, but
the designed surface expresses misery independently of its depictive content.
Our approach is similar to that offered by Lopes for representational
painting in his Sight and Sensibility. Lopes writes, “The physical configura-
tion of a picture’s design or the figure or scene a picture depicts expresses
E if and only if (1) it is an expression-look that (2) has the function, in the
circumstances, of indicating E” (2005, 78).21 Adapting Lopes’s account of
pictoral expression, we are committed to
Minimal Contour: An artwork or element of it (design, depicted figure, or

depicted scene) expresses an emotion if, and only if, it indicates the emo-
tion and has the function of indicating it.
The minimal contour theory is an impersonal account of expression in

artwork. An artwork may express an emotion without there being a person
20 Here we oversimplify. There are instances of abstract pictorial art that do contain depicted
figures or scenes. For example, Willem de Kooning’s Women I depicts a female figure, and J. M.
W. Turner’s Snowstorm depicts a steamboat. How might this be? One potential explanation for this is
that paintings can be more or less abstract. Fewer and fewer properties of an object can be depicted
in the works. Second, Richard Wollheim considers concepts like boy, dancer, and torso as figurative
and concepts like irregular solid, sphere, and space as abstract (Wollheim 1987, 62). Hence, insofar
as a spectator I am object-aware of those properties or fact-aware of those properties deploying the
associated concepts, the work is abstract.
21 The notion of an “expression-look” is unhelpful in part because that is precisely what we are
trying to analyze; when does something appear to have an emotion? Thus, we have dropped it from
the account.
or persona to whom the emotion is attributable and without there being a

person whose emotion constitutes (even in part) successful expression.
The minimal contour theory is a teleosemantic account of expression in
artwork. An artwork or an element of an artwork—its design, depicted fig-
ures, or depicted scene—need not resemble the emotions expressed. An
artwork or element of it expresses an emotion if it indicates and has the
function of indicating that emotion; it need not resemble the emotion in
order to so indicate. Indicator-functions are contingent. In the case of fa-
cial expressions and other embodied states that indicate emotion, the
mechanisms that ground indication are a matter of the evolutionary history
and environment of the species along with cultural and social conventions.
In the case of expressions in artworks, the mechanisms are often cultural and
social and far more contextual. For example, in the United States, the color
yellow is expressive of happiness, but in Egypt, it is expressive of sadness
and is associated with mourning. Embodied expressions and expressions in
artworks have a history that could have been otherwise than they are: smiles
could have indicated aggression, and dissonant colors could have indicated
delight. As Lopes writes, “No single factor explains what gives . . . phys-
ical configuration the function of indicating one or another emotion. . . .
Nevertheless, it is their having emotion-indicating functions that makes
them expressions” (2005, 81).
The minimal contour theory of expression in artworks is continuous
with a teleosemantic account of expression in embodied states: anger
as expressed in tone of voice, fear as expressed in trembling, and so on.
Embodied states express emotions because they have the function of indi-
cating these emotions. Artworks and their elements—design, depicted
figure, depicted scene—express fear, sadness, anger, and other emotions be-
cause they have the function of indicating these emotions. In neither case
does the expression of emotion require the emotion expressed to be pre-
sent. Your friend’s smile expresses happiness even as she struggles in si-
lent sorrow; Picasso’s Weeping Woman expresses sorrow even as no woman
weeps. Whether a feature expresses an emotion is a matter of it having the
function of indicating the emotion, not a matter of the presence of the emo-
tion indicated.
Our view is inspired by Lopes, and it is useful to juxtapose it with another
theory similar to his: the acccount developed by Mitchell Green in his book
Self-Expression. According to Green, an emotion is expressed if and only if
that emotion is signaled and shown. We signal emotions by functionally
indicating them as Lopes suggests. We show emotions in one of three ways: we

show-that, show-α, and show-how. One shows-that one is angry by providing
evidence that allows another to perceive-that one is angry. Jay’s email with its
terse wording and no signoff allows Rebecca to form a perceptual belief that
Jay is angry. One shows-α by making one’s anger perceptible. Jay’s clinched
fist, lowered eyebrows, and tight, straight lips make his anger perceptible.
Finally, one shows-how one’s anger by enabling or inacting in another the
ability to be empathetically aware of one’s anger. In seeing Jay’s expression of
anger, Rebecca may come to know how he feels by feeling anger as well, empa-
thetically. We think Green’s notion of showing is suggestive but that it makes
expression more opaque than it need be. As with Lopes’s notion of expresssion-
look, the notion of showing simply stands for that which functionally indicates
an emotion and is availabe to the senses such that a person may form percep-
tual beliefs, perceive, or be empathically aware of another’s emotion.
One objection to the minimal contour approach is the following. If ex-
pression is functional indication, then some abstract paintings represent
emotions in virtue of expressive devices that have the function of indicating
emotions. However, abstract expressive paintings are not representational.
Therefore, expression is not functional indication. In response, we would
like to make two points. First, abstract artists like Mark Rothko and Adolf
Gottlieb were adamant that their paintings had subjects. They famously
commented in their abstract expressionist “statement,”
It is a widely accepted notion among painters that it does not matter what
one paints as long as it is well painted. This is the essence of academism.
There is no such thing as good painting about nothing. We assert that
only that subject matter is valid which is tragic and timeless. (Rothko and
Gottlieb 1943)
Rothko also wrote,
I’m not an abstractionist. I’m not interested in the relationship of color

or form or anything else. I’m interested only in expressing basic human
emotions: tragedy, ecstasy, doom, and so on. (Quoted in Rodman 1957, 93)
According to these abstractionists their paintings have a subject: in this

instance, the emotions. Second, the earlier objection by modus tollens
equivocates between two different senses of being representational:
Rd: a painting represents sadness by a depictive property (figure or scene).

Ra: a painting represents sadness by a design property.
Non-representational (abstract) paintings may represent emotions. They do

not represent by way depicted figures or scenes, but they represent nonethe-
less, via design properties that functionally indicate emotions. We claim that
Rothko’s Black on Maroon is representational in the latter sense.
9.5. Experiencing Emotions
Finally, let us apply our view to an example—Théodore Géricault’s The

Raft of the Medusa—in order to illustrate how we experience emotions as
representated properties of a painted canvas.
Suppose two art afficianados, Rebecca and Jay, find themselves at the
Louvre. Rebecca is a fan of classical art, whereas Jay is enamored with
modern art. Contrary to his impulses, they head to the first floor and stand
before Théodore Géricault’s The Raft of the Medusa (1818–1819; figure 9.1),
which she very much admires.
Jay is a novice with respect to classical art, but he immediately perceives
several things. He perceives the facial and gestural expressions of the raft’s
passengers: the sadness of the figure on the bottom who hovers over near-
dead fellow passengers, the surprise of the figures on the right waving to the
distance below the darkened mast, the fear of the figures in the center. Jay
experiences these emotions: they are represented in the contents of his experi-
ence as represented properties of the figures and scene depicted in the painting.
He experiences sadness, surprise, and fear, even as he is not sad, surprised, or
frightened. However, Rebecca knows there is more that Jay may perceive in the
The Raft of the Medusa than this, so she offers Jay a lesson in art history.
Rebecca tells Jay about the events on which the painting is based. In 1816
the French Royal Navy frigate Méduse set sail for Senegal, captained by an
officer who had not sailed for 20 years and who ran the ship aground. The
ship had few lifeboats, so 146 of the 400 passengers were forced to build an
impromptu raft: the raft of the Méduse. They drifted for nearly two weeks,
during which those on the raft starved, fought, and resorted to cannabilism.
Fifteen remained when the raft was rescued by mere chance by the Argus.
Next Rebecca describes the artistic skills Géricault brings to the canvas,
realized after Géricault interviewed two passengers, prepared models, and
Figure 9.1 Théodore Géricault’s The Raft of the Medusa, 1818–1819
drew sketches in preparation for the final painting. Jay is already sensitive
to the figure expression in the paining, but Rebecca highlights the scene
and design expression. She draws Jay’s attention to the diagonal movement
from the dead bodies on the bottom left to those waving in hope at the ho-
rizon on the upper right. She points to the speck in the upper right corner,
where Jay notices a small ship in the distance, a detail he had overlooked.
Where Jay perceived despair and surprise in distinct depicted figures, he now
perceives an emotional transformation from sadness and despair to hope—
hope rising from loss. Jay perceives in the painting the temporal unfolding
of emotions over the 13 days the passengers awaited their fate. Rebecca asks
Jay to consider the color palette used by Géricault: dark tones that are natu-
rally or conventionally expressive of negatively valenced emotions such as
sadness and fear. Géricault uses chiaroscuro for the passengers, clouds, and
water: a somber design. And she asks him to consider the size of the painting.
It is enormous: 491 cm × 716 cm (16′1″ × 23′6″). The figures at the base
of the painting are larger than those who would view the painting—larger
than Jay and Rebecca—a design that renders the emotions expressed urgent,
impending, looming.
In talking with Rebecca, Jay has learned some history, but he also learned
to perceive in The Raft of the Medusa emotions and features to which he was
initially blind. He has learned something about how to approach artworks, to
be sure, but he has also become more sensitive to the features in his environ-
ment that indicate emotions. He has gained propositional knowledge, but he
has also gained recognitional and discriminatory abilities that alter the kinds
of representational contents that figure in his experiences of emotion, rend-
ering those possible contents richer, more fine-grained, and discerning. He
has learned facts, and he has made new judgments, but he has also learned
something about how to look and what to look for, as well as what sorts of
things he might perceive in such looking.
Recently, Dominic Lopes (2011) has argued that pictures are “percep-
tual prostheses” for developing empathic skill. Pictures evoke experiences
as of the scenes they represent. However, their contribution to empathetic
skill is different from face-to-face viewings of scenes, and the skills they de-
velop carry over to life outside pictures. Pictures are different precisely be-
cause they are able to express emotion impersonally. The skills developed
by experiencing pictures such as The Raft of the Medusa carry over be-
cause pictures socially reference which emotions are appropriate in various
circumstances. By seeing the grief-stricken faces in The Raft of the Medusa,
we determine which emotional responses are warranted. The artistic value of
expressive pictures consists, in part, in their unique ability to provide us with
opportunities to develop empathic skills.
The theories of expression we have considered as alternatives to ours ex-
plain what is going on with artworks like this painting in the Louvre dif-
ferently. The resemblance theory holds that Jay and Rebecca experience
a resemblance between the human figures depicted and the gestures of ac-
tual humans, but it has difficulty explaining how scene and design expres-
sion contribute to what Jay and Rebecca experience when they perceive the
painting. The persona theory holds that Jay and Rebecca experience the
emotions expressed by The Raft of the Medusa by make-believe: they imagine
personae that have the emotions expressed. The arousal view holds that Jay
and Rebecca experience the emotions expressed by The Raft of the Medusa
by having those emotions themselves, or by having emotions suitable to
witnessing misery on such a scale. But there need be no persons—real or
imagined—who have the emotions expressed in The Raft of the Medusa in
order for the painting to express those emotions, or for Jay and Rebecca to
perceive them. And while Jay and Rebecca may have emotions upon per-
ceiving the emotions expressed in the painting, they need not. They may ex-
perience the emotions expressed in The Raft of the Medusa without thereby
having those emotions or having emotions suitable to the knowledge of what
the passengers of the raft endured.
9.6. Conclusion
On our representationalist account of the basic emotions, emotions are

representations. Emotions are experiences that indicate and have the func-
tion of indicating how our body is fairing and how we are fairing in our en-
vironment. And on our representationalist account of experience, emotions
are among the features presented in the contents of experience. Emotions
are objects of experience. Among the features that figure in the contents of
experience are features that express emotions by indicating and have the
function of indicating emotions. Thus there are two senses in which we may
be said to experience emotions: we may have emotions, and we may experi-
ence emotions as representated properties of the environment. In both cases,
experiencing emotions is a matter of experiencing how things are in the
world: experiencing how one is fairing in the environment, or experiencing
emotional features as features of the environment.
We examined one way of experiencing emotional features: emotional ex-
pression in art. Artworks and their elements express fear, sadness, anger, and
other emotions because they have the function of indicating these emotions.
In this way, our account of emotional expression is continuous with our rep-
resentationalist account of the basic emotions. We experience the emotions
expressed in artworks not by having those emotions, but by perceiving them.
We perceive them they same way we perceive the emotions expressed by the
embodied states of humans (and nonhuman animals): by being attuned to
properties that indicate emotions. In neither artworks nor humans does the
expression of emotion require the emotion expressed to be present. Whether
a feature expresses an emotion is a matter of it having the function of indi-
cating the emotion, not a matter of the presence of the emotion indicated.
In this way, our account of emotional expression is impersonal. Features in
the environment (including artworks and their elements) express emotions
even in conditions in which there is no person to whom the emotion is
attributable.
Acknowledgments
We would like to thank Dominic Lopes and Kent Bach for their very helpful
comments on earlier drafts.
References
Armstrong, D. M. 1968. A Materialist Theory of the Mind. London: Rouledge and
Kegan Paul.
Barlassina, L., and A. Newen. 2014. The role of bodily perception in emotion: In defense of
an impure somatic theory. Philosophy and Phenomenological Research 89(3): 637–678.
Bayne, T. 2009. Perception and the reach of phenomenal content. Philosophical Quarterly
59(236): 385–404.
Bechara, A., and N. Naqvi. 2004. Listening to your heart: Interoceptive awareness as a
gateway to feeling. Nature Neuroscience 7(2): 102–103.
Bermúdez, J., and A. Cahen. 2015. Nonconceptual mental content. In E. N. Zalta, ed.,
The Stanford Encyclopedia of Philosophy. Spring 2015 ed. http://plato.stanford.edu/
archives/spr2015/entries/content-nonconceptual/.
Byrne, A. 2001. Intentionalism defended. Philosophical Review 110(2): 19–239.
Cannon, W. B. 1927. The James-Lange theory of emotions: A critical examination and an
alternative theory. American Journal of Psychology 39: 106–124.
tion of the body. Nature Reviews Neuroscience 3(8): 655–666.
Crane, T.. 1998. Intentionality as the mark of the mental. In A. O’Hear, ed., Contemporary
Issues in the Philosophy of Mind. Cambridge: Cambridge University Press. 229–252.
Damasio, A. 1994. Descartes’ Error: Emotion, Reason and the Human Brain. New York:
Random House.
Damasio, A. 1999. The feeling of what happens. Nature 401(6756): 847.
Darwin, C. 1998. The Expression of the Emotions in Man and Animals. Oxford: Oxford
University Press.
Davies, S. 2011. Infectious music: Music-listener emotional contagion. In A. Coplan and
P. Goldie, eds., Empathy: Philosophical and Psychological Perspectives. Oxford: Oxford
Dretske, F. 1981. Knowledge and the Flow of Information. Cambridge, MA: MIT Press.
Dretske, F. 1993. Conscious experience. Mind 102(406): 263–283.
Dretske, F. 1988. Explaining Behavior: Reasons in a World of Causes. Cambridge: Cambridge
University Press.
Ekman, P. 1984. Expression and the nature of emotion. In K. Scherer and P. Ekman, eds.,
Approaches to Emotion. Hillsdale, NJ: Lawrence Erlbaum. 319–343.
Ekman, P. 1992. An argument for basic emotions. Cognition & Emotion 6(3–4): 169–200.
Ekman, P., and W. V. Friesen. 1971. Constants across cultures in the face and emotion.
Journal of Personality and Social Psychology 17(2): 124.
Ekman, P., and W. V. Friesen. 2003. Unmasking the Face: A Guide to Recognizing Emotions
from Facial Clues. Malor Books.
Ekman, P., R. W. Levenson, and W. V. Friesen. 1983. Autonomic nervous system activity
distinguishes among emotions. Science 221(4616): 1208–1210.
Frank, R. H. 1988. Passions within Reason: The Strategic Role of the Emotions.
New York: Norton.
Frank, R. H. 2001. Cooperation through emotional commitment. In R. M. Nesse, ed.,
Evolution and the Capacity for Commitment. New York: Russell Sage Foundation. 57–76.
Goldie, P. 2002. Emotions, feelings and intentionality. Phenomenology and the Cognitive
Sciences 1(3): 235–254.
Goodman, N. 1976. Languages of Art. 2nd ed. Indianapolis: Hackett.

