In Other Shoes

In Other Shoes
Music, Metaphor, Empathy, Existence

K E N D A L L L . WA LTO N

1
 1
Oxford University Press is a department of the University of
Oxford. It furthers the Universitys objective of excellence in research,
scholarship, and education by publishing worldwide.

OxfordNewYork
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala LumpurMadridMelbourneMexico CityNairobi
New DelhiShanghaiTaipeiToronto

With offices in
ArgentinaAustriaBrazilChileCzech RepublicFranceGreece
GuatemalaHungaryItalyJapanPolandPortugalSingapore
South KoreaSwitzerlandThailandTurkeyUkraineVietnam

Oxford is a registered trademark 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, NewYork, NY 10016

Oxford University Press 2015

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 Cataloging-in-Publication Data

Walton, Kendall L., 1939-
[Essays. Selections]
In other shoes : music, metaphor, empathy, existence / Kendall L. Walton.
pages cm
Includes index.
ISBN 9780195098716 (hardcover : alk. paper)ISBN 9780195098723
(pbk. : alk. paper) 1. Aesthetics. 2. MusicPhilosophy and
aesthetics. 3.ArtPhilosophy. I.Title.
BH39.W327 2015
781.17dc23
2014021506

135798642
Printed in the United States of America
on acid-free paper
 CONTENTS

Preface vii

1. Empathy, Imagination, and Phenomenal Concepts 1

2. Fictionality and ImaginationMind the Gap 17

3. Two Kinds of Physicality in Electronic and TraditionalMusic 36

Postscripts to Two Kinds of Physicality in Electronic and

Traditional Music 52

4. ThoughtwritingIn Poetry and Music 54

5. Its Only a Game!:Sports as Fiction 75

6. Restricted Quantification, Negative Existentials, and Fiction 84

7. Existence as Metaphor? 89

Postscript to Existence as Metaphor? 114

8. Projectivism, Empathy, and Musical Tension 118

9. Listening with Imagination:Is Music Representational? 151

10. Metaphor and Prop Oriented Make-Believe 175

11. Understanding Humor and Understanding Music 196

v
vi C o n t e n t s

12. What Is Abstract about the Art of Music? 208

13. The Presentation and Portrayal of Sound Patterns 230

Postscripts to The Presentation and Portrayal of

SoundPatterns 248

14. Fearing Fictions 252

15. Spelunking, Simulation, and Slime:On Being Moved by

Fiction 273

Acknowledgments 289
Index 291
 P R E FA C E

Many of my thoughts about the disparate topics listed in my subtitle have some-
thing to do, positively or negatively, peripherally if not centrally, with what
Icall other-shoe imaginative experiences. Empathy will probably seem to most
of us to involve such experiences; the other topics may not. But in Empathy,
Imagination, and Phenomenal Concepts, Iargue that to empathize is not nec-
essarily to engage in other-shoe imagining, nor indeed to engage in imagining
of any kind. In Thoughtwritingin Poetry and Music, Iclaim that the appre-
ciation of expressive qualities of much poetry and music, which has often been
said to consist in something like other-shoe experiences, does not necessar-
ily, and also need not involve imagining at all. This volume might better have
been titled, In and Out of Other Shoes. Moreover, although Fictionality and
Imagination:Mind the Gap concerns other-shoe experiences only indirectly, it
too treats imagination negatively. It argues, contrary to the position Itook in pre-
vious writings, that prescriptions to imagine a proposition are not sufficient for
its being fictional, true in a fictional world. It will be evident, nevertheless, from
other essays in this collection, that Ithink imaginings, in particular other-shoe
imaginings, play important though varied roles in many areas where they may
not be suspected.
Although shoes are designed to be worn by persons, many stand empty for
much of their lives; some are never worn at all. My title relies on the metaphor of
imagining oneself in, imaginatively occupying, someone elses shoes, but it focuses
on the shoes themselves, not their wearers. Some of the essays do examine expe-
riences of imaginative identification with a person (actual or fictitious) or other
imaginings targeting a person. But these are special cases of other-shoe experi-
encesby which Imean imagining oneself in a certain situation, or having cer-
tain experiences, or being a certain way, whether or not one has in mind another
person (actual or fictional or imaginary or hypothetical) who is in that situation
or undergoes those experiences or is that way. Arecurring theme in much of my

vii
viii P r eface

writing has been resistance to the often gratuitous and distorting postulation of
fictional or imagined beings whose shoes appreciators imagine occupying, and
insistence on the prevalence and importance of mere other-shoe experiences.
I present the following essays, with one exception, in reverse chronological
order,1 beginning with essays that are wholly are partly new in this volume. Any
arrangement by topic would have highlighted some connections among them
while obscuring other equally significant ones. So I opted for the more neu-
tral chronological order. The exception is Spelunking, Simulation and Slime,
which Iplace at the end of the volume immediately following the much earlier
Fearing Fictions. Fearing Fictions, as well as chapters5 and 7 of my Mimesis as
Make-Believe,2 has been widely misunderstood. Spelunking was designed partly
as an antidote, clarifying my earlier discussions and also expanding on them. Iurge
readers interested in the nature of appreciators emotional responses to fiction to
consult Spelunking alongside Fearing Fictions and the related parts of Mimesis.
Readers who want to pursue particular issues according to their own interests
will find discussions of the following topics in the chapters listed below (in some
cases without using these words):

Empathy:Chapters1, 3, 4, 8, 11.
Existence and fictitious entities:Chapters6, 7.
Expression in the arts: Chapters3, 4, 8, 9, 10, 12.
Fictionalist metaphysical theories:Chapters7, 10.
Fictionality (truth in fiction): Chapters2, 7, 8, 9, 10, 14.
Ineffability:Chapters1, 8, 10, 12.
Mental simulation and emotional contagion:Chapters1, 3, 4, 8, 15.
Metaphor:Chapters7, 8, 9, 10, 12.
Music:Chapters3, 4, 8, 9, 10, 11, 12, 13.
Narrators, apparent (fictional, implied) artists, musical personae:Chapters2, 3,
4, 8, 9.
Prop oriented make-believe:Chapters7, 10.
Responses (especially emotional responses) to fiction:Chapters1, 2, 3, 4,
5, 8, 9, 14, 15.

I have added postscripts to some of the previously published essays and inserted
new notes [in square brackets] in many of them, some pointing out connections
or conflicts with more recent writings by other theorists, or links among my
own. Iapologize for occasional overlaps among the essays. Each is designed to
stand alone.

1
An approximation thereof anyway; the date of origin of some of them is ambiguous.
2
Mimesis as Make-Believe (Cambridge, MA:Harvard University Press, 1990).
In Other Shoes
 1

Empathy, Imagination, and

Phenomenal Concepts

I. What Is Empathy?
Definitions of empathy are all over the map.1 But one ingredient usually
included is the idea that empathy is or essentially involves a special kind of imag-
ining, an imaginative experience described variously as role taking, perspective
taking, imaginative identification, or imagining oneself in another persons
shoes. I shall argue that, although many or most empathetic experiences do
involve some such imaginative experience, empathy is best understood as not
requiring this, indeed not requiring any imagining at all. The work imagination
is supposed to do is actually accomplished instead by the deployment, in empa-
thetic experiences, of phenomenal concepts.
Empathy, as Iunderstand it, always has an object, a target; it is like sympathy
in this respect. To empathize is necessarily to empathize with someone or some-
thing. Gregory Currie (2004: 181184) has identified tricky questions about
the nature of this object directedness.2 My account will provide a simple and
satisfying answer to them.
Alvin Goldman (2006: 201ff.) more or less identifies empathy with men-
tal simulation. Ido not. Itake simulation not necessarily to have a target. One
may simulate being in a certain situation, for instance, without simulating the

1
Along with many philosophers and psychologists, Idistinguish empathy from sympathy (in its
modern sense), i.e., from feeling or being sorry for someones misfortune. Empathy often leads to
sympathy, but doesnt necessarily, and one can sympathize without empathy. Also one can empathize
with a person who does not suffer misfortune and so is not a candidate for sympathy. My interest is
not, of course, in the proper use of the word, empathy. Iaim to understand better what happens and
what matters in many or most cases that are frequently counted as instances of empathy.
2
Currie points out that these questions are similar to ones concerning what makes joint atten-
tion joint. He proposes a functional account of the link between the empathizer and her target. My
proposal is very different.

1
2 In Other Shoes

experience of any particular person in a situation of that kind. The simulator

need not be thinking even of a fictional or imaginary or hypothetical or merely
possible person in the situation in question, even if she is aware that it is pos-
sible that someone should be in that situation. Simulation, or a close relative,
frequently targetless, is the fundamental kind of experience one enjoys when
one is caught up in a story or a fiction of one kind or another. One simulates an
experience of learning about or knowing about or (in the case of visual fictions)
observing the characters or events or activities of the fiction. Sometimes there is
a narrator or character whom one simulates, with whom one might empathize,
but often there is not.3 Empathy must have a target; simulation neednt. (There
may be other differences as well.) My interest now is in empathy.
Empathy is often said to be a source of knowledge about the mental state or
activities of the person empathized withknowledge of a special kind, Verstehen,
or knowledge of what it is like for the target, not, it is said, mere propositional
knowledge. My account will explain how empathy might provide (or constitute)
such knowledge. But Iwill argue that it is best understood as propositional, as a
special kind of propositional knowledge.
We must, of course, allow that one who seems to know or understand a target
by empathizing with him may get him wrong. If we take empathy to be a suc-
cess term, as Iprefer to do, this will be a case of merely apparent empathy; it may
amount to an illusion about how it is with the target. Imake no claims about how
often empathizers are right about their targets, nor how likely one is to be right in
any particular case. We often have the impression of empathizing successfully in
any case, and such impressions, illusory or not, figure importantly in our think-
ing about others and our interactions with them.4
Tolstoy (1899) thought that the primary function of art is to bring people
together, to foster a sense of community. He might have, though he didnt, put
this by saying that art enables appreciators to empathize with the artist and
with other people who experience the same work, sharing emotions with both.
Notice that the sense of community, of togetherness, that Tolstoy thought so
important can be fostered by art even if the apparent empathy is illusory, even if
appreciators are mistaken in thinking they share the emotions of the artist and
other appreciators.
Nancy Eisenberg and Janet Strayer (1987: 5) define empathy as an emo-
tional response that stems from anothers emotional state or condition and that

Iargued for this in Walton (1997:3749).

3

4
Even when empathizers are right about their targets, one might consider the justification for
their judgments insufficient for their beliefs to qualify as knowledge. (Thanks to Robert Stecker.)
Iwont worry about this here. For a discussion of how reliable a source of knowledge empathy is, see
Matravers 2011.
 Empathy 3

is congruent with the others emotional state or situation. This is one definition
that does not mention anything like role taking or other-shoe imagining. It is
clearly inadequate as it stands, for reasons to be found in Currie (2004:181).
It doesnt require a sufficiently intimate connection between the empathizers
experience and her target. Emilys psychological state might be similar to (or
congruent with) Oscars and might have been caused, in one way or another,
by his (or by his situation as she understands it), without her empathizing with
him. It wont help to add that she is aware of the similarity and realizes that his
state caused hers. Her experience might still fail to be one of empathy.
Is other-shoe imagining what is needed? What makes Emilys experience
count as empathy with Oscar, it may seem, is that it results from imaginatively
occupying Oscars shoes. This is not a satisfactory answer. For one thing, it is
not at all clear what kind of imagining, perspective taking, or whatever, empathy
involves, what it is to imaginatively occupy anothers shoes in the required sense.
As we shall see, some possibilities dont connect the empathizer and her object
sufficiently or in the right way, and others are difficult to make sense of. There is
a better way to understand the link between the empathizer and the target.

II. Parallel and Fancy Imagining

On learning that Oscar is exploring a cave, Emily imagines doing so herself. She
then finds herself experiencing a feeling that she identifies as panic. Thinking
that Oscars psychological make-up is like hers in relevant respects, she infers
that he experiences panic; she describes him as feeling panicked.5 Emily is
engaging in what Icall parallel imagining, imagining parallel to Oscars actual
experience. She imagines herself, Emily, to be in a certain situation and to be
performing certain activities, ones like those she takes Oscar to be in and to be
performing. Oscar is not in the content of her imagining.6
Emilys experience does not amount to empathy (even assuming she is cor-
rect in thinking he feels panic). Empathy, central instances of empathy at least,
or the most important kinds of empathy, require a closer, more intimate link
between the empathizers experience and her impression of or judgment about

5
Cf. Walton 1997:3749.
6
What Amy Coplan (2011) calls self-oriented perspective-taking and Peter Goldie (2011)
in-his-shoes perspective shifting are parallel imaginings. We need not restrict parallel imagining
to imaginings about oneself. If Emily imagines exploring a cave without imagining herself doing so
(if this is possible), her imagining will count as paralleling Oscars actual experience. What makes
it merely parallel is that it is not imagining about Oscar. The instances of parallel imagining Iwill
consider, however, are ones in which one does imagine about oneself; one imagines oneself doing or
experiencing what another person does or experiences.
4 In Other Shoes

the target persons experience. Moreover, Emilys experience as described does

not involve anything like Verstehen, knowing what is like for Oscar in the cave.
She acquires mere ordinary propositional knowledge about him, knowledge that
he feels panic.
Maybe what is required is a stronger sense of imagining, imagining that
does not merely parallel the experience of the target. Simon Baron-Cohen
(2003:24)offers a colorful characterization that seems to point to something
more interesting:

Empathy involves a leap of imagination into someone elses head. While

you can try to figure out another persons thoughts and feelings by read-
ing their face, their voice and their posture, ultimately their internal
world is not transparent, and in order to climb inside someones head
one must imagine what it is like to be them.

What sense can be made of this? Shall we say that to empathize with Oscar
Emily must imagine herself being Oscar, not just being similar to him and/
or in a similar situation and/or performing similar actions? If this means
imagining an identity between Emily and Oscar, what is imagined (that
Emily = Oscar) is metaphysically impossible, something that, some claim,
cannot be imagined.
Coplan (2011) and Goldie (2011) both take imagining being another person
to be necessary for empathizing with her (in their preferred senses of empathy).
What is crucial for them seems to be that one take on, in imagination, relevant
aspects of the targets personality, character traits, emotional dispositions,
desires and inclinations, etc. This would not seem to require imagining being
(literally) the other person, however. Imight imagine myself with inclinations
and interests like those of my target, imagine myself being extroverted, as my
target is, even if Iam introverted. This would be parallel imagining, in my sense;
Goldie (2011) calls it ambitious in-his-shoes perspective-shifting. In fact, it is
not clear to me how to distinguish between imagining being her and possessing
many of her properties, and imagining just being myself and possessing all of the
same properties, while knowing that they are properties of her. Of course, there
is the question, in either case, of whether or to what extent people are capable of
imagining having desires or personalities different from their actual ones. Goldie
is very skeptical about this; Coplan much less so.7
Richard Wollheim (1984:75)questions the intelligibility of imagining one-
self to be identical to someone else and introduces instead a notion of imagining

7
Goldie (2011) argues interestingly that empathetic perspective-shifting is conceptually impossi-
ble, not just ruled out by contingent limitations of our imaginative abilities. Iam not (yet) convinced.
 Empathy 5

another person centrally. According to Goldie in an earlier essay (2000), to

empathize with someone is simply to imagine her centrally, in Wollheims sense.8
Ido not have a very good grip on the notion of imagining centrally, however, and
Wollheims explanation of it is inadequate.
In imagining the Sultan Mahomet IIs entry into Constantinople in 1453,
Wollheim says, he might imagine the Sultan centrally. This is not to be con-
fused, he thinks, with centrally imagining myself in the Sultans shoes ... and
therefore doing what Iknow he did there and then. ... Imagining myself in the
Sultans shoes [what Icall parallel imagining] ... leaves it open to me at any
moment to imagine myself brought face to face with the Sultan. And that is
something that [imagining the Sultan centrally] clearly rules out (Wollheim
1984:7576).
I dont see that this is clearly ruled out. To imagine the Sultan, centrally or
otherwise, coming face to face with the Sultan would be to imagine something
incoherent. But it seems that we do sometimes imagine incoherencieswhen
we dream, for instance.9 Moreover, it is not clear that parallel imagining, imag-
ining myself in the Sultans shoes, does leave it open to me ... to imagine myself
brought face to face with the Sultan. If my project is to imagine myself doing and
experiencing things as the Sultan actually did, I surely wont imagine coming
face to face with the Sultan.
What to do? Itake no stand here on whether it is possible to imagine an iden-
tity between oneself and another person. Idont rule out making sense of imag-
ining someone centrally. And perhaps another stronger-than-parallel kind of
imagining or imaginative identification is the key to the notion of empathy. But
there is an easier way. We can do an end run around the mysteries of imagina-
tion, for the purpose of explaining empathy. Mere parallel imagining will turn
out to be enough. Indeed, as Imentioned, empathy does not require any imagin-
ing at all.

III. Phenomenal Concepts; Samples

In Walton (1999) Inoted that in many instances in which we learn about others
by imaginatively occupying their shoes, what we learn contains a demonstra-
tive element. If, imagining myself in the shoes of a person who misses his plane
or a person crawling in a cave, Ifeel upset or panicked and judge him to be upset
or panicked, the content of my judgment, in the first instance, is that he feels like

8
Cf. also Smith 2011:100.
9
For what it is worth, this incoherence strikes me as less unimaginable than imagining that
Kendall Walton is identical with the Sultan.
6 In Other Shoes

this (or something like this), where this refers to an aspect of my own current
state of mind. Iam using my feeling of upset or panic to represent his.10 Imight
also find a predicate to characterize his state; Imight describe him as upset or
as panicked. But the predicate is likely to be considerably less specific than what
Irepresent to myself about him using my own mental state. It does not exhaust
the content of my thought when Isay to myself, He feels like this. Judgments
of this kind employ what have been called phenomenal concepts, although this
notion was introduced to serve an entirely different purpose.11
Sometimes this, in judgments of the form, She feels like this, refers not to
ones actual mental state but to the content of one or another of ones intentional
attitudes, the content of ones imagining, for instance. More about this presently.
Iwill focus, now, on cases in which it is an aspect of ones actual state of mind
that one uses as a sample.
To judge that a person feels like this is to use ones own mental state as a sam-
ple, indicating a property that one then attributes to her. It will be useful to look
briefly at the work samples in general do, in our thinking and speaking.12 Awitness
to a bank robbery uses a sample to describe the getaway car:They drove off in
that kind of car, or ... in a car like this one, she says, pointing to a vehicle parked
in front of the police station. To modify an example from Jane Heal:Imight say,
She sang thus, followed by a warbly, out of tune rendition of Yankee Doodle.13
My vocalization serves as a sample of a manner of singing that Iattribute to the
other person, and perhaps also, though not necessarily, a sample of a way of sing-
ing Yankee Doodle. Samples can be picked out by means of descriptions or proper
names, of course, as well as by demonstratives. The witness might describe the
getaway car as being like the one her mother drives. We might describe a child as
another Mozart or another Einstein. Fictional objects are often used as sam-
ples:We speak of a Cinderella team, a catch 22, a Trojan horse.
In all such cases, a particular thing (actual or fictional14) is used to call
to mind or indicate a kind, a property, one that the particular saliently

10
Heal (2003a, b) made similar suggestions and explored the more general notion of indexical
predication.
11
Brian Loar (1997) introduced it first, so far as Iknow, in connection with the mind-body prob-
lem. Cf. also Lycan 1996, Papineau 2002, Tye 2003, and others. Loar and others appeal to phenom-
enal concepts to support materialism, the idea being (roughly) that although we have both mental
concepts (phenomenal concepts) and physical ones, they pick out properties of only one kind, physi-
cal properties. Itake no stand on whether this defense of materialism is successful. Nor will Iattempt
to choose among the various definitions of phenomenal concepts in the literature.
12
What Isay about samples owes much to Nelson Goodmans (1968) notion of exemplification.
Asomewhat similar notion of exemplarization is central in Keith Lehrers (2012) theory of art.
13
Heal 2003b:206. Cf. also Herbert Clarks (1996:172174) discussion of demonstrations.
14
Here, as so often is the case, it is convenient to speak or write as though fictional entities exist.
No ontological commitment is implied.
 Empathy 7

possesses (or a property it is thought saliently to possess or one commonly

associated with it). The sample thus functions as a predicate:Like linguistic
predicates, it picks out a property, which a speaker may then attribute to
something else.15
To use a particular as a sample in this way is not to say something about the
sample, any more than using a linguistic predicate to attribute a property to an
object is to assert something about the predicate. What the robbery witness
asserts is not that the vehicle she points to and the getaway vehicle are similar or
of the same kind. She asserts that the robbers made their getaway in a vehicle of
a certain sort (e.g., a recent model red pickup truck)using the sample vehicle
rather than a linguistic predicate to indicate what sort it is.
A given particular possesses many properties any of which it could serve as
a sample of. Which one (or ones) does it pick out on a given occasion?16 This is
obviously a highly context dependent matter. The relevant property will be one
that is salient in the context, where the context includes background informa-
tion possessed by the speaker and her hearers, the point of the conversation,
etc. Sometimes a speaker will specify what kind of property the relevant one
is (This shade of color, That kind of architecture, That make of automobile),
while still relying on the sample to indicate which property of the specified
kind it is. But often this is unnecessary. We can think of a sample as a highly
ambiguous predicate, disambiguated (partially or wholly) in a given conver-
sational situation in roughly the ways ambiguous linguistic predicates are
disambiguated.17
I emphasize that we neednt have a linguistic predicate or phrase or any
way of saying, with words, what the property is (even when the sample does
pick out a fairly definite property); using the sample may be our only way of
identifying it. This is obviously true of shades of color, most of which have
no names. (Even if we do have a name, e.g., burnt sienna, we cant reliably
ascertain whether something possesses the property without comparing it to
a sample.)18

15
Many of our beliefs have the form:The color of her hair is ___, or The song he was singing
went ___, where the blanks are filled with images, sensory impressions, or what have you, but cer-
tainly not words (Kaplan 1968:208).
16
A sample is a sample of some of its properties but not others (Goodman 1978:64). Clark
(1996:173174) points out that in using demonstrations for communicative purposes, the speaker
must rely on the hearers recognizing, somehow, which aspects of the demonstration are intended to
depict the subject of discussion and which are not.
17
Sometimes the ambiguity is left unresolved, perhaps deliberately. The speakers point may be
simply to induce listeners to think about what features of the sample might also characterize the sub-
ject. This is the point also of some metaphorical attributions.
18
Cf. Heal 2003b:196222.
8 In Other Shoes

IV. Empathy Again

Back to the special case in which one uses ones own psychological state as a
sample, as when Emily empathizes with Oscar. We access our psychological
states by means of introspection (or anyway, I assume, in some first person
privileged manner). So although Emilys mental state serves as a sample in
conceptualizing her judgment about Oscar, it will be of limited usefulness in
explaining to a third party what she knows or believes about him. If she identi-
fies her own state as an instance of panic, she can tell another person, Thomas,
that she thinks Oscar feels panic. But she wont thereby be telling Thomas all
that she knows or believes about Oscar. Her judgment is that Oscar experiences
a particular sort of panic (panic like this), one she may not be able to express
in words.19
It is open to Thomas to declare that Oscar feels like that, referring to Emilys
state of mind. He might have good reason to think her state is a fair sample of
Oscars. But there is a sense in which he wont know what property it is that
he is thereby attributing to Oscar. (Compare:Imight assert, truly and on good
authority, that a person suffers from ribose-5-phosphate isomerase deficiency with-
out having the foggiest idea what this is.) It is not clear that Thomas will even
have disambiguated the predicative sample. In saying Oscar feels like that he
might mean just that Oscars and Emilys mental states are similar, that there is
a property that they share. This is not using her state as a sample, a predicate, to
specify a particular property.
There remain difficult questions, which Iwont attempt to answer, about how
it is that Emily picks out one aspect or property of her current mental state, as
opposed to others, which she attributes to Oscar, and what it is for her to do so.
Presumably Emily attends to one property rather than others, one that is salient
to her in the context. These are questions also for those who use the notion of
phenomenal concepts for other purposes, of course. 20 Iassume that they can be
answered satisfactorily.