Green, M. S. 2007. Self-Expression. Oxford: Oxford University Press.
Hawley, K., and F. Macpherson, eds. 2011. The Admissible Contents of Experience. Malden,
MA: Wiley-Blackwell.
Jackson, F. 2004. Representation and experience. In H. Clapin, P. Slezack, and P.
Staines, eds., Representation in Mind: New Approaches to Mental Representation.
Amsterdam: Elsevier. 107–124.
James, W. 1884. What is an emotion? Mind 34: 188–205.
Kind, A. 2013. The case against representationalism about moods. In U. Kriegel, ed.,
Current Controversies in the Philosophy of Mind. London: Routledge. 113–134.
LeDoux, J. 1998. The Emotional Brain: The Mysterious Underpinnings of Emotional Life.
New York: Simon and Schuster.
Levinson, J. 1996. Musical expressiveness. In Jerrold Levinson, ed., The Pleasures of
Aesthetics. Ithaca. 90–125.
Levinson, J. 2006. Contemplating Art: Essays in Aesthetics: Essays in Aesthetics. Oxford:
Lopes, D. 2005. Sight and Sensibility: Evaluating Pictures. Cambridge: Cambridge
University Press.
Lopes, D. 2011. An empathic eye. In Amy Coplan and Peter Goldie, eds., Empathy:
Philosophical and Psychological Perspectives. Oxford: Oxford University Press. 118–133.
Lycan, W. 1996. Consciousness and Experience. Cambridge, MA: MIT Press.
Macpherson, F. 2011. Introduction: The admissible contents of experience. In K.
Hawley and F. Macpherson, eds., The Admissible Contents of Experience. Malden,
MA: Wiley-Blackwell. 1–15.
Matravers, D. 1998. Art and Emotion. Oxford: Clarendon Press.
Mendelovici, A. 2013. Pure intentionalism about emotions and moods. In U. Kriegel, ed.,
Current Controversies in the Philosophy of Mind. London: Routledge. 135–157.
Millikan, R. G. 1984. Language, Thought, and Other Biological Categories: New Foundations
for Realism. Cambridge, MA: MIT Press.
Millikan, R. G.1995. Pushmi- pullyu representations. Philosophical Perspectives
9: 185–200.
O’Callaghan, C. 2012. Perception and multimodality. In E. Margolis, R. Samuels, and S.
P. Stich, eds., The Oxford Handbook of Philosophy of Cognitive Science. Oxford: Oxford
Panksepp, J. 1998. Affective Neuroscience: The Foundations of Human and Animal
Emotions. Oxford: Oxford University Press.
Plutchik, R. 1980. A general psychoevolutionary theory of emotion. Emotion: Theory,
Research, and Experience 1(3): 3–33.
Price, C. 2006. Fearing fluffy: The content of an emotional appraisal. In G. MacDonald
and D. Papineau, eds., Teleosemantics. Oxford: Oxford University Press. 208–228.
Prinz, J. 2004a. Gut Reactions: A Perceptual Theory of Emotion. Oxford: Oxford
University Press.
Prinz, J. 2004b. Embodied emotions. In R. E. Solomon, ed., Thinking about Feeling:
Raffman, D. 1995. On the persistence of phenomenology. In T. Metzinger, ed., Conscious
Experience. Paderborn: Ferdinand-Schoningh. 293–302.
Rey, G., 1991. Sensations in a Language of Thought. Philosophical Issues 1: 73–112.
Ritchie, J. B., and P. Carruthers. 2015. The bodily senses. In M. Matthen, ed., The Oxford
Handbook of Philosophy of Perception. Oxford: Oxford University Press. 353–370.
Rodman, S. 1957. Conversations with Artists. New York: Devin-Adair.
Rothko, M., and A. Gottleib. 1943. Manifesto. New York Times, June 13.
Siegel, S. 2010. The Contents of Visual Experience. Oxford: Oxford University Press.
Siewert, C. 1998. The Significance of Consciousness. Princeton, NJ: Princeton
University Press.
Smith, C. A., and R. S. Lazarus. 1993. Appraisal components, core relational themes, and
the emotions. Cognition & Emotion 7(3–4): 233–269.
Sober, E. 1993. The Nature of Selection: Evolutionary Theory in Philosophical Focus.
Chicago: University of Chicago Press.
Thau, M. 2002. Consciousness and Cognition. Oxford: Oxford University Press.
Tooby, J., and L. Cosmides. 1990. The past explains the present: Emotional adaptations
and the structure of ancestral environments. Ethology and Sociobiology 11(4): 375–424.
Trivers, R. L. 1971. The evolution of reciprocal altruism. Quarterly Review of Biology
46(1): 35–57.
Tye, M. 1997. Ten Problems of Consciousness: A Representational Theory of the Phenomenal
Mind. Cambridge, MA: MIT Press.
Tye, M. 2000. Consciousness, Color, and Content. Cambridge, MA: MIT Press.
Tye, M. 2008a. Consciousness Revisited: Materialism without Phenomenal Concepts.
Tye, M. 2008b. The experience of emotion: An intentionalist theory. Revue Internationale
de Philosophie 243(1): 25–50.
Tye, M. 2009. Consciousness Revisited: Materialism without Phenomenal Concepts.
Vermazen, B. 1986. Expression as expression. Pacific Philosophical Quarterly 67(3):
196–224.
Wollheim, R. 1987. Painting as an Art. London: Thames and Hudson.
Zahavi, A. 1997. The Handicap Principle: A Missing Piece of Darwin’s Puzzle. Oxford:
Zajonc, R. B. 1980. Feeling and thinking: Preferences need no inferences. American
Psychologist 35(2): 151.
Zaki, J., J. I. Davis, and K. N. Ochsner. 2012. Overlapping activity in anterior insula during
interoception and emotional experience. Neuroimage 62(1): 493–499.
10
The Emotional Dimension
to Sensory Perception
Lana Kühle
10.1. Introduction
Suppose you had a long, tiring, and very stressful day at work. You aren’t in
a very good mood—you’re irritable, tense, and anxious. You arrive home,
open the front door, and immediately hear your two toddlers laughing,
screaming, and banging away at some pots and pans as they happily play in
the kitchen. The sounds grate on your already ragged nerves—they are loud
and high-pitched and resonate deeply. Contrast this case with the following
variation: you had a nice, relaxing day—you had coffee with a close friend,
followed by a full-body massage—and you’re feeling calm, rested, relaxed,
and happy. You arrive home, open the front door, and immediately hear your
two toddlers laughing, screaming, and banging away at some pots and pans
as they happily play in the kitchen. The sounds make you smile—they are
loud, but not high-pitched and resonating in a painful way; rather they are
pleasant and warm, inviting even.
Call the first case the Grating Sound Case (GSC) and the second case the
Pleasant Sound Case (PSC). The sounds produced by the children are the
same,1 yet the perceptual experience of these sounds is very different. The
experiences differ insofar as the auditory experience is affected by the emo-
tional state that you’re in.2 Put another way, your emotional state affects how
you perceive the world. This observation is not as trivial as it seems at first
blush. In most instances we are not fully aware of the emotional state that
we’re in, nor of the influence that such a state has on our perception of the
1The same amplitude and frequency are stimulating your auditory receptors.
2Or so this is the claim I’m defending here. For an alternative account on how emotions affect per-
ception, see Aydede and Fulkerson 2014 and Fulkerson 2014.
Lana Kühle, The Emotional Dimension to Sensory Perception In: The Epistemology of Non-Visual
DOI: 10.1093/oso/9780190648916.003.0011
The Emotional Dimension to Sensory Perception 237
world. In GSC, I experience the sounds being made by the children as being
of a certain quality—loud, high-pitched, resonating. In PSC, I experience
the sounds as being different. What differs is not the sounds, though, but
a state of my being. If states of my being, such as emotions, can affect how
I perceive the world, and I’m not fully aware of which emotional states I’m
in or how they affect my sensory perception, then we have a problem on our
hands. We rely on sensory perception to generate beliefs about the world
and justify claims to knowledge on the basis of these beliefs, yet there is an
influence on our sensory perception that we are unaware of or unclear about
that affects these epistemic moves. Which states affect sensory perception,
how they do so, and what consequences these effects have for the justifi-
cation of perceptual beliefs are the three main questions at the root of the
issue. These are large questions that require careful treatment that is beyond
what I can accomplish here. My focus instead will be to give a rudimentary
map of the landscape of discussion—the kinds of non-cognitive states that
can affect sensory perception, how we might understand the influence they
have, and what this influence means for our epistemological concerns.3
In recent years there’s been a strong move away from vision-centric
approaches to understanding sensory perception and a turn to the other ex-
teroceptive modalities—audition, touch, olfaction, and gustation. One result
from this move away from vision has been a broader view of what constitutes
sensory perception. First, once we move away from vision as the paradigm
sense, we find that defining a sense and distinguishing one sense from another
are no easy tasks. Moreover, in broadening our view we find that considering
the senses in isolation from one another isn’t true to how we experience the
world. The senses work together, and our sensory perception results from
multisensory integration. Indeed, it appears that our sensory perception is
influenced not only by what each sense brings to the table but also by other
mental states, cognitive states in particular. To be sure, the non-vision-centric
approach complicates the discussion, but clearly the discussion needs to be
broadened. All of these considerations move us forward in understanding
sense perception, and there is still far more to be explored. However, there is
one clear area of investigation that’s been left untouched by those concerned
to look beyond vision. As mentioned, the move away from the vision-centric
3 My focus in what follows will be on conscious perception. There is certainly an equally important
and similar discussion to be had about unconscious perception and what influences it succumbs to,
but I set that discussion aside here.
paradigm involved turning to the other exteroceptive senses. Accordingly,

there has been very little consideration of the non-exteroceptive elements
to sensory perception—interoception and proprioception. Proprioception,
broadly construed, involves our sense4 of balance and bodily position in
space—the sense that you are now upright and that your legs are beneath
your arms, for example. Interoception, broadly construed, is our sense of
the inner, visceral body—feeling hungry, tired, ill, for example. I’m not here
making a claim as to whether these are senses in their own right—the debate
is ongoing as to what constitutes a sense modality. But regardless of whether
we grant interoception and proprioception “sense” status, there remains very
good reason to include them in our discussion of sensory perception. If I’m
sick with the flu and a fever, my sensory perception of the world is affected by
my body’s interoceptive state. As we will see, interoception is also associated
with emotions—in particular with the feeling of an emotion. Given the GSC
and PSC previously considered, this gives us further reason to take a serious
look at the non-exteroceptive senses.
My focus here will be with interoception. Specifically, I consider how
interoception influences perceptual content and what consequences this in-
fluence has for issues of perceptual justification. There are two main types
of influence on sensory perception: cognitive penetration and multisen-
sory integration. First, I’ll show there is a clear argument to be made for
interoception being an influence on sensory perception. Second, I’ll consider
reasons for thinking it might be best understood as a form of penetration,
and then reasons for thinking it might be best understood as a sense mo-
dality in its own right, and thus a form of multisensory integration. However,
regardless of which type of influence it turns out to be—penetration or mul-
tisensory integration—we remain faced with epistemological consequences.
So I’ll end with a look at what those consequences are.
10.2. Types of Influence on the Content

of Perception
For ease of discussion, I assume here a representational view of perception,

though nothing that I will discuss hinges on this view. Accordingly, the con-
tent of perception is the content represented in a perceptual state, usually
4 I’m using the term “sense” here loosely and only to refer to what we are aware of on the basis of
proprioception. I’m not making any claims about its status as a sense modality.
as a result of a sensory interaction with the environment. For example, if

I hear a bird chirping, then the content of my sensory perception will be the
represented sounds of a bird chirping. My perceptual experience is said to be
veridical if indeed my auditory system picked up the chirping of a bird and
thus correctly represented the environment as involving a bird chirping. I’ll
return to considerations of veridicality in section 10.4. For now it suffices
to understand the content of perception as the way the environment is per-
ceptually represented as being. The content of perception is typically com-
plex and influenced in a number of ways. Broadly speaking, there are two
ways in which the content of perception can be influenced: non-perceptually
and perceptually. The paradigmatic form of a non-perceptual influence is
cognitive penetration. In short, cognitive penetration is when a cognitive
state penetrates the perceptual representation in such a way as to affect the
represented content of the perception.
What we experience isn’t always solely constituted by the sensory stim-
ulus we receive. Often the contents of our perceptual experience are shaped
by other mental states—most commonly, cognitive states. Take Siegel’s ex-
ample of a sheet of Cyrillic script. To someone who doesn’t read Russian, it
will appear as a sheet containing various meaningless symbols. However, to
someone who can read Russian, the symbols will appear different—they will
make sense and convey meaning. The perceptual experience will differ for the
reader of Russian, even though the perceptual stimulus remains the same for
both subjects. The difference in perceptual experience results from the subject’s
cognitive state—here, an understanding of what Cyrillic script means—pen-
etrating the content of perception in a way that changes how the content is
represented. Note that it’s not that the reader of Russian has the same percep-
tual experience as the non-reader but then deploys her knowledge to interpret
the perceptual content. Rather, the reader of Russian sees the scripts differently.
Her perceptual experience is causally influenced by her cognitive state.
There are all sorts of potential cognitive penetrators: beliefs, hypotheses,
knowledge, desires, and so on. Importantly, for the influence to be a case of
cognitive penetration, the influencing state must be a non-perceptual state
that lies outside the perceptual system.5 Moreover, the difference in percep-
tual experience between the reader and the non-reader of Russian cannot be
the result of a difference in attention, sensory organ, or stimulus—proximal
or distal. That’s to say that if the case is to be made for the difference being due
5 Silins 2016, 24. If it were a perceptual state then we’d be dealing with multisensory integration
instead.
to cognitive penetration, then all other variables need to be held the same, else
the difference could be attributable to a difference in one of these variables.
To date, the best approach to defending cognitive penetration is Siegel’s (2007)
phenomenal contrast method. As Stokes explains, in using this method one
will “consider a pair of perceptual phenomena that contrast in some impor-
tant way (two perceptual experiences with apparently contrasting phenomenal
character, two contrasting perceptual reports, two distinct actions in response
to the same perceptible stimulus) and then infer some hypothesis about per-
ception on the basis of its best explaining the contrast” (2014, 6).
If we return to our case of Cyrillic script, we have a pair of perceptual
experiences with apparently contrasting phenomenal character. Both subjects
are related to the Cyrillic text in the same way, yet their experiences differ. “[If]
two people are the same with respect to their sensory inputs, the state of their
sensory organs, and the orientation of their attention, and they are still different
with respect to what their experience is like, [then, by inference to the best ex-
planation, it’s] because of their beliefs, desires, or other cognitive states” (Silins
2016, 27). In other words, the only difference between the subjects is a differ-
ence in reading knowledge of Russian. This difference in cognitive states, then,
is the best explanation for the difference in perceptual experience. Cognitive
penetration involves a causal relation between a cognitive state (C) and a per-
ceptual state (E) such that “if C did not occur (antecedent to E), then E would
not occur. Thus, the phenomenal character of one’s visual or auditory or other
perceptual experience depends non-trivially upon a background belief, desire,
or other cognitive state” (Stokes 2013, 650).
The important takeaways about cognitive penetration at this point are the
following:
a. One internal type of influence on perceptual content is a non-

perceptual causal influence, typically from a cognitive state.
b. The best method by which to determine if a difference in phenomenal
character or perceptual content is due to this type of influence is by
phenomenal contrast.
c. The phenomenal contrast argument runs as follows: if two subjects can
have perceptual experiences with different contents even though they
are seeing and attending to the same object under the same external
conditions, then the best explanation for the difference is that it’s the
result of penetration from a non-perceptual—commonly a cognitive—
state (Siegel 2012).
As mentioned, there are two types of influence, non-perceptual and percep-