19
Peter Goldie (2000: 181) claims that to understand anothers emotions, we must be able
. . . to say what the emotion is which that person is experiencing. According to Alvin Goldman
(2006:especially 127, 224, 259), we classify our mental state, then attribute it to the other person.
If saying what the state is, or classifying it, means having words for it, neither is necessary for under-
standing the target person. One can use ones own mental state as a sample.
20
A mental sample that exemplifies one phenomenal property will exemplify many. ... Which of
the exemplified properties is the one to which the demonstrative concept THAT PHENOMENAL
PROPERTY refers? It seems that appealing to a mental sample does not help to fix the reference of
the phenomenal concept at all (Tye 2003:95). Tyes conclusion is too strong. Yes, pointing to the
sample does not by itself fix the reference. But this does not mean that pointing to it doesnt help; it
may even be necessary.
 Empathy 9

In paradigm or standard instances of empathy, the empathizer uses some

aspect of her current mental state to understand the targets, in the way Ihave
described. Icount this as a necessary condition, at least, for empathy.
Is to empathize necessarily to judge, or believe, or know, that the other person
is like this? That seems too strong. Lets count as instances of empathy cases
in which a person experiences a target as feeling like this or is under the impres-
sion that she does, without requiring that she say to herself anything like He
feels like this.21 This will help to accommodate what has sometimes been called
automatic empathy, empathy based on emotional contagion, for instance.22
The empathizers use of her own current mental state as a sample constitutes
an especially intimate link between her state and her judgment about or impres-
sion of the targets experience. This, Isubmit, is the link we are after, the ground
of the object directedness of empathy. It is by virtue of this that Emilys experi-
ence counts as empathy with Oscar.
Emilys judgment or impression is not merely that I am panicked, and so is
Oscar, but rather, Oscar is as Iam, like this. She can appropriately say, I know
how it is with him or I know how he feels, where know carries a connota-
tion of intimacy, acquaintance.23 This is close to what one might call Verstehen,
or knowing what it is like for Oscar. Notice that the content of what she knows
is in propositional form:She knows that Oscar feels like this. But this is proposi-
tional knowledge of a special kind, with the sample taking the place of a linguistic
predicate in the formulation of what she knows. What Emily knows, the proposi-
tion she knows to be true, is not special or unusual. The very same proposition
could be expressed in other ways, as (a) He feels X, where X is a predicate
indicating the same property that like this does (we might not possess such a
predicate, but there could be one); or as (b) He feels as Patricia did on such and
such occasion. What is special is Emilys way of understanding this proposition,
her way of accessing and thinking about it. She does so by means of a demon-
stratively indicated sample or a sample identified in some other way, rather than
a linguistic predicate. Her knowledge of this proposition, her knowing that he
feels like this, is special. It is knowledge of a (not so special) proposition that she
represents to herself in a special way.
It looks as though much of the explanatory work we expect of a notion of
empathy is done by the role of phenomenal concepts in the empathizers experi-
ence, rather than some sort of in-her-shoes imagining or imaginative identifica-
tion. As a first stab, Ipropose to define empathy as, simply, using some aspect
of ones current mental state as a sample to understand another person, in the

21
Merely apparent empathy, if the supposed empathizers impression is mistaken.
22
Cf. Goldman 2006:40, 207208.
23
Assuming that her judgment or impression of him is correct.
10 In Other Shoes

way Ihave described, i.e., judging or experiencing the target person to be feeling
like this.
Refinements are needed. Suppose I learn by some non-empathetic means
that Sadie is sad:She tells me that she is, or her shrink does, or Iapply a theory,
inferring from her actions or facial expressions that she is sad. Suppose also that,
as it happens, Iam sad as well. Now Iam in a position to say, She feels like this,
referring to my state of mind and using it as a sample indicating (just) the prop-
erty of being sad. This hardly qualifies as an instance of empathy, for two distinct
reasons. (a) Sadness is a very unspecific mental state. We might want to require
that, to count as empathizing, one must use ones mental state as a sample of a
much more specific property. (An empathizer is likely to say not merely, I know
how you feel, but I know just how you feelalthough this is an exaggeration
if it means that the empathizer knows exactly how the target feels.) (b) My men-
tal state, my sadness, is not the source of my knowledge about Sadie (or of my
impression of how it is with her). It is because Iknew already that she was sad,
that Iwas able to use my state in characterizing hers. In paradigmatic instances
of empathy, the empathizer judges or has the impression that the target feels a
certain way, because she (the empathizer) does.

V. Empathy without Imagination

What, then, does imagining have to do with empathy? Certainly imagining is an
important ingredient of many empathetic experiences, perhaps most of them,
including Emilys as Idescribed it. Her imagining crawling in a cave, when she
learns about Oscars adventure, makes two important contributions to her
experience:It induces in her the panicked feeling that she uses as a sample in
understanding Oscars state of mind. And (on the assumption that Oscars psy-
chological makeup, etc., is sufficiently like hers in relevant respects) the fact that
her feeling was generated by the imaginingby imagining being in a situation
like the one Oscar is inarguably gives her some reason to think that it is a fair
sample of Oscars state of mind, that her judgment that Oscar feels like this has
a reasonable chance of being true. This is mere parallel imagining. Ididnt postu-
late any fancier variety, and Isee no reason to think empathy requires anything
fancier.
Moreover, circumstances other than imagining can make the contributions
imagining makes in Emilys case.
Sometimes an empathizer is actually in a situation similar in crucial respects
to that of her target. If you and Ihave a mutual friend who has suffered misfor-
tune, we may both grieve for him. Imight then judge that you feel like this,
referring to my own feeling of grief (a fairly specific kind of grief). My actual
 Empathy 11

situation generates the state Iuse as a sample, and the fact that my situation is
similar to yours may give me some reason to think it is a fair sample, may help
to justify my judgment. My grief is not caused by yours, or by my knowledge of
the situation you are in. Yet Imay think of you as feeling like this. And Ican say
that Iknow how you feel or what it is like for you. This should count as an
instance of empathy.
Alternatively, you and Imay both experience grief but for different persons.
Igrieve for a friend of mine, and judge that you, in grieving for your friend, feel
like this. Then the property my statement serves to indicate is different; it
is not grief for so-and-so, but just grief (probably a particular sort of grief) for
someone or other. Still, in using my statement as a sample, Iam empathizing
with you.
That being in a situation similar to the targets helps one to empathize with
him is certainly not news. But the usual idea is that being in a similar situation
makes it easier than it would be otherwise to put oneself in his shoes or to
imaginatively identify with him. My proposal is that this last step is unneces-
sary. Actually being in a parallel situation is enough.
What about contagion? Entering a jolly gathering, Carol forgets her troubles
and finds herself in a cheerful mood. Some count such contagion as a kind of
empathy; some do not.24 Carol might be entirely unaware that she caught her
mood from the others, if she even notices their good spirits. In that case she
wont judge or have the impression that they feel like this, and her experience
wont qualify as empathy. But if she is aware of the contagion (implicitly at least),
and does judge or have the impression that they feel like this, attributing an
aspect of her mental state to them, she is empathizing with them. Imagination
seems not to be involved.25
A science fiction example: By manipulating my brain, doctors produce in
me an affective (and/or perceptual) state that Ican use as a sample. There may
be good scientific evidence that my artificially acquired brain state is correlated
with an experience of a certain kind, an experience of severe claustrophobia, or
of drowning, or undergoing a near death experience, or smelling durian, or feel-
ing ostracized. Knowing this, Iuse my state as a sample to understand the experi-
ence of a target person who suffers claustrophobia, or drowns, or smells durian,
... Ijudge that she feels like this. If my judgment is correct, Iam empathizing
with her. This is empathy without imagining. It is not imagining, neither parallel

24
Hoffman (2001) does; Goldman (2006:207208) apparently does, citing Hodges and Wegner
(1997); Coplan (2002, 2011)doesnt.
25
At least it is not obvious that imagination is involved. Perhaps the mechanism of emotional
contagion includes (implicitly) imagining behaving as the infecting persons are behaving. This is not
imagining oneself in their situation, as in the instances of parallel imagining Ihave considered. It is in
a different sense imagining oneself in their shoes.
12 In Other Shoes

nor fancy imagining, that causes my experience or justifies my taking it as a fair

sample of the targets experience.
Memory may be responsible, in either of two ways, for a state of mind an
empathizer might use as a sample. Iwould count both of them instances of imag-
ining, but some would not. In any case, only parallel imagining is needed.
A person might re-live a previous automobile accident, feeling panic or
terror again, and judge that a target person experiencing an accident feels like
thisi.e., in certain respects like she now feels. The empathizer imagines
(Iwould say) a huge SUV bearing down on her now, imagines facing a clear and
present danger. (She does not really think she is in danger, for a second time,
while she is re-living the past experience.)
In other cases, memory traces left from previous experiences, rather than
re-living or recalling any particular one or ones, may affect a persons experience
as she imagines being in a similar situation. (Having been inebriated previously,
one might feel nauseous on imagining downing a bottle of whiskey.) The imag-
iner may then judge that a target person feels like this.26

VI. Extensions, Variations, Relatives

Empathy, as Ihave explained it, is an occurrent state (or experience, or activity),
in which one uses ones own current mental state as a sample. Call this the pri-
mary sense of empathy.
We can, of course, recognize a dispositional sense as well. One may be dis-
posed to empathize, occurrently, with a particular target or targets of a certain
kind, or able to do so if we set our mind to it. We can describe someone as an
empathic person, if she is uncommonly capable of empathizing with others or
does so uncommonly often.27
I have concentrated on empathy with a targets phenomenological states. One
can empathize also with a persons beliefs, desires, hopes, intentions, etc., and
with his coming to believe something, changing his mind, forming intentions,
seeing his hopes dashed, etc. The extension of my account in these directions
is not entirely straightforward. If an empathizer does not believe what the tar-
get person believes, for instance, she cannot judge, He believes like this, with
the demonstrative referring simply to a belief of her own (nor can she judge,
He came to believe like this, where this refers to her own coming to believe
something).

26
Cf. Walton 1999: VII.
27
Thanks to Sarah Buss.
 Empathy 13

I mentioned that one sometimes uses, as a sample, not ones actual mental
state but the content of one or another of ones intentional attitudes, the content
of an imagining, for instance.28 This is true of viewers experiences of point of
view shots in film. Following a shot of a character looking out a window, there is
a shot of a scene outside. Watching the second shot, we imagine observing the
scene, and we judge that the character looking out the window has an experience
like this, like the one we imagine enjoying. We do not attribute to the character
an experience (much) like our actual visual experience, a visual experience of
a film shot, of a depiction of the scene outside the window. The experience we
attribute to the character is like our actual one only insofar as imagining seeing
is like actually seeing.
Emily might, possibly by putting herself in Oscars shoes as he crawls in the
cave, imagine deciding (after deliberation, or spontaneously, in a spirit of des-
peration or one of studied calmness) to back slowly through the passageway,
hoping to find a place where she can turn around. Or she may imagine planning
to sue a tour operator who assured her that the cave was perfectly safe for novice
spelunkers. She does not actually decide to back out or to sue the company, and
she does not really hope that she will find a place to turn around; she is not in a
cave and no tour operator assured her of a caves safety. But she might judge (or
speculate) that Oscar did or will make decisions or plans or entertain a hope
like this, i.e., like the decisions or plans she finds herself imagining making or
the hope she entertains in imagination.
This probably should count as empathy only if Emilys imagining is of an
appropriately experiential sort (like that of the viewer of the point of view shot).
Imagining that Idecide to back out, and judging that Oscar decides like this,
i.e., as Iimagine that Idecide, may be just to judge that he decides to back out.
Itake imagining deciding to back out to be richer and more experiential, prob-
ably including phenomenological elements, than merely imagining that Idecide
to back out. Judging (or having the impression) that another person decides as
Iimagine deciding, in this richer sense of imagining, may reasonably count as

28
Compare what Herbert Clark calls demonstrations:Aperson who, pretending to drink tea in a
certain manner, declares, She drinks tea like this, attributes to her what he pretends to do, not just
aspects of what he actually does (Clark 1996:172174). The pretenders action is an icon, Clark will
say, of a way of drinking tea. It is not itself a sample of that; the pretender is not (actually) drinking tea.
But if the pretender uses an actual teacup in his demonstration, it is a sample of a way of bringing a
teacup to ones mouth; otherwise it is a sample at least of a way of moving ones hands toward ones
mouth. Many or most icons used in demonstrations, in Clarks sense, are or include samples. Iam
not sure whether all things used as samples should count as icons. In any case, many uses of samples
are not demonstrations in Clarks sense, i.e., many of them are not communicative signals, the speaker
having Gricean intentions to produce an effect in a hearer. Iam interested especially in cases in which
samples enable one to articulate a thought, but dont (or cant) serve a communicative purpose.
14 In Other Shoes

a case of empathy, of empathizing with the targets deciding. The same goes for
empathizing with a persons hoping, believing, intending, desiring, etc.
To empathize in the primary sense, again, is to use an aspect of ones current
mental state as a sample. (We can now add that this may include the content
of a current imagining.) In the automobile accident example, the empathizers
sample is her present experience of re-living the accident she had previously. If
she remembers the experience but does not re-live it, she could use her previous
state, her actual experience of the accident, as a sample, judging that the target
feels like that, like I did then. Her previous experience may, of course, have
been more like the targets experience than a present re-living would be. It will
include actually fearing the SUV, not just experiencing a feeling of panic, and
perhaps (not necessarily) a more intense feeling of panic. But judging that the
target feels like that wont be empathizing with him, in the primary sense.
One might, however, have something like empathy consisting primarily in
recollections of a previous experience, without using ones current mental state
or the content of a current imagining as a sample.29 Iwill call this sort-of empa-
thy (introducing a new technical term). Suppose that Iremember an experience
Ionce had, without in any way re-living it, without now feeling anything like
Idid then; Imay be incapable even of imagining being in that state. But, lets say,
Ihave good reason to think that my previous state is a fair sample of how another
person, Joan, now feels. (A perceptive mutual friend tells me, You know the
way you felt last April? Thats how Joan feels now.) This is not empathy in the
primary sense. Imay not even be empathizing dispositionally (in the primary
manner) with Joan. But Imay remember some of the consequences of my pre-
vious experience, without experiencing even an echo of its phenomenology.
Imight remember what Iwas inclined to do or to believe as a result of feeling
as Idid:that Iwas really upset, or strangely calm, that Itended to notice certain
things and was oblivious to others, that Ifelt close to, or alienated from certain
people, etc. Given that Itake Joan to be having a phenomenal experience like my
previous one, Iexpect her to have inclinations to believe and act and feel similar
to the ones Ihadall without my now experiencing anything like the phenome-
nology of the experience. These are expectations that Imight have had as a result
of re-living my previous experience. If Iam right about Joan, about what she is
inclined to do or to believe or feel, we can count this as knowing, in a way, what
it is like for her, even though this knowledgemy sort-of empathyis based
just on memories of my experience, not my re-living it. (A variation: I might
re-live some aspects of the experience, some of my previous inclinations to do
or believe or feel.)

29
Thanks to Peter Railton.
 Empathy 15

Jonathan Weinberg has emphasized what he calls situational empathyby

which we know what its like to be in love, or to go on the job market for the first
time, or to go through the breakup of a relationship.30 Situational empathy, he
says, concerns phenomena that are extended in time, often involving distinct
qualitative characters at different times; it includes not just the phenomenol-
ogy, but affective, practical, behavioral, cognitive, etc. elements; and it can be
primarily past-tense, without needing to be produced via a re-imagining of a
past experience. Situational empathy comes in many varieties, and there is a
lot to be said about it.31 Isubmit that combinations of empathy in the primary
sense, empathy in which one uses contents of ones imaginings as samples, sort-of
empathy, and dispositions to experience the above, as well as awareness of ones
possession of such dispositions, will go a long way toward accounting for these
cases.32

References
Baron-Cohen, Simon (2003). The Essential Difference:The Truth about the Male and Female Brain.
NewYork:Basic Books.
Clark, Herbert H. (1996). Using Language. Cambridge:Cambridge University Press.
Coplan, Amy (2002). Empathetic Engagement with Narrative Fiction. Journal of Aesthetics and
Art Criticism 62(2):141152.
Coplan, Amy (2011). Understanding Empathy: Its Features and Effects. In Amy Coplan and
Peter Goldie (eds.), Empathy: Philosophical and Psychological Perspectives. Oxford: Oxford
University Press.
Currie, Gregory (2004). Arts and Minds. Oxford:Oxford University Press.
Eisenberg, Nancy, and Strayer, J. (1987). Empathy and Its Development. Cambridge:Cambridge
University Press.
Goldie, Peter (2000). The Emotions:APhilosophical Exploration. Oxford:Oxford University Press.
Goldie, Peter (2011). Anti-Empathy. In Amy Coplan and Peter Goldie (eds.),
Empathy:Philosophical and Psychological Perspectives. Oxford:Oxford University Press.
Goldman, Alvin (2006). Simulating Minds: The Philosophy, Psychology, and Neuroscience of
Mindreading. Oxford:Oxford University Press.
Goodman, Nelson (1968). Languages of Art. Indianapolis:Bobbs-Merrill.
Goodman, Nelson (1978). Ways of World-Making. Indianapolis:Hackett.
Heal, Jane (1995). How to Think about Thinking. In M. Davies and T. Stone (eds.), Mental
Simulation. Oxford:Blackwell.

30
In comments on an earlier version of this paper, presented at the 2011 Chapel Hill Colloquium
in Philosophy.
31
Aterminological preference:Idont call it empathy unless there is a target, a person in love or
on the job market or ... with whom one empathizes; otherwise it is simulation.
32
This essay began as a talk at a conference on Empathy, organized by Amy Coplan and Peter
Goldie, in 2006. Iam much indebted to Coplan and Goldie and to audiences on this and later occa-
sions for helpful discussion, and especially to Sarah Buss, Gregory Currie, Ron Endicott, Iskra Fileva,
Stacie Friend, David Hills, Mitchell Green, Fred Kroon, Bence Nanay, Peter Railton, Robert Stecker,
Damian Wassel, and Jonathan Weinberg.
16 In Other Shoes

Heal, Jane (2003a). Indexical Predicates and their Uses. In Mind, Reason and Imagination.
Cambridge:Cambridge University Press.
Heal, Jane (2003b). Lagadonian Kinds and Psychological Concepts. In Mind, Reason and
Imagination. Cambridge:Cambridge University Press.
Hodges, Sara D., and Wegner, Daniel M. (1997). Automatic and Controlled Empathy. In W. Ickes
(ed.), Empathic Accuracy. NewYork:Guilford.
Hoffman, Martin L. (2001). Empathy and Moral Development:Implications for Caring and Justice.
Cambridge:Cambridge University Press.
Kaplan, David (1968). Quantifying In. Synthese 19(1/2):178214.
Lehrer, Keith (2012). Art, Self and Knowledge. Oxford:Oxford University Press.
Loar, Brian (1997). Phenomenal States. In N. Block, O. Flanagan, and G. Gzeldere (eds.), The
Nature of Consciousness. Cambridge, MA:MIT Press.
Lycan, William (1996). Consciousnes and Experience. Cambridge, MA:MIT Press.
Matravers, Derek (2011). Empathy as a Route to Knowledge. In Amy Coplan and Peter Goldie
(eds.), Empathy:Philosophical and Psychological Perspectives. Oxford:Oxford Univerity Press
Papineau, David (2002). Thinking About Consciousness. Oxford:Oxford University Press.
Smith, Murray (2011). Empathy, Expansionism, and the Extended Mind. In Amy Coplan and
Peter Goldie (eds.), Empathy: Philosophical and Psychological Perspectives. Oxford: Oxford
University Press.
Tolstoy, Leo (1899). What Is Art? Trans. Louise and Aylmer Maude. London:Oxford University
Press.
Tye, Michael (2003). A Theory of Phenomenal Concepts. In A. OHear (ed.), Minds and Persons.
Cambridge:Cambridge University Press.
Walton, Kendall L. (1997). Spelunking, Simulation, and Slime: On Being Moved by Fiction.
In M. Hjort and S. Laver (eds.), Emotion and the Arts. NewYork:Oxford University Press.
Reprinted in this volume.
Walton, Kendall L. (1999). Projectivism, Empathy, and Musical Tension. Philosophical Topics
26(1/2):407440. Reprinted in this volume.
Wollheim, Richard (1984). The Thread of Life. Cambridge, MA:Harvard University Press.
 2

Fictionality and ImaginationMind

the Gap

The notion of truth in fiction, of a propositions being true in a fictional world

fictional, in my terminologyis absolutely central to my theory of fiction, to my
proposals about what fiction is, how we understand and appreciate and respond
to works of fiction, and how they are important in our lives.1 This notion is far
more important than that of works of fiction, the distinction between fiction and
nonfiction. Many theorists begin with and emphasize the latter. My strategy, in
Walton (1990), is the opposite.
In Walton (1990) and elsewhere, Ioffered a reductive account of fictional-
ity in terms of imagining:Aproposition is fictional, Isaid, just in case there is a
prescription to the effect that it is to be imagined. More precisely, a proposition
is fictional in (the world of) a particular work, W, just in case appreciators of
that work are to imagine it, just in case full appreciation of W requires imagin-
ing it.2 Ihave come to realize, belatedly, that this is only half right. Prescriptions
to imagine are necessary but not sufficient for fictionality.3 This means that the
notion of fictionality is even more fundamental than Ithought it was, and more
difficult to pin down.
Before explaining, it will be helpful to remind ourselves of an important fea-
ture of imagining, a respect in which imaginings differ from beliefs.

1
Thanks for discussion and comments to Carola Barbero, Gregory Currie, Daniel Groll, Stacie
Friend, David Hills, Fred Kroon, Patrick Maynard, Jerome Pelletier, Shaun Nichols, Dawn Phillips,
Denis Robinson, Richard Woodward, Steven Yablo. Thanks especially to David Braddon-Mitchell for
numerous very helpful conversations during an extended visit at Sydney University..
2
Walton 1990: 3941, 5761. I added an important qualification which neednt concern us
now:A proposition is fictional ... if it is to be imagined (in the relevant context) should the question
arise (1990:40, emphasis in original).
3
This realization was provoked first by conversations with Jerome Pelletier. Stacie Friend and
Patrick Maynard also called my attention to counterexamples.

17
18 In Other Shoes

1. Imagination (and Belief)

Several theorists have remarked that imaginings (propositional imaginings) are
much like beliefs with respect to their functional role. One similarity is supposed
to be that imagination preserves the inferential patterns of belief (Currie and
Ravenscroft 2002:1213). Inferences from a set of imaginings (together with
some beliefs) to further imaginings correspond to inferences from a set of beliefs
to further beliefs.4 This mirroring of inference patterns is supposed to be true not
only for free standing imaginative experiences, e.g. daydreams, but also for imag-
inative responses to fiction (Currie and Ravenscroft 2002:1314). We imagine
what is made explicit in a work of fiction. Then we draw inferences from it; we
infer that certain other propositions are also true in the fiction, and we imagine
them. These inferences are supposed to parallel inferences we would make from
propositions we believe initially to others which we come to believe as well.
Inferences from imaginings to imaginings do sometimes parallel inferences
from beliefs to beliefs, but very frequently they do not. In the case of imaginings
in response to works of fiction, they do (roughly) when what Icall the Reality
Principle of implication is operative. But the applicability of the Reality Principle
is very limited (Walton 1990:ch. 4). Iwill focus now on an especially fundamen-
tal way in which imaginings differ from beliefs, one that makes for differences
in inference patterns that do not result from limitations of the Reality Principle.
Fictional truths come in clusters, and so do ones imaginings of the propo-
sitions that are fictional. Different clusters correspond to different fictional
worlds, the worlds of different works of fiction, for instance, or different fanta-
sies or daydreams. Fictionality is always relative to a particular fictional world:a
given proposition is not fictional simpliciter, but fictional in the world of a par-
ticular novel, story, movie or daydream (Walton 1973: 287288; 1978: 10;
1990:3536). Much of the recent literature on the imagination ignores this clus-
tering, or pays insufficient attention to it.5 Nichols and Stichs (2003:36)boxo-
logical diagram of mental architecture features one box for beliefs, and another
one for imaginingsall of them.
The importance of the clustering lies partly in how imaginings combine with
one another. Contents of different clusters dont combine to justify inferences
in the way that contents of the same clusters do. If p and q are both fictional,
and belong to the same cluster, usually the conjunction, p&q, is fictional also,
and is to be imagined. This is not so if p and q belong to different clusters, dif-
ferent fictional worlds. Reading Kafkas Metamorphosis Iimagine that a boy was

Currie and Ravenscroft 2002:1214; Nichols and Stich 2003:2932. All agree that imaginings
4

differ functionally from beliefs with respect to their connections with action.
5
An exception is Skolnick and Bloom 2006.
 Fic tionalit y and Imag inati on 19

transformed into a bug. Reading War and Peace, Iimagine that things like that
just dont happen. (Or Iwould imagine this should the question arise.) There is
no pressure at all to imagine the conjunction, to imagine that someone turned
into a bug and people never turn into bugs, nor is there any tension between
the conflicting imaginings. While watching a performance of Othello, Ifantasize
about taking Othello aside, telling him about Iagos treachery and forestalling
the threatened disaster. Iimagine doing this, but Ialso imagine, in accordance
with the events on stage, that no one intervenes and that the tragedy unfolds as
scheduled. Icertainly do not imagine that Ido and do not reveal Iagos treachery
to Othello, and there is no tension at all in the fact that Ihave imagined incom-
patible propositions. These imaginings belong to different clusters.
Nothing quite like this clustering is true of beliefs. Any beliefs that I pos-
sess will combine with any others to justify the inference to their conjunction.
If I find myself believing contradictory propositions, I have a problem. I feel
obliged to change one or the other of my beliefs to avoid being committed to
the conjunction.6

2.Fictionality
My original account of fictionality, again, is this: a proposition is fictional in
(the world of) a particular work, W, just in case appreciators of that work are to
imagine it, just in case full appreciation of W requires imagining it. This proposal
hasnt been especially controversial. Many writers have gone along with it, some-
times changing the terminology.7 However, it simply will not do, and not just
because it is a little fuzzy, which of course it is. It gives us a necessary condition
for fictionality in a particular world, but not a sufficient one. 8
Counterexamples to its sufficiency, cases in which appreciators of a given
work are to imagine propositions that are not fictional in it, come in several

6
Acertain kind of clustering of beliefs does occur. One might entertain different sets of beliefs
in different contexts without paying attention to how they are related, and without actively believing
conjunctions of propositions believed in the different contexts. Imay not notice that in one context
Ibelieve p, and in another q. But if someone points this out to me, Iwill certainly expect to be com-
mitted to p&q. And if Ishould notice that Ibelieve p in one context and not-p in another, Iwill feel
obliged to revise one or the other of the beliefs.
7
Lewis (1983) account of truth in fiction is very different, but his objectives are also different
from mine. His definition is meant to capture the circumstances by virtue of which propositions
are fictional (what Icall the mechanics of generation), whereas mine is based on the function that
fictional truths serve. Adefinition of the kind he proposes will turn out to be drastically disjunctive,
it seems to me, given my conclusions in Walton (1990:ch. 4). Iam aiming for a more univocal one.
8
It might be true that a proposition is fictional if and only if it is to be imagined, if this means that
it is fictional in some world or other, if and only if it is to be imagined.
20 In Other Shoes

Figure1 Vermeer (van Delft), Jan (16321675). A Young Woman Standing at a Virginal,
about 16702. Oil on canvas, 51.7 x 45.2cm. Bought, 1892 (NG 1383). National Gallery,
London, Great Britain. National Gallery, London/Art Resource, NY.

varieties. Iwill present more than are needed to demonstrate the insufficiency,
in order to block some tempting but inadequate fixes, fixes that work for some
kinds of cases but not for others. Also, some of the examples are interesting in
their own right.
The most obvious counterexamples are what some call iconic
meta-representations. Vermeers A Young Woman Standing at a Virginal (fig. 1)
depicts a framed picture of Cupid on the wall behind the woman. Viewers are to
imagine a picture of Cupid. But they are also to imagine Cupid, a naked, winged
child with a bow; they are to imagine that there is such a child.9 Full appreciation
of the painting includes looking at the part of the canvas that depicts the picture,
and being induced to imagine Cupid, or anyway a child with wings and a bow,
to imagine that there is such a being. Yet it is not fictional in Woman at a Virginal
that there is a winged child with a bow.

Iam assuming that, in the cases we are interested in at least, imagining a entails or implies
9

imagining that there is a . Obviously the reverse entailment doesnt hold.

 Fic tionalit y and Imag inati on 21

Figure2 Photograph of a Doll. Photograph Kendall Walton 2014.

The point of imagining Cupid is, of course, to discover what the picture on the
wall depicts. We learn that it is fictional in Woman at a Virginal that the picture
on the wall is a picture of Cupid, when we find ourselves imagining Cupid.
The depicted frame lets us know that we are to imagine that there is a picture of
Cupid, and that it is fictional in Woman at a Virginal that there is only a picture
of Cupid there.
We can think of the small portion of the canvas inside the depicted frame as
having its own fictional world, one in which it is fictional that there is a child with
wings. That part of the canvas illustrates the content of the depicted picture, in
the world of the larger picture, but it remains true that spectators, qua viewers
of Vermeers painting as a whole, are to imagine that there is a child with wings,
although this is not fictional in Vermeers painting.
We cant always recognize a part of a work and attribute to it its own fictional
world. It is fictional in fig. 2 that there is a doll, a representation of a child with
red hair wearing a frilly pink dress. How do we know that that is what the doll
represents (and that it is a doll)? Looking at the photographic depiction, we
imagine a child with red hair wearing a frilly pink dress; we imagine that there is
one. There is a prescription, to viewers of the photograph, to imagine this; qua
22 In Other Shoes

appreciators of the picture, they are to do so. But it is not fictional in the pho-
tograph nor in any part of it considered alone, that there is a child with red hair
wearing a frilly pink dress.
Many other iconic meta-representations are counterexamples to the suffi-
ciency of my original account of fictionality. There are stories within stories
(e.g. One Thousand and One Nights), dream sequences in film, Hamlets play
about Gonzagos murder in Shakespeares Hamlet, and the film, Rashomon
(Kurasawa, 1950).
So far, the problematic examples are meta-representations, works represent-
ing other representations and their contents, but we shouldnt rush to find a
solution specific to meta-representations. Counterexamples of other kinds are
on the way.
Some fictions represent illusions and the contents of the illusions. When the
illusion is suffered by a character, the work will qualify as a meta-representation,
a representation of the characters mental representation. I mentioned dream
sequences in film, and there are ordinary point-of-view shots:Ashot of a char-
acter eating mushrooms, then wobbling around, stoned, with glazed eyes, is fol-
lowed by a shot of a purple elephant flying through the air. It is fictional that
the character hallucinates a purple elephant, that he seems to see a purple ele-
phant, but it isnt fictional in the film that there is a purple elephant. In order to
ascertain what the character seems to see, the viewer must, in the second shot,
imagine seeing a purple elephant, imagine that there is a purple elephant. Is the
viewer to imagine merely seeming to see a purple elephant, and not that there is
one? How does she figure out what it is that she is to imagine seeming to be the
case? She finds herself imagining seeing a purple elephant, and there being one.
Sometimes a work represents simply an illusory situation, without portraying
anyone suffering from the illusion. It is fictional in the photograph, fig. 3, that the
cactus looks soft and cuddly, but (by virtue of obvious background information)
fictionally it is actually prickly, not soft and cuddly. The viewer is to imagine the
cactus being soft and cuddly, although it is fictional only that it looks soft and
cuddly. It is by engaging in this imagining that they discover how, fictionally,
the cactus looks (how, fictionally, it would look were someone to see it from the
right point of view).
Since it is not fictional, in the world of the picture, that anyone experiences
the illusion, it is a stretch to call this a meta-representation. It is not fictional
even that someone has a non-veridical mental representation. But the fuzzy cac-
tus photograph does involve what we might call a secondary content. Now for
counter examples to the prescribed imagining account of fictionality that dont
even have a secondary content. The imaginings of propositions that are not fic-
tional, in these cases, do not in general help one to ascertain what is fictional;
they serve different purposes.
 Fic tionalit y and Imag inati on 23

Figure 3 Fuzzy Cactus. Photograph Kendall Walton 2014.

Appreciators sometimes experience a (real) illusion, an illusion that a work

makes it fictional that p, even when it is not fictional that there is such an illusion
(not even an illusory situation). Awork may be designed to give the impression
of, to hint at, its being fictional that p without making it so. Appreciating the
work fully may require experiencing this illusion. Appreciators may be expected
to notice that it seems to them as though p is fictional (and to recognize that it
is not), and this may involve their being induced to imagine p. It need not be
fictional in the work that it seems that p, in such cases, or that there is an illu-
sion that p, or that there is any kind of a representation that p; there may be no
mandate to imagine this. (It probably wont be fictional that there is an illusion,
insofar as the actual illusion or impression appreciators are subject to is an arti-
fact of the mode of representation; hence, likely not part of the content of the
representation.)
It seems as though the guy on the left, in fig. 4, has a golf ball for a nose, as
though this is fictional. We can hardly help imagining that he does. But it is argu-
able that this is not fictional, not true in the picture world, that our impression
is illusory. Nevertheless, to appreciate the photograph fully (whether or not the
photographer intended this) arguably requires imagining a person with a golf ball
nose. One misses something important about the picture if one does not imagine
this. The misleading impression, in this case, is largely an artifact of the medium:of
the photographs lack of stereoscopic depth cues, and the fact that it captures a
single moment of the depicted event and presents it to the viewer indefinitely.
24 In Other Shoes

Figure4 03062802043Vicente Fernandez from Argentina hits from the sand to the
second hole at Inverness Club in Toledo, Ohio, during the third round of the U.S. Senior
Open, Saturday, June 28, 2003. Fernandez bogeyed the hole. (AP Photo/Mark Duncan)
2014 The Associated Press.

Jumonjis photograph (fig. 5) is disturbing. The explanation may be that view-

ers imagine, implicitly, a headless person. But it isnt fictional in the picture that
anyone is missing a head, nor, Ithink, is it fictional that there is an illusion, or
illusory situation, that a person is headless.
A couple of more obvious examples:Background music in film can easily cre-
ate an illusion, give the impression that a certain proposition is fictional which
turns out not to be.10 Amurder mystery may trick readers into thinking that the
butler is the villain and to imagine that he is, though they discover in the end
that, say, the UPS deliveryman, not the butler, committed the crime. If the mis-
leading impression is due to conventions of the literary genre, conventions as to
who can be the villain in this kind of story, or the tendencies of this particular
author, or the fact that at a certain point readers think there arent enough pages
left in the book to make the UPS man rather than the butler the villain, it is likely

10
Thanks to Paul Bloom.
 Fic tionalit y and Imag inati on 25

Figure5 Jumonji, Bishin (b. 1947) Copyright. Untitled. 1973. Gelatin silver print,
915/16 x 9 15/16 (25.3 x 25.3cm). Gift of the photographer. The Museum of Modern
Art, NewYork, NY, U.S.A. Digital Image The Museum of Modern Art/Licensed by
SCALA/Art Resource, NY.

not to be fictional, true in the story world, that there is any sort of illusion that
the butler did it.
A final group of examples:Metaphors often (not always) elicit imaginings.
And in literary works they often prescribe the imaginings they elicit; that is typi-
cally why authors choose them. On reading, in Katherine Mansfields story, A
Dill Pickle, She felt the strange beast that had slumbered so long within her
bosom stir, stretch itself, yawn, prick up its ears, and suddenly bound to its feet,
and fix its longing hungry stare upon those far away places, one is to imagine,
perhaps, the proposition expressed by this sentence taken literally, or at least that
there is a true proposition expressed by these words, or anyway a sleeping beast
stretching, yawning, and jumping up, etc.that such an event occurs. None of
these propositions is fictional in the story. What the sentence makes fictional
is that the character referred to, Vera, has certain feelings, including a desire for
adventures in far off places with her former lover, Thomas. Iwont try to decide
whether cases like this are instances of meta-representation, or involve illusions
about what is fictional, or whether readers do, or must, engage in the prescribed
26 In Other Shoes

imaginings in order to determine what is fictional. The answers are not obvious,
and will not be the same for all literary metaphors. 11

3. Tempting Solutions
The murder mystery example will suggest a solution, one that fails to generalize
to several of the other cases. We are expected to imagine, on reading Chapter3,
that the butler did the dastardly deed, but by the end of the novel, when all is
said and done, we realize that we are to imagine not this, but that it is the UPS
delivery man who is guilty. The suggestion is that only what is to be imagined at
the end, after we have experienced and absorbed all relevant aspects of the work,
is fictional in the work.
This solution does not work for the Woman at a Virginal, for instance, or for
the Jumonji photograph. As long as we see Vermeers entire painting, including
the depiction of Cupid, we are to imagine seeing a child with wings, that there is
a child with wings. We are not supposed to stop seeing the marks as a child with
wings, or seeing a child with wings in the marks, when we notice the depicted
frame. Jumonjis photograph doesnt stop being disturbing when we figure out
that it is not fictional that the man is decapitated. The hypothesis (which Itake
to be plausible at least) is that we continue to imagine his being headless as long
as we find the picture disturbing.
Gregory Currie suggested another solution (though he didnt claim that it
works for all of the examples).12 It rests on a distinction between imaginings that
are mandated or prescribed, and what appreciators must imagine in order to
engage in the mandatory ones, i.e. imaginings which, although not themselves
prescribed, are necessary for full appreciation of the work in question. The idea
is that only the content of the former imaginings count as fictional, true in the
fictional world. Viewers of Woman at a Virginal are to imagine a picture of a child
with wings. In order to do so, they must imagine a child with wings. But, accord-
ing to this suggestion, this imagining is not itself prescribed. So it is fictional in
Woman at a Virginal that there is a picture of a child with wings, but not that
there is a child with wings.
This is not a viable general solution. It wont help with the non-meta-
representational cases, and there are serious worries concerning the meta-

11
Some of the answers will depend on whether the words of the text are, fictionally, those of
a narrator or another character or just words the author uses to tell the story, or on whether the
make-believe the metaphor involves is partly content oriented, as well as prop oriented. See my
Metaphor and Prop Oriented Make-Believe (this volume).
12
Currie, personal communication.
 Fic tionalit y and Imag inati on 27

representational ones.13 It is not clear that there is a non-question-begging way

of distinguishing between imaginings that are prescribed, and ones that are nec-
essary in order to ascertain which ones are prescribed but not themselves pre-
scribed. What grounds are there for holding that imagining a picture of a child
with wings is prescribed and imagining a child with wings is not, other than that
the content of the former is fictional and that of the latter is not? Vermeer cer-
tainly intended and expected viewers to imagine a child with wingsthe marks
he made on the portion of the canvas within the depicted frame are designed to
get viewers to imagine this. (Such intentions are not necessary for propositions
to be fictional, but they help to make some of them fictional.)
There could be a mandate to imagine a child with wings, in any case, a man-
date to imagine a picture of a child with wings (partly) by imagining a child
with wings. Suppose that there is. Suppose that whatever social conventions or
facts about the artist or viewers natural propensities (or combination thereof)
is needed to establish this prescription do in fact obtain. Surely this would not
make it fictional in Woman at a Virginal that there is a child with wings.
The primary interest, or a very substantial one, of some meta-representational
storiesstories representing the telling of other storiesis in the internal
story or stories, rather than the frame story. It would be strange indeed to deny
that readers of One Thousand and One Nights are prescribed to imagine the con-
tent of the stories Scheherazade tells. Yet it is fictional in One Thousand and One
Nights itself, the frame tale, only that Scheherazade tells these stories, not that
they are true.
As Imentioned, this proposed solution fails our non-meta-representational
examples, several of them anyway. It probably doesnt apply to the whodunit
story if the readers impression that the butler is the villain derives from genre
conventions or the number of pages remaining in the book. Imagining the but-
lers guilt is probably not necessary in order to ascertain what other propositions
are fictional. The same may well be true of the Jumonji photographimagining
a headless person is not needed to discover what else we are to imagine, so it will
be hard to deny that this imagining is prescribed.
The solution also fails to account for the metaphor examples (even if they
qualify as meta-representations). If imagining a slumbering beast coming to
life is needed to discover what is fictional about Veras state of mind, it surely
is still true that this imagining is itself prescribed. To learn about Veras state
of mind in some other way and imagine accordingly, without imagining a

13
We dont need a single general solution. Different propositions whose imagining is prescribed
may fail to be fictional for different reasons. (Thanks to Richard Woodward.) But we do need to
account for all such failures, and we want to do so without having to adopt an awkward disjunctive
or ad hoc notion of fictionality.
28 In Other Shoes

beast, would not be enough; one would not be responding appropriately to

the words of the story.

I dont know how to fill the gap; Idont know what, in addition to a prescription
to imagine, is required to make a proposition fictional in the world of a given
work. Iam not sure what sort of account of fictionality we should expect. Iand
others have mostly relied simply on intuitions about what is fictional in par-
ticular cases, usually without invoking this or any definition of fictionality. We
can continue doing this for some purposes, at least insofar as our intuitions are
shared, but of course we would like to know as well as we can what lies behind
the intuitions.
What Iliked about my original account was that it seemed to capture what
is important in our experience of fictions, viz. the imaginings we actually
engage in, and our judgments or impressions about which imaginings are pre-
scribed, and which are optional. We now see that appreciators also judge, on
some basis or other, which of the propositions whose imagining is prescribed
are fictional in the work, and which are not. This too is an important aspect
of our experience of fictions, important in part because our judgments about
which propositions are fictional affect what else we imagine, including imagin-
ings that are not themselves prescribed.14 We somehow construct a fictional
world, recognizing a subset of the to-be-imagined propositions as constitut-
ing it. We deploy a more substantial notion of fictionality than I previously
thought, one that is not in any obvious way reducible to or explainable in
terms of imaginings.
(When daydreaming, Iimagine certain things, sometimes deliberately, some-
times spontaneously, more or less at random. But then I deciderather than
discover, in the case of daydreamingwhich imaginings to accept for my day-
dream, which of their contents to count as fictional in the daydream.)

4. Seeing the Unseen; Reporting the Unreported

Whatever fictionality exactly is, the fact that the content of prescribed imagin-
ings need not be fictional will help us to deal with a couple of awkward kinds of
representations, one in the visual arts and a rough analogue in literature.
Fig. 6 is a picture of a Rhamphorynchus, a Pterosaur from the Jurassic period.
Iwill call him Ralph. Given obvious background informationthe absence of

14
Thanks to David Braddon-Mitchell.
 Fic tionalit y and Imag inati on 29

Figure6 Rhamphorhynchus, artist John Klausmeyer, University of Michigan Museum

of Natural History.

humans 150million years agoI take it to be fictional in the picture world that
no one is observing Ralph.15 Viewers of the picture imagine that this is so. But in
looking at pictures, Iclaim, viewers imagine seeing the objects or kinds of objects
that are depicted, and Iunderstand imagining seeing something to entail imagin-
ing that it is seen.16 Do we, then, observing Ralphs portrait, imagine that Ralph
is and is not seen? No, but we do imagine that he is seen and also imagine that he
is unseen. Is this a problem? No.17
A partial analogue in literature of the seeing-the-unseen problem is the
reporting-the-unreported problem. A story ends with the words, and no one
lived to tell the tale, or less explicitly, the narrative indicates that all of the char-
acters and all witnesses to the story events die off on a remote island or planet,
implying that (fictionally) no one was able to report the events. In either case,
readers are to imagine that the events go unreported. However, it is often claimed

15
Iignore the fact that other Jurassic beasts see Ralph.
16
Iwont argue for either of these claims here. We neednt suppose that in imagining seeing Ralph
Iimagine that I, Kendall Walton, see him. We have the puzzle if Iimagine merely that Ralph is seen.
Also, those who do not agree with me that to see Ralph in the picture is necessarily to imagine that
Ralph is seen will nevertheless find it hard to deny that the viewer might well imagine this.
17
Currie and Ravenscroft (2002:3031) take the latter to be about as problematic as the former is.
30 In Other Shoes

that all or most stories and novels have narrators; many do in any case.18 Readers
imagine the words of the text being uttered (or having been written) by a person,
the narrator, who thereby reports the events of the story. Do readers of stories like
those Ijust described imagine that someone reported the story events and no one
did? No, but readers do, in some instances, imagine that someone reported the
events, and also imagine that no one reported them. Is this a problem? No. But
there are more complications in this case than in the case of seeing the unseen.19
Currie and Ian Ravenscroft (2002:2.2) deny that to view a picture is to imag-
ine that what it depicts is seen. Rather than imagining seeing Ralph, with seeing
in the content of the imagining, they will say, we imagine Ralph and the rest of
the scene in a visual manner. This nicely sidesteps the seeing-the-unseen prob-
lem, if it is right, but no analogous resolution of the reporting-the-unreported
story is available. (Do we imagine Huck Finns adventures in a reported on man-
ner, rather than imagining receiving the narrators, Hucks, reports about them?)
We might expect that the seeing-the-unseen case is to be resolved in whatever
way works for the reporting-the-unreported case, making postulation of a visual
mode of imagination unnecessary (for this purpose at least). As a matter of fact,
seeing-the-unseen will turn out to be more easily treated (without appealing to a
visual mode of imagination) than reporting-the-unreported is.
It is probably obvious how Iwill defuse these puzzles, the first one at least.
In our previous examples, the counterexamples to the prescribed imagining
account of fictionality, appreciators do and are expected to imagine propositions
inconsistent with one another. We imagine that there is a child with wings, and
we imagine that there is only a picture of a child with wings. We imagine both:

... that there is a purple elephant, and also that there is no such thing.
... that a person has a spherical white nose, and also that he doesnt.
... that the cactus is soft and fuzzy, and also that it is prickly, not soft
and fuzzy.
... that someones head is missing, and also that it isnt.
... that the butler did it, and also that he didnt.

There is nothing at all paradoxical or problematic about these pairs of imagin-

ings, nothing strained or strange about imagining each of the two conflicting
propositions. There is no pressure at all to imagine their contradictory conjunc-
tions:that there is and is not a purple elephant, for example, or that a person
does and does not have a spherical white nose. Why? Because the different

What Icall reporting narrators, in contrast to storytelling narrators. See Walton 1990:9.6.
18

Lewis (1983:274275) understands and no one lived to tell the tale cases as instances of
19

impossible fictions. This makes them more problematic than many of them need to be.
 Fic tionalit y and Imag inati on 31

imaginings with conflicting contents are not associated with the same fictional
world; they dont belong to the same cluster. It is fictional in the world of the movie
that there are no purple elephants. Viewers imagine this. They also imagine that
there is a purple elephant, but it is not fictional in that world that there is a purple
elephant. Viewers manage somehow to exclude this proposition from the world
of the movie.