tual. The paradigmatic perceptual influence on perceptual content comes
from multisensory integration. For a long time, the prevailing view of per-
ception was that it is unisensory. To be sure, we don’t experience from a
single modality at a time, but the idea is that each modality generates its own
perceptual content and these come together to occur co-consciously, causing
our complex sensory perception. Perception, in short, is modality specific.
As Nudds put it, the prevailing view was that perception simply is the “sum
total of what each sense alone provides” (2001, 224). Giving an account of
how perception works in each modality would thereby give us a complete
account of perception. Recent work in neuroscience and interesting expe-
riential cases have put pressure on this view. Perception appears to involve
more than just different senses working alongside each other. Instead, cases
like the McGurk effect (McGurk and McDonald 1976), ventriloquism effect
(Bertelson and de Gelder 2004), parchment skin effect (Guest et al. 2002),
rubber hands illusion (Botvinick and Cohen 1998), and so on push us toward
a multisensory view of perception.6
Multisensory perception involves a representational integration across
different sensory modalities. What each sense brings to the table may differ
from what ends up being experienced.7 This is because the sensory infor-
mation is integrated—the experience associated with one sense is shaped by
input from another sense, and, thus, the sensory information is not merely
co-consciously presented. The integration is not merely the collection of
multisensory content, but the creation of one multimodal representation.8
Take, for example, the McGurk effect. You perceive a video of someone
mouthing [ga], along with a soundtrack playing the sounding of [ba]. When
you are presented with both of these simultaneously, you experience the
person saying [da]. However, if you perceived each of these inputs sepa-
rately—just the video or just the audio—you’d perceive either [ga] or [ba],
respectively. “The response produced by [both] of the senses differs from that
which would have occurred had the two stimuli not been presented together”
6 See Bayne and Spence 2015 for a discussion of these cases in relation to multisensory perception.
7 At this point I’m not committed to any particular account of how best to distinguish the senses.
Depending on which view you take, your interpretation of which senses bring what to the table may
differ. For my purposes it suffices to use broad, common sense ways of thinking about the senses as
I’m only here proposing that perceptual content is influenced by multisensory integration.
8 As one delves more deeply into the multisensory account, one can find variant degrees to which
sensory input is integrated—O’Callaghan in his 2015 paper describes 6 distinct types of multisensory
perception, each specifying a different level of integration. But, it is beyond my purposes here to con-
sider that discussion.
(Bayne and Spence 2015, 7). If perception were unisensory, then we should
have a co-conscious, simultaneous experience of [ga] and [ba]. But this is not
what happens. Indeed, the experience is of an entirely different sound—[da].
There is no [da] in either the auditory or visual stimuli; [da] arises out of the
combined experience of the auditory and visual stimuli. The conclusion is
that this is due to multisensory integration, and that the integration of mul-
tisensory stimuli beings forth a novel element in the perceptual content.9 “In
multisensory integration, the processing of input in one (or more) sensory
modality (modalities) is sensitive in content-respecting ways to information
about stimuli that have been registered in another sensory modality” (Bayne
and Spence 2015, 7).
The argument for multisensory integration, again, relies on drawing a
phenomenal contrast. There is a phenomenal difference between what the
content of perception would be in each individual sense, and what it is in the
integrated experience. As O’Callaghan explains it, the lesson here is the fol-
lowing: when you have consciously perceptible feature instances and feature
types that could not be perceptually experienced through the use of indi-
vidual sense modalities working on their own or simply in parallel, then this
can only be explained by perception being richly multimodally integrated
(O’Callaghan 2015, 2).
The important takeaways about multisensory integration at this point are
the following:
a. One internal type of influence on perceptual content is a perceptual in-

fluence, typically from multisensory integration.
b. The best method by which to determine if a difference in phenomenal
character or perceptual content is due to this type of influence is by
phenomenal contrast.
c. A phenomenal contrast argument might run as follows10: if the per-
ceptual content of a multisensory experience cannot be fully reduced
and explained by the perceptual content of each participating sensory
modality, then there is something in addition to what each sense mo-
dality brings to the table that produces the multisensory content of the
9 Note that there are different models of integration, and not all take it to involve blending.
10 I suggest here one phenomenal contrast argument, but there may be more. Which phenomenal
contrast argument one prefers will depend on the view taken with respect to how we ought to indi-
viduate the senses, and on the view taken with respect to what mechanisms are at play in multisen-
sory integration. What I argue for here doesn’t hinge on which phenomenal argument I make, but
only on the ability to make such an argument.
experience. The addition must be the result of the integration of the sen-
sory stimuli.
What should become clear here is that a similar argument, a phenomenal

contrast argument, is deployed for each type of influence on the content
of perception: there is some relevant phenomenal contrast that cannot be
explained by the external signal—as this is the same in each instance—and so
our best explanation for the difference is an influence by an internal process,
cognitive state, or input from another set of transducers.11 If there is a differ-
ence in the content of experience, then, using this argument, we can show
that the difference is the result of an influence on perception. Once we’ve de-
termined that an influence is in play, we need only ascertain whether it is
a non-perceptual influence—and thus cognitive penetration—or a percep-
tual influence—and thus multisensory integration. I now consider whether
interoception is an influence on perception by means of a phenomenal con-
trast argument and then evaluate if it is best counted as a non-perceptual or
perceptual influence.
10.3. Interoception and the Content of Perception
10.3.1. Is Interoception an Influence?
One of the goals for this chapter is to consider whether interoception is an

influence on the content of perception. To do this, I’ll consider two cases and
see whether there is a phenomenal contrast between them that reflects an in-
fluence by interoception. But first we must look closer at what interoception
involves.
As mentioned, interoception is the sense12 of the inner, visceral body. It
comprises a variety of receptors—pain, chemical, stretch, temperature, and
so on—and monitors the state of the body, seeking to maintain the body
in a state of health. Digestion, pain, fatigue, feeling hot or cold, breathing,
heart rate, and so on are all aspects of interoception—turn your gaze to
the workings of the visceral body and you’ll be turning your gaze toward
interoception. Emotions are in large part interoceptive states; they involve the
11 I thank Mohan Matthen for help in formulating the argument.

12 I’m using the term “sense” here loosely and only to refer to what we are aware of on the basis of
proprioception. I’m not making any claims about its status as a sense modality.
body.13 An emotion is a complex state that comprises various elements. Take,

for example, an instance of anger. Suppose I suffer from road rage because
I’ve just been cut off by another driver. I immediately feel certain changes in
my body: I get excited and worked up, my face flushes, my heart races, I feel
a tightness in my chest. I express my anger with some choice words and hand
signals. I understand that I’m angry—I can categorize what I’m feeling as
anger, rather than sadness, say. What we see here are the three main elements
of an emotion: (1) bodily changes, (2) behavioral changes, and (3) cognitive
understanding.14 All three elements form an emotion. Simply stating the
same choice words and making the same hand signals on their own would
not put me in an emotional state of anger. Similarly, thinking about anger or
recalling a previous experience of anger will not, in itself, make me feel angry.
The behavioral and cognitive components are part of an emotion, to be sure,
but they do not form the feeling of an emotion. If I want to feel angry—if
I want to experience an emotion—then I need to feel the bodily state changes
that are associated with that emotion. As William James put it,
Without the bodily states following on the perception, the latter would be
purely cognitive in form, pale, colourless, destitute of emotional warmth. . . .
Can one fancy the state of rage and picture no ebullition of it in the chest, no
flushing of the face, no dilation of the nostrils, no clenching of the teeth, no
impulse to vigorous action, but in their stead limp muscles, calm breathing,
and a placid face? . . . If we fancy some strong emotion, and then try to ab-
stract from our consciousness of it all the feelings of its characteristic bodily
symptoms, we find we have nothing left behind, no “mind-stuff ” out of
which the emotion can be constituted, and that a cold and neutral state of
intellectual perception is all that remains. (James 1884, 190–194)
Insofar as the feeling of an emotion involves these bodily state changes, then
the feeling aspect of an emotion experience involves interoception.15
13 For many, this is a bold and contentious claim. To be clear, it is still hotly debated whether
emotions are interoceptive at all. I, following others such as Damasio (1999), Craig (2002, 2003,
2008), Prinz (2003, 2004), and so on, take emotions—in part—to be interoceptive states, as I go on to
explain.
14 There is much discussion that remains about how these work together, and how these elements
related to normative and motivational aspects associated with emotions. I set these discussions aside
for now.
15 As mentioned, this is a contentious claim. I refer the reader to a growing area of research in the
neurosciences that looks directly at interoception and the emotions—of particular importance is the
work of Craig (2002, 2003, 2008) and Damasio (1999).
Let us now return to the two cases stated at the outset:
Grating Sound Case (GSC): you’ve had a long, tiring, and very stressful day
at work. You aren’t in a very good mood—you’re irritable, tense, and anx-
ious. You arrive home, open the front door, and immediately hear your two
toddlers laughing, screaming, and banging away at some pots and pans as
they happily play in the kitchen. The sounds grate on your already ragged
nerves—they are loud, high-pitched, and resonate deeply.
Pleasant Sound Case (PSC): you had a nice relaxing day—you had coffee
with a close friend followed by a full-body massage—and you’re feeling
calm, rested, relaxed, and happy. You arrive home, open the front door,
and immediately hear your two toddlers laughing, screaming, and banging
away at some pots and pans as they happily play in the kitchen. The sounds
make you smile—they are loud, but not high-pitched and resonating in a
painful way; rather they are pleasant and warm, inviting even.
Let us assume that in both cases all other sensory stimuli at the moment you
enter the door are the same, for example, same body position in the doorway,
same visual input of the foyer, same non-emotional physical state (such as
hunger, pain, etc.), same olfactory stimulus, and so on. For simplicity, let us
focus on two elements of the experience in both cases, namely the auditory
stimulus and the emotional state. In both GSC and PSC the auditory stimulus
is exactly the same. Your attention to the auditory stimulus is also the same.16
What differs is the emotional state that you’re in. All else being equal, you’re
in a negative mood in GSC (stressed), and you’re in a positive mood in PSC
(happy). Now there is clearly a phenomenal contrast between the cases: in
GSC you perceive the sounds as loud, high-pitched, and resonating—in short,
the sounds are unpleasant. In PSC you perceive the sounds as having a normal
loudness and pitch—in short, the sounds are pleasant. But the difference is
a bit more complex than simply a difference in emotional state. The sounds
are perceived differently. That’s to say, the environmental input of amplitude
16 To be sure, there is a complex and important discussion to be had on the role attention plays in
shaping the content of perception. Moreover, given that attention is an influence on perception, we
must be sure that the influence under consideration is not one that can be reduced to a difference in
attention. However, for the purposes of the analysis here I set aside the discussion of what role atten-
tion plays in shaping the content of the perceptual experience, and I assume that in both cases your
attention is the same, i.e. your attention shifts to and is focused on the same sensory stimulus when
you open your front door: the sound of the kids playing. That is not to say that attention doesn’t play a
role here, but simply that the role it plays will not account for the difference that I’m interested in.
and frequency on the auditory receptors is the same, but the auditory experi-
ence varies in that the amplitude is experienced as louder and the frequency
as higher-pitched in GSC. You don’t hear the sounds objectively and then in-
terpret them consciously as one way or another depending on your mood.
Rather, the moment you perceive the sounds, they are experienced as having
the phenomenal characteristics of being loud, high-pitched, and resonating
(GSC), or of normal loudness and pitch (PSC).17 It is an immediate difference
in the content of experience, and one that produces a phenomenal difference.
Among all the contributing factors to the content of experience in each of
these cases there is one that varies between GSC and PSC: the emotional state
that you’re in. The emotional state you’re in is the only differing contributing
factor and clearly influences the content of your perception.
As noted, the phenomenal contrast argument holds that if two perceptual
experiences differ and the difference is not explained by looking at each in-
dividual contributing sensory modality in turn, or by a difference in atten-
tion, external conditions, or sensory organs, then you have a genuine instance
of influence. In GSC and PSC, my attention, my sensory organs, and the ex-
ternal conditions are all the same, yet my perceptual experience differs. If we
consider interoception as a contributing factor, then there is a clear difference
in the contributed content insofar as in GSC interoception contributes the
feeling of stress, and in PSC it contributes the feeling of happiness. But the dif-
ference between the two cases isn’t simply a difference in emotional feelings,
but in the auditory experience. The auditory content differs, and not because
of a difference in auditory stimulus.18 All else being equal, the best explana-
tion for the difference in the auditory experience must be that interoception
has influenced the perceptual content. What the nature of that influence is re-
mains to be determined; however, it seems clear that whichever way you look
at it, interoception is a causal influence on the content of perception.
Granted, this is only a theoretical analysis, and one might want more ro-
bust empirical evidence to support it. However, on the face of it, this isn’t a
stunning claim and there is much anecdotal evidence to support it—our
common metaphors are rife with emotion/ perception connections: the
warmth of anger, the coldness of being lonely, the weight of sadness, the gray
17 See Fulkerson’s entry in this volume— in particular, the BOMB and LUTEFISK cases
discussion—for a nice argument distinguishing an affective character to experience that’s the result
of background knowledge or other conditions, and an affective character that’s primarily sensory.
(See also Fulkerson 2016.) In the cases I consider here, the affective element is a sensory one.
18 Recall that the environmental stimulus to your auditory receptors—amplitude and frequency—
are the same in both cases.

hue of depression, and so on.19 Beyond anecdotal evidence, there is growing

empirical evidence that also supports the claim that there is a clear influence
of emotions on the content of conscious perception. Depression, for example,
causes alterations in sensory perception on many levels (Fitzgerald 2013). In
a 2010 study, Bubl et al. show that depression leads to visual alterations, in
particular to a significant reduction in contrast gain, which correlates with
the subjective reports that when one is depressed everything looks darker or
gray. Depression has also been shown to alter auditory perception (Schwenzer
et al. 2012) and touch perception (Adler and Gattaz 1993). Beyond depres-
sion, anxiety has been shown to affect color contrast perception, in particular
processing along the red-green spectrum (Hosono et al. 2014). Feelings of
love make distilled water taste sweeter (Chan et al. 2013), and other emotions
and moods have been shown to affect local versus global attention, the per-
ception of elements in the environment—hills appear steeper if you’re sad—
and reduce the effect of visual illusions (Zadra and Clore 2011). In short, our
emotional states are contributing factors to our perceptual experience of the
world—they are clear influences on the content of our perception.
Importantly, though, it is not only via the feeling of emotions that
interoception has been shown to influence sensory perception. There is also
empirical evidence showing that various other interoceptive states affect how
we perceive the world. As Zadra and Clore review (2011), being in a state of
thirst will cause you to perceive a glass of water as taller, being in a state of
nicotine craving will cause you to perceive the cigarette as longer, and being
in a state of fatigue will cause you to perceive the distance between you and
point B as longer or a hill as steeper. In short, it is clear that interoception, via
emotions or other bodily states, has a causal influence on sensory perception.
Traditionally, the study of perception has been quite distinct from the study
of emotion. Psychologists have tacitly viewed perception, cognition, emo-
tion, and other basic processes as separable phenomena to be studied in
isolation. Increasingly, however, we are coming to see relevant areas of the
brain and the processes they support as highly interactive. . . . Not only is it
possible for emotion to influence perception, but in fact it seems to happen
quite frequently—across many levels of [sensory] perception and in re-
sponse to a variety of affective stimuli. (Zadra and Clore 2011, 10)
19 See the work of Lakoff and Johnson (1980) for a discussion on the relation between embodiment
and perception as portrayed through linguistic metaphor.

The answer to our question—is interoception an influence?—is clearly yes.