Observing the Rhamphorynchus picture, one imagines Ralph unseen. One also
imagines seeing Ralph. These imaginings do not belong to the same cluster. The
content only of the former belongs to the world of the picture, even though both
imaginings are prescribed for appreciators of it. So the two imaginings live hap-
pily together in the viewers experience. This is like the meta-representational
cases:One imagines seeing Ralph, thereby ascertaining what it is that one is to
imagine occurring unseen.20
Compare:Iread a story which offers a vivid visual description of an exotic
animal, then says that no one has ever seen it. I imagine seeing the beast,
I visualize it. But I also imagine that it isnt seen. In my visualizing I learn,
come to understand, what it looks like. Then Iimagine that a beast that looks
like that has never been seen. This is not the least bit problematic. Neither is
the Ralph case.
Untold tales are a little more complicated, and they come in several varieties
which need to be distinguished. We do experience tension, in some cases more
than others.
Consider a straightforward instance of a novel with an explicit narrator,
Mark Twains The Adventures of Huckleberry Finn. Huck Finn is the narrator. It
is fictional in the novel world that Huck and his friends, Tom Sawyer and Jim,
had various adventures, and that he reports them by means of the words of the
text.21 We would have a jolting paradox if at the end of the novel Huck declared,
no one lived to tell the tale, or if Twain had made it obvious for one reason or
another that Huck couldnt have reported his adventures. (Suppose the novel
ended with, The posse chasing Jim shot me dead and dumped my body in the
river. It was never found and no one ever knew.)22 It would be hard to deny, in
20
My treatment in Walton (1990:4.5, 6.6) of the Seeing the Unseen problem was based on
a notion of silly questions. What Isay here can be construed as an explanation of why the relevant
questions are silly. Stories about untold events seem especially amenable to silly questions treatment.
21
The famous opening sentence of The Adventures of Huckleberry Finn:You dont know about
me, without you have read a book by the name of The Adventures of Tom Sawyer; but that aint no
matter.
22
The novel could have taken a fantastic turn, however, and had Huck reporting from the dead
through a medium (Cp. Rashomon).
32 In Other Shoes

that case, that it is fictional in the novel world that Huck recounted his adven-
tures and also that he couldnt have done so and didnt. There would be pressure
to imagine that he did and did not recount them.
But no one lived to tell stories can be much less jolting, especially when the
narrator is (as some say) not a character. What does this mean? Anarrator is a
fictional person. Doesnt that make him or her a character? Not all narrators have
names. Many dont refer to themselves in the first person (or at all), and many do
not participate in the actions that they report; they just report them. But none of
this disqualifies them as characters.
Lets not worry about whether narrators count as characters; consider
instead what fictional worlds they belong to. In the case of stories with name-
less, omniscient, narrators, who do not refer to themselves and do not par-
ticipate in the action, Ithink it is often reasonable to recognize a world, call
it the primary story world, containing the events of the story but not the
narrator (perhaps this is what is meant when a narrator is said not to be a
character). It is fictional in this world that the events occur but not that the
narrator reports them. Readers do imagine the narrators reporting them (and
probably expressing attitudes about them in doing so), but this imagining does
not belong to the cluster associated with the primary story world. We can rec-
ognize a secondary story world, in which the narrator does report the events
of the story.23
This imagining does, however, help readers determine what is fictional in the
primary story world. The kinds of events the reader imagines the narrator report-
ing are the ones that (fictionally) occur unreported, in the primary story world.
In the special case of a no one lived to tell story, it is fictional in the primary
story world that the events are unreported, not reported by anyone, and readers
imagine that this is so. There is no tension between this imagining as part of the
primary story world cluster, and readers imagining the narrators reports, since
the latter imagining does not belong to this cluster. (The readers imaginings, in
the primary story world cluster, wont include any imaginings about the narrator,
neither that she reported the events, nor that she did not.)
So far, this is much like the Rhamphorhynchus case (except that the reporting
of the story events occurs in a work-world, whereas the seeing of Ralph occurs
only in what Ihave called the spectators game-world). There is an interesting
difference, however, an added complexity in the story case. The narrator reports
that the events were unreported. We imagine this, though not as part of the pri-
mary story world cluster. Viewers of the Rhamphorhynchus picture, although
they imagine that Ralph is seen, do not imagine his being seen to be unseen; they

In Walton (1990:285287, 9.6) Iproposed recognizing two distinct work worlds for a single
23

work, in somewhat similar instances.

 Fic tionalit y and Imag inati on 33

dont imagine seeing that he is not seen. When the narrator is omniscient (and
probably effaced) it will be fictional, in the secondary work world, that what they
report is true, hence fictional that they report unreported events. So readers are
expected to imagine that these events are and are not reported, this imagining
belonging to the secondary work world. There remains the primary work world
cluster, however, in which readers imagine only that the events are unreported.
And readers are likely to be more interested in this world, than in the secondary
world in which it is fictional that the events are reported.
When narrators are explicit, as in Huckleberry Finn, and also One Thousand
and One Nights, we will be less inclined to recognize a primary work world which
does not include the narrators reporting, and more likely to be interested in the
work world that contains the narrator. That there is no sharp line between these
two kinds of cases should be of no concern.

5. Game Worlds
The discussion above concerning the meta-representational examples sug-
gests that we should think about game worlds differently from how Idid in pre-
vious writings. Agame world is the fictional world of a game of make-believe
that an appreciator engages in using a work of fiction as a prop. It includes
what is fictional in the world of the work, plus fictional truths generated by
what the appreciator does and thinks and feels as she observes or reads it
(Walton 1978:5; 1990:5861). It is fictional in my game world, as Iwatch
a performance of Romeo and Juliet, that Romeo and Juliet meet, fall in love,
and eventually come to grief. It is fictional also, in the game world but not
the work world, that Ilearn about or observe these events, and think and feel
about them in various ways, etc.
In Walton (1978 and 1990)Imore or less assumed that a person appre-
ciating a particular work of fiction on a particular occasion would have just
one game world, a mostly coherent, extended fictional world paralleling the
world of the work. Inow think that it is better to think of game worlds as frag-
mented. The appreciators experience involves various shifting clusters of imag-
inings which sometimes combine and sometimes dont, as she works out what
is fictional in the work worldwhat, fictionally, characters believe, what they
dream, the content of their illusions, what (fictional) pictures depict and what
stories are about, etc. In the spectators experience of the point of view shot, there
will be at least two game worlds, one in which it is fictional that she sees a purple
elephant, and another in which it is fictional that she watches the character as he
hallucinates a purple elephant. It may also be fictional, in the second game world,
that the viewer has certain attitudes or feelings about the character because of the
34 In Other Shoes

content of his hallucination. It may be fictional in the first game world that she is
frightened or startled by the purple elephant.
I mentioned that Nichols and Stichs boxological diagram contains only a
single box for imaginings, apparently ignoring the clustering that Ihave empha-
sized. Ishould point out, in their defense, that they describe this box as a tem-
porary workspace (2003: 2829). Presumably it contains different clusters
of imaginings at different times, which we use to work out what is fictional in
one or another fictional world. Imaginings tend to be temporary, short-lived,
more so than beliefs are. We have a very large stock of relatively permanent,
mostly non-occurrent beliefs which, when occasion arises, get combined with
others and/or desires in drawing inferences and motivating actions. Some have
claimed that imaginings, unlike beliefs, are necessarily occurrent, never disposi-
tional (Currie 1995:160). Iargued, in Walton (1990:1618), that imaginings
are sometimes non-occurrent. In any case imaginings dont seem to hang around
more or less indefinitely, in dispositional form, as beliefs tend to do. So Nichols
and Stichs possible worlds box can accommodate, alternately, different clusters
of imaginings; it is a workspace in which, at different times, we work out what is
fictional in different work worlds.
We probably do need more than one workspace or imagination box, however,
or some way of accommodating different clusters of more or less simultaneous
imaginings, and not only because we may have in mind more than one imagi-
native project at pretty much the same time. Observing Woman at a Virginal,
Iimagine not only that the picture on the wall depicts Cupid; Iimagine that it
depicts a scene of a much more specific kind:Cupid with his hand raised in a
certain manner, a certain expression on his face, a certain pattern of clouds in the
background, etc. Idont and probably cannot describe, in words, the scene in all
its detail. Irepresent it to myself as what the part of the canvas inside the depicted
frame induces me to imagine seeing. Ijudge it to be fictional, and Iimagine, that
the picture on the wall depicts a scene like this, the demonstrative referring to
the content of my imagined seeing.24
In imagining this propositionthat the picture on the wall depicts a scene like
thisI also imagine the scene; Iimagine seeing it, seeing Cupid raise his hand
in such and such manner, etc. These two imaginings occur simultaneously or
nearly so,25 but they belong to different clusters; Idont imagine the conjunc-
tion of their contents, that both:the picture depicts Cupid raising his hand in
such-and-such manner, etc. and Isee him actually doing so.

See my discussion of point of view shots in Empathy, Imagination, and Phenomenal Concepts,
24

this volume, p. 22.

25
We cant rule out, a priori, the possibility that they occur alternately, imperceptibly, in quick
succession.
 Fic tionalit y and Imag inati on 35

References
Currie, Gregory (1995). Imagination and Simulation: Aesthetics Meets Cognitive Science.
In Martin Davies and T. Stone (eds.), Mental Simulation: Evaluations and Applications.
Oxford:Blackwells.
Currie, Gregory, and Ravenscroft, Ian (2002). Recreative Minds: Imagination in Philosophy and
Psychology. Oxford:Oxford University Press.
Lewis, David (1983). Truth in Fiction. In David Lewis (ed.), Philosophical Papers, Vol. 1, 261
280. NewYork:Oxford University Press.
Nichols, Shaun, and Stich, Stephen (2003). Mindreading. Oxford:Oxford University Press.
Skolnick, Deena, and Bloom, Paul (2006). The Intuitive Cosmology of Fictional Worlds. In
Shaun Nichols (ed.), The Architecture of the Imagination, 7386. Oxford:Oxford University
Press.
Walton, Kendall L. (1978). Fearing Fictions. The Journal of Philosophy 75(1):527. Reprinted
in this volume.
Walton, Kendall L. (1973). Pictures and Make-Believe. The Philosophical Review 82(3):283319.
Walton, Kendall L. (1990). Mimesis as Make-Believe:On the Foundations of the Representational
Arts. Cambridge, MA:Harvard University Press.
 3

Two Kinds of Physicality

inElectronicand Traditional Music
Suddenly a violent noise leaped at them from no source that he could
identify. He gasped in terror at what sounded like a man trying to gar-
gle while fighting off a pack of wolves.1

I will examine two very different ways in which listeners experiences of musical
works sometimes involve physical actions or eventsin many instances emo-
tions and other mental states as well. Both concern music that is often described
as expressive. But they are so different that Iquestion whether it is reasonable
to subsume them under any single category, whether, for instance, they are aptly
characterized as two kinds of expression or expressiveness. The first consists in
the impressions appreciators have of physical activities or events that produced
the sounds they hear. The second is the tendency of listeners to engage in physi-
cal activity themselves in response to music.

I
How a work of art, or a performance, appears to have come about often has a
lot to do with its interest, its aesthetic character, and the experiences of appre-
ciators.2 This includes psychological facts about the artisther apparent objec-
tives in creating a work, and facts about her attitudes, personality, or mood that
it seems to reflect. It also includes (in many cases) apparent physical causes; a
work or performance may seem to have resulted from actions or events of one
or another sort.

1
Douglas Adams, The Hitchhikers Guide to the Galaxy (NewYork:Simon and Schuster, 1979), 56.
2
I have more to say about apparent circumstances of production in the arts in Style and the
Products and Processes of Art, reprinted in Kendall Walton, Marvelous Images:On Values and the Arts
(NewYork:Oxford University Press, 2008).

36
 P hys icalit y in Mu si c 37

A few quick examples:whether a story seems to have been meant to be funny,

or not, is often important. Astory told with a straight face may be funnier than
it would be otherwise. Or its humor might be so subtle that audiences miss it.
Astory with a serious moral might undercut its purpose if the storyteller gives
the impression of trying to be humorous. Alternatively, the apparent jocular-
ity might make palatable to readers an unpleasant, serious message that the
story aims to get across. That a picture was painted in a childish style, as though
by a child, may give it a kind of charm that it wouldnt have otherwise ( Jean
Dubuffet).
In musical performances (and in ordinary life as well), there are crash-
ing, clanging, scraping, caressing, and wailing soundssounds that sound
as though they were made by crashing, clanging, scraping, caressing, wailing
actions or events. Violent or gentle sounds seem to be sounds of violent or gen-
tle happenings.
The difference between the impression that sounds we hear were produced
by a persons deliberate actions, actions of banging or bowing or blowing or vocal-
izing, and the impression just of sounds caused by physical events of certain
kinds, objects scraping or banging against one another, whistling wind, and so
on, this difference is of first importance. Insofar as our impression is of the for-
mer sort, we might be said to be recognizing something like a fictional character
in the music. It may be reasonable to say that the music is (minimally at least)
representational, that the sounds of a performance represent or depict them-
selves as having been made by a person of a certain kind.
Appearances need not correspond to reality, of course; the impression a work
gives of its genesis may not be veridical. But appreciators are often interested
in the appearance itself, regardless. (I do not rule out their being interested,
also, in how a work actually did come about, nor do Iassume that this is not an
aesthetic interest. But interest in its apparent sources is not itself interest in its
actual sources.) The score of William Krafts percussion piece Momentum con-
tains the notation run amok. Idoubt that this instruction is meant to be taken
literally. More likely the performers are expected to produce sounds that merely
seem to be made by people running amok. Perhaps they are expected also, in
a live performance, to behave so as to look as though they are running amok.
Interest in appearances regardless of their veridicality is an interest of a special
and somewhat unusual kind. In many or most ordinary (non-artistic) contexts,
appearances are important primarily for what they might indicate about reality.
Along tradition in philosophy, going back to Platos Cave, wrestles with ques-
tions about the legitimacy of inferring how things are from how they appearas
though that is why appearances are important. Aesthetic contexts are different.
Appreciators notice and enjoy appearances more or less for their own sake, without
necessarily even wondering whether or not things actually are as they appear.
38 In Other Shoes

This fact has two important consequences. First, we recognize, attend to, care
about conflicting appearances in works of art without feeling any particular need to
resolve the conflict. Consider the joke told with a straight face. The impression of
seriousness and that of joking are both important; to fully appreciate the story, we
must be sensitive to both. It may be obvious, in the end, that the storyteller was not
serious, that she meant only to be joking. But we miss something important if we
ignore or fail to detect the appearance of seriousness. What appreciators are likely
not to do is to weigh the two contrary appearances and replace them with a single
all-things-considered appearance:All in all, she seems to have been joking rather
than being serious.
Given that multiple conflicting appearances are to be noticed and often relished,
things can be fascinatingly complicated. Ajoke told with a straight face might also
be intended to make a serious point. Ones first, most immediate impression may
be that the speaker did not mean to be funny (the straight face). The air of serious-
ness may be exaggerated, however, making it seem that the speaker did intend to
be funny after all. Finally, it may be apparent on reflection that the speaker meant
to be making a serious point, that she intends the story to have a serious moral. The
speaker tells an apparently serious story in a way that makes it seem to have been
meant to be funny, evidently intending thereby to make a serious point. Jonathan
Swifts A Modest Proposal may be an example of exactly this.
The second consequence of the fact that appreciators attend to appearances
for their own sakes is that the appearances that matter include ones that, on pain
of circularity, could not constitute a reason for thinking that things actually are as
they appear to be. Lying in a tent in the wilderness, someone tells me that there
is a grizzly bear outside. That is a bear! she says, referring to what until then
sounded to me like a squirrel scrambling up a tree or branches blowing in the
wind. These sounds now have a sinister, scary quality. They seem to me to have
been made by a bear; my impression is of a bear prowling around looking for food,
bumping into vegetation, and so on. Insofar as my prior, independently acquired
belief that there is a bear in the camp is responsible for the (auditory) appearance
of a bear, that appearance gives me no reason to think that the sounds are those
of a bear. (Imay or may not realize that this is so.) Nevertheless, the impression
that they are bear sounds is part of my auditory experience. If, perchance, Iam not
especially interested in what actually is the case, whether the sounds Ihear were
or were not actually made by a bear, Imay be interested in, Imay notice and enjoy
the bearish quality of the sounds more or less for its own sake.3
The point Ijust made is a special case of a more obvious one. What we know
or believe or have internalized about how certain kinds of things are actually

See my discussion of a similar visual example in Walton, Style and the Products and Processes
3

of Art, passim, pp. 2456.

 P hys icalit y in Mu si c 39

produced has a lot to do with how a particular one seems to have been produced.
Aloud sound seems to be the result of more violent happenings than a softer one
does because we realize (implicitly) that violent actions or events actually do,
ordinarily, produce louder sounds than gentler ones do. It is largely because we
have observed the antics of pianists in making piano sounds, not to mention our
own attempts to play the piano, that we have the impressions we do, when hear-
ing a recording of a piano sonata, of the performers antics in making the sounds
we hear. My present point is that how a particular sound sequence seems to have
come about may depend on ones prior knowledge of how those very sounds did
come about, as well as on prior knowledge of how other similar sounds normally
doalthough the appearance counts as evidence for the reality in one case but
not the other.
Michel Chion distinguishes between causal listeninglistening for the pur-
pose of gaining information about the sounds source, and reduced listening
listening for the purpose of focusing on the qualities of the sound itself (pitch,
timbre, etc.) independent of its source or meaning.4 Listening to sounds notic-
ing or attending only to how they appear to have been produced is neither of
these. It isnt paying attention to the actual cause of the sounds, but neither is it
focussing just on sonic properties.

It is obvious that the apparent genesis of much electronic music is very different
from that of traditional acoustic music. One might suppose, however, that there
isnt much difference in the case of electronic music that mimics traditional
acoustic instruments. Dont electronically generated sounds that successfully
reproduce sounds like ones made by musicians playing violins or accordions or
whatever sound as though they were made in those ways? This ignores the influ-
ence awareness of how sounds are actually made has on how they seem to have
been made. Listeners experiences may include an impression that the sounds
were made by violins or accordions, in any case, but this impression will mix
with other contrary ones, for listeners who have some awareness of their actual
genesis (even if little more than a realization that violins or accordions were
not involved). Isuspect that interactions between the different impressions are
important. (I am not considering, now, recordings of actual violin or accordion
music, or electronic music that makes recognizable use of such recordings.)

4
Michel Chion, Audio-Vision:Sound on Screen, trans. Claudian Gorbman (NewYork:Columbia
University Press, 1994), 2223. Reduced listening is approximately what some call acousmatic lis-
tening. Chion recognizes a third mode of listening as well, semanticlistening for the purpose of
gaining information about what is communicated in the sound (usually language) (p.224).
40 In Other Shoes

What about entirely unrecognizable sounds generated electronically? The

ethereal, disembodied, mysterious character often attributed to such sounds is
no doubt largely a matter of the impressions listeners have or do not have of their
sources. At the premier of his Hymnen, Stockhausen remarkedthat:

Many listeners have projected that strange new music which they expe-
riencedespecially in the realm of electronic musicinto extrater-
restrial space. . . . Several have commented that my electronic music
sounds like on a different star, or like in outer space. Many have said
that when hearing this music, they have sensations as if flying at an infi-
nitely high speed, and then again, as if immobile in an immense space.5

It may be unclear whether listenersI am thinking now of nave listeners, with only
the foggiest idea of actual electronic music productionsimply lack any sense of
how the sounds were generated, or whether they do have an impression of their
origin, an impression of their having come from nowhere, out of the blue, of their
lacking any physical cause (or perhaps any cause at all). There may be no fact of the
matter about this. In either case, Ithink it would be a mistake, misleading anyway,
to deny that the music is expressive. The ethereal, disembodied, mysterious qualities
are aesthetically important whether or not they count as expressive properties. The
music may inspire awe if not empathy. It might be especially appropriate in religious
contexts (compare church organs). There may or may not be such a thing as totally
disembodied music. But music may be strikingly lacking in one particular kind of
embodiment, and this absence may be immensely important aesthetically.
Some justification for denying that the music is expressive, insofar as expressive-
ness depends on the impression it gives of the generation of its sounds, might come
from the fact that the apparent genesis of the sounds doesnt depend much, if at all, on
their specific sonic qualitiesbeyond their merely lacking sonic qualities indicative of
familiar sound sources. Changes of timbre, pitch, volume, attack and decay properties,
and so on make little or no difference (so it seems to me, anyway) in this respect:the
sounds do not seem to have any particular physical sources or seem not to have
physical sources at allno matter. Idoubt that listeners, nave ones anyway, follow
the musics expressive, emotional development as they often seem to do (sometimes
with empathy) in the case of acoustic musicinsofar as the emotional development
consists in impressions the music gives of the expressive behavior of a creator of the
sounds. They may of course follow the progression of sounds intently for other rea-
sons, and there may be expressive or emotional development of other kinds.

Thomas B.Holmes, Electronic and Experimental Music:Pioneers in Technology and Composition

5

(NewYork:Routledge, 2002), 145. Hymnen does contain recognizable recordings of a human voice,
as well as many unfamiliar sounds.
 P hys icalit y in Mu si c 41

Again, Iwould expect that the situation is somewhat different for sophisti-
cated listeners with more or less detailed knowledge, more or less internalized,
either of how electronic sounds of the kind one hears are ordinarily created, or
how the very sounds in question actually were. But, still, the sonic properties of
the various sounds will not make much difference to what physical actions seem
to have produced themmaybe a difference in which keys are pushed, or in
what order (on an organ, which stops are pulled out).
As always, there are complications. Dont louder sounds, even unfamiliar elec-
tronically generated ones, seem to result from more violent events than softer
ones do, and quick successions of different sounds from faster moving successions
of events? Probably. But in my experience, at least, these impressions are vague,
muted, at best, when Ihave no sense of what kinds of events produce the louder
or softer sounds, or the sounds that succeed one another more or less quickly.
Even nave listeners are likely to have an impression of psychological states of
the music makerwhether she was trying to be funny, for instance, or wanted
to fool listeners about where the music would be going nextor whether the
music is improvised or composed. No doubt sophisticated listeners will have
more detailed impressions of this kind than nave ones do.

II
I feel intimate with the music, more intimate than I feel with the
world of a painting. The world of a painting ... is out there, something
Iobserve from an external perspective. But it is as though Iam inside
the music, or it is inside me.
Kendall Walton, 19946

Waltons impressionistic observation is hardly a paradigm of philosophical rigor!