We know, and have a wealth of empirical evidence to show, that a perceptual
stimulus can affect our moods or emotions. But it’s now time to grant the
reverse causal influence—the effect of our moods and emotions on our per-
ceptual experience. With good reason to accept that there is genuine causal
influence by interoceptive states on the content of perception, I explore the
next key question: what kind of influence is it?
10.3.2. What Kind of Influence Is Interoception?
As considered in section 10.2, there are two main types of influence on per-
ception: cognitive penetration and multisensory integration. I now look, in
turn, at reasons for understanding interoception’s influence as one or the
other. Note that it is beyond the scope of this chapter to argue for a definitive
position on this issue. Instead I set myself the task of laying out a preliminary
landscape for ways to argue one view or the other.
First, let’s look at reasons to consider interoception’s influence a form of
penetration, akin to cognitive penetration. There is already precedent in
taking emotions to be cognitive penetrators. As Siegel opens her 2012 paper,
“It is sometimes said that in depression, everything looks grey. If this is true,
then mood can influence the character of perceptual experience: depending
only on whether a viewer is depressed or not, how a scene looks to that
viewer can differ even if all other conditions stay the same. This would be
an example of cognitive penetration of visual experience by another mental
state” (201). It is intuitive to think of emotions as cognitive states, and it is
intuitive to accept that moods and emotions influence our perceptual experi-
ence. Most who discuss cognitive penetration accept emotions as one among
many potential cognitive penetrators. But as I have described, emotions are
much more than cognitive states—they are also felt. Moreover, we often feel
an emotion well before we are able to cognitively grasp it—this is especially
clear in the case of background, long-standing moods. I often don’t realize
that I’m stressed or anxious until well after I’ve begun feeling stressed or
anxious. To be sure, my cognitive grasp of my mood can go on to influence
my perception. However, prior to my understanding what mood I’m in, the
feeling of stress or anxiety affects my perception. In this way, it might not
be cognitive, but nevertheless is a penetrator—call it an interoceptive pen-
etrator. Opening the discussion of penetration to include non-cognitive
penetrators is one way to integrate the evidence discussed regarding the ef-
fect of interoception on perceptual experience, whether it be via emotions
or various bodily states. And this might be appealing to those who don’t
endorse the claim that interoception is a sense modality because penetra-
tion accounts for the evidence of interoception’s influence on perception by
counting it a non-perceptual influence. Although there are good preliminary
reasons to consider interoception’s influence on perception a form of pen-
etration, there are also preliminary reasons for thinking it is better under-
stood as part of multisensory integration.
Recall that for an influence on perceptual experience to be part of mul-
tisensory integration it needs to be a perpetual influence. That is to say that
integration occurs across sense modalities, and thus only sense modalities
participate in integration.20 Thus, we must first see whether there are any
reasons to accept that interoception is a sense modality.
It must be stated that we do not yet have a definitive account of what makes
something a sense or how to clearly distinguish the senses.21 However, al-
though the details are still to be worked out, there are two clear lines along
which we would distinguish between the senses: the anatomical/functional
level and the experiential level. At the anatomical/functional level there is
good reason to suggest that interoception is its own sense. Interoception
comprises various types of receptor, each specific to one kind of stimuli
within the body.22 In primates, interoception has been shown to be strictly
associated with specific anatomical structures.23 At the experiential level
there is also good reason to suggest that interoception is its own sense. There
20 There are proposal that something needn’t be a sense to be part of multisensory integration—see
work by Campos et al. (2012). I set this aside for now but thank Matthew Fulkerson for pointing this
out to me.
21 There are a few proposals, though, the common ones being the following: Gray (2005), to
use typical criteria: a. by the proper objects of the senses (colors, sounds, etc.), b. by the distinc-
tive characters of experience, c. by the physical stimulus (light waves, sound waves, etc.), d. by the
distinctive organs. Fulkerson (2014), to use feature binding: a collection of sensory subsystems that
function to assign a unique set of qualitative features to individual objects. Keeley (2002), “to possess
a genuine sensory modality is to possess an appropriately wired-up sense organ that is historically
dedicated to facilitating behavior with respect to an identifiable physical class of energy.” Bayne and
Spence (2015), the senses differ in terms of the nature of their input, and the bodily organs and neural
pathways involved in processing that input, and others such as Matthen 2015.
22 There are chemoreceptors (chemical stimuli), nociceptors (pain stimuli), mechanoreceptors
(pressure stimuli), thermoreceptors (temperature stimuli), visceroreceptors (visceral organ stimuli),

to name a few.
23 There is a direct sensory pathway (lamina I) that provides the primary interoceptive representa-
tion of the physiological condition of the entire body and that projects to the insular cortex, and the
somatosensory cortices I and II. See Craig 2002, 2003, 2008.
is a clear, distinctive perceptual experience of the inner body. I can feel nau-
seous, I can feel an urge to run to the bathroom, butterflies in my stomach
before a first date, depressed, angry, and so on. These experiences are not at-
tributable to any of the standard five senses. They are experiences of a distinct
environment, with its own distinct experiential properties and states—the
visceral body.
In short, although it is still contentious whether to consider interoception a
proper sense, given that we don’t yet have an agreed-upon view of how to dis-
tinguish the senses, and given that there is reason to claim that interoception
is a sense along anatomical, functional, and experiential lines, I will grant it
“sense” status for now and consider what this means in relation to multisen-
sory perception. As Bayne and Spence grant:
Although uncertainty about how to individuate the senses “problematizes”

the discussion of multisensory perception, we believe that it is possible to
make some progress here even without a full account of how the senses are
to be individuated. . . . Arguably, the task of providing a taxonomy of the
senses ought to proceed in tandem with the task of giving an account of the
multisensory nature of perception. And if that’s right, then one’s model of
multisensory perception ought to constrain—and, in turn, be constrained
by—one’s taxonomy of the senses. (2015, 3)
The model we have is that multisensory perception occurs when perceptual

content cannot be clearly divided and reduced to the sensory input from each
participating sense. When the experience associated with a sense is shaped
by input from another sense. As stated, the feeling of emotions just is the
feeling of the inner body and certain sensations and state changes therein.
Importantly, by “emotion” we not only mean explicit, strong emotions such
as anger, sadness, and joy, but also more subtle emotions and background
moods, such as anxiety, tension, stress, and depression. If this is so, then the
feeling of an emotion is an interoceptive perceptual experience insofar as it
is the perpetual experience of the visceral body state changes associated with
that particular emotion. Moreover, we are never in an emotion-less state of
being. We may not be experiencing a strong, explicit emotion, but we are al-
ways in some state of implicit, subtle emotion or mood. Emotions affect how
we experience the world—recall GSC and PSC. Instances such as these are
explainable on the multisensory account of perception if we include intero-
ceptive input into the integration account. In short, if interoception is a sense
modality, then we can make sense of the influence it has on perception by

including it in our accounts of multisensory integration.
In any case, given that interoception is a causal influence on perception, it
will either turn out to be a form of penetration or a part of multisensory inte-
gration. Regardless of which one it turns out to be, it can no longer be ignored
in discussion of perception and discussion of what shapes our perceptual ex-
perience. In light of this, and because discussions of cognitive penetration
and multisensory integration have had epistemological repercussions, I end
by turning to concerns having to do with the epistemology of perception that
arise as a result of adding interoception to the perceptual picture.
10.4. Interoception and the Epistemology

of Perception
Both discussions of cognitive penetration and multisensory integration

bring up concerns about perceptual justification. These concerns only grow
if we add interoception to either of those discussions. The main concern has
to do with the veridicality of our perceptual representations. If the content of
perception doesn’t accurately represent the environment we are perceiving,
then it is no longer a reliable source of information on which to build and
justify our beliefs. This is problematic, as most theories of perceptual justifi-
cation take it that, absent defeaters, if you experience seeing an apple—that
is, the content of your visual experience is “there is an apple”—then you’re
justified in believing that there’s an apple.
It’s important to note that not all cases of influence on the content of per-
ception are bad. As Siegel explains, some cases of cognitive penetration can
be epistemically beneficial: “If an x-ray looks different to a radiologist from
the way it looks to someone lacking radiological expertise, then the radiol-
ogist gets more information about the world from her experience (such as
whether there’s a tumor) than the non-expert does from looking at the same
x-ray” (2012, 201). However, there are clear cases where the influence on per-
ception is bad and leads to circular justification.
We’ve all had something like this happen: I’m angry and irritable. I ask
my friend if she can help me with something. She responds no. I experience
her response as slightly aggressive and I infer that she must be annoyed or
upset with me. I take my belief to be justified by my perceptual experience of
her response. However, her response, objectively speaking, is not aggressive,
I simply perceived it to be because of the emotional state that I was in. “Your
own mind sometimes unwittingly causes you to experience the world to be
the way you antecedently believed or expected it to be. . . . If your experi-
ence is influenced in this way, it’s not so clear whether your experience is in a
good position to support your belief ” (Silins 2016, 24). This is precisely a case
of the negative epistemic outcome that certain influences on perception can
have. It can lead to a circular justification where your belief in p causes you to
see the world as p, which in turn justifies your belief in p.
The problem isn’t resigned to instances of cognitive penetration; the same
epistemic concerns come forth in multisensory integration as well. Recall
the McGurk effect. The content of the perceptual experience is a feature that
is not in the world. The visual stimulus [ga] and auditory stimulus [ba] each
on its own accurately represents what is produced in the environment, but
the content of perception—what one experiences the world as being—turns
out to be something entirely different. The belief formed on the basis of this
perception will be that the person said [da], and one will take himself to be
justified in this belief on the basis of his perceptual experience.
The problem only gets complicated by adding interoceptive influence to
the picture. Emotions are pervasive. We are never in a completely neutral
state; rather our conscious lives are always valenced in some way—however
slight or strong. This makes sense given that emotions lend an affective and
motivational dimension to experience. Whatever is perceived is perceived
by an embodied subject that is in a particular bodily state. If, as in the case
of PSC, I’m tired and stressed, stimulating sensory input would not be wel-
come. So there is, at least on the face of it, good reason why I perceive the
sounds made by the kids as being a certain way—loud, piercing, grating—a
way that will motivate me to move away from them. But, although this makes
intuitive sense, it is epistemically problematic as well. If I’m suffering from
severe depression, my perception of the world will be influenced in a manner
that is not to my advantage—my perception will be skewed.
These issues suggest that threats to the standard view of perception and
its role in justifying our beliefs and generating knowledge are not only rel-
egated to atypical cases like the McGurk effect. Instead, the threat is perva-
sive—as pervasive as emotions are. If this is correct, how are we to handle
the problem? “Our knowledge about the world will, ultimately, be based
on experience, with perception providing the terminus in the long chain of
reasons that support one’s putative knowledge about the world. However,
if perception can be infected with [multisensory integration,] background
beliefs and other cognitive states, then the supposed epistemic role of per-
ception is threatened. . . . Experience cannot justify a belief or provide
knowledge” (Stokes 2013, 651).
Perhaps we need to rethink how to rationally ground perception and per-
ceptual justification. The current picture of perception—a relation between
sensory systems and the environment whereby any further thoughts resulting
from the systems’ representation of the environment are justified only if the
representation accurately and objectively mirrors the environment—needs
revision. Interoception brings to the table the affective and motivational
elements to perception. Maybe, then, the picture of perception and percep-
tual justification would be more rational if taken as a relation between an
embodied agent and its environment via its sensory systems, and an accurate
representation is one that accounts for what the environment means to the
agent. That is to say that, depending on the agent, the states it is in, and the
environment, meaning will be fluid rather than objective.
To be sure, this suggestion doesn’t resolve the epistemic issues I’ve
highlighted here. But it might move us in a more fruitful direction. Thus far,
we’ve wanted perception to remain objective and unbiased—therein lies its
rational grounding. Perception is supposed to provide knowledge, of both
the everyday and scientific kinds, and if our epistemic practices are to remain
rational, they must give accurate and unbiased information about the world.
But maybe we need to broaden our understanding of accuracy and avoid re-
jection of any subjectivity. In some sense, the X-ray technician is biased, but
he is biased in a way that has meaning for him and his work, and has benefits
for those whose X-rays he’s analyzing. As conceded, not all influences on the
content of perception are negative. Although interoceptive influences might
at first seem to distort perception in a way that challenges perceptual justifi-
cation, there is a reason emotions pervade our lives, and they may turn out to
be beneficial influences in most instances—and even strengthen perceptual
justification.
In the end, I don’t have solutions for the epistemological issues that arise
from the discussion here. But by turning toward them I hope to avoid falling
into the same problems that we take previous views to have fallen prey to. As
Bayne and Spence put it in the context of their discussion of multisensory per-
ception, “Although philosophers, psychologists, and neuroscientists have tra-
ditionally taken what can be characterized as a unisensory approach to the
study of perception, it is increasingly clear that such an approach leaves us with
a view of perception that is at best partial and at worst positively distorted”
(2015, 17). I fear we are repeating the same mistake by focusing our attention
only on the exteroceptive senses in discussing what shapes the content of per-
ception and the consequences this has for perceptual justification—leaving
us “with a view of perception that is at best partial and at worst positively dis-
torted.” I have shown that there is another important factor in the generation
of perceptual content—interoception—and that it is now time we take it seri-
ously in our discussions of perception and epistemology.
Acknowledgments
I am very grateful to Clare Batty, Matthew Fulkerson, and Mohan Matthen