It may or may not seem, initially, to be somehow apt, as a characterization of
ones experience of some kinds of music. Iwill suggest that there is more to it
than one might think. Indeed, it may be more nearly literally true than even
Walton thought possible.
There is nothing especially special about music (in general), as far as apparent
circumstances of production are concerned. Apparent painters and sculptors are
no less important than apparent music makers. The same goes for our impres-
sions of physical activities or events that give rise to musical performances, and
to visual works of art. We certainly have a sense of the physical movements that
resulted in the paintings of Wassily Kandinsky and Jackson Pollock, and an

6
Kendall Walton, Listening with Imagination: Is Music Representational? The Journal of
Aesthetics and Art Criticism 52 (1994), 4761, 54. Reprinted in this volume.
42 In Other Shoes

impression of Van Goghs brushstrokes, as well as those of typical Japanese ink

painters. Photographs, like much electronic music, are likely not to give viewers
a very vivid impression of the activity of the photographerperhaps little more
than the clicking of a shutter. This impression will be the same in many instances,
no matter what the photograph depicts. (Sometimes, however, it will seem as
though the photographer was hanging from the ceiling, or staring into a rhinoc-
eross mouth, when she clicked the shutter.)
Music clearly is special in some respects, howevera diverse conglomeration
of aestheticians and musical scholars have thought so, in any case. Imentioned
that apparent music makers are, in some ways, like characters in (visual and liter-
ary) works of fiction. The apparent creator of a musical work or performance,
like characters in a movie, for example, may be in a certain emotional state, and
may engage in certain physical activities. Watching the movie or listening to the
music, we have a third-person relation to the character or apparent music maker
and her emotions and activities.
The contrasts between music and the visual and literary arts (the contrasts
Ihave in mind) have more to do with appreciators own affective states and phys-
ical responses than with emotions and activities of characters or apparent artists.
Music seems to exert an especially powerful and direct influence on listeners
emotions or moods, and to have a unique tendency to evoke certain kinds of
physical responses in listeners. We can begin to clarify and explain this differ-
ence by thinking about differences between visual and auditory perception,
between sounds and sights themselves, apart from particular uses in works of
art. Iwont be concerned with the physics or the physiology and neurology of
visual and auditory perception, rather with our experiences when we see and
hear. For that is what matters in our appreciation of works of art. (Our intuitive,
internalized conceptions, or misconceptions, of the relevant physics and physi-
ology color our experiences, however.) Nor am Imuch concerned now with see-
ing and hearing as sources of information.
I will focus on two related differences between our experiences of sights and
sounds:reification and physicality.

Reification
Italo Calvino, in A King Listens, introduces the first of these, reification:

The music comes and goes, in gusts, it oscillates, down in the rumbling
groove of the streets, or it rises high with the wind that spins the vanes
of the chimneys.
And when in the darkness a womans voice is released in singing, ...
What is it? ... That voice comes certainly from a person ... ; a voice,
 P hys icalit y in Mu si c 43

however, is not a person, it is something suspended in the air, detached

from the solidity of things.7

Calvino brings out our tendency to think of sounds, to experience them, as things
(or stuff) that travel to us, to our ears, from objects or events that emit them, from
bells and trains and gurgling brooks and people speaking or singing. Sounds have
lives of their own. They are distinct from and independent of their sources, of the
things or events that emit, cause them. They have their own spatial locations, as
they move away from their sources. They have their own temporal dimensions
aswell:a sound arrives at our ears after the event that gave rise to it; in some cases
perceptibly after the event. Sounds fill rooms, and travel across streets. Sights
dont do that.
But smells do. Asmell might fill a room, or be carried by the wind away from
the skunk or the incinerator that emitted it. Asmell, like a sound, may be in a
different place, and also a different time, from that of its source.
What about sights? What we call sights are things like the Grand Canyon, the
Eiffel Tower, a sunset.8 We think of ourselves as seeing objects out there, ones that
stay out there while we examine them visually. We do, sometimes, speak of seeing
glimmers, flashes, reflections. These are distinct from their sources. But they are
on or next to their sources, even if they are not parts of them; they are out there, at
a distance from us. And usually it is the object itself that we think of ourselves as
seeing. (We know that light travels from the objects we see to our eyes. It enables
us to see objects. But we dont usually think of ourselves as seeing light.)
It isnt clear that there is a mode of visual perception analogous to reduced
listening, in Chions sense (what some call acousmatic hearing).9 If we subtract
from our usual visual experiences the physical objects that we see, what would
be left to count as objects of our visual experiences?

Physicality
If sounds come to us, what do they do when they get there? Here is a thought:they
enter our bodies and animate usthat is, we think of or experience sounds,
some sounds, as doing this.10

7
Italo Calvino, A King Listens, in Under the Jaguar Sun, trans. William Weaver (San
Diego:Harcourt Brace Jovanovich, 1988), 50, 53.
8
Stanley Cavell, The World Viewed:Reflections on the Ontology of Film (NewYork:The Viking
Press, 1970), 20.
9
Thanks to Alicyn Warren.
10
Seeing is like touching, hearing like being touched. But the touch of sound does not stop at
the skin. It seems to reach inside of us and to attenuate along with the distinction between here and
there the still more basic distinction between inner and outer (David Burrows, On Hearing Things,
44 In Other Shoes

That they animate us, that some musical sounds stimulate or encourage physi-
cal behavior is obvious (though until recently musicologists have not paid much
attention).11 One of the main objections some people had to jazz in its early days
was based on its tendency to make people move. After impudence comes the
determination to surprise:you shall not be gradually moved to the depths [by
jazz], you shall be given such a start as makes you jigger all over.12 The National
Dancing Masters Association adopted the rule:Dont permit vulgar cheap jazz
music to be played. Such music almost forces dancers to use jerky half-steps and
invites immoral variations.13
It is not just jazz (or rock n roll, etc.) that makes people move. We march
and dance to music of more sedate sorts; we tap our feet and sway with it. We
are supposed to control ourselves in concert halls, but the inclination to move
is there; our muscles contract. The pianist and composer Oscar Levant tried to
explain his way out of a speeding ticket by pointing out that You cant possibly
hear the last movement of Beethovens Seventh [Symphony] and go slow.14
Levant was speeding to a steady, insistent rhythm, a hurried one. You will
remember the irregular, unpredictable crashes in Stravinskys Rite of Spring. This
passage makes me feel off balancelike walking downstairs without knowing how
many steps there are.
That sounds tend to produce immediate physical reactions is an obvious com-
monplace; Ididnt need to belabor the point as much as Ijust did. But it is remark-
able. We can begin to appreciate how remarkable it is by noting that nothing very
similar occurs in the visual realm. Visual objects have rhythms alsoat least we
describe them thus. Figure 1 shows a regular visual rhythm.
There is some point in this comparison. We will be reminded of the idea
that architecture is frozen music. But the comparison is limited:architecture
doesnt cause foot tapping, not even frozen foot tapping. We dont march or
dance or tap our feet to Paul Klees Camel in Rhythmic Landscape (Figure 2),

Musical Quarterly 66/2 [1980], 18091; 1834). Musical sound has direct access to the soul. It
finds there an echo, for man hath music in himself (Wassily Kandinsky, Complete Writings on Art
[NewYork:Da Capo Press, 1994], 161).
11
We dont just experience music as animating us; it really does. Daniel Levitin has emphasized
this. (This Is Your Brain on Music:The Science of a Human Obsession [NewYork:Penguin, 2007]).
There is plenty of empirical confirmation, much of it usefully summarized by Jenefer Robinson,
Deeper Than Reason:Emotion and its Role in Literature, Music, and Art (Oxford:Oxford University
Press, 2005), 3958. Idiscussed some implications of musics tendency to induce physical behavior
in Walton, Listening with Imagination, passim, pp. 167169.
12
Clive Bell, Plus De jazz, New Republic 21 (September 1921), 93. Thanks to Mark Katz.
13
Quoted in Anne Shaw Faulkner, Does Jazz Put the Sin in Syncopation?, Ladies Home Journal
38 (August 1921), 16.
14
Kathleen Kimball, Robin Peterson, and Kathleen Johnson, The Music Lovers Quotation
Book:ALyrical Companion (Toronto:Sound and Vision, 1990), 44.
 P hys icalit y in Mu si c 45

Figure1 Agra Fort. Photograph Kendall Walton 2014.

although we see the analogy between it and the Beethoven; we see that there
is one.
Figure 3 has an irregular rhythm. Again, we see something of an analogy with the
Rite of Spring passage. But we dont feel the rhythm in the same physical way.
These are still pictures. Perhaps moving ones will approximate the physical
effects of musical motion? Take a rhythmically regular screen saver. We can appreci-
ate the analogy with regular musical rhythms. But we dont march or tap our feet.
A quick experiment: compare two video clips (thanks to Alicyn Warren),
available here:www.oup.com/us/inothershoes.15
Clip #1 is a regular rhythmic alternation between a blue-black movie and a
red-white one, without sound. This doesnt make me want to march or dance or
tap my feet.16 Clip #2 adds sounda clicking metronome, in 2/4 time. Iat least
find this a little more encouraging of movement.
Why the difference? And can we give any sense to the idea that it is by enter-
ing our bodies that sounds make us move, animate us? This is an empirical ques-
tion calling for empirical research. But Iwill speculate. Aclue may come in the

Both clips Alicyn Warren (2008).

15

I found it quite hard to pick up a beator rather, to stay with the beatjust from watching
16

(Alicyn Warren, personal communication).

Figure2 Paul Klee, Camel in Rhythmic Landscape (1920). Photograph Walter Klein.
Reprinted with permission from Kunstsammlung Nordrhein-Westfalen.

Figure3 Paul Klee, Rhythmical, More Rigorous and Freer (1930). Photograph VBK,
Wien, 2010. Reprinted with permission from Stdtische Galerie im Lenbachhaus und
Kunstbau, Mnchen.
 P hys icalit y in Mu si c 47

observation that low frequencies have a greater tendency to get us moving than
higher ones do. We are more likely to move and groove with the thumping of a
bass than the twittering, even rhythmic twittering, of a piccolo. In the case of
low-pitched tones especially, we feel the vibrations. We experience sounds somat-
ically, as well as through our ears, especially if we are in contact with a vibrating
solid object linking us to the sound production, such as a dance floor. You will
remember Evelyn Glennie, the percussionist who lost most of her hearing by the
time she was a teenager. She is deaf, but not insensitive to sound. Hearing is a
form of touch, she said, you feel it through your body, and sometimes it almost
hits your face.17
One doesnt have to be deaf to feel sounds, to experience them somatically.
Iexpect that most of us fail to realize how much of our detection of sounds
even not very low-pitched onesis somatic rather than aural. Experiments have
shown that much of what we call taste is actually smell. (People have trouble dis-
tinguishing potatoes and apples by taste, when their noses are pinched.) Idont
know of similar experiments in the case of sound. But Iwouldnt be surprised
if a significant part of what we think of as hearing is actually feeling, experienc-
ing sounds somatically. We cant help noticing our somatic experiences of very
low-frequency sounds. We may be only vaguely aware of feeling the vibrations
of mid-range sounds. Perhaps very high-pitched sounds are felt almost not at all.
These somatic experiences, however vague our awareness of them, help to
explain the impression that sounds not only come to us, but are experienced as
being inside us, in a way that visual objects arent. Perhaps we identify sounds
with physical sensations, on a par with itches, pains, and adrenalin rushes, sensa-
tions of muscles tensing or relaxing, and so on, all of which are inside us and
often give rise to physical behavior. Or maybe we (implicitly) identify sounds
with the vibrations that we feel, the sensations, the feelings in our muscles, these
being sensations of movement or incipient movement, rather than causes of
movement. Probably our experiences are simply not definite enough to admit
of precise characterization. But that is not needed, Ithink, to see that listeners
somatic experiences encourage the idea that sounds enter their bodies and ani-
mate them.
My speculative hypothesisto be slightly more explicitis that somatic
experiences of sound, feelings of vibrations in our bodies, cause tendencies to
tap our feet, march, dance, and so forth. Iwont speculate about the mechanisms
involved beyond proposing that the causation is relatively direct and mechanical
(physiological, neurological), and does not involve anything like (even implicit)
cognition, or imaginings or deliberate action.

Quoted in Stephen Holden, How Sound Feels to a Musician Who Lost Her Hearing, NewYork
17

Times, September 7, 2005.

48 In Other Shoes

This, Isaid, is speculation. But it is not entirely without empirical support. If

we are willing to grant, on the basis of our own experience, that low-frequency
sounds have a greater tendency to encourage foot tapping or marching than
higher-frequency ones do, and also that the vibrations of low-frequency sounds
are felt more strongly than those of high-frequency ones, we have a correla-
tion between somatic experiences of sounds, relatively intense ones anyway,
and physical responses. This correlation suggests a causal relation. The hypoth-
esis also nicely explains the factwhich I assume is obvious enoughthat
music has a much greater tendency to elicit physical activity than the visual
arts do, absent other explanations (I will sketch some in a moment). We expe-
rience sounds, but not sights, somatically. Again, a correlation suggestive of
causation.18
The somatic perception of sounds goes some way toward explaining what has
seemed to many the extraordinarily direct and immediate way in which music
affects our emotions. Somatic sensations themselves qualify as what psycholo-
gists call affective responses. Part of what we feel, when we feel rage or anger or
fear or love or contentment, is sensations of muscles tensing or relaxing. Physical
sensations constituting the perception of sounds are easily imagined to be feel-
ings associated with full-blown emotions, when we imagine a situation in which
the emotions in question are called for. They are easily imagined to be feelings of
rage, when one imagines something to be angry at, or fear when there is some-
thing to be afraid of. They may actually be feelings of rage when one actually is
angry at something. An important function of music accompanying narrative
worksfilm, theater, operais to generate somatic sensations, sensations con-
stituting our perception of the sounds, which we then construe as feelings of
anger at the bad guys, or disgust at their evil doings, or worried concern for the
good guys. This is a contribution that the piano or organ player in silent mov-
ies makes, a job taken over by sound tracks, often electronic ones, in latter-day
talkies.
One more speculation:why do rock musicians so often want their music to be
as loud as possible? Surely not to make for dramatic contrasts between loud and
soft passages; what needs explaining is their dialing up the volume of an entire
performance, soft passages (if there are any) included. The objective may be to

Alicyn Warren (personal communication) pointed out that participants in Silent Raves dance
18

wildly to music transmitted through individual headphones. If the music, which they hear but
scarcely experience somatically, nevertheless provokes movement, this would seem to suggest that
my speculation is wrong. But it is possible that suggestion is at work in this caserecognizing what
they hear as something that ordinarily would encourage movement, the Silent Rave dancers expect
to be moved, and so are moved. (Compare:the sight of a glass of whiskey might make an experienced
heavy drinker feel sick.) Also the heard but not felt music might not stimulate behavior as strongly, or
as automatically, as somatically experienced music would.
 P hys icalit y in Mu si c 49

maximize the intensity of listeners somatic sensations, and so the intensity of

their affective/emotional responses.
If these speculations are on the right track, it would be a serious mistake
to think of feeling sounds, experiencing them somatically, as merely a backup
system, a convenient alternative means of access to sounds for the hearing
impaired, but redundant for the rest of us. Isuggest that somatic perception is
an integral part of ordinary experiences of sound, one that supplements in cru-
cial ways, rather than duplicates, auditory perception. Daniel Levitin is seriously
mistaken, Ibelieve, when he claims that our access to the auditory world is
almost entirely through the eardrum.19
If somatic perception of sounds is crucial in listeners experiences in the ways
Ihave described, this would seem to apply equally to electronic and acoustic
music. It shouldnt matter how the sounds are produced, or whether they are
familiar ones with recognizable sources. It shouldnt matter what impression we
have, if any, of a music maker or sound creator. But the contrast between music,
whether electronic or acoustic, and the visual arts (not to mention literature) is
profound. We do not experience sights somatically, in anything like the way we
do sounds.
Visual and literary works of art do frequently cause affective, emotional states
in appreciators, and also elicit physical responses (tensing or relaxing of mus-
cles, for instance, if not foot tapping and marching). Music is not unique in this
respect. But the mechanisms by which pictures and novels and movies (apart
from their sound tracks) have these effects do not seem to involve anything com-
parable to somatic perception, and most of them seem to be indirect in ways
that somatically perceived music is not.20 Many of these responses are to char-
acters recognized in works of fiction, or to apparent artists. Watching a film or
reading a novel, we may respond emotionally and physically to what we under-
stand about the characters or apparent artists and their situations. Our responses
are not always cognitively mediated, however. There is more or less automatic
motor mimicry:we may unwittingly adopt the posture of a depicted character,
or mimic her facial expression, or tense our muscles as we watch a dancer using
hers. Emotions or moods of characters can be infectious. Appreciators can catch
them from characters or apparent artists, via whatever mechanisms, Isuppose,
are involved when we catch them from real people. There is also something like
motor mimicry in response to inanimate objects, rather than persons (fictional

19
Virtually all of your impressions of the auditory world come from the way in which [the ear-
drum] wiggles back and forth in response to air molecules hitting it (Levitin, This Is Your Brain on
Music, passim, p.102).
20
Visual works no doubt have some relatively direct influence on viewers physical states (if
not their behavior). Certain colors may be exciting, or calming, for instance. Thanks to Richard
Daniel Blim.
50 In Other Shoes

or actual) or anything we think of as being sentientmimicry of the shape or

movements of objects. Observing the Tower of Pisa, one might find oneself
diagonally inclined, in sympathy.21

Music employs a variety of devices for inducing physical and affective responses
in listeners, including most of the ones just mentioned, which it shares with the
visual arts. There are responses to apparent music makers:listening to sounds
seemingly made by a person acting in an overly aggressive manner, a listener
may feel intimidated. The impression of a calm producer of sounds may be calm-
ing. Apompous apparent music maker may be disgusting. These affective emo-
tional responses are likely to have physical manifestations, in the case of music
as in those of the visual arts and literature; ones muscles tense or relax. Probably
there is something like infectious emotions or moods, and motor mimicry.
My hypothesis is that somatic sensitivity to sounds, feeling vibrations, is
responsible for physical reactions in a very direct way, one not involving any
(even implicit) cognitive processes, or recognition of characters or apparent
music makers, or anything like motor mimicry or empathy with either a sentient
being or an inanimate object.
Some have argued that proprioception, the perception of the position and
movements of ones limbs, is important aestheticallyin appreciating dance,
for instance, or architecture. Somatic experiencing of sounds, feeling vibrations,
is not proprioception (although it might aid us in detecting the positions and
movements of our limbs). Those who regard proprioception as part of aesthetic
experiences often take it to result from or consist in a kind of empathy with (for
example) a dancer or a building; one mimics with ones body the movement or
stance of the object of appreciation.22 Feeling vibrations in our muscles when
we listen to music, Iam suggesting, is more direct and automatic than this. It
is not that we observe and somehow take on the physical stance or motions of
something outside of us.

21
This probably would be an instance of empathy, in its original sense, when it was introduced
into English as a translation of Einfhlung. Cf. Theodor Lipps, sthetik (Leipzig:Leopold Voss Verlag,
1903). Vernon Lee, The Beautiful:An Introduction to Psychological Aesthetics (Cambridge:Cambridge
University Press, 1913).
22
Cf. Gregory Currie, Empathy for Objects, in Amy Coplan and Peter Goldie (eds.),
Empathy: Philosophical and Psychological Perspectives (Oxford: Oxford University Press, 2011);
Barbara Montero, Proprioception as an Aesthetic Sense, The Journal of Aesthetics and Art Criticism
64/2 (2006), 23142; Richard Shusterman, Somaesthetics: A Disciplinary Proposal, The Journal
of Aesthetics and Art Criticism 57/3 (1999), 299314; Kendall Walton, Projectivism, Empathy, and
Musical Tension, Philosophical Topics 26/1 & 2, (1999), IX. Reprinted in this volume.
 P hys icalit y in Mu si c 51

One final note:Daniel Levitin wrote that Music communicates to us emo-

tionally through systematic violations of expectations, obviously taking a
cue from Leonard Meyer.23 I suggest that somatic perception of soundan
important feature, Ibelieve, of our experiences of both acoustic and electronic
musicis a powerful vehicle of emotional communication that has nothing
to do with expectations, letalone violations of expectations. This is important.
Traditions and standard formulae internalized by listeners are largely respon-
sible (arguably) for the expectations operative in music of the common practice
period. Electronic music (and contemporary music of various other kinds) does
not often exploit these traditions or formulae, and listeners may not have inter-
nalized other newer ones that the music might utilize. Experiencing the music
somatically can work emotionally on them nonetheless.24

23
Levitin, This Is Your Brain on Music, passim, p.172.
24
This essay benefited from helpful observations by Alicyn Warren.
 Postscripts to Chapter3:Two Kinds of Physicality
inElectronic and Traditional Music

A. Impressions of Production: Music

andPainting
In Two Kinds of Physicality Ipointed out that an important feature of both
musical performances and many visual works of art is the impression they give
of how they were produced. But Iwas too quick to declare that there is nothing
especially special about music (in general), as far as apparent circumstances of
production are concerned (p.41). There is something especially special about
music in this regard. The sounds of a musical performance seem to have been
produced by events or actions occurring now, as we listen. In the visual arts, it
seems that the events causing the visual object occurred sometime in the past.
The listeners impression is of being in the presence of, witnessing or experi-
encing, someone behaving in a certain wayviolently or calmly, frantically,
awkwardly, deliberately, or with abandon, perhaps running amok. On viewing
a drawing or painting, our impression is of experiencing the result of a previ-
ous actiona calm or frantic or aggressive or nervous one, for instance. The
difference is akin to that between coming upon a grizzly bears footprints and
seeing or hearing the grizzly itself, as it prowls around the campground. This is
an important respect in which music has a kind of immediacy that visual art
(and literature) lack.
A musical performance may also seem to be the product of past composi-
tional acts of certain sorts, of course, in addition to giving listeners an impres-
sion of currently occurring actions or events. The music may seem to have been
written (on some indeterminate past occasion) by a composer who was trying
to be funny, or to show off, or who was then cheerful or morose or pompous or
self-indulgent.

52
 P hys icalit y in Mu si c 53

B. Julian Henriques:Sonic Dominance

Julian Henriques describes what he calls sonic dominance in ways that suggest
both the reification of sounds, and the physicality of our experiences of them.
The Jamaican Reggae Sound System is, for him, a paradigmatic instance of sonic
dominance:

The first thing that strikes you in a Reggae sound system session is the
sound itself. The sheer physical force, volume, weight and mass of it.
Sonic dominance is hard, extreme and excessive. At the same time the
sound is also soft and embracing and it makes for an enveloping, immer-
sive and intense experience. The sound pervades, or even invades the
body, like smell. Sonic dominance is both a near over-load of sound and
a super saturation of sound. Youre lost inside it, submerged under it.
This volume of sound crashes down on you like an ocean wave, you feel
the pressure of the weight of the air like diving deep underwater. Theres
no escape, no cut off, no choice but to be there. ... Sonic dominance
is visceral, stuff and guts. Sound at this level cannot but touch you and
connect you to your body. Its not just heard in the ears, but felt over the
entire surface of the skin. The bass line beats on your chest, vibrating
the flesh, playing on the bone and resonating the genitals.1

Henriques recognizes that visual stimuli sometimes dominate ones conscious-

ness. But it is hard to imagine a visual experience that one would want to charac-
terize in anything like the way Henriques describes sonic dominance.

1
Julian Henriques, Sonic Dominance and the Reggae Sound System Session, in The Auditory
Culture Reader, ed. Michael Bull and Les Back (NewYork:Berg, 2003), 451452.
 4

ThoughtwritingIn Poetry
andMusic
But to return to myself, Iwas thinking about my book in more modest
terms, and it would even be a mistake to say that Iwas thinking of those
who would read it as my readers. For they were not, as Isaw it, my read-
ers, so much as readers of their own selves, my book being merely one
of those magnifying glasses of the sort the optician at Combray used to
offer his customers; my book, but a book thanks to which Iwould be
providing them with the means of reading within themselves. With the
result that Iwould not ask them to praise me or to denigrate me, only
to tell me if it was right, if the words they were reading in themselves
were really the ones Ihad written (possible divergences in this regard
not necessarily always originating, it should be said, in my having been
wrong, but sometimes in the fact that the readers eyes might not be of
a type for which my book was suitable as an aid for self-reading).
Marcel Proust1

Speechwriters compose speeches for others to deliver; speechwritings are

speeches composed for this purpose. By thoughtwriters Ishall mean writers who
compose texts for others to use in expressing their thoughts (feelings, attitudes).
The texts thoughtwriters compose are thoughtwritings.
I shall argue that poets are often plausibly regarded as thoughtwriters, and
their poems as thoughtwritings. Music and music making can frequently be
understood in an analogous way.

I. Poetry and Music

The arts are customarily divided into several broad categories, two of which are litera-
ture and music. Athird is the visual arts, whose species include painting and sculpture.