for guidance and comments that greatly helped shape the arguments in this
chapter.
References
Adler, G., and W. F. Gattaz. 1993. Pain perception threshold in major depression.
Biological Psychiatry 34: 687–689.
Aydede, M., and M. Fulkerson. 2014. Affect: Representationalists’ headache. Philosophical
Studies 170(2): 175–198.
Bayne, T., and C. Spence. 2015. Multisensory perception. In M. Matthen, ed., The Oxford
Bertelson, P., and B. de Gelder. 2004. The psychology of multimodal perception. In C.
Spence and J. Driver, eds., Crossmodal Space and Crossmodal Attention. Oxford: Oxford
Botvinick, M., and J. Cohen. 1998. Rubber hands “feel” touch that eyes see. Nature
391: 756.
Bubl, E., E. Kern, D. Ebert, M. Bach, and L. Tebartz van Elst. 2010. Seeing gray when
feeling blue? Depression can be measured in the eye of the diseased. Biological
Psychiatry 68: 205–208.
Campos, J. L., J. S. Butler, and H. H. Bülthoff. 2012. Multisensory integration in the esti-
mation of walked distances. Experimental Brain Research 218(4): 551–565.
Chan, K. Q., E. M. W. Tong, D. H. Tan, and A. H. Q. Koh. 2013. What do love and jealousy
taste like? Emotion 13(6): 1142–1149.
tion of the body. Nature Reviews Neuroscience 3: 655–666.
Craig, A. D. 2003. A new view of pain as a homeostatic emotion. Trends in Neuroscience
26(6): 303–307.
Craig, A. D. 2008. Interoception and emotion: A neuroanatomical perspective. In M.
Lewis, J. M. Haviland-Jones, and L. Feldman Barrett, eds., The Handbook of Emotions.
3rd ed. New York: Guilford Press. 215–234.
Damasio, A. 1999. The Feeling of What Happens: Body and Emotion in the Making of
Consciousness. Orlando: Harcourt Books.
Fitzgerald, P. J. 2013. Gray colored glasses: Is major depression partially a sensory percep-
tual disorder? Journal of Affective Disorders 151: 418–422.
Fulkerson, M. 2014. The First Sense: A Philosophical Study of Human Touch. Cambridge,
MA: MIT Press.
Fulkerson, M. 2016. Affective touch from a philosophical standpoint. In H. Olausson, J.
Wessberg, I. Morrison, and F. McGlone, eds., Affective Touch and the Neurophysiology
of CT Afferents. New York: Springer. 323–340.
Guest, S., C. Catmur, D. Lloyd, and C. Spence. 2002. Audiotactile interactions in rough-
ness perception. Experimental Brain Research 146: 161–171.
Gray, R. 2005. On the concept of a sense. Synthese 147(3): 461–475.
Hosono, Y., K. Kitaoka, R. Urushihara, H. Séi, and Y. Kinouchi. 2014. Anxiety affects the
James, W. 1894. The physical basis of emotions. Psychological Review 1(5): 516–529.
Keeley, B. L. 2002. Making sense of the senses. Journal of Philosophy 99: 5–28.
Lakoff, G., and M. Johnson. 1980. Metaphors We Live By. Chicago: University of
Chicago Press.
Matthen, M. 2015. The individuation of the senses. In M. Matthen, ed., The Oxford
McGurk, H., and J. W. MacDonald. 1976. Hearing lips and seeing voices. Nature
264: 746–748.
Nudds, M. 2001. Experiencing the production of sounds. European Journal of Philosophy
9: 210–229.
O’Callaghan, C. 2015. The multisensory character of perception. Journal of Philosophy
112(10): 551–569.
Prinz, J. 2003. Emotional Perception. New York: Oxford University Press.
Prinz, J. 2004. Embodied Emotions. In R. C. Solomon, ed., Thinking about Feeling:
Schwenzer, M., E. Zattarin, M. Grözinger, and K. Mathiak. 2012. Impaired pitch identifi-
cation as a potential marker for depression. BMC Psychiatry 12: 32.
Siegel, S. 2007. How can we discover the contents of experience? Southern Journal of
Silins, N. 2016. Cognitive penetration and the epistemology of perception. Philosophy
Compass 11(1): 24–42.
Stokes, D. 2013. Cognitive penetrability of perception. Philosophy Compass 8(7): 646–663.
Stokes, D. 2014. Cognitive penetration and the perception of art. Dialectica 68(1): 1–34.
Zadra, J. R., and G. L. Clore. 2011. Emotion and perception: The role of affective informa-
tion. Wiley Interdisciplinary Review of Cognitive Science 2(6): 676–685.
11
The Perception of Virtue
Jennifer Matey
11.1. Introduction
It is good to be virtuous. But there are different types of virtues. They differ
in the types of ends they are valued relative to. For example, whereas com-
passion is widely taken to be a moral virtue in terms of which one might be
judged to be a good person, open-mindedness is valued for making one a
good knower. Other virtues, like persistence, can’t easily be categorized as
either moral or epistemic because they make us both good persons and good
knowers. I’m interested in a different kind of case in which virtues count as
both moral and epistemic. In the case I am interested in, virtues that are more
often thought to be moral virtues turn out to also play a special kind of ep-
istemic role. Having good character, I want to show, can make one a better
perceiver and in turn a better knower.
The specific type of knowledge that I’m interested in is knowledge about
another person’s character. Cultivating virtuous moral character can make us
better perceivers of the moral character of others. One way to be a better per-
ceiver of something involves being better at seeing it in terms of the qualities
that are distinctive of it qua the type of thing that it is. Virtues are character
traits that are set apart from other character traits at least in part by the fact
that they have positive value. Insofar as a trait like courage is a virtue, then,
part of what it is to be courageous is to have a valuable trait. Here I argue that
emotions play a perceptual or quasi-perceptual role in representing virtue
and that the structure of our character helps us to have the right emotional
responses to the virtuous character of others.
The argument is developed in four stages. Section 11.1 presents the view
that emotions give us a kind of perceptual or quasi-perceptual access to eval-
uative properties. Section 11.2 discusses the specific category of esteeming
emotions, which some philosophers have argued represent that toward
which they are felt as having positive value. Some studies show that we
Jennifer Matey, The Perception of Virtue In: The Epistemology of Non-Visual Perception. Edited by:
Berit Brogaard and Dimitria Electra Gatzia, Oxford University Press (2020). © Oxford University
Press.
DOI: 10.1093/oso/9780190648916.003.0012
The Perception of Virtue 257
esteem people when we recognize that they have virtuous character traits.
Assuming that virtuous character traits are traits that have positive value, in
esteeming in respect of good character we represent those character traits to
be virtues. Section 11.3 grounds this in a realist account of value that is con-
sistent with a naturalist account of the perception of that value. Sections 11.4
and 11.5 discuss the epistemic role that the perceiver’s character can play in
the ability to perceive virtue.
11.2. Emotion and Perception
Emotion and perception share enough relevant similarities that some pro-
pose emotion to be a type of perceptual or quasi-perceptual state.1 According
to the perceptual view of emotion, just as vision is a modality sensitive to
color and hearing for sound and pitch, emotion is a modality sensitive to
evaluative properties, properties that evaluate in a normative, prudential,
or aesthetic way. Some might find this initially counterintuitive since there
isn’t a specific perceptual organ devoted to emotion. And unlike perceptual
experiences in other modalities such as vision, emotions often depend on
other perceptual experiences. Being afraid of a spider in the shower depends
on your experience of the spider. Nonetheless, perception and emotion share
some important commonalities, and whether or not we want to call them
perceptions proper, emotions are similar enough in the relevant ways to play
an epistemic role akin to perceptual experiences.
For instance, emotions are intentional states, which is to say that emotions
represent what they are directed toward as being some way. They are assess-
able for accuracy. A common view among philosophers of emotion is that
emotions represent what they are directed toward as having some evalua-
tive property. Fear evaluates what it is felt toward as being dangerous, disgust
evaluates its object as corrupt. But we typically recognize just two catego-
ries of intentional mental states: perceptions and cognitive states (beliefs,
judgments). And in terms of their functional profiles, emotions resemble
perceptions a lot more than they do cognitive states.
1 For defense of the perceptual theory of emotion see Deonna 2006; Doring 2007; Johnston 2001;
Prinz 2004; Tye 2008; Goldie 2009.

The fact that emotions more closely resemble perceptual than cog-
nitive states can be brought out by considering the category of irrational
emotions.2 Irrational emotions are those that persist even while we have
beliefs that the object they are directed toward does not have whatever eval-
uative property the emotion represents the object to have. Phobias are one
example. The arachnophobe may continue to feel terrified of an ordinary
house spider despite knowing that the house spider poses no threat. And
this is similar to the way that, in the Muller-Lyer illusion, our visual experi-
ence of the two lines as being unequal lengths persists despite the fact that
we know the lines are the same length. If emotions were belief-like rather
than perception-like, we would expect irrational emotions to resolve once
we come to the contrary belief that the object doesn’t actually have the eval-
uative property that we experience it to have. In fact, emotions do persist
despite changes in evidence and background beliefs, at least far more fre-
quently than beliefs do.
Emotions also seem to play the same role as perceptual experiences in
causing evaluative beliefs, as well as in providing their contents. For instance,
moral beliefs are often accompanied by motivations to act in certain ways.
This makes sense if moral beliefs are in many cases caused by emotions since
emotions, but not beliefs, are thought to be inherently motivational. We
wouldn’t expect moral beliefs that aren’t accompanied by emotions to have
the same motivational force. Evidence about moral motivation from people
who rate high in psychopathic traits is consistent with this difference in mo-
tivation. Those with such traits have emotional deficits, and particularly with
“moral emotions” such as the ability to feel empathy for others. And people
lacking in these emotions have difficulty forming their own moral beliefs.
When they do, they rely on other means such as testimony. But even though
people with psychopathic traits can profess to have moral beliefs, when they
do they are not typically motivated by them.3 Presumably this is because
these beliefs do not have their typical emotional causal source.
People high in psychopathic traits also have problems grasping moral
concepts such as “right” and “wrong” (Hare 1993; Eichler 1965; Joyce 2006;
Kennet and Fine 2008). For instance, these individuals have difficulty
drawing the distinction between moral wrongs and mere conventional
2 For more in-depth discussion of irrational emotions in relation to the perceptual and cognitive
theories of emotion see Deonna and Teroni 2012.

3 See discussion in Nichols 2002.
wrongs (Blair et al. 2001). Moral wrongs are those that cause harm to indi-
viduals, whereas conventional harms involve mere violations of social
norms. Most people easily make distinctions between moral wrongs and
mere conventional violations by early childhood, but psychopathic indi-
viduals take all wrongs to be of the conventional sort. It is thought that the
lack of so-called moral emotions underlies the psychopathic person’s ina-
bility to grasp the moral understanding of “right” and “wrong.” If emotions
cause and provide some of the evaluative content for our moral beliefs,
then it would make sense that those with psychopathic traits have the cog-
nitive deficits regarding moral beliefs that they have. The difficulties that
psychopathic individuals have with grasping and using moral concepts
could be seen as analogous to the difficulty that colorblind people have
with color concepts.4
The role of emotion in causing and providing content for evaluative beliefs
is also demonstrated in work by Haidt and colleagues (Haidt 2001, 814).
In one study they asked subjects to judge whether it is morally acceptable
for a brother and sister to have sex on vacation if all of the potential nega-
tive consequences were mitigated (i.e., they use birth control, don’t tell an-
yone . . . ). Even when subjects were not able to point out any reasons in virtue
of which it would be wrong for the brother and sister to have sex, subjects al-
most always claimed that it was still wrong (see also Prinz 2006). It is natural
to assume that it is the reaction of disgust that causes subjects to make the
moral evaluation that they do and that the content of the emotion of disgust
also explains the content of the subjects’ moral evaluations. Moreover, this
study reveals that, as in the case of perception, people trust the deliverance
of their emotional experience even when they don’t have cognitive sources
of support for their beliefs such as other beliefs. The fact that people lacked
evidence in the form of beliefs for their moral judgments did not have any
bearing on their faith in them. And this suggests that we afford emotions and
other perceptions a similar epistemic role as basic evidential sources in justi-
fying our moral beliefs.
I think this suffices to show that emotions are intentional experiences
that play an analogous role to other perceptual experiences vis-à-vis
beliefs. For instance, emotions are sensitive to evaluative properties,
which they give us experiential awareness of. They also seem to cause and
4 As Neil Sinhababu observes, “Without experiencing color one can only achieve a partial grasp of
color concepts by sharing a public language with those who can see” (2017, 78).
provide content for our evaluative beliefs about their intentional objects
and this suggests that they might well be considered adequate justifica-
tion for those beliefs. The next section focuses on the specific case of
esteeming emotions.5
11.3. Esteem as the Representation of Value
Esteeming emotions are often directed toward people we admire and respect.
They vary in intensity and quality and include fondness, liking, caring for,
and loving, as well as others. Among philosophers who discuss emotions,
there is a tradition of taking esteeming emotions to evaluate objects they are
felt toward to have value.
For instance, in On the Origin of Right and Wrong, Franz Brentano claimed
that love and hate are ways of experiencing the goodness and badness of per-
ceptual objects: “We call a thing good when the love relating to it is correct.
In the broadest sense of the term, the good is that which is worthy of love,
that which can be loved with a love that is correct” (Brentano 1969, 11). Later,
in The Foundation and Construction of Ethics, he wrote, “When we call cer-
tain objects good and others bad we are merely saying that whoever loves the
former and hates the latter has taken the right stand” (Brentano 1973, 90).
David Velleman, in Love as a Moral Emotion, claims that love is the proper
experience when confronted with the dignity of another, a special kind of in-
trinsic value particular to persons: “I am inclined to say that love is likewise
the awareness of a value inhering in its object; and I am inclined to describe
love as an arresting awareness of that value” (1999, 360). Graham Oddie
follows this tradition in Value, Reality and Desire, nominating desire as the
source of value data, but adds that emotions either involve desire or can be
analyzed entirely in terms of desire.
It is plausible that emotions all have a desirative component; so the value

theory of experience I am advocating here can happily appropriate the
insights of the broad tradition which identifies a particular emotion or
emotions in general as a source of value data. . . . experiencing the value
of a person might take the form of loving rather than desiring (although
5 Alternatively, some identify related emotions with moral judgments. For instance, see Prinz 2006.
as we will see the nature of love might be analyzed in terms of desire).

(Oddie 2005, 77)
I follow this tradition in taking esteem to be a way of finding value in the

object of our esteem. One trait in respect of which persons become objects
of our esteem is for their character. That we like people with good character
seems intuitively correct. Moreover, this makes evolutionary sense since
most virtues are prosocial traits in respect of which a person would be a val-
uable ally or mate. But in addition to the intuitive and theoretical reasons,
empirical studies confirm that we tend to like people when they appear to
be virtuous.
For instance, in one study by Lewandowski et al. (2007), individuals were
asked to rate a series of positive, negative, and neutral character traits in
terms of their desirability. The rated traits were then paired with photographs
of random people. In this latter condition the subjects were asked, “How
much would you like to be friends with this person?” and “How much would
you like to date this person?” Positive trait information was highly correlated
with desire for friendship and then dating. It seems reasonable to assume that
we want to be friends with people whom we like and that esteem for them
also leads us to find them more attractive as dating partners.
In a similar study by Knitten et al. (2004), members of a rowing team
were asked to rate each other on a 0–99 scale for talent, effort, liking, and
physical attractiveness at the beginning and conclusion of an 18-month
period over which teammates got to know one another through training
and competition. Ratings of physical attractiveness, talent, effort, liking,
and respect were all highly correlated. Better “team players” received
better ratings on all of the measures. The same correlations were found in a
follow-up study in which students in an archaeology course were asked to
rate one another both before and at the end of the course on familiarity, in-
telligence, effort, liking, and physical attractiveness. The same correlations
were found in this study. Here a woman who was uncooperative and lazy
was universally disliked, and a woman who was considered to be hard-
working was universally liked.
What I’m proposing is that esteeming emotions are perception-like states
that represent what they are directed toward as being valuable. When it
comes to persons, one of the traits in respect of which we esteem someone
is good character. And in liking people on behalf of their good character, we
take the traits and the people who possess them to be good.
11.4. Naturalism and Value Perception
We can be better or worse at perceiving. One way to be a better perceiver

of something involves representing it in more detail. The ideal case involves
representing it in respect of the specific qualities that are distinctive to it qua
the type of thing that it is. Virtues are character traits that are distinguished at
least in part by the fact that they have positive value. So representing virtuous
traits as valuable, as when we feel esteem for others in respect of their moral
courage, is one way of being better perceivers of those traits.
This claim relies on the assumption that character traits can have real
value and that in esteeming a person in respect of their virtue we are put
into contact with that value. On a widespread picture, perception involves
being put into contact with a perceptible feature of the natural world via a
process that reliably tracks that feature. Moreover, the epistemic value of per-
ception depends on the reliability of the process and that is usually grounded
in a causal relation between the represented property and an experience that
represents the property. So if emotion really is epistemically on par with
other modes of perception, then there should be an account of value that is
consistent with naturalism. And since causation is involved, it would most
likely be an account that locates value in the material nature of the perceptual
object. Moreover, the causal process leading from the evaluative property in
the object to the perceiver’s emotional response should enable us to explain
why the perceiver has the emotional response that they have by appeal to the
appropriate features of the object.
Naturalist accounts of value are probably most familiarly attributed to
those working within Aristotelian virtue traditions, such as Hursthouse and
Foot.6 This is the view that for every evaluative property there will ultimately
be some material feature or set of material features that the evaluative pro-
perty is determined by.7 So whenever the material conditions for the evalua-
tive property are present, the evaluative property will be instantiated through
those properties. In the case of virtue, one reasonable view is that the value
of virtuous character traits is grounded in the material features of the person
and the behavioral manifestation of the character trait in question. It is these
material features that our responses of esteem track and that determine the
value that our esteem represents.
6 See, for example, Foot’s Natural Goodness (2001) and Hursthouse’s On Virtue Ethics (1999).
7 This view is put forward by Deonna and Teroni (2012). See especially chapter 4.
A reliable causal process between those material conditions for the evalu-
ative property and our emotional response can be explained by appeal to the
emotion’s “cognitive base.”8 According to Deonna and Teroni, the cognitive
base of an emotion includes any mental states that act as reasons for the emo-
tion, including perceptual representations of the person and his or her vir-
tuous actions. These perceptual representations will themselves be reliable.
The cognitive base acts as a causal intermediary between the value in the per-
ceptual object and our emotional response to it. Moreover, the cognitive base
not only causes the emotional response but also supplies the emotion with its
intentional properties; the fear we experience is about the spider, the percep-
tion of which gives rise to our fear. When we perceive an act of compassion
and respond with favorable feelings of esteem, the esteem is directed toward
the action and actor whose behavior we have perceived, and the emotion is
correct when the object that is provided by way of the cognitive base exem-
plifies the evaluative property that the emotion represents it to have.9
What I am claiming is that we perceive the virtue of another when we re-
spond to manifestations of virtuous character by liking the person in respect
of that character. Perceiving character traits as valuable is a way of taking
them to be virtues. And perceiving a virtue as a virtue is a way of more clearly
seeing that character trait for what it is. The better we get at this, the better
our epistemic position with respect to our understanding of others. The re-
mainder of this chapter concerns the epistemic role that cultivating good
character can play in making perceivers into better perceivers, and perhaps
better knowers, about the character of the people we encounter.
11.5. Character and Virtue
I have argued that esteeming people in respect of their character can be a

way of recognizing the quality of their character. Now I aim to show that cul-
tivating virtuous character traits in ourselves can make us more sensitive
perceivers of others’ character.
Although I have defined virtue as a type of character trait that is valuable,
I have not said much about what character traits are. I take character traits to
be relatively stable, enduring, and coherent sets of related dispositions. For
8Deonna and Teroni 2012. See especially c hapters 6 and 8.

9Although this is not meant to be an exhaustive defense of a naturalistic account of value and its
emotional perception, it suggests that a promising account is available.
a disposition to be stable means that while it may not always manifest in the
same behavior from situation to situation, a person possessing the character
trait will consistently exhibit behaviors that conform to the trait. An honest
person may not tell the truth in every circumstance but will nonetheless be-
have in a way consistent with honesty by revealing as much of the truth as is
appropriate to the circumstance. A character trait’s being enduring means
that the person who possesses it will manifest a relatively fixed pattern of be-
havior that persists across the lifespan. I am not saying that character traits
are unchangeable, but they are deeply ingrained. Finally, people with a spe-
cific character trait will manifest a set of three types of dispositions for en-
gaging with the world: they will be disposed to be perceptually attentive to
and to apprehend situations a certain way, they will be disposed to feel and
desire certain things relative to each situation, and, based on these emotions,
they will be disposed to behave in certain ways.
Having a character trait implies that there is integrity among these dif-
ferent sets of dispositions insofar as different dispositions associated with
each set will reflect the same character trait. A person who has a virtuous
character trait will therefore manifest these related dispositions with re-
spect to that trait. The person will be disposed to notice aspects of the so-
cial and behavioral context that are relevant to the manifestation of the
virtue and which call for a certain response. For instance, the courageous
person will be attuned to situations in a way that is sensitive to danger but
will nonetheless notice where and which action is called for. The person
will also be disposed to have a specific kind of emotional response to those
features of circumstances. For instance, the virtue of courage involves not
only experiencing the amount of fear proper to the circumstance but also
having the desire for a certain end that the courageous person recognizes as
good. The third dispositional category concerns response and behavior. To
be a courageous person it is not enough to have the attentive and emotional
dispositions particular to courage. The courageous person will not only feel
the proper amount of fear, but will act in spite of this fear in accord with the
desired end.
The psychological structure constituted by these related dispositions
should also manifest in an appropriate attunement to, and responsiveness
to, virtue in others. For instance, insofar as virtuous agents are attentive to
situations that call for virtuous responses in themselves, they will be likely
to recognize when virtuous responses are called for in others. And insofar
as virtuous agents know which responses would be virtuous for themselves,
they also know which behaviors are required of others. For instance, the
honest person will know when a situation calls for revealing the truth and
will be attuned to features that would distinguish a tactless comment from
an honest one. Finally, as I mentioned, a person with the character structure
associated with virtue will be disposed to have certain emotional responses
consistent with virtue. Among these are feelings of esteem for and approval
of virtuous behavior in others, and disapproval of the manifestations of vice.
The honest person is disposed to esteem witnessed acts of honesty and will
likewise regard dishonesty unfavorably.
Crucially, notice that the perceptual attunement to virtuous behavior in
others provides the perceptual component that becomes the cognitive base
for the appropriate emotional response of esteem. For the possessor of a
virtue, the affective response of esteem should arise naturally from the con-
frontation with the virtue in another. The esteem is directed toward the per-
ceived in respect of the natural properties in persons that become manifest in
their virtuous acts. Those natural properties constitute the value that the vir-
tuous perceiver, in esteeming the virtue of that person, represents that char-
acter trait to have.
I am not saying that recognizing virtue in others requires that one pos-
sess the trait in question. There are other ways to come to know about virtue.
Another way that we might recognize virtue is as a result of evaluative beliefs
that we form through rational methods such as inference or testimony.
For instance, one can come to believe through observing the testimony of
others that dishonesty is morally problematic or come to believe that hon-
esty consists in a certain type of behavior. These judgments can have top-
down influences on our emotional states, causing us to feel disapproval at
dishonest actions or to feel esteem for honest people. But forming evaluative
beliefs about character in a way that brings into play our own character, as on
the model that I have described here, confers several advantages over the ap-
plication of moral beliefs. Namely, it increases the likelihood that character
traits will be represented as virtues and the reliability of those representations
when they arise.
Some have observed that normative judgments about value, specifically
moral value, are intrinsically motivating and so depend in some sense on
emotions.10 In the strongest version of this view, what it is to make a moral
judgment is just to have the appropriate feelings of approval or disapproval
10 For example, see Prinz 2006; Nichols 2002.

toward the intended target. On a weaker version of the view, moral judgments
involve emotional dispositions. Although beliefs about virtue can influence
us to have the appropriate emotional responses to virtue, developing moral
beliefs does not guarantee that the corresponding emotion will be manifest.
Recall that psychopathic people appear to develop moral beliefs based on tes-
timony. But they do not appear to have the appropriate emotions to match
these beliefs. Psychopaths can categorize actions as right or wrong and may
believe that it is “wrong” to exploit other people but simultaneously feel no
aversion to doing so. They lack the motivational disposition that is a feature of
moral judgments because they lack the required emotional representations.
One might say that in these cases what appear like normative judgments are
actually not genuine moral judgments at all. Possessing the character struc-
ture associated with virtue, on the other hand, increases the likelihood that
one has the right emotional stances toward objects of moral evaluation and
therefore increases the likelihood that one will recognize virtue as virtue
when it is confronted.
Moreover, which specific acts will count as virtuous or vicious is highly
context dependent.11 A behavior that might count as honest in one situation
may qualify as tactless in a different context. And as situations increase in
complexity, it will be harder to accurately determine, by application of rules
and principles, which actions count as virtuous or vicious. This becomes
particularly obvious as our cognitive resources become loaded by other
demands or when we are pressured to make quick decisions. On the other
hand, insofar as character traits are associated with automatic dispositions,
they are routinely taken to be responsive to the complexities that arise in
the context of practical action. And there is reason to think that emotional
responses grounded in these dispositions will be more reliable, as there will
be an increased likelihood that they will arise when a virtuous trait is encoun-
tered, and these representations, when they arise, will be more accurate.
11.6. Liking the Vicious?
In this final section, I want to discuss a potential counterexample. I have argued

that we tend to esteem people in virtue of their good character; in doing so we
11 I am assuming that the trait of honesty is not reducible to truthfulness. The honest person reveals
only as much of the truth as is called for.

represent their character as valuable and so as virtuous. Furthermore, such

cases count as providing perception-like contact with virtue. And to the ex-
tent that we cultivate good character in ourselves, we can also become better
at detecting virtue in others. However, it might also be noted that people are
often drawn to individuals with questionable character. If it turns out that
people are as attracted to vice as much as to virtue, then esteem may not in the
end provide the reliable kind of access to the value of character traits that the
perceptual model advanced here requires. I will consider two versions of this
counterexample and argue that in both the cases actually lend support for the
view I’ve developed, rather than evidence against it.
One way of developing the counterexample is to appeal to common cul-
tural ideas about what people find appealing. Consider the common trope
in film and popular culture of the outsider character, a “bad boy.”12 These
characters are outsiders to social norms, often norms regarding common
morality. Moreover, the outsider archetype is simultaneously taken to be a
sex symbol precisely on account of his or her outsider status. Although it isn’t
entirely clear how heavily we want to rely on common tropes in the context
of philosophical theory construction or disconfirmation, I nonetheless take
it to count in a theory’s favor if it can make sense of the cultural data. And the
view that I have developed can do this.
The first thing to point out about this counterexample is that it rests on
the false premise that when we like such characters, we like them in respect
of their moral vices (though I will allow that in some cases we may like them
in spite of them). Examples of characters who easily fit the example might
include James Dean, Elvis Presley, and Muhammad Ali. When we think of
alluring rebels, we do not think of people like Ted Bundy or Joseph Stalin,
who would more aptly count as villains. An obvious difference between
these sets is that, while the former are violators of social norms, the latter
group violated not merely social but also genuine, uncontroversial, and con-
sidered moral beliefs about right and wrong. We like those who defy social
convention, particularly when those conventions stifle or impede genuine
moral progress. Challenging social conventions also seems to employ virtues
such as courage, authenticity, and open-mindedness. It is these qualities and
not moral vices, I suspect, we are drawn to in liking those who transgress
social norms.
12 I am taking this concept to be gender neutral. Perhaps one female manifestation is the femme
fatale.
Others have appealed to more scientific empirical findings to argue that

people are attracted to vice.13 “Dark triad” personality styles, which include
narcissism, psychopathy, and Machiavellianism, are characterized by the
presence of traits such as: callousness, deceit, exploitativeness, manipula-
tiveness, entitlement, and low empathy. And results of two types of studies
seem to link these traits to increased desirability. In the first set of studies,
photographs of people high in these traits turned out to be rated more attrac-
tive than others low in these traits. In another set of studies, people rating
high in these traits had more sexual partners than people who rated lower in
them (Jonason et al. 2009).
In order for these empirical findings to support a counterexample, they
must support the view that the gains in mating success and desirability that
the traits associated with the “dark triad” personality styles confer occur be-
cause those vicious traits themselves are the object of our esteem; the causal
connection between the esteem felt for the person must align with the causal
account about the perception of character that I have developed. But the
empirical findings do not support that reading. Rather, when we look more
comprehensively at what we know about the dark-triad personalities, it is
clear that these traits facilitate desirability and mating success in quite a dif-
ferent way. It is not the vicious traits themselves that are the object of appeal.
For instance, people who rate high in narcissism and psychopathy do tend
to put more effort into physical presentation and grooming, on average. So
it makes sense that in photographs they would get higher ratings on attrac-
tiveness. In fact, when people are shown photographs of the same individuals
without their self-adornment, they are judged as no more physically attrac-
tive than typical subjects (Holtzman and Strube 2012). I think it is unlikely
that here vicious traits are being represented as valuable. Rather, I think that
if character traits are being tracked in these evaluations at all, they are value-
neutral traits such as vanity. But there is also overwhelming evidence that
people with vicious personalities seek more short-term relationships than
long-term relationships, are less discriminating in partners, and also engage
in other deceptive behaviors that make them more successful at securing
partners (Jonason et al. 2009). These facts go a long way in explaining why
they tend to have more sexual partners than neurotypicals.
13 This conclusion is drawn by Steve Connor in an article published in The Independent, “Why
Women Really Do Love Self-Obsessed Psychopaths,” June 19, 2008.

For instance, the manipulative exploitative social style and superficial

charm characteristic of the dark triad helps these individuals to make pos-
itive initial impressions (Jonason et al. 2009). On the one hand, we wouldn’t
expect vice to be on full display in the earliest stages of getting to know
someone. Even neurotypicals are on their best behavior, so to speak, when
they meet a new potential partner. Those skilled in deception are likely to be
particularly good at concealing vicious traits. Add to this that the person-
ality styles we are referring to often come coupled with positive traits like
extraversion and low neuroticism, and the appeal becomes more apparent.
Deceptive individuals can also claim feelings or intentions that are insincere
but effective in seduction. Some vicious traits might even initially present
as virtues: narcissistic entitlement and self-aggrandizement can falsely come
off as confidence, impulsive behavior and lack of concern for consequences
can look like courage, and desire for social dominance can initially seem to
be strength or fortitude.14 All of this seems to be helpful for securing short-
term relationships, though the appeal of such individuals is not likely to be
long-lasting.
In fact, rather than presenting a counter-example to the view I have
offered, considering the appeal of dark-triad character traits can help us to
home in on the ways that our own character can facilitate, or in some cases
impede, our perception of virtue in others. I have argued that possession of
virtuous character traits can make us better detectors of virtue in others. But
possession of perfect virtue is uncommon. Character traits come in degrees.
Those who are honest may be generally truthful but fail to have the appro-
priate emotional response on some occasions and so sometimes fail to see in
others what they themselves possess only imperfectly. But the more perfectly
we embody our virtuous traits, this view predicts, the better we should be-
come at detecting those traits in others. And on the flip side, the view predicts
that the further from virtue our character deviates, the worse we will be at
recognizing virtue. There is some evidence that this prediction bears out.
Studies show that psychopathic individuals s are the only individuals who
prefer long-term partners who also rate high in psychopathy (Jonason et al.
2015). One way of reading this is that people in this group do not adequately
distinguish the virtuous from the vicious when selecting mating partners. In
14 Positive traits associated with dark-triad personality styles are discussed in Jonason et al. 2009.
It seems to be an open question which positive traits associated with dark-triad personalities are gen-
uinely separate positive traits (i.e., extroversion) and which are vices that we mistake for virtues in
conditions of insufficient knowledge (i.e., where the drive for social dominance confers power).
lacking virtue, they are poorer than average at detecting it in others. But their
mistakes go beyond a failure of recognition. In esteeming others in respect
of their vice, they mistakenly represent negative traits as having positive
value. Perhaps the further from perfect virtue our own character deviates,
the bigger mistakes we are wont to make.
11.7. Conclusion
In this chapter, I have argued that virtues are a category of character traits
that have positive value. When we esteem something, we represent it to have
positive value. So esteeming persons in respect of their character traits is a
way of apprehending virtuous character traits as virtues.15 Moreover, these
feelings of esteem directed at virtue are epistemically perception-like in their
role in causing, providing content for, and justifying judgments about the
moral properties associated with character. Finally, I have argued, our own
moral character can facilitate our perception of virtue by causing us to at-
tend to the right things, recognize the right behaviors, and have the right
emotional responses to virtue in others. And that makes having moral virtue
epistemically as well as morally advantageous.
References
Blair, R. J., E. Colledge, and D. G. Mitchell. 2001. Somatic markers and response re-
versal: Is there orbitofrontal cortex dysfunction in boys with psychopathic tendencies?
Journal of Abnormal Child Psychology 29, 499–511.
Brentano, F. 1969. On the Origin of Right and Wrong. Trans. Roderick Chisholm and E.
Schneewind, Routledge: London.
Brentano, F. 1973. The Foundation and Construction of Ethics. Trans. E. Schneewind.
Humanities Press: New York.
Connor, S. 2008. Why women really do love self-obsessed psychopaths. The Independent,
June 18.
Deonna, J. 2006. Emotion, perception, and perspective. Dialectica 60(1): 29–46.
Deonna. J., and F. Teroni. 2012. The Emotions: A Philosophical Introduction. New York:
Routledge.
Doring, S. 2007. Affective perception and rational motivation. Dialectica 61(3): 363–394.
Eichler, M. 1965. The application of verbal behavior analysis to the study of psychological
defense mechanisms: Speech patterns associated with sociopathic behavior. Journal of
Nervous and Mental Disease 141(6): 658–63.
15 Again, I am happy to call this a case of quasi-perception. I only intend to highlight the way in
which emotion functions analogously to perception in the ways that I have discussed.
Foot, P. 2001. Natural Goodness. Oxford: Clarendon Press.