1
Marcel Proust, Finding Time Again, trans. Ian Patterson (1927; NewYork:Penguin, 2002), 34243.
(Thanks to Michael Gorowitz.)

54
 Tho ught w r it ing In Poetr y and Mus i c 55

Literaturethe word artsincludes novels, stories, and poetry. In Mimesis as

Make-Believe and elsewhere Itreated these literary forms mostly together, usu-
ally using novels or stories as my examples but thinking, assuming, that what
Isaid about them would, for the most part, go for poetry as well.2 Ishould have
been alert to significant differences between poetry, some poetry at least, and
most novels and stories. Poetry, some of it, belongs more naturally with music
in important respects than with other literary forms. This should not be surpris-
ing given the intertwined histories of these arts. It is arguable that in the murky
ancient past when poetry was almost always spoken, it was hardly distinguish-
able from song, then probably the dominant form of music.3 Sound is important
in poetry as it is in music, of course, and much less so, ordinarily, in the novel.
Likewise for meter and rhythm. Indeed, the particular meter and rhythms of a
poem may match those of a musical work, even if the sounds are very different.4
Hence the felicitous combination of poetry and instrumental music in opera,
oratorio, musicals, lieder, popular song, jingles, etc.5
I am interested now in another respect in which poetry is like music and
unlike the novel and other forms of literature, one that is less obvious than those
Ijust mentioned, but no less important. Bringing out this similarity will require
rethinking the nature and function of (some) poetry, and resisting a recent trend
in philosophers and music theorists understanding of musical expressiveness.

II. Narrators and Personae

I begin with a sweeping claim about all of the arts, advanced by a well-known
musicologist and music theorist, Edward Cone:[E]ach art in its own way proj-
ects the illusion of the existence of a personal subject through whose conscious-
ness [a certain kind of experience] is made known to the rest of us. That is the
role of the character in a play, of the narrator in a novel, of the persona in a lyric.6

2
Mimesis as Make-Believe: On the Foundations of the Representational Arts (Cambridge,
MA:Harvard Univ. Press, 1990) (hereafter cited as MMB).
3
Cf. Andrew Barker, ed. Greek Musical Writings, vol. 1. The Musician and His Art
(Cambridge:Cambridge Univ. Press, 1984); James Winn, Unsuspected Eloquence: A History of the
Relations between Poetry and Music (New Haven, CT:Yale Univ. Press, 1981); and Lawrence Kramer,
Music and Poetry (Berkeley and Los Angeles:Univ. of California Press, 1984).
4
Cf. C. L.Stevenson, The Rhythm of English Verse, The Journal of Aesthetics and Art Criticism
28, no. 3 (1970):32644.
5
Those who bemoan the lack of interest in poetry in our fast paced, digitally pumped up
twenty-first century, are forgetting that sung poetry is poetry. Probably no other art form is more
pervasive in contemporary culture.
6
Edward Cone, Some Thoughts on Erlknig, in The Composers Voice (Berkeley and Los
Angeles:Univ. of California Press, 1974), 3.
56 In Other Shoes

Cones main point is that this applies to music, that there are musical personae
that function somewhat as narrators in novels and speakers of lyric poems
do. He doesnt have in mind just the vocal line of songs, like Franz Schuberts
Erlknig; personae are to be found in all music. The piano accompaniment of
Erlknig has its own persona, he thinks. And there are personae in all purely
instrumental music. He mentions a fugue in J.S. Bachs Second Brandenburg
Concerto.7 On this picture, appreciation of all of the artsincluding instrumen-
tal music and all literaturecentrally involves recognizing, engaging with, and
responding to another person, a fictional person if not an actual one.
Most of us will have little trouble with Cones claim about literature, the
notion that most if not all literary works have narrators. But, special cases aside,
it is by no means obvious that there are anything like narrators in instrumental
music. Ihave worries about bothabout literature as well as music.8 Iwill pro-
pose an alternative to recognizing narrators in literary works, even in instances in
which they seem most obvious. This will suggest a way of understanding instru-
mental music that does not postulate musical personae, one that is, as it happens,
an approximation of one Iproposed previously.9

Literature
In nonfictional literature, a real person (typically) uses the words of the text
to communicate with readers, asserting the declarative sentences it contains.
Novels, stories, and poems are standardly understood as modelled on such ordi-
nary nonfictional uses of language, as fictional or pretend instances of what we
might call (with tongues partly in cheeks) serious discourse. Literary fictions
are created by real persons, of course. But the reader may be more interested

Cone, The Composers Voice, 9495.

7

Cf. Robert Stecker, Artworks:Definition, Meaning, Value (University Park:Pennsylvania State

8

Univ. Press, 1997), 208.

9
Cf. my Listening with Imagination:Is Music Representational? in The Journal of Aesthetics and
Art Criticism 52, no. 1 (1994):4761 and reprinted in this volume. Iraised objectionsdifferent from
those in the present essayto profligate postulation of musical personae in Projectivism, Empathy,
and Musical Tension, Philosophical Topics 26, nos. 12 (1999):40740 and reprinted in this volume.
Stephen Davies offered other objections in Contra the Hypothetical Persona in Music, in Emotion
and the Arts, ed. Mette Hjort and Sue Laver (NewYork:Oxford Univ. Press, 1997), 95109. See also
Derek Matravers, The Experience of Emotion in Music, The Journal of Aesthetics and, Art Criticism
61, no. 4 (2003):35363. Skeptics about the wholesale postulation of narrators in literary fictions
include Robert Stecker, Apparent, Implied, and Postulated Authors, Philosophy and Literature 11,
no. 2 (1987):25871; Andrew Kania, Against the Ubiquity of Fictional Narrators, The Journal of
Aesthetics and Art Criticism 63, no. 1 (2005):4754; Thomas E.Wartenberg, Need There Be Implicit
Narrators of Literary Fictions? Philosophical Studies 135 (2007): 8994; and Gregory Currie,
Narratives and Narrators:APhilosophy of Stories (Oxford:Oxford Univ. Press, 2010). See also George
Wilson, Elusive Narrators in Literature and Film, Philosophical Studies 135 (2007):7388.
 Tho ught w r it ing In Poetr y and Mus i c 57

in fictional narrators than in the actual authors. Narrators (what Ihave called
reporting narrators [MMB 36869]) are characters who, in the fictional world of
the work, speak or write the words of the text seriously, asserting the declara-
tive sentences it contains, thereby recounting the events of the story as known
fact. (In some instances narrators merely think the words of the text, think them
seriously.)
The readers experience of novels, stories, and poems, on this familiar picture,
is in the first instance an experience of recognizing, engaging with, responding to
this fictional persondoing so in imagination, at leastsometimes empathiz-
ing with her. The readers experience is akin to being addressed by or overhear-
ing a real person speaking or writing seriously.
Narrators are more prominent and more important in some works than in
others. Sometimes they are effaced, some would say to the point of disappear-
ance. In Mimesis as Make-Believe Iexpressed a preference for recognizing narra-
tors, effaced ones at least, in virtually all literary fictions, partly to leave plenty
of room for detecting even the slightest hint of expressions of attitude or point
of view on the part of the narrator (MMB 36869).10
But that was then. There is an important possibility that Idid not think of, a
way of understanding and experiencing literary worksand not just special or
unusual onesthat does not involve recognizing narrators, and need not con-
sist in anything like recognizing and responding to another person.11 This, Isay,
is a possible way of understanding literary works. Whether and to what extent
the possibility is actual is another question. Idoubt that readers experiences are
ever fully or exclusively of the kind Ishall describe. But it will be obvious, Ithink,
that in the case of poetry they are often partly so. This is an important dimen-
sion of many experiences of literature, one that has been seriously neglected by
philosophers and theorists of literature.

Poetry
Poetry, especially lyric poetry, is a form of literature in which narrators tend
not to be effaced, in which they are especially prominentso it seems anyway.
10
Idid suggest that in very special cases, for example, Manuel Puigs Heartbreak Tango, literary
works might be best understood as lacking narrators. Another special case is Lydia Daviss short story
Wife One in Country. (Thanks to Sarah Buss.) Ihave always emphasized that nonliterary works of
fiction often do not have anything like narrators.
11
Readers do, of course, recognize the actual authors of poems and other literary works, and
respond to them in various ways. Arguably they must do so in order to identify a work as an instance
of thoughtwriting. But one need not identify a work as an instance of thoughtwriting or even recog-
nize its author, in order to treat or experience it as such, that is to appropriate its words for ones own
purposes. The actual author (or musician or artist) is obviously not what Cone means by a personal
subject, to be found in any work of art.
58 In Other Shoes

Jenefer Robinson develops what she calls a new Romantic theory of expression
in the arts, on which works expressive of emotions (or ideas or attitudes) are
works in which there is a narrator or persona who expresses the emotions (ideas,
attitudes) in question. The theory applies to all of the arts, she thinks, at least
to all those artworks that have some claim to be called expressions, but paradig-
matically to lyric poetrymost obviously lyric poems in the first person. Her
main examples are John Keats and Percy Bysshe Shelley.12
But it is poetry especiallymore than novels and storiesin particular lyric
poetry, where my alternative proposal is most plausible! And Imost certainly
include lyric poetry in the first person.
How could we not think of the words, My heart aches, and a drowsy numb-
ness pains / My sense, as though of hemlock Ihad drunk13 as someones words,
used seriously? How could we not, when we read them, imagine a person using
them to express his or her heartaches? By understanding them to be an instance
of thoughtwriting. Iwill explain shortly.
Robinson allows for several familiar varieties of narrators (personae, poetic
speakers) in poetry, and Iwill as well. Sometimes the actual poet is the narrator.
Sometimes the narrator is a fictionalized version of the poet. And some narrators
are just plain fictions. Lets not worry about these differences. What is important
now is the notion that the words of lyric poems are understood to be written or
spoken or thought, seriously, by someone (either fictional or actual) other than
the reader or listener.

Music
What about music? Edward Cone is not alone in taking narrators in literary
works to have analogues in music, pure instrumental music included. Robinson
does also, and so does Jerrold Levinson. They do so largely in order to account
for musical expressions of emotion. All three are committed to something like
this:Just as narrators in literary works express beliefs, attitudes, intentions, emo-
tions, by means of the words of the text, musical personae express emotions
(feelings, attitudes) by means of the sounds (or gestures) of the music. So
listeners experiences, like those of readers, involve something like recognizing
and responding to another person, one who experiences and expresses the emo-
tion in question. Levinson proposed a general account of musical expression,

12
Jenefer Robinson, Deeper than Reason: Emotion and Its Role in Literature, Music and Art
(Oxford: Oxford Univ. Press, 2005), 234, 25455 (her emphasis) (hereafter cited as DR). Anna
Christina Ribeiro takes the first person in lyric poetry to encourage readers to identify with the speaker
of the poem. Toward a Philosophy of Poetry, Midwest Studies in Philosophy 33 (2009):6970.
13
John Keats, Ode to a Nightingale, in The Golden Treasury, ed. Francis T. Palgrave, rev. ed.
(NewYork:Macmillan, 1960), 296.
 Tho ught w r it ing In Poetr y and Mus i c 59

a predecessor of Robinsons new Romantic theory, which we can gloss as fol-

lows (omitting qualifications that need not concern us now):a passage of music
expresses a given emotion just in case it is aptly heard as the expression of that
emotion by the musics persona.14
I also made a suggestion in the direction of musical personae some years ago.
Iproposed that to describe music as anguished is to say that it is fictional that
the music expresses someones or somethings anguish.15 Icontinue to think that
some instrumental musica lyrical passage of a piano sonata, for instanceis
naturally and appropriately heard in this way.16 But it now seems to me that the
prevalence and importance of personae in music have been seriously exagger-
ated. The insight Ithought Ihad about anguished music was very limited.
Here is an initial worry:On the persona hypothesis, expression turns out to
be a species of representation:music represents itself as someones expression
of feelings, as a story represents itself as someones reports of a series of events.
(Levinson contributes to this impression by comparing hearing-as to seeing-as,
which he substitutes for Richard Wollheims seeing-in.17) Iexpect that others will
share my impression that this fails to do justice to the depth of the difference
between stories and (instrumental) music, between literary or pictorial repre-
sentation and musical expression.

III. Experiences From Within

An essay by R. K. Elliott, Aesthetic Theory and the Experience of Art, has
always struck me as uncommonly perceptive, yet frustratingly obscure.18 Elliott
distinguished between what he called experiencing a work of art from without
and experiencing it from within, and applied these notions to poetry, music,
and painting. The clearest application of the former, indeed the latter as well, is
to lyric poetry. A poem can be perceived ... as if it were the speech or thought of

14
Jerrold Levinson, Musical Expressiveness, in The Pleasures of Aesthetics (Ithaca, NY:Cornell
Univ. Press, 1996), 107.
15
Pictures and Make-Believe, The Philosophical Review 32, no. 3 (1973): 299. See also my
Style and the Products and Processes of Art, reprinted in Marvelous Images:On Values and the Arts
(NewYork:Oxford Univ. Press, 2008).
16
Cf. my Listening with Imagination:Is Music Representational? The Journal of Aesthetics and
Art Criticism 52, no. 1 (1994):49. Reprinted in this volume, p.155. It is not easy to see how certain
expressive features of music, for example, harmonic progressions, dissonances and consonances,
major and minor modes, can be heard as a persons expression of emotion.
17
Levinson, Musical Expressiveness, in The Pleasures of Aesthetics (Ithaca, NY:Cornell Univ.
Press, 1996), 109.
18
R. K.Elliott, Aesthetic Theory and the Experience of Art, in Aesthetics, ed. Harold Osborne
(Oxford:Oxford Univ. Press, 1972), 14557 (hereafter cited as AT).
60 In Other Shoes

another person, he wrote, and it is possible for us to make this expression our
own (AT 146). When we do make the expression our own, we are experiencing
the poem from within; otherwise we experience it from without.
This and other formulations suggest that, for Elliott, in both cases we per-
ceive the poem as the expression of another person, a person with whom we
have a third-person relation. The idea may be (although he doesnt put it this
way) that in the from within case we not only recognize this other person
the poet (qua speaker of the poem)and her expressive behavior; we also
empathize with her, imaginatively occupying her shoes and feeling (somewhat)
as she does:We experience an emotion ... through an imaginative assumption
of the expression and situation of another person (real or imaginary) ... Ido not
merely recognize that the poet is expressing, for example, sadness, but actually
feel this sadness (AT 147).19
If empathy is what Elliott has in mind, the reader empathizes not just with
the poets situation and her feelings, attitudes, etc., but also with her expressing
these feelings or attitudes by means of the words of the text. This means, Itake
it, that the reader imagines using the words of the poem to express her own feel-
ings or attitudes. One can empathize with a persons feelings, or those of another
sentient being, without empathizing with her way of expressing them. Imay feel
contentment, in imagination or actually, as I take a purring cat to feel, or joy
while watching a dog wagging its tail. But it is unlikely that I should imagine
expressing contentment by purring, or joy by wagging my tail.
There is no challenge yet to the idea that experiencing a poem is fundamentally,
in the first instance, a matter of recognizing and responding to another person, a
(possibly fictional) speaker of the poem. The readers experience from within is
triggered by, comes after, and is dependent on, the experience from without.20

IV.Speechwriters
Another way of understanding readers experiencesthe alternative I prom-
iseddoes challenge this idea.

But the appreciators experience is not an ordinary instance of feeling the emotion. When
19

we experience an emotion ... through an imaginative assumption of the expression and situation of
another person (real or imaginary) we need not and commonly do not experience it as we would if
the situation were unequivocally our own. ... [T]he emotion that Ifeel in experiencing a work of
art from within (and that which Ifeel as another persons in real life) may be present in me without
being predicable of me. ... [I]t would be false to say that Iam sad or even, unqualifiedly, that Ifeel
sad (147).
20
Robinson, who also claims, following R. G. Collingwood, that readers or listeners take on,
experience for themselves, or recreate in imagination the emotion they observe in the narrator or
 Tho ught w r it ing In Poetr y and Mus i c 61

Rather than understanding a poem on the model of an ordinary assertive or

expressive utterance, addressed to or overheard by a listener, we might under-
stand it on the model of a speech written by a speechwriter, for use by another
person. Speechwriters dont use the words they inscribe (not insofar as they
are simply speechwriters); they mention them.21 They produce a text for use
by someone else. The speechwriter doesnt assert the declarative sentences he
comes up with; his client asserts them when she delivers the speech.
The client knows where the words came from. She may appreciate and admire
the speechwriters skill and pay him well for his services. But recognizing him is
not necessary in order to make use of his words in her speech. The words might
have grown on trees or appeared in driftwood patterns on a beach. Or she might
think they did. Once she has the words before her, no matter how they got there,
she can use them assertively (or to perform other illocutionary actions) in her
address.
There are lots of disanalogies between poets and speechwriters, of course.
Poems and the speeches speechwriters prepare, the ways they are understood
and used and appreciated, differ hugely in obvious ways. But the differences are
obvious. The similarities I will point out are less so, but striking nonetheless
once noticed. So there is more to learn by focusing on the similarities than on
the differences. (We will take note of some differences in Section X.)
Lest we suppose that relying on speechwriters is rare or special or unusual,
keep in mind that we learn the language we use from others. Not only do we
master its syntax and semantics largely by witnessing others using it, we also pick
up, from other speakers and writers, pithy phrases, apt metaphors, allegories,
hypothetical examples, telling anecdotes, rhetorical tricks, and so on. We dont
usually acquire complete speeches from other language users, extended texts of
the kind a politician might pay a speechwriter to supply. But we do learn from
others combinations of words that may come in handy for our own use, as we
search for ways to express ourselves in our language. Some examples:Nothing
succeeds like success, It aint over til its over, Boys will be boys. I have

persona of an expressive work, emphasizes that these emotional responses result from recognition
of the narrators or personas emotion. An emotional expression will evoke emotion in those who
observe or hear or feel it because of what it signifies about the emotional state of the person expressing the
emotion (DR, 29091, her emphasis. See also 255, 265, 27071, 27677, 28889.)
21
Aspeechwriter does make use of the words he produces, but not in the standard waynot by
asserting the declarative sentences the words express, for instance. He uses them, nonstandardly, to
recommend that the client use them assertively in her speech. In quoting someone directly, Ialso
make use of the words Iproduce, the ones Imention. Ispeak or inscribe them in order to report that
the person Iam quoting used them, for example, assertively. Both speechwriting and direct quota-
tion essentially involve calling attention specifically to the words one produces. So Icount both as
mentioning the words.
62 In Other Shoes

learned from several politicians that if you want to admit that you screwed up
without admitting that you screwed up, the thing to say is Mistakes were made.
Our teachersother speakersdont usually speak for the purpose of pro-
viding listeners with words to use. They simply use words themselves, assertively
or otherwise. We neednt be concerned with what these speakers are up to. The
pithy phrases and so on are there, in what might as well be mentions rather than
uses, available to be added to our working repertoire for later use.
There are more formal cases also. Ajudge who says, to people getting married
or to courtroom witnesses taking the oath, Say after me ..., feeds words to the
bride and groom, or the witness, for them to use seriously.

V. Poets as Thoughtwriters
Poems arent trees. But they do contain phrases, sentences, paragraphs, verses
which readers can, if they wish, use themselves. The words are there ripe for
picking, no matter what the poet was doing in writing them down, and no matter
what the reader takes her to have been doing. People sometimes borrow phrases
or sentences from Shakespeare in conversation and in formal speech:To be or
not to be, Methinks she protests too much,22 Alls well that ends well, You
cant take it with you.23
We use words in thinking, in formulating thoughts for ourselves, as well as in
communicating with others. Words found in a poem or learned from other speak-
ers may come in handy for this purpose alsoto articulate thoughts, to express to
ourselves opinions, feelings, attitudes. Apoets words might strike me as just the
right way of expressing a thought Ithought 1 had.24 Or, on reading a poem Imight
decide then and there to endorse the thought they express, and use the words of
the poem to assert it to myself. Alternatively, Imay think of the words as clarify-
ing my thoughts, as well as providing a means of expressing them.25 So Iwill say
that poets sometimes serve not exactly as speechwriters, but as thoughtwriters.
(By thoughts Iinclude not just intellectual ideas, but any feelings, emotions,
sentiments, attitudes, etc., that might be expressed by means of words.)

Shakespeares exact words, in Hamlet, are:The lady protests too much, methinks.
22

Anna Christina Ribeiro remarks on the common practice of appropriation, where we use
23

poems written by others to express our own ideas or feelings. See Toward a Philosophy of Poetry,
Midwest Studies in Philosophy 33 (2009):70.
24
One of our most ordinary reactions to a good piece of literary art is expressed in the for-
mula:This is what Ihave always felt and thought, but have never been able to put clearly into words,
even for myself. Aldous Huxley, Tragedy and the Whole Truth, in Music at Night:and Other Essays
(NewYork:Fountain Press, 1931), 6. Thanks to Tilmann Koeppe.
25
There is a clear affinity here with the views of Benedetto Croce and R.G. Collingwood.
 Tho ught w r it ing In Poetr y and Mus i c 63

If Iagree with the sentiments expressed by the words of a poem or find them
especially apropos on a particular occasion, Imight recite the entire poem, seri-
ously asserting (if only to myself) those sentiments. Think of a person who, find-
ing herself in a tight spot, recites the twenty-third Psalm.
It is not just while repeating the words of a poem after reading it that Imay
make use of them in thought. Ican think or say them assertively to myself, even
as Iread. When Ido, Ihave the text of my internal speech in front of my eyes.
It is not unlikely that poets sometimes have, as at least part of their purpose in
composing a poem, the objective of making words available for use by their read-
ers. They may think of themselves, maybe not very explicitly, as thoughtwriters.
This could be the poets main or primary purpose. Insofar as it is her purpose,
she need not expect readers to imagine a fictional narrator using the words of
the poem, or to suppose that she herself, the author, meant them seriously. The
poet might expect the reader merely to recognize her invitation to use the words
himself, to recognize her role as a thoughtwriter. Or she might not expect even
this. The words are there for readers to use, no matter what the poet was doing
in writing them down, and no matter what readers think she was doing. For this
purpose, poems might as well be trees.26

Fiction or Nonfiction?
Suppose this is all that is going on. Consider an unlikely pure and simple case
in which the poet thinks of herself as just a thoughtwriter; she intends simply
to put words into readers mouths or minds, to give them a text by means of
which they can express, articulate sincerely held opinions. And suppose that this
is understood to be the proper, appropriate function of the poem. The poet does
not use the words or mean them seriously. And she does not expect readers
to recognize a fictional narrator who does. Appreciators then use the text in the
expected manner; some do anyway. Finding the words suitable for his purposes,
the reader genuinely expresses genuinely felt convictions by means of them.
Does the poem, in this pure case, count as a work of fiction, or of nonfiction?
There need be no imagining at all, and no prescriptions to imagine. Nothing is
true in the world of the poem; there is no fictional world. The poem isnt a work
of fiction, any more than a speech written by a speechwriter is.
Is it nonfiction? Yes, if that just means that it is not fiction. But it isnt a typical
work of nonfiction, an ordinary instance of nonfictional literature. The poet, the
author of the work, didnt use the words in the usual manner, but only mentioned

26
Words growing on trees and words of poems are equally available for use by readers. But readers
may have much better reason to expect that using the latter will be rewarding in one way or another
than that using the former would be.
64 In Other Shoes

them. She didnt assert its declarative sentences. The reader alone uses the words
(the word types)if he chooses to do so. The poem doesnt serve as an actual
vehicle of communication, not the usual kind at leastnot even a pretended or
attempted one.