Goldie, P. 2009. Getting feelings into emotional experience in the right way. Emotion
Review 1(3): 232–239.
Haidt, J. 2001. The emotional dog and its rational tail: A social intuitionist approach to
moral judgment. Psychological Review 108: 814–834.
Hare, R. D. 1993. Without Conscience: The Disturbing World of the Psychopaths among Us.
New York: Guilford Press.
Holtzman, N., and N. Strube. 2012. People with dark personalities tend to create a physi-
cally attractive veneer. Social Psychological and Personality Science 4(4): 461–467.
Hursthouse, R. 1999. On Virtue Ethics. Oxford: Clarendon Press.
Johnston, M. 2001. The authority of affect. Philosophy and Phenomenological Research
53: 181–214.
Jonason, P., M. Lyons, and A Blanchard. 2015. Birds of a “bad” feather flock together: The
dark triad and mate choice. Personality and Individual Differences 78: 34–38.
Jonason, P., N. P. Li, G. D. Webster, and D. P. Schmitt. 2009. The dark triad: Facilitating a
short-term mating strategy in men. European Journal of Personality 23: 5–18.
Joyce, R. 2006. The Evolution of Morality. Cambridge, MA: MIT Press.
Kant, I. 1996. The Metaphysics of Morals. Trans. M. J. Gregor. Cambridge: Cambridge
University Press.
Kennett, J., and C. Fine. 2008. Internalism and the evidence from psychopathy and ac-
quired sociopathy. In W. Sinnott- Armstrong, ed., The Neuroscience of Morality:
Emotion, Disease, and Development. Cambridge, MA: MIT Press. 173–190.
Kniffin, K. M., and D. S. Wilson. 2004. The effect of nonphysical traits on the perception of
physical attractiveness. Evolution and Human Behavior 25: 88–101.
Lewandowski, G., A. Aron, and J. Gee. 2007. Personality goes a long way: The malleability
of opposite-sex physical attractiveness. Personal Relationships 14: 571–585.
Nichols, S. 2002. How psychopaths threaten moral rationalism. The Monist 85(2): 285–303.
Oddie, G. 2005. Value, Reality and Desire. New York: Oxford University Press.
Prinz, J. 2004. Gut Reactions: A Perceptual Theory of Emotions. New York: Oxford
University Press.
Prinz, J. 2006. The emotional basis of moral judgment. Philosophical Explorations.
9(1): 29–43.
Sinhababu, N. 2017. Humean Nature: How Desire Explains Action, Thought, and Feeling.
New York: Oxford University Press.
Tye, M. 2008. The experience of emotion: An intentionalist theory. Revue Internationale
de Philosophie 243: 25–50.
Velleman, J. D. 1999. Love as a moral emotion. Ethics 109(2): 338–374.
Index
For the benefit of digital users, indexed terms that span two pages (e.g., 52–53) may,
on occasion, appear on only one of those pages.
Figures are indicated by f following the page number
Adler, G. (with W. F. Gattaz), 246–47 Billon, A., 181–83

agential perception, 1 Blair, R. J. (with E. Colledge, and
Akins, K., 71–72 D. G. Mitchell), 258–59
Allais, L., 168–69, 174 blind spot, 10–11, 85, 89, 91, 92
Allman, M. (with S. Teki, T. Griffiths, and Block, N. J., 122
W. Meck), 194–95, 196, 197, 198–99 Block, N., 14, 148, 154–55
Alston, W., 2, 3, 4 Block, R. (with S. Grondin), 196
amodal completion, 78, 86 bodily location, 159, 183–84
The Anarchic Hand, 180–81 Bolano, R., 77
appearances of reality BonJour, L., 1, 2
the question of, 168–69 Boswell, J., 168–69
Aristotle, 37–38, 60n.11, 262 Botvinick, M. (with J. Cohen), 241
Armstrong, D. M., 167, 173 Brentano, F., 260
aspect perception, 14, 15, 19, 118–19, 122, Brillat-Savarin, J., 34
124–25, 126, 128–29, 132, 133, 134, Briscoe, R. E., 7–8
138, 141, 142 Brogaard, B., 5–6, 19
Audi, R., 2, 7–8 Brogaard, B. (with B.
auditory experience, 13–14, 19, 21, 22, Chomanski), 21–22
220–21, 236–37, 245–46 Brogaard, B. (with D. E. Gatzia), 7–8,
Auvray, M. (with C. Spence), 8, 9, 37–38 17, 21–22
Awater, H. (with J. R. Kerlin, K. K. Evans, Brogaard, B. (with E. Chudnoff),
and F. Tong), 85 9, 21–22
Aydede, M. (with M. Fulkerson), 68–69 Bubl, E. (with E. Kern, D. Ebert, M. Bach,
and L. T. van Elst), 21, 246–47
Barlassina, L. (with A. Newen), 221 Budek, T. (with K. Farkas), 100,
Batty, C., 12, 100, 102–3, 104, 106 102–3, 104
Bayne, T. (with C. Spence), 241–42, Buhusi, C. (with W. Meck), 192, 194–95,
250, 253–54 196, 198–99
Bechara, A. (N. Naqvi), 220 Bullier, J., 81, 83–84
Bennett, J., 167–68 Bult, J. (with R. Wijk, and T. Hummel), 37
Bergmann Tiest, W. M. Burge, T., 11–12n.4, 14, 147
(with A. M. L. Kappers), 57–58 Byrne, A. (with D. R. Hilbert), 9, 214–15
Berkeley, G., 168–69, 174
Bertelson, P., 59, 87 cadential formula, 135f 5.2, 136f 5.3,
Bertelson, P. (with B. de Gelder), 87, 241 136–37f 5.4, 137f 5.5
Berto, F., 173 Calvert, G. (with T. Thesen), 58
274 Index
Calvert, G. (with E. T. Bullmore, M. J. Damasio, A., 1, 218

Brammer, R. Campbell, S. C. R. Darwin, C., 218
Williams, P. K. McGuire, P. W. R. Davidson, D., 1, 189–90
Woodruff, S. D. Iversen, and A. S. Davies, Stephen, 223
David), 87, 89 DeBellis, M., 133
Campbell, J., 16 Delboeuf, J., 178–79
Campbell, J. (with Cassam), 168–69 Deonna, J. (with F. Teroni), 263
Campos, J. L. (with J. S. Butler, and Deroy, O. (with C. Spence), 10, 89
H. H. Bulthoff), 56–57 Deroy, O. (with Y. C. Chen and
Casati, R. (with J. Dokic), 120 C. Spence), 10, 89
Casati, R. (with A. Varzi), 189–90 Descartes, R., 3–4, 16, 166, 176
Casey H. (with R. D. Rogers, T. Burns, and Destutt de Tracy, A., 169–70
J. Yiend), 22 Deutsch, D., 15
Cassam, Q. (with J. Campbell), 174 De Weerd, P. (with R. Desimone, and
Cavedon-Taylor, D., 100, 102–3 L. G. Ungerleider), 86
Chalmers, D. J., 3–4, 11–12, 14 De Weerd, P. (with R. Gattass, R.
Chan, K. Q. (with E. M. Tong, D. H. Tan, Desimone, and L. G. Ungerleider), 86
and A. H. Koh), 21, 87, 246–47 De Weerd, P. (with E. Smith, and P.
Cheng, R. K. (with C. MacDonald, and Greenberg), 86
W. Meck), 198 diamond figure, 129–30f 5.4
Chisholm, R., 3, 5–6 diminished 7th chord, 134–35f 5.1
Chuard, P., 200–1 Dokic, J., 183–84
Chudnoff, E., 6 dot arrays, 129f 5.3
Churchland, P. S., 121 Dretske, F., 1–2, 122, 219
Clark, Andy, 192–93 Droit-Volet, S. (with W. Meck), 198
Clark, Austen, 121–22 duck-rabbit, 124–25f 5.1
cognitive penetration, 7, 21–22, 59, 77–78, Dugas, L. (with F. Moutier), 181–82
80, 81–82, 123, 139, 142, 188, 202, duration, 17–19, 120, 124, 188–92, 193,
203, 205–6, 207, 238–40, 243, 248– 194–96, 197–98, 199–200, 201–2,
49, 251, 252 203–4, 205–6, 207, 208–9
cognitive penetrability, 18–19, 78–79, the puzzle of, 190, 192
123, 188, 189, 192, 201–2, 203–4,
208–9 (see also cognitive penetration) Eagleman, D. (with P. Tse, D. Buonomano,
cognitivist, 221 P. Janssen, A. Nobre, and
coherentism, 2 A. Holcombe), 196
common sensibles, 37–38 egocentric location, 184
complimentary overlap, 60, 62 Eichler, M., 258–59
Condillac, 169–70 Ekman, P., 217, 218
Conee, E. (with R. Feldman), 2 emotion, 19–23
content of experience, 243, 245–46 emotion experience, 216–17, 220, 221,
content of perception, 175, 238–39, 242, 244, 259
243, 245–46, 248, 251, 252, 253–54 emotion perception, 1, 55–56, 257
Craig, A. D., 220 encode, 39–40, 192–93
Crane, T., 168–69, 174 analog encoding, 215–16
Crisnel, A. S. (with C. Spence), 9, 41–42 digital encoding, 215–16
epistemic conservatism, 5–6
Dainton, B., 191, 199–200 epistemic privilege, 4, 16–17, 165, 166,
Dalton, P. (with N. Doolittle, H. Nagata, 167–68, 169–70, 173, 178
and P. A. S. Breslin), 37 Ernst, M. O. (with M. S. Banks), 57–58
Index 275
error of commission, 107–8, 109 Goldman, A. I., 1–2

error of omission, 107–8, 109 Gombrich, E. H., 118, 124–25, 128, 138
Essick, G. K. (F. McGlone, C. Dancer, grating sound case, 236–37, 246, 250–51
D. Fabricant, Y. Ragin, N. Phillips, Graziano, M. S., 56–57
T. Jones, and S. Guest), 70–71 Greco, J., 1–2
evaluative belief, 22, 258, 259–60, 265 Green, B. (with P. George), 37
event perception, 188, 189–90, 203 Green, M. S., 227–28
event segmentation, 192–93 Gregory, R. L., 86
evidentialism, 2 Grice, H. P., 131–32, 141
experience of duration, 188, Grill-Spector, K. (with R. Malach),
190–92, 194–95 81, 83–84
experimentum crucis, 151, 156, Grondin, S., 194–95, 196, 198
158–59, 161 GSC. See grating sound case
exteroceptive senses, 142, Guest, S. (with C. Catmur, D. Lloyd, and
237–38, 253–54 C. Spence), 241
Farennikova, A., 13–14, 120 Hagura, N. (with H. Barber, and

feeling of presence, 84–85, 182–84 P. Haggard), 37
Feldman, H. (with K. Friston), 47–48 Haidt, J., 259
Fiorani, J. M. (with M. G. Rosa, R. Gattass, Hansen, T. (with L. Pracejus, and
C. E. Rocha-Miranda), 85 K. R. Gegenfurtner), 86
Firestone, C. (with B. J. Scholl), 79 Hanson, N. R., 121
Fitzgerald, P. J., 246–47 haptic touch, 54–55, 56–58, 61–62
flash-induced sound effect, 56–57n.4 Hare, R. D., 258–59
Fodor, J., 123, 204 Harman, G., 5–6, 163
Foley, R., 5–6 Harrar, V. (with B. Smith, O. Deroy, and
Foot, P., 262 C. Spence), 9, 41–42
foundationalism, 2–3 Heider, F., 166
Frissen, I. (with J. L. Campos, J. L. Helbig, H. B. (with M. O. Ernst), 57–58
Souman, and M. O. Ernst), 56–57 Hertrich, I. (with S. Dietrich, and H.
Friston, K. (with K. Stephan), 192–93 Ackermann), 87, 89
fugitive sensations, 154–55 Hohwy, J., 192–93
Fulkerson, Matthew, 9–10, 11, 56–57, Holtzman, N. (with N. Strube), 268
166, 254 Hort, J. (with T. Hollowood), 37
Fumerton, R. A., 1, 2–3, 4, 5 Hosono, Y. (with K. Kitaoka, R.
Urushihara, H. Sei, and Y. Kinouchi),
Gallagher, S., 200–1 21, 246–47
Gandhi, S. P., et al., 80 Huemer, M., 1, 2–3, 5–6, 168–69
genuine perception, 40 Hume, David, 9, 29–30, 174
Ghazanfar, A. A. (with C. E. Schroeder), 87 Hursthouse, R., 262
Ghijsen, H., 78 Husserl, E., 18, 191
Gibbon, J., 196
Gibbon, J. (with R. Church, and independency from perception, 173
W. Meck), 196 independency from the will, 173
Gibbon, J. (with C. Malapani, C. Dale, intentionality, 106, 110
and C. Gallistel), 196 intentionalism (see intentionality)
Goldberg, S., 72–73 intentionalist theories, 2
Goldie, P., 221 See also coherentism; foundationalism
276 Index
internalism, 2, 168–69 Lederman, S. J. (with R. L. Klatzky), 58

interoception, 3–4, 7–8, 20, 21, 22, 213, LeDoux, J., 221
220–21, 237–38, 243–44, 246, 247, Lehrer, K., 2
248–51, 253–54 Lerdahl, F. (with R. Jackendoff), 134
introspection, 2–4, 7–8, 79–80, 107, Levinson, J., 224
109–10, 111, 121–22 Lewandowski, G. (with A. Aron, and
Iurilli, G. (with D. Ghezzi, U. Olcese, G. J. Gee), 261
Lassi, C. Nazzaro, R. Tonini, V. Tucci, Lim, J. (with M. Johnson), 36, 45
F. Benfenati, and P. Medini), 87 Loken, L. S. (with J. Wessberg, I. Morrison,
F. McGlone, and H. Olausson), 70
James’s subtraction argument, 221 lone smell experiences, 104–5
James, T. W. (with G. K. Humphrey, long-term body image, 160–61
J. S. Gati, P. Servos, R. S. Menon, Lopes, D., 213, 225–26, 227–28, 231
and M. A. Goodale), 87 Lycan, W., 14–15, 100, 102–3,
James, William, 191, 219, 221, 243–44 121–22, 123
James, W., 17–18, 19, 169–70, Lyons, J., 78, 207
218–19, 244 Lyons, J. C., 203
Jonason, P. (with M. Lyons, and A.
Blanchard), 269–70 Macpherson, F., 7–8, 79, 87, 202,
Jonason, P. (with N. P. Li, G. D. Webster, 203–4, 216
and D. P. Schmitt), 268–69 Maine de Biran, F. P. G., 169–70, 178–79
Jones, L. A. (with S. J. Lederman), 9–10, 54 Malebranche, N., 169–70
Joyce, R., 258–59 Maravita, A. (with C. Spence, and
J. Driver), 56–57
Kania, A., 119–20 Marr, D., 104
Kant, Emmanuel, 166 Martin, M. G. F., 166, 183–84
Katzner, S. (with S. Weigelt), 81, 83–84 Martuzzi, R. (with M. M. Murray, C. M.
Kelly, S., 17–18, 21–22, 190–92, Michel, J. P. Thiran, P. P. Maeder,
195–96, 200–1 S. Clarke, and R. A. Meuli), 87
Kelp, C., 1–2 Massin, O., 171–72
Kim, J., 189–90 Massin, O. (with J. M. Monnoyer), 166
Kivy, P., 138–39 Massin, O. (with Vignemont), 16–17
Klatzky, R. (with S. J. Lederman), material-weight illusion, 54–55
57–58, 61 Matravers, D., 225
Kleinschmidt, A. (with C. Buchel, S. Zeki, Matthen, M., 68–69, 100, 183–84, 254
and R. S. Frackowiak), 128 Matthews, W. (W. Meck), 195, 198
Knox, J. (with P. Cordo, R. Skoss, S. maximum likelihood estimation
Durrant, and P. Hodges), 56–57 model, 57–58
Koffa, K., 14 McCain, K., 5–6
Komatsu, H. (with M. Kinoshita, and McDowell, J., 1, 101, 181
I. Murakami), 85 McGlone, F. (with A. B. Vallbo,
Krueger, L. E., 146 H. Olausson, L. Loken, and
Kuhn, T., 121, 123 J. Wessberg), 70
Kurby, (with J. M. Zacks), 192–93 McGrath, M., 1, 2–3
McGurk, H., 8, 22, 241–42, 252–53
Lapid, H. (with S. Shushan, A. Plotkin, McGurk effect, 56–57n.4, 201–2n.17,
H. Voet, Y. Roth, T. Hummel, 241–42, 252–53
E. Schneidman, and N. Sobel), 69 McGurk illusion, 8, 22
Index 277
Mendelovici, A., 12–13, 102–3, 104–5, non-belief states, 6