Imagination, Pretense
It is not always possible for a reader to use the words of a poem seriously; she
might not be in a position to do so. Sometimes she will be unwilling to.
Consider thesewords:

Whenever Richard Cory went downtown,

We people on the pavement looked at him:
He was a gentleman from sole to crown,
Clean favored, and imperially slim.27

I am not a homeless street person, a pavement person. Nor have Iever known
a distinguished gentleman named Richard Cory. So Ican hardly assert these
words sincerely. Imight read a poem, written in the first person, about a death-
bed experience, when I am not facing imminent death ( John Donnes Holy
Sonnets), or a love poem when Iam not in love, or not with the person or the sort
of person mentioned in the poem. Few of us are likely to use the words Twas
the night before Christmas ... to recount, seriously, seeing a miniature sleigh
and eight tiny reindeer, and the rest.
If the ideas or attitudes expressed in a poem are ones the reader doesnt
accept, she may be unwilling to think or utter them assertively, and unable to do
so sincerely. She might not share the religious convictions expressed in one of
Donnes Holy Sonnets, for instance.
Does the reader, in these cases, have no choice but to read the poem either as
another persons serious utterance, or as the unusable and perhaps inappropri-
ate handiwork of a thoughtwriter? Enter the imagination, pretense, role-playing.
The reader may imagine uttering or thinking the words of the poem seriously;
he may pretend to do so. Ican imagine being a person of the pavement, observ-
ing from that lowly perspective a well-dressed gentleman named Richard Cory,
and Imay use, in imagination, the words Edwin Arlington Robinson supplied to
describe him. Imight pretend to curse death as Iprepare to die or, employing
John Donnes words, pretend to praise God and express confidence in an after-
life. We use the words of The Night Before Christmas to express a delightful

Edwin Arlington Robinson, Richard Cory, in Modern American Poetry, ed. Louis Untermeyer
27

(NewYork:Harcourt, Brace, and World, 1950), 123.

 Tho ught w r it ing In Poetr y and Mus i c 65

fantasy. No doubt karaoke singers often pretend to endorse the attitudes or feel-
ings they are expressing.
If I disagree with the sentiments expressed in a poem, I may try them on,
imagining uttering the words seriously to see what it feels like to express such
thoughts or attitudesand probably what it feels like to endorse or accept or
adopt them.
I need not, in any of these cases, empathize with anyone, with a fictional
pavement person, for instance, or someone who wholeheartedly endorses the
sentiments or accepts the ideas that Imerely try on. Ineednt recognize anyone
(either fictional or actual) to empathize with.28
Does this imagining or pretense make the work a work of fiction? No. The text
itself doesnt make anything fictional, doesnt prescribe any imaginings. It is not
fictional, by virtue of the words on the page, that anyone uses themseriously
or at all.29 What the text does is to invite readers to do soin speech or thought,
actually or in pretense. If a reader does use the words in pretense, he will be
engaging in what Icall a game of make-believe. It will be fictional in his game that
he asserts the declarative sentences in the poem. But nothing is fictional unless
and until he does so. So there may be what Ihave called a game world, but not a
work world. (This is exactly the situation Ienvisioned for music, insofar as it is
expressive just in the sense that does not involve musical personae.30)

Imaginative Resistance
If Idisagree strongly enough with the sentiments Iwould be expressing should
I assert the words of a poem seriously, if I find them sufficiently repug-
nant,Imight refuse to utter or think them even in pretense, refuse to try them
out. Imay be unable to bring myself to sing the praises of the Nazis even in
pretense, when a fascist poem invites me to do so, just as Imight be unable to
bring myself to stick pins into the portrait of a loved one. This is an instance

28
It would be misleading to say that Iempathize with a hypothetical or possible person. Ineed not
be thinking even of a hypothetical or possible individual, whom Iimagine to share the sentiments
Iimagine expressing, even if Iam well aware that it is possible that there should be such a person.
29
I am relying here on the account of fiction and nonfiction that I present in Mimesis as
Make-Believe, chapter2. Very roughly, a work counts as fiction if it prescribes propositional imagin-
ings, thereby making the imagined propositions fictional, i.e., true in the fictional world of the work.
[This account is useful for my theoretical purposes, but is not meant to capture any close approxi-
mation of the ordinary fiction/nonfiction distinction; that, Iargued, is massively confused. But
Iwould expect those inclined to take the ordinary distinction seriously to count prescribing imagin-
ings as a necessary (though not sufficient) condition for being a work of fiction.]
30
Cf. Listening with Imagination, IV. See also What Is Abstract about the Art of Music?
The Journal of Aesthetics and Art Criticism 46, no. 3 (1988):35164, reprinted in this volume; and
Projectivism, Empathy, and Musical Tension.
66 In Other Shoes

of imaginative resistance (properly so-called), an instance of the imaginative

puzzle.31 Hearing someone else (a fictional or actual person) express obnox-
ious sentiments might be unpleasant, but Iwould expect to be able to tolerate
this a lot more easily than expressing those sentiments myself, pretending to
express them sincerely.
We have here a partial explanation of the asymmetry there seems to be
between moral and descriptive contents of fictions. There may be no very great
difference between our willingness to imagine another person claiming that
time travel is possible and our willingness to imagine another person claiming
that infanticide for purposes of gender selection is right and proper. But Iam
much less willing to pretend to make the latter claim myself than Iam the former.
Ihave no trouble pretending to assert that time travel is possible, but Iwill resist
pretending to assert that it is perfectly all right to kill babies because of their
gender.
If, or insofar as, the point of a work is to put words in readers mouths, rather
than to have the reader observe a fictional speaker, what is a reader to do if he
is unwilling or unable to use the words even in pretense? He might still read
the poem, of course, and he may recognize the poems attempt to get readers
to go along, or its invitation to do so. He might also admire, and even appreci-
ate, the authors skill in expressing obnoxious ideas or sentiments, her skill as a
thoughtwriter.

VI. Is the Possibility Actual?

Appropriating the words of a poem or other text for ones own use, seriously
or in pretense, is not incompatible with recognizing a narrator, a poetic speaker.
Readers probably do both in many instances. Pure cases of thoughtwriting are
probably rare, maybe even nonexistent. But it seems to me that the thoughtwrit-
ing function of a poem is sometimes by far the more important one. It may be
important to the reader, and intended to be so by the poet, even if somewhere in
the readers experience there is an implicit, if not explicit, imagining of or recog-
nition of a serious speaker.
If we recognize a narrator of the twenty-third Psalm, a poetic speaker, it
would probably be its presumed actual author, a historical figure named David.
Alternatively, we might recognize a fictional narrator whom we imagine to be
using the words seriously, without presuming that it is David; the real David

Thanks to Jessica Wilson. This is not an instance of the fictionality puzzle. That puzzle cant arise
31

if the work doesnt purport to generate any fictional truths. Cf. my On the (So-Called) Puzzle of
Imaginative Resistance, reprinted in Marvelous Images:On Values and the Arts.
 Tho ught w r it ing In Poetr y and Mus i c 67

might have been just a thoughtwriter. But the adventurer lost hopelessly in a
trackless desert who recites the Psalm in serious desperation may have not the
slightest interest in David or whatever predicament he might have been in or
what he might have meant by his words, or in any fictitious speaker. The words
are there in the adventurers memory. He deems them appropriate in his situa-
tion, and uses them. Notice that the Psalm is in the first person. (Yea though
Iwalk through the valley of the shadow of death, Iwill fear no evil. ...) But
the adventurer neednt take its first-person pronouns to refer to David or to a
fictional narrator. I, in his recitation of the Psalm, refers to himself.
Think also of political and religious songs (for example, We Shall Overcome),
prayers, chants, mantras. It is unlikely that in singing We Shall Overcome at a
political rally people have a substantial interest in another person, the real life
composer of the lines or a fictional speaker, whom they believe or imagine to
express her genuine sentiments by means of them. Idoubt that readers or recit-
ers of Twas the Night Before Christmas, enjoying their fantasy, pay any atten-
tion to a fictional character who had an amazing experience one Christmas eve,
if they even recognize such a character. (The Night Before Christmas is also in the
first person).
What about the purposes of the authors of these textsthe twenty-third
Psalm, We Shall Overcome, Twas the Night Before Christmas? Did
they regard themselves as thoughtwriters? Who knows? The origins of the
twenty-third Psalm are murky. Those of We Shall Overcome are complicated.
(I gather that the latter arose gradually out of something like an oral tradition,
though not a very long one.) But if authors of these texts meant them to be
recited or sung repeatedly in religious services or political rallies or ceremo-
nies or ritual events, it would not be unreasonable to speculate that they meant
the reciters or singers to use them in expressing their own feelings or ideas or
attitudes.
One final group of examples: T-shirt and bumper-sticker slogans: Buy
Local, Vote for X, Support our Troops, Bring them Home, or on T-shirts
at the Ann Arbor Art Fair, It Aint Art, and it Aint Fair! The composer of such
slogans can hardly have anything in mind but their being used, on automobile
bumpers and T-shirts, to express thoughts of the driver or the wearer. The com-
poser may or may not agree with the sentiments of the slogans he produces. He
may be producing them just for the money, knowing that slogans other people
want to use are the ones that will sell. In any case surely he doesnt expect drivers
and T-shirt wearers to think of him, the slogan composer, as using the slogans
seriously, or to imagine a fictional narrator who does. (Compare a speechwriter
for a politician of an opposing political party, who needs the money, and hopes
to cash in on his wordsmithing talent.) Poems on commercial greeting cards are
not addressed by their authors to friends whose birthdays or anniversaries are to
68 In Other Shoes

be celebrated, or to whom they mean to offer sympathy for a loss.32 The authors
are paid to produce lines for others to use in these ways.
The twenty-third Psalm, We Shall Overcome, The Night Before
Christmas, bumper stickers, commercial greeting cards, dont say anything
directly about the early nineteenth-century lyric poems that I mentioned
earlier (Keats and Shelley). But they do show that we have it in us to employ
words primarily, even exclusively, in thoughtwriting or speechwriting
modes. We shouldnt be surprised if thoughtwriting is, in some instances,
the primary function of literary works, more important than their serving
as vehicles for the authors actual assertions or for fictional assertions by a
narrator.
Am Ibeing disrespectful to poets by comparing them to speechwriters, slo-
gan composers, etc.? Not at all. The analogy between poets and speechwriters
goes only so far. We will note differences in Section X that should put this worry
to rest.

VII. Clarification:Two Kinds of Appropriation

The appropriation Iam interested in is not to be confused with another kind.
Areader with an agenda or an axe to grind might read or interpret a poem as
endorsing or supporting his own personal views, ignoring what the poet meant
or might have meant or what by some other standard the poem, in its context,
actually does mean. The appropriation Iam interested in is not a matter of what
the words are taken to mean, but rather who it is who means something by them.
To appropriate the words of a poem, in my sense, is to use them oneself, not to make
them mean what one wants them to mean.
The first kind of appropriation is misreading, misinterpreting. To appropriate
a poem in my sense is not to misread it. Areader may be careful to understand the
poem correctly (by whatever standard one considers relevant), when he thinks or
utters the words assertively. Idont rule out the possibility that he misconstrues
the poem in a different way. If the poet meant her words to be understood only as
the words of another person, not to be used by readers, or if for some other reason
such appropriation is understood to be inappropriate, doing so may amount to
misusing the poem. But if the author intended readers to use her words them-
selves, and/or if that is understood to be the function or point of the poem, read-
ers who do not appropriate it in this way will, arguably, not be appreciating it
properly and fully.

32
Thanks to Jennifer Neilson.
 Tho ught w r it ing In Poetr y and Mus i c 69

VIII. Why Poetry in Particular?

All literary works contain words, words the reader can appropriate for her own
use if she chooses. Any text could conceivably be used by a reader to express her
thoughts or feelings. Is there anything special about poetry, or lyric poetry, in
this regard?
In poetry there is often special emphasis on how things are said, rather than
just what is said.33 (Speechwriters are charged with finding the best ways of say-
ing things, usually what the client wants to say.) So poems, in contrast to novels,
for instance, are especially likely to express thoughts or ideas in a manner that
the reader will find particularly apt for her expression.
There are indications that readers do in fact appropriate the words of poems
more often than those of other literary forms. We tend to perform poems as we
read them, to pronounce the words ourselves, sometimes to read them aloud;
arguably this is the best way to read them. People memorize poems or parts of
them, and recite them on other occasions. We are generally much less inclined
to repeat and recite passages from novels. (Some stories are told repeatedly, to be
sure, but often not in the same words.)
It is well-known that verse, with rhyme, alliteration, regular meters, tends
to be easier to remember, to memorize, than other linguistic forms. Hence the
usefulness of ditties to recall, for instance, the number of days in each month.
More seriously, stories, fables, myths are typically transmitted in oral traditions
by means of poetic language. Using words learned from his ancestors, an elder
of a tribe passes on to his offspring the wisdom and traditions of the culture.34
Peter Kivys arguments in The Performance of Reading are somewhat helpful to
my cause, although his cause is very different from mine.35 He points out analo-
gies between reading literary fictions and performing music, and he proposes
counting the readings, silent ones included, as performances. Insofar as these
analogies hold, it seems to me, the performing reader is likely to be using the
words himself, rather than (just) observing or recognizing another persons use
of them.
Kivys focus is on novels and stories, although he thinks his point applies
equally to poetry. But he excludes nonliterary texts entirely; no silent readings of
nonliterary texts are performances, he claims (PR 9093). Ithink the analogies

33
Thanks to Gregg Crane.
34
Thanks to Susan Pratt Walton.
35
Peter Kivy, The Performance of Reading: An Essay in the Philosophy of Literature (Malden,
MA: Blackwell, 2006) (hereafter cited as PR). See also J. O. Urmson, Literature, in
Aesthetics:ACritical Anthology, ed. George Dickie and Richard K.Sclafani (NewYork:St. Martins
Press, 1977), 33441.
70 In Other Shoes

between reading and musical performance are far more impressive in the case
of poetry than that of novels. And Iam unpersuaded by his reasons for exclud-
ing nonliterary texts (by which he means texts that are not works of art). In any
case, readers can appropriate the words of nonliterary texts for their own use
about as naturally as they can those of novelsdepending heavily of course, in
both cases, on what the words are. We do frequently, as Imentioned, add pithy
sayings and other combinations of words heard in ordinary conversation to our
own repertoires. But it seems clear that readers appropriate words of poems far
more often, and far more of them, especially while reading, than they do words
of either novels or nonliterary texts.36
Live theatrical events are almost inevitably experienced from without,
that is, we almost inevitably regard the words as spoken by another person, the
character, even if the character uses poetic verse. Seeing the actor pronounce
the words, we can hardly avoid thinking of them as being used, assertively, by a
person other than ourselves. This doesnt prevent us from also experiencing the
words from within, of course, thinking them assertively ourselves; we may very
well do both. What is unlikely is understanding the words merely as resources
made available for our use. This will be less likely, also, when we listen to a recita-
tion of a poem, than when we read it to ourselves. But the reciter, who, unlike
the stage actor, does not act (and dress) convincingly like a serious user of the
words, could be understood as a thoughtwriter offering words to the audience
verbally rather than in writing. We can hardly avoid recognizing narrators of
texts written in the second person, speakers or writers distinct from ourselves
addressing us.37

IX.Music
Several of the features that distinguish poetry rather sharply from the novel and
other literary forms are ones poetry shares with music (with much common
practice period music, at least). Imentioned our tendency to perform poems
as we read, to mouth the words ourselves, and the tendency to remember and
recite them on later occasions. Listeners sing along with music as they hear it, or
tap their feet, or sway with the music, dance, or march. Even if we dont actually

36
The performing Kivy thinks readers of novels and stories engage in is storytelling. To tell a story
is, Isuppose, to pretend to assert the declarative sentences of the text. My interest is also in readers
serious uses of the words of a poem, in their actually asserting its declarative sentences. Kivy claims
that the reader impersonates the teller of the story, the author, or fictional author. He declines to
call this impersonation make-believe, but it counts as make-believe in the sense of the term Ihave
employed.
37
Thanks to Steve Campbell.
 Tho ught w r it ing In Poetr y and Mus i c 71

voice a melody, we may follow along, thinking it as we listen. Idoubt that we do

anything much like this following along very often when we read a novel, or
watch a film. Listeners also remember previously heard melodies and, on later
occasions, hum or sing, or just think them, or tap out rhythms.
In performing music in these ways, listeners are likely to be using the sounds
(the sound types, anyway, or tokens of the types they hear) to express their own
feelings or emotions, or feelings or emotions they try on, ones they experience
at least in imagination. If musical sounds can be understood as someone elses
expression, that of a musical persona, as Levinson and Robinson claim, there is no
reason why they shouldnt be appropriated by listeners for their own expression.
Nor is there any reason why they could not function only in the latter way.
Using musical sounds or gestures to express ones own feelings is not incompat-
ible with regarding them as expressions of another person, a musical persona.
We often do both, no doubt. But we dont need to recognize a persona in order
to use the music for our own expressive purposes; we neednt take the music to
depict or represent itself as someone elses expression. Appropriating musical
sounds for ones own use need not involve anything like recognizing or empa-
thizing with another persons expressive behavior.38 This suggests that musical
personae, even when there are such, may not be very central in listeners experi-
ences. Even if at some level we do recognize a persona, what is important may be
just our own expressive use of the sounds (sound types).
So musical performances can be understood to present tunes, rhythmic pat-
terns, musical gestures for possible use by the listener, whether or not they also
present personae. Musicians (composers and performers) may intend or expect
their music to be so understood. They may serve as something like thoughtwrit-
ersas gesture- or expressive-behavior indicators. In some instances we may
not want to or may be unwilling to make music our ownmusic that we find
disgustingly pompous, for instance.39
Jenefer Robinson mounts an interesting and elaborate defense of musical
personae in (much) instrumental music. (She rejects Cones view that all music
should be experienced as an utterance by a persona [DR325].) Apsychologi-
cal reading of certain musical works is needed, she thinks, to make sense of
some apparently anomalous musical events, ones that would not make sense in
formal or purely structural terms (DR 329).40 She may be right about this. But a

38
Elliott seems to allow that music, unlike poetry, can be experienced from within without being
experienced from without, although he doesnt make this explicit.
39
Thanks to Alicyn Warren. She suggested that one might resist experiencing Krzysztof
Pendereckis Threnody for the Victims of Hiroshima as ones own expression.
40
She relies on discussions by Anthony Newcomb, Robinson and Gregory Karl, and Charles Fisk,
of works by Schuman, Shostakovich, and Schubert, respectively.
72 In Other Shoes

psychological reading need not include personae. One can hear a musical work
as a kind of psychological drama without hearing it as expressing a succession of
another persons psychological states. The psychology of the psychological read-
ing may be the listeners own, as she performs the music. The listener, rather
than a persona, may be the protagonist of the drama.41 Moreover, it seems to me
that just recognizing personae and the progression of their psychological states,
in the Robert Schuman, Dmitri Shostakovich, Franz Schubert, and Johannes
Brahms works she discusses, experiencing performances of these compositions
only from without, would not begin to account for the experiences apprecia-
tive listeners enjoy, for the intimacy that, Ihave suggested, listeners often feel
with music.42 Robinsons new Romantic theory of expression does have appre-
ciators recreating in themselves the emotions they recognize in personae.43
But the appreciators emotions, and their expressing them by means of musi-
cal sounds and gestures, need not be recreations of another persons emotion
or expression. Whether or not listeners do experience music from without,
it is their experience from within that is central to appreciation, at least of the
Romantic music Robinson focuses on.

X. DifferencesAmidst the Similarities

Most useful analogies can be misleading if pressed too far or in the wrong direc-
tions. The analogies Ihave proposed, including that between speechwriters and
thoughtwriting poets or musicians, are no exceptions. We need to acknowledge
differences, including obvious ones, if only to prevent them from obscuring the
similarities. Recognition of one especially striking difference will also deflect the
worry that my comparisons are disrespectful to poets.
Poets who serve as thoughtwriters are rarely if ever mere wordsmiths, and
it is not for their wordsmithing alone that we admire and appreciate them and
their poetry. They have wonderful things to say, as well as wonderful ways of say-
ing them. Poets contribute original ideas, fresh insights, perspectives, points of
view that may be new to the reader, as well as giving readers words with which to
understand and explore them.

The work itself isnt a psychological drama, with the listener as protagonist. There is no fictional
41

work world at all in the pure case, and in impure cases the listener does not belong to the work
world. The dramatic psychological events unfold in the listeners game world.
Robinson adduces several other considerations in favor of recognizing musical personae (DR
32629). But it is clear, Ithink, that they count equally in favor of the hypothesis that the music is
heard from within but not from without. They do count against certain other alternatives, however.
42
Listening with Imagination, 5455.
43
DR, 255, 265, 27071, 27677.
 Tho ught w r it ing In Poetr y and Mus i c 73

Speechwriters employed by politicians are often more than wordsmiths also.

The words they supply inevitably suggest clarifications and refinements, at least,
of the clients ideas. Aides hired as speechwriters frequently double as advis-
ers expected to make substantive proposals about the content of a speech. But
the speechwriter qua speechwriter is charged primarily with inventing effective
ways of expressing, of communicating to audiences, what his client has already
decided to say.
Poets are under no such restriction. They are free to compose words express-
ing thoughts that diverge as radically as they like from the preconceptions of
likely readers. Contributing new ideas doesnt have to consist in offering them
assertively, however. Poets neednt mean what they say any more than speech-
writers must, although they often do. Acontribution might consist just in making
words available for readers to use; in using them, readers may achieve drastically
new insights or find themselves adopting or trying on previously unimagined
points of view. The insights along with the words to express them would be avail-
able even if the words grew on trees.
The thoughtwriting poet, by her choices of what ideas to offer readers the
means of expressing, is likely to reveal much about herself, notably her attitudes
about what ideas are worth expressing, worth (at least) trying on. Insofar as the
speechwriters client dictates the content of the speech he is to write, his text
will not convey information of this kind about him. Speechwriters (qua speech-
writers) are likely to reveal much less of themselves than thoughtwriters do. Of
course readers can use a poets words themselves even if they dont learn any-
thing much from them about the poet.
When a musician offers listeners means of expressing feelings or emotions,
these may be feelings or emotions new to them, ones that differ at least to some
extent from any the listener has experienced previously. (Or they may be ones the
listener is not aware of or whose expression she has repressed.) Combined with
a narrativein film, theater, song, operamusic may encourage the appreciator
to feel or imagine feeling about depicted situations or the actions of characters
very differently from how she feels about similar real-life situations or actions.
Appropriating the music to express her (real or imagined) pleasure, for instance,
she may find herself taking pleasure, actually or in imagination, in (fictional)
events of a kind she would not take pleasure in in real life. (She might resist, of
course.) Aspeechwriter who attempted to manipulate his clients feelings or atti-
tudes in an analogous way would be dramatically exceeding his job description.

These differences, important though they are, do not conflict with or call into
question my claim that poetry and music are often instances of thoughtwriting,
74 In Other Shoes

thoughtwriting understood as analogous to speechwriting in the ways I have

insisted on. There can be no doubt that readers do, frequently, use words they
find in poems for their own expressive purposes (and sometimes communica-
tive ones), and that listeners use melodies and other musical materials similarly.
It is more than likely, also, that poets and musicians, expecting their works to be
so used, often craft them specifically so as to serve this purpose. Appropriation
is a hugely important aspect of our experience of music and poetryincluding
sung poetry, of coursea substantial part of why we value them and why they
matter in our lives.
Understanding poems and musical works as thoughtwritings (or
expressive-behavior indicators) nicely explains much of what counts as their
expressive qualities. Doing so thus undercuts a primary motivation for recogniz-
ing narrators in poetry and personae in music. It does not follow that narrators
and personae are absent when poems and music are so understood. They can be
present even if they are not needed to account for a works expressive properties.
But narrators and personae are far less important, if not less pervasive, than phi-
losophers of literature and music have made them out to be.
Poetry and music are strikingly similar in their propensity to function as
thoughtwriting. In this respect poetry is more like music than like other forms
of literature.44

Thanks to many people with whom Ihave discussed the ideas in this paper, beginning in 2007
44

and 2008 with audiences at the University of Southern California, Victoria University in Wellington
(New Zealand), the University of Sussex (U.K.), and the University of Warwick (U.K.). Aconversa-
tion with Eileen John first inspired me to try turning vague intuitions into something more coherent.
 5

Its Only a Game!