112–13, 115–16 See also perceptual experience
Mennella, J. A. (with J. A. Jagnow, and non-human animals, 215
G. K. Beauchamp), 44–45 non-relationalism, 3
mentalism, 2 See also non-belief states; perceptual
See also internalism experience
mental imagery, 7, 10–11, 83, 84–85, Nozick, R., 7
86, 87–93 Nudds, M., 120, 241
Merino-Rajme, C., 195–96
Merleau-Ponty, M., 14 O’Callaghan, C., 9, 13–14, 87, 119, 140,
Meyer, L., 138 220–21, 242
Millikan, R., 146–47, 218 O’Connor, D. H. (with M. M. Fukui,
mind-independent reality, 173 M. A. Pinsk, and S. Kastner), 80
Mizrahi, V., 100, 102–3 Oddie, G., 260
MLE. See maximum likelihood olfactory experience, 11–13, 35–36, 65,
estimation model 69, 70, 99–101, 102–5, 106, 107–14,
moral emotions, 258–59 115–16, 220–21
motion-bounce illusion, 56–57n.4 Orlandi, N., 128
motor effort, 171–72f 7.1 orthonasal olfaction, 35–36
multimodality, 87 O’Shaughnessy, B., 120, 160–61, 166, 184
multimodal experience, 1, 103
multimodal mental imagery, 7, 84, 86, Pacherie, E., 180–81
87, 88, 89–92 Pappas, G., 2
multimodal perception, 87, Parker, A., 64
90–91, 220–21 Peacocke, C., 14, 100, 123
multisensory integration, 21–22, 37, Pearson, J. (with F. Westbrook), 10–11
45–46, 61, 237–38, 241, 242, 243, Pearson, J. (with T. Naselaris, E. A.
248–49, 250–51, 252–53 Holmes, and S. M. Kosslyn),
multisensory perception, 1, 9, 44–45, 10–11, 83
48–49, 87, 241–42, 250–51, 253–54 Pekkola, J. (with V. Ojanen, T, Autti,
Murray, S. O. (with D. Kersten, B. A. I. P. Jaaskelainen, R. Mottonen,
Olshausen, P. Schrater, and D. L. A. Tarkainen, and M. Sams), 87, 89
Woods), 80 perceptual affect, 65, 68–69
muscular thesis, 173, 176 perceptual beliefs, 1–2, 4–5, 11–12, 18–19,
22, 54, 55–56, 123, 189, 205–6,
Nanay, B., 7, 10–11, 80, 85, 87, 91 215–16, 227–28, 236–37
Nanay, B. (with Teufel), 81, 82 perceptual experience, 1, 3–4, 7–8, 9–11,
naturalist, 262 16, 19, 20, 21–22, 55–57, 59, 62–64,
naturalised epistemology, 78, 93 65–66, 67, 68–69, 71–73, 79, 80–81,
naturalism, 262 100, 101, 105, 115–16, 123, 146–47,
nature of emotions, 19 148, 168–69, 174, 175, 190–91, 192,
Positions on, 221 201–2, 203, 213, 214–15, 216, 220–
necker cube, 125–26f 5.2 21, 236–37, 238–40, 245–46n.16,
Newtson, D. (with G. Engquist, and 246–47, 248–50, 251–52, 257,
J. Bois), 192 258, 259–60
Nickel, B., 129–30, 131 perceptual justification, 10–11, 19, 21, 22,
nociception, 7–8 55–56, 72, 78, 82, 83, 86, 89, 90, 91,
Noe, A., 123, 183–84 92–93, 238, 251, 253–54
278 Index
perceptual knowledge, 31, 40, 55–56 representationalism, 11–12, 213, 214–15,

perceptual phenomenology, 79–81, 82, 88, 216, 221n.16, 232
90, 239–40 representational, 14, 19, 22, 61, 119,
peripheral vision, 86 126, 147, 163, 205, 217, 220–21, 226,
Perkins, M., 100, 120 228, 229, 238–39, 241–42
Pessoa, L. (with L. G. Ungerleider), 80 representational content, 11–12, 14, 19,
phenomenal conservatism, 5–6, 61, 119, 147, 217, 219–20, 230–31
168–70, 175 retention theory, 18, 191
phenomenology, 18, 19, 34, 38, 66, 79–81, Retronasal olfaction, 32–33, 35–36, 45
82, 85n.3, 90, 107, 125–26, 127, 128, Reynolds, J. (with J.Zacks, and T.
140n.34, 174–75, 179, 183–84, 204, Braver), 194–95
208–9, 214–15, 216–17, 219 Ricci, C. (with C. Blundo), 128
Phillips, I., 17–18, 193 Richardson, L., 100, 102–3
physicalism, 106, 107–8, 109–10, 120 Roberts, T., 100
Piqueras-Fiszman (with Spence, C), 41–42 Rodman, S., 228
plato, 165, 169–70, 173 Rolls, E., 37, 43
pleasant sound case, 236–38, 245–46, Rolls, E. T. (with M. J. Tovee), 37
250–51, 252 Rosch, E. (with C. Mervis, W. Gray, D.
predictive coding, 47–48, 192–93, Johnson, and P. Boyes-Braem), 121
204, 205–6 Ross, P. W., 8
Prescott, J., 29–30, 37, 39–40 Rothko, Mark, 223, 228, 229
primitivism, 12–13, 106, 109–10 Rothko, M. (with A. Gottleib), 228
Prinz, J., 19, 22–23, 219–20, 221, 222, 259 Rozin, P., 8, 35–36
proprioception, 7–8, 56–57, 58, 62–63, Russell, B., 2–3, 4
170, 170n.5, 237–38
Pryor, J., 1, 2–3, 5–6, 78 safety, 7, 91–92
PSC. See pleasant sound case the safety strap, 180
Pylyshyn, Z. W., 21–22, 77 See also the anarchic hand
Scandola, M. (with E. Tidoni, R. Avesani,
quietism, 168–69 G. Brunelli, S. M. Aglioti, and
Quine, W. V. O., 5–6, 72, 189–90, 192 V. Moro), 180–81
Schaal, B. (with P. Orgeur), 44–45
Raffman, D., 215–16 Schellenberg, S., 11–12, 123
The Raft of the Medusa, 229f 9.1, 230–31 Schopenhauer, 166
Ramachandran, V. S., 85 Schwenkler, J., 7–8
Ramachandran, V. S. (with R. Gregory), 86 Scruton, R., 120
reality of appearances senses, 7–8, 16, 29, 30, 37–38, 44, 45, 49,
the question of, 168–69 53–56, 57–58, 60, 61–63, 64, 71–74,
redundant overlap, 57–58, 60–61 122, 165, 166, 167, 174, 227–28,
Reid, T., 100, 151–52, 153, 154–56, 232, 237–38, 241–42, 241–42n.7,
158–59, 161, 162–64, 174–75 249–50, 253–54
Reidian foundationalism, 175 chemical senses, 11–12n.4
relationalism, 3, 106, 108, 109–10 sensitivity, 7, 66, 70, 72–73, 91, 158, 170,
reliability, 1–2, 55–56, 71, 72–73, 78, 172, 179
79–80, 92, 93, 112–13, 166, 168, sensory discounting, 9–10, 54, 59
262, 265 sensory reinforcing, 9–10, 11, 54–55
reliable misrepresentation, 112–13, sensory stimulation, 9, 10–11, 78, 82–86,
114, 115–16 87–89, 91, 92, 203–4
Index 279
Shams, L. (with Y. Kamitani, and tactual sensation, 148–53, 154–56, 157–58,

S. Shimojo), 88 159, 160–61, 162–64
Shea, N., 203–4, 205–6 See also touch
Shea’s top down/bottom-up Talavage, T. M. (with M. I. Sereno, J. R.
distinction, 205–6 Melcher, P. J. Ledden, B. R. Rosen,
Sheppard, G., 9, 30, 39–40 and A. M. Dale), 83–84
Sherrington, C. S., 170 taste, 7–9, 21, 29–34, 35–37, 38–40, 41–43,
Shipley, T. (with J. Zacks), 192 44–49, 66, 67–68, 154, 216, 246–47
Shoemaker, S., 3, 4 Taylor-Clarke, M. (with P. Jacobsen, and
Sibley, F., 33–34 P. Haggard), 16, 166
Siegel, S., 1, 7, 22, 78, 79, 121–22, 124, 133, Teigen, K. (with G. Bohm, S. Bruckmiller,
168–69, 174–75, 202, 203, 239–40, P. Hegarty, O. Luminet), 198
248–49, 251 teleosemantics, 213, 217, 227
Siegel, S. (with Silins, N), 201–2 temporal experience, 188, 189, 190, 200–1
Sierra, M., 182 thermal-weight illusion, 54–55
Silins, N., 1, 78, 202, 240, 251–52 Thompson, E., 64
size-weight illusion, 9–10, 54 Titchner, E., 34
Small, D., 9, 39–40 top-down influences, 18, 77, 78–82, 89, 90,
Smith, A. D., 9, 30–31, 34 197, 206, 208–9
Smith, B., 9 top-down influences on perception, 77,
Smithies, D., 4 78–79, 80, 82, 89, 90
Sober, E., 218 touch, 7–8, 9–10, 16, 17, 29, 30, 32, 33,
Solar, M. (with H. Kumru, 34, 36, 37–38, 39, 41, 44, 45–46, 49,
J. Vidal, R. Pelayo, J. Tormos, 54–55, 56–58, 60–62, 69, 70–71,
F. Fregni, X. Navarro, and 86, 146, 147–49, 150–52, 154–55,
A. Pascual-Leone), 35–36 156, 158, 163, 165–70, 171, 172–74,
somatic theory, 19, 219 175–76, 178, 179, 182–84, 184n.13,
of emotions, 19 237–38, 246–47
Sorensen, R., 13–14, 120 effortful touch, 17, 165, 169–70, 171f,
Sosa, E., 1–2, 7, 91–92 172–74, 175–76, 178, 179, 182–83
specious present theory, 17–18, 191 passive touch, 158, 173–74
Spence, C., 9, 37–38, 41–42, 44 Treisman, M., 17–18, 196
Spence, C. (with J. Driver), 87 Trivers, R. L., 218
Spence, C. (with M. Shankar, and Tucker, C., 6, 7, 78
H. Blumenthal), 41–42 Tye, M., 11–12, 14, 221
Spence, C. (with M. Auvray, and
B. Smith), 44–45 value perception, 262
square, 129–30f 5.5 Van Cleve, J., 158, 163
Steup, M., 2 Van Essen, D. C., 81, 83–84
Stevenson, R, (with C. Tomiczek), 8–9, Vance, J., 202
32–33, 37 Velleman, J. D., 260
Stokes, D., 79, 80, 200, 239–40, 252–53 ventriloquist effect, 59–60, 72–73, 87, 241
Strawson, G., 168–69, 174 Vermazen, B., 224
subjectivism, 9, 31–33, 38–39, 43, 49 Vetter, P. (with F. W. Smith, and
subjective awareness, 68–69 L. Muckl), 87
subjective durations (see duration) Vignemont, F., 177–78, 183–84
sui generis, 12–13, 109–10, 191 Vignemont, F. (with O. Massin), 16–17,
Swinburne, R., 1 166, 167, 170
280 Index
virtue, 2, 14, 20, 22–23, 121, 171–72, 176, Yeomans, M. (with L. Chambers, H.
205, 225, 228, 256–57, 259, 261, 262, Blumenthal, and A. Blake), 41–42, 48
263–67, 269–70 Young, N., 13–14
visual sensory experience, 1, 11–12
Vroomen, J. (with P. Bertelson, and Zacks, J. M. (with K. M. Swallow), 188, 192
B. de Gelder), 87 Zacks, J. M. (with B. Tversky), 192, 198
Zacks, J. M. (with B. Tversky, and G. Iyer),
Waterman, Ian 18, 192–93
Case of, 62–63 Zacks, J. M. (with N. Speer, K. Swallow,
Watkins, S. (with L. Shams, S. Tanaka, T. Braver, and J. Reynolds), 193–94
J. D. Haynes, and G. Rees), 87, 88 Zadra, J. R. (with G. L. Clore), 21, 246–47
Weil, R. S. (with J. M. Kilner, J. D. Haynes, Zajonc, R. B., 221
and G. Rees), 86 Zaki, J. (with J. I. Davis, and
Weil, R. S. (with S. Watkins, and K. N. Ochsner), 220
G. Rees), 86 Zampini, M. (with C. Spence), 41–42
Williamson, T., 4, 7 Zampini, M. (with D. Sanabria, N. Phillips,
wishful seeing, 77 and C. Spence), 41–42
Wittgenstein, L., 7–8, 118, 122, 126, 127, Zangaladze, A. (with V. D. Weisser, E.
128, 131–32 Stilla, S.C. Prather, and K. Sathian), 87
Wollheim, R., 124–25

The Epistemology of Non-Visual Perception

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Epistemology of Non-Visual Perception

Uploaded by

Copyright:

Available Formats

The Epistemology of Non-​Visual Perception

PHILOSOPHY OF MIND SERIES

The Conscious Brain Attention is Cognitive Unison

Phenomenal Intentionality Perceptual Learning

The Character of Consciousness Combining Minds

Published in the United States of America by Oxford University Press

© Oxford University Press 2020

All rights reserved. No part of this publication may be reproduced, stored in

You must not circulate this work in any other form

Library of Congress Control Number: 2019954225

Printed by Integrated Books International, United States of America

Carrie Figdor is Professor of Philosophy at the University of Iowa with appointments

Philosophy of Language: A Contemporary Introduction, Real Conditionals (Routledge,

Oxford has created a website to accompany The Epistemology of Non-​Visual

I.1. The Epistemic Landscape

This book is an anthology of new essays by top researchers in epistemology

landscape within epistemology. The answers to the questions relating to the

of redness contains the red experience itself (Chalmers 2003). Infallibility

Principle of Inferential Justification

To be justified in believing one proposition p on the basis of another propo-

I.2. Multisensory Experience and Multimodal

The standard discussion on perception includes only five exterocep-

include proprioception (our sense of the body’s position in space, move-

All of these experiences are instances of multisensory experiences (for

strengthened when sensory inputs are added, it is weakened in other cases.

I.3. Representation and Accuracy

Sensory reinforcement and discounting occur not only between sense

By comparison to visual experiences, which are typically thought to repre-

experience systematically misrepresents the world around us, but it does so

I.4. Audition and Aspect Perception

Auditory experiences pertain to the sense of hearing. As with visual experi-

(Young 2016),6 or that sound absences can be perceived as such (Sorensen

I.5. Epistemic Privilege and Perceptual Experience

I.6. The Epistemology of Event Perception

stages. This suggests a need to distinguish epistemic (personal) and cognitive

I.7. Emotional Experience and Interoception

Emotions, on Copenhaver and Odenbaugh’s account, are experiences

Given that emotions depend significantly on interoception, the question

realize that one is feeling stressed or anxious, the influence of emotions is

This volume provides novel approaches to questions about justification with

Harman, G. 1986. Change in View. Cambridge, MA: MIT Press.

O’Callaghan, C. 2015b. Not all perceptual experience is modality specific. In D. Stokes, M.

Sorensen, R. 2009. Hearing silence: The perception and introspection of absences. In M.

1.1. The Subjective Nature of Tastes and Tasting

afterimage, an aftertaste as we say, and that’s not a paradigm case of tasting

1.2. Taste Sensations Alone?

The early 20th-​century psychologist E. B. Titchner beautifully captures the

And the nineteenth-​ century physiologist Jean-​ Anthelme Brillat-​ Savarin

2 Thanks to Janice Wang for drawing this passage to my attention.

In a now-​classic 1982 article, psychologist and biologist Paul Rozin first

despite this erroneous localization (and classification), that we are aware of a

experience. No such possibility is available on Aristotle’s framework, and

4 This appears to be the position adopted by Richardson 2013.

1.3. Flavors as Constructs of the Brain

A more sophisticated and scientifically respectable subjectivism about taste

that our perceptions of both of them are extremely discriminating though we

1.4. Multisensory Perception and

It is easy to understand the motivation of neuroscientists who see flavors

Arguably, multi-​sensory integration may be at its most extreme in the case

flavor experience, and it doesn’t distinguish between veridical experiences

1.5. The Multisensory Perception

To be entitled to claim that tasting provides us with genuine knowledge of

of multisensory integration, and there are signs of just such superadditivity

8 For more on this point see B. Smith 2007.

The Epistemology of Non-Visual Perception

Oxford has created a website to accompany The Epistemology of Non-Visual

The early 20th-century psychologist E. B. Titchner beautifully captures the

And the nineteenth- century physiologist Jean- Anthelme Brillat- Savarin

In a now-classic 1982 article, psychologist and biologist Paul Rozin first

Arguably, multi-sensory integration may be at its most extreme in the case

An extensive empirical literature exists on cross-modal interactions, and

Although our results imply a hard-wired aspect of odorant pleasantness,

A subclass of unmyelinated afferents (C-tactile) provided us with a candi-

Macpherson, F. 2011. Cross-modal experiences. Proceedings of the Aristotelian Society

3.2. The First Problem: Top-Down Influences

I argued in section 3.2 that top-down influences on perception should not be

The auditory mental imagery in this case clearly depends on bottom-up

Bertelson, P. 1999. Ventriloquism: A case of cross-modal perceptual grouping. In G.

Physicalism: Olfactory properties are mind-independent physical prop-