Sports as Fiction

Its official. Contrary to what college football fans might believe, the
earth will rotate as normal and not stop for todays Michigan-Colorado
game. You cant blame people for feeling that way, though. After all, this
is the first Game of the Season for 1997.
Tim Robinson, Michigan-Colorado Your Best Bet,
AnnArbor News, September 13, 1997
[The Florida Gators and the Ohio State Buckeyes] have developed a
healthy hatred for each other in record time.
Lee Jenkins, At Ohio State, Basketball Team
HopesToDo What Football Team Could Not,
NewYorkTimes, April 2, 2007
Baseball is make-believe, its fantasy, not real life.
Sparky Anderson, manager of the Detroit Tigers,
quotedintheAnnArbor News, May 10, 1993

Sarahs Dad is reading a scary story to her. She shows inordinate distress, so he
reassures her:Its just a story. Agroup of children are playing tag. Sam bursts
into tears when he is tagged. Dont worry, his Mom says, its only a game.1
Sports and competitive games of many kindsfrom tag to chess to base-
ballare often occasions for make-believe. To participate either as a competitor
or as a spectator is frequently to engage in pretense. The activities of playing

1
Thanks to audiences for talks in which Iincluded various of the ideas presented in this paper, and
to Patrick Maynard, Aaron Meskin, Nils-Hennes Stear, Paul A.Taylor, and Eric Walton. Aportion of
a version of this essay appeared in French translation as part of Le Sport comme fiction:Quand
fiction et realit concident (presque), in Les Arts visuels, le web et la fiction, ed. Bernard Guelton
(Paris:Publications de la Sorbonne, 2009). Translated by Bernard Guelton.
The discussion in this essay intersects with observations about competitive games, from a differ-
ent perspective, in my How Marvelous:Toward a Theory of Aesthetic Value, in Walton, Marvelous
Images:On Values and the Arts (NewYork:Oxford University Press, 2008), especially pp. 69.

75
76 In Other Shoes

and watching games have this in common with appreciating works of fiction and
participating in childrens make-believe activities, although the make-believe in
sports, masked by real interests and concerns, is less obvious than it is in the
other cases. What is most interesting about tag and chess and baseball, however,
are the ways in which the make-believe they involve differs from other varieties.
Make-believe is more prevalent and more blatant in some kinds of cases
than in others, and some peoplecompetitors or fans or observersare more
inclined than others to engage in pretense. Lets look first at spectators of staged
sports events, baseball fans, for instance.

In watching a stage play you root for the hero and boo the villain. You care about
characters you like and wish them well. Spectators feel badly when Romeo and
Juliet come to their tragic ends; some even shed tears. Likewise, sports fans root
for the home team, or for a team or player they like. Fans of the Boston Red Sox
or the NewYork Yankees cheer their victories and bemoan their losses. Alumni
follow the fortunes of their schools athletic teams.
Romeo and Juliet dont exist, and the spectator knows they dont. How,
then, can she care about them? This is a puzzle.2 Sports events do not present
an equally pointed puzzle. The Red Sox and the Yankees exist and they really do
win and lose baseball games. They are there to be cared about, and people do,
sometimes, really care whether they win or lose.
There is a lot to explain about sports, however. Why should people care about
the Yankees or the Red Sox? Their fortunes on the field have no obvious bear-
ing on the welfare of most fans. Why does it matter whether the home team
wins or loses? Life will go on afterwards just as it did before, regardless. But the
spectators, some of them, scream their hearts out during the game, as though it
is a matter of life and death. Some people pick which teams or players to like,
which ones to root for, more or less arbitrarily, on whimsbecause they find the
team logo or uniforms attractive; because a players name is the same as that of
an old flame, whatever. They may choose to root for the side their friends root
for, or decide to root against their friends favorites, setting up a friendly rivalry. 3

My answer, very roughly, is that she is caught up in a game of make-believe in which she imag-
2

ines that they exist; her experience is one of imaginatively following their fortunes, imaginatively
hoping for the best while fearing the worst, and grieving, in imagination, when the worst comes to
pass. These imaginings accompany and interact with often intense emotional affect. (Cf. Fearing
Fictions and Spelunking, Simulation, and Slime, this volume.)
3
Reasons for rooting for one side or another need not be reasons for thinking it would be bet-
ter that it win, or even reasons for preferring it to win. And one might grasp at even feeble reasons,
excuses, for choosing which to root for, thinking that the event will be more fun if one has a favorite.
 Spor ts a s Fic tion 77

Yet they may let themselves be carried away during the game, as though genuine
and substantial values or self-interest is at stake.
Are fans irrational? Do they believe, falsely but sincerely, that it really is a matter
of life and death? Have they lost their minds? This hypothesis is no more attractive
than the idea that readers of a story lose their senses, temporarily, and believe in
goblins or hobbits or magic rings. Many sports fans, like many readers of stories,
are otherwise sensible people who know what matters and what doesnt. Some
will tell you, if you take them aside and break the spell of the game, that it doesnt
really matter a whit who wins, or not much anyway, allowing that they dont and
didnt really care as much as they seemed to. Many forget the game quickly after it
is over, much too quickly for people who care as much as they seem to care during
the gamefor people whose hearts leap to their throats as they spring to their feet
to watch a long fly ball that may or may not be caught before it clears the fence.4
It is hard to resist comparing the avid sports fan to the playgoer who sheds bitter
and voluminous tears over the tragic fate of Romeo and Juliet, and twenty minutes
later has a jolly good time with her friends at an espresso bar. The fan imagines
that the outcome matters immensely and imagines caring immenselywhile (in
many cases) realizing that it doesnt actually matter much, if at all. She is caught up
in the world of the game, as the spectator at the theater is caught up in the story.
Afterwards, like the playgoer, she steps outside of the make-believe and goes back
to living her life as though nothing much had happenedeven if the home team
suffered a devastating and humiliating defeat. Its just a story; its just a game.

It isnt always just a game, however; sometimes it is not a game at all. There
remains the fact that, unlike Romeo and Juliet, teams and players exist and really

The choice may itself give one a reason to want it to win, however; we take pleasure in wins by the
team or player we happen to be rooting for.
4
Nils Stear reminded me that we sometimes quickly forget news reports of disasters in far off
places, even if our concern for the victims is entirely genuine and deeply felt. This kind of case seems
to me very different from typical instances in which fans get over sports disasters quickly. The news-
paper reader is concerned for others, for the misfortunes of strangers with whom he has little connec-
tion. The sports fans concern (actual or pretended) is selective, directed toward the players or teams
she roots for; she is likely to be unmoved by others who suffer equally horrible defeats in competitive
games. And her attitude may not be especially one of (actual or pretended) sympathy or pity, even
for the objects of her rooting; her experience is more like suffering a misfortune herself or as part of
a group to which she belongs. (She might bemoan the fact that we lost miserably, identifying with
her favored team.)
On putting down a newspaper, it is not hard to put out of mind the far off strangers whose suf-
fering one nevertheless cares about. It is not so easy to forget ones own misfortunes or misfortunes
closer to homeunless they are to some extent only pretended.
78 In Other Shoes

do fare well and ill in competition. So we can genuinely care about them, and
sometimes we do; sometimes it really matters. It usually matters to the competi-
tors; the salaries and careers of professionals are on the line, and so are the egos
of amateurs. Spectators also may care about the competitors welfare, especially
if they are friends or classmates. And one might expect that a winning home
team will shake loose large alumni donations to the fund that supplies ones
scholarship.
But these grounds for caring are often blatantly insufficient to account for
the intensity of spectators reactions during the game. And considerations such
as the prospect of alumni contributions are likely not to be on ones mind while
one is caught up in the game; they are likely not to be reasons one tells oneself
for wanting the home team to win. Superimposed on a modest genuine inter-
est in the outcome, there is, frequently, a pretense of much greater concern, and
of concern that is not, in ones pretense, of the kind one actually has. It is typi-
cally indeterminate in the pretense what kind of concern this is, why it matters
who wins and why one cares; it is fictional just that it matters a lot and that one
cares a lot. In games of tag, there is a pretense that being IT is undesirable, but
there is no answer to the question what, in the pretense, is undesirable about
being IT. This is another respect in which sports and competitive games differ
from literary and other fictions. We can give reasons why, fictionally, Romeo and
Juliet dont deserve their fate and why we care. 5
A spectators actual interest in the outcome of a sports event and the interest
she fictionally has in it, when both are present, do not merely coexist; usually
they interact, reinforcing one another in various ways. The spectator is likely to
experience sensations of excitement, pleasure, and disappointment, as the game
proceeds, because of her genuine concern, quite apart from any make-believe
(although her participation in the make-believe also plays a role in generating
such sensations). These sensations can then serve as props in the make-believe.6
She imagines them to be sensations of excitement concerning something that
matters greatly, and in (probably unspecified) ways different from the ways it
actually matters.7 Fans who place bets on the outcome make it really matter

5
Drastically indeterminate fictions are not unique to sports and games. They are especially
important in music and in some poetry and visual art. Cf. my What Is Abstract about the Art of
Music? and Listening with Imagination:Is Music Representational? (both in this volume). Some
games, Monopoly, for instance, do involve more specific sorts of make-believe. It is fictional, in a game
of Monopoly, that the participants financial well-being is at stake.
6
As reflexive props. See my Mimesis as Make-Believe:On the Foundations of the Representational
Arts (Cambridge, MA:Harvard University Press, 1990), 3.6, pp. 210213.
7
Compare:a dreamer imagines the sound of his alarm clock to be the sound of a school bell, or
Rodney, pretending to be driving a stage coach, imagines his actual swelling sensations to be swell-
ings of pride in his responsibility for getting a coach to its destination. Cf. Pictures and Hobby
Horses:Make-Believe Beyond Childhood, in Walton, Marvelous Images:On Values and the Arts, 7374.
 Spor ts a s Fic tion 79

to them more than it would otherwise, and they probably let themselves in
for more thrills and chills, or more intense ones, which then figure in their
make-believe in the manner Idescribed. Betting can be just business, like play-
ing the stock market; one hopes to make a profit. But it can also be a way of
enhancing make-believe, a way of making the make-believe more realistic (in
one sense). If the bet is a large one, it may be true as well as fictional that the out-
come matters greatly to the fan, although he may imagine that it matters in a way
that is not simply financial (even if there is no specific way in which it matters,
in his imagination). His attitude may not be simply that of a cold businessman.
Our make-believe involvement with a sports event may itself give us a reason
for genuinely wanting our favored team to win. We look forward to the plea-
sure of experiencing, in imagination, a victory of the good guys over the bad
guyswhether or not we have a special interest in the egos or salaries of the
competitors on one side or expect a windfall in alumni donations. Playgoers and
readers of stories sometimes take a similar pleasure in the fictional victory of
good over bad.

But tragic works of fiction have their appeal as welland now we come to an
especially striking difference between sports fictions and those of theater and
other arts. Tragedies can be deeply moving, even satisfying, if not exactly plea-
surable. So we sometimes want the bad guys to win, i.e. we want the work to have
a tragic endingeven while we are, fictionally, rooting for the good guys. We
may be pleased, to be displeased in the world of our pretense.8 This is rarely our
attitude concerning sports. Idoubt that fans are often moved by their favorite
teams losses in anything much like the way people are moved by the deaths of
Juliet and Romeo. The vaunted Paradox of Tragedy seems not to have much
of an analogue in sports. Some of us are fair weather fans. We tolerate a few fail-
ures by our favorite teams or players, but after a few more we either change the
object of our affection, find someone else to root for, or simply lose interest. It is
convenient to be able to tell ourselves that it doesnt really matter who wins and
forget about the whole thing, or to simply step out of the make-believe, when we
are denied the pleasure of experiencing, in imagination, victories of the side we
favor. People do sometimes, in some moods, decline to experience tragic works
of fiction, preferring fictions with happy endings. But for many of us, loyalty to
fictional characters and willingness to feel with them empathetically, through
thin as well as thick, far exceeds our willingness to stand by sports heroes.
(A representational work of art, a theatrical play, for instance, that portrays a

8
Cf. my Mimesis as Make-Believe, pp.193194, 258259.
80 In Other Shoes

team or player losing sports events may be a tragedy and may be appreciated as
such. Ernest Lawrence Thayers poem, Casey at the Bat is at least a mini-tragedy.)
In theatrical tragedies, it is partly because the good guy, the tragic hero whom
we root for, comes to grief that the work is moving. We may appreciate sports
events partly independently of the outcome; a game in which our favored side
loses can be enjoyable. But we dont appreciate it because our side lost.
Part of the reason for the absence of an interest in sports tragedies is probably
the indeterminacy Imentioned. What makes tragedy moving is not just the fact
that, fictionally, bad things happen, but also the fact that they happen for such
and such reasons, because the tragic hero has such and such flaws despite being
basically good, and faces circumstances of certain kinds. There is no answer,
typically, to the question of why, fictionally, the competitors in a tragic sports
event do or do not deserve the fate they receive, or to other questions concern-
ing the circumstances surrounding the disaster.
Indeed, there is probably no answer in the sports event itself as to what fates any
of the characters deserve; there are no ready-made good guys and bad guys in
sports. In the case of theater and other works of art, a controlling author or artist
typically decides who are the good guys and who the bad guys (and who are the
ambiguous ones), and manipulates us into rooting for the former and against the
latter. But sports fans are free to choose for themselves; each has his or her own
personal heroes and villains. To root for Iago and revel in Desdemonas death is
to misunderstand Shakespeares play. But you are not getting anything wrong
if you root for the Tigers instead of the Blue Jays, or the Blue Jays instead of
the Tigers. If your choice suffers miserably in the competition, you may regard
the event as something of a tragedy (though probably without appreciating it
as such), but for other fans it will have a wonderful happy ending. Tragedy in
sports is in the eyes of the beholding fan. (Make-believe games such as Cops and
Robbers are like sports in this respect.)
Sports events do not generally have anything like a controlling author or artist at
all. They are not anyones creation in the way that Romeo and Juliet is Shakespeares,
and they do not qualify as works in the sense that theatrical productions and
other works of art do. In some ways they are more like natural objects (the Grand
Canyon, Niagara Falls) than works of art, despite being products of human activ-
ity. Many sports events are not meant for audiences at all (dominoes in the park,
tag, pickup basketball). But even in the case of spectator sports like professional
baseball games and track meets, no one arranges the events of the game to best
advantage for appreciationat least no one is supposed to. The participants play
to win, not to put on a good show. (Professional wrestling is a bizarre exception.)
The resulting spectacle is largely a by-product of their competitive actions.
Some sports events do turn out to be good shows and others do not.
There are great games and sorry ones; ugly games and ones that are remarkable,
 Spor ts a s Fic tion 81

wonderful, memorable, if not beautiful. But the quality of the gamethe game
as a whole, as opposed to the play of individual teams or competitorsis
something of an accident, not something that anyone can take direct credit for.
Aclose score helps to make a game good or great; so do multiple lead changes,
and a result that is deemed an upset. But the competitors try to produce these
circumstances only insofar as doing so serves their interest in winning. They will
be eager to make the game close when they are behind, but once in the lead they
aim for the opposite resultthe pleasure of the fans, the opposing ones at least,
be damned.

In the bottom of the twelfth inning of the sixth game of the 1975 world
series, probably the greatest baseball game ever played, Carlton Fisk ...
hit a long ball toward left field in Fenway Park. It seemed to curve foul,
but Fisk gyrated his body, put some English on the air space between
home plate and the arching ball, and bent its trajectory right into the
left field foul polethus winning the game as he jigged around the
bases.9

Fisks ambition was not to create a great game, for the amazement of the specta-
tors and the wonderment of sports historians. Arguably it would have counted
as even closer than it was and even greater had it gone to a 13th inning. And Fisk
was not aiming for the foul pole.
On one occasion, as Iwas preparing a version of this essay to give as a talk,
Iwas more or less following on the Internet a Detroit Tigers versus Minnesota
Twins baseball game. The Tigers, whom Imore or less root for, were leading by
110 after the fourth inning, and 171 after the 7th. The final score was 181.
Was Ipleased? Well, it was a really awful baseball game, as a game. It bored me
(so I wasnt much distracted from my preparation for the talk). Do I fault the
Tiger players and manager for winning so big and so boringly? Do Iwish they had
deliberately made the score close, so the game would be exciting? Of course not.
Spectator sports are not quite show business, even if spectators pick up the
tab. This, of course, is the way we want it, even when our interest in who wins is
partly or largely make-believe.

There is a great deal of variety, more than Ihave indicated so far, in the atti-
tudes and behavior of participants in and spectators of competitive games.

Stephen Jay Gould, The Virtues of Nakedness, The NewYork Review of Books, Vol. 37, No. 15
9

(October 11, 1990), p.3.

82 In Other Shoes

Lets look more closely at participants, at the competitors themselves. Some,

we notedprofessional athletes, bettorsare likely to care immensely
about outcomes, leaving little or no room for make-believe (only, perhaps,
for make-believe about ones reasons for caring, if even that). The attitudes of
participants in casual contests with family or friendsa pleasant evening of
Chinese checkers or Scrabble or croquet or ping pongare often very differ-
ent. So also, sometimes, is a gentlemanly round of golf with business associ-
ates. Aperson may try hard to win in such instances, wanting perhaps to show
ones determination and show off ones skill or just enjoying the challenge,
yet hardly be disappointed if she loses. Acroquet player jokingly lambastes
a friendly competitor for blasting her ball to the boonies. (What did I do
to deserve this? And Ijust took you and your friends on a camping trip!
Just for that, no ice cream for you tonight!). AScrabble player rails at a rival
across the table (You rat!) for (inadvertently!) filling a space on the board
where she wanted to play, pretending to express venom that she doesnt feel.
An impassioned Darn! might express either fully genuine dismay, or dis-
may partly pretended. Alimiting case:a parent, playing tag or hide-and-seek
or a board game with a young child, deliberately, explicitly, consciously pre-
tends to be trying hard to win and win big, while actually making sure that
the child comes out ahead. The pretense might fool the child; but then again
it might not. Often, I expect, the parents pretense is obvious to the child,
and the parent expects it to be; the two knowingly play at engaging in serious
competition.
Joking can be partial. The reaction of a croquet victim may be only half in
jest, half pretended. But like the sports writers implication that the earth might
have stopped rotating for a football game (cf. the epigraph above), blatant exag-
geration betrays pretense and makes it obvious, and is probably intended to do
so. Friendly competitors play at engaging in serious, high stakes competitions,
while engaging genuinely in less serious, lower stakes competition.
Sometimes, no doubt, friends and family are not even half joking. Achild may
be thoroughly heartbroken, genuinely upset, when her brother knocks her cro-
quet ball into the bushes. It may be appropriate then to tell her, remind her, Its
only a game, encouraging her to adopt a make-believe stance and to recognize
others make-believe. Idont doubt that some adults care a lot about winning,
really care, even in competitions with friends and family in otherwise casual cir-
cumstances. Some of us are more seriously competitive than others.
It will not always be obvious whether and to what extent a competitor or
spectator engages in make-believe. Pretense is often not signaled by exagger-
ated shows of competitiveness like those of the croquet and Scrabble players
or clearly manifested in other ways. And we must, of course, leave room for
tacit, implicit pretense, pretense that may not be evident even to the pretender
 Spor ts a s Fic tion 83

herself.10 Perhaps in some instances there is no fact of the matter about whether
a person is engaging in pretense. In any case, many of us often do, to one extent
or another, engage in make-believe as we observe or participate in competitive
games, make-believe that is similar in some respects to that of childrens games
and theater, for instance, but intriguingly different in others.

10
Philosophers and psychologists generally agree, with support from empirical studies, that
desires, intentions, beliefs, attitudes can be tacit or implicit, unrecognized and unacknowledged by
the person who possesses them. Some theorists assumestrangely, without argumentthat this is
not true of imaginings or pretendings or engagings in make-believe. See, e.g., Peter Lamarque and
Stein Olsen, Truth, Fiction, and Literature (Oxford:Oxford University Press, 1994), p.47. First person
testimony about any of these mental states or activities can be overridden by inferences to the best
explanation. For a sampling of relevant empirical studies see John A.Bargh and Tanya L.Chartrand,
The Unbearable Automaticity of Being, American Psychologist, Vol. 54, No. 7 (1999): 462479;
Richard E.Nisbett and Timothy DeCamp Wilson, Telling More than We Can Know:Verbal Reports
on Mental Processes, Psychological Review, Vol. 84, No. 3 (May 1977): 231259; and Timothy
D.Wilson, Strangers to Ourselves:Discovering the Adaptive Unconscious (Cambridge, MA:Harvard
University Press, 2002). See also Brock, Stuart, The Phenomenological Objection to Fictionalism,
Philosophy and Phenomenological Research (2013): 119.
 6

Restricted Quantification, Negative

Existentials, and Fiction

Although much of our discourse concerning fiction is to be understood in a spirit

of pretense, Ihold that when storytellers assure children that they neednt worry
about monsters and bad guys in story books because they dont exist, and when
parents finally fess up to having fibbed about Santa Claus, they are speaking the
pure unvarnished, honest-to-god truth. (1)and (2), taken literally, aretrue.

(1) There is no Big Bad Wolf.

(2) There is no Santa Claus.

If this is not a tenet of common sense, it falls short only because literality is not
exactly a common sense notion.
What is said by means of (1)and (2)might well be rephrasedas

(3) The Big Bad Wolf doesnt exist.

(4) Santa Claus doesnt exist.

which also seem clearly to be true when understood literally, as do There are no
unicorns and Unicorns dont exist.1
Common sense isnt sacrosanct, and we sometimes have reason to
abandon our initial intuitions. But the fact that people commonly voice
denials like (1)(4) in a straightfaced and assertive tone and seemingly

1
To understand predications of existence and nonexistence such as (3)and (4)literally is not
to take them at face value; it is not to take the speaker to be attributing a property expressed by
the predicate to an object referred to by the subject expression. Rather (qualifications aside) the
speaker pretends to be doing this, as a means of asserting something of a different form. Such pretense
involving construals of statements like (3)and (4)are the central or standard ways of understanding
singular existence claims, and qualify as literal interpretations of the predicates exist(s) if anything
does. Cf. Walton 1990 and 2000.

84
 Negative Ex istential s and Fi c ti on 85

accept them as true, is part of the data that any theory of fiction must take
into account. Defenders of realist theoriesaccording to which The Big
Bad Wolf and Santa Claus along with their fellow fictions are constituents
of the universemust do some serious wiggling to accommodate it.1a As
Peter Van Inwagen admits in a footnote to Creatures of Fiction, what to do
about Mr.Pickwick does not exist and the fact that it can be used to express
a truth, given his realist position, is a very complicated question. Amie
Thomasson finds herself having to say that claims that fictional characters do
not exist are literally false.2
Probably the most common wigglewhich is not serious enoughconsists in
appealing to the familiar phenomenon of context dependent domain restrictions
on quantifiers. In a conversation about the exhibits in the Taronga Zoo in Sydney,
one might remark, There are no cheetahs, meaning not that there are no such
animals anywhere at all, but just that none of the animals in the zoo are cheetahs
limiting the domain of quantification to residents of the zoo. In other conversa-
tional contexts, of course, to say There are no cheetahs would be to say that there
are none in Australia, or that there are none on earth, or none anywhere. Realists
sometimes propose that (1)and (2), when true, concern a domain restricted to
real or actual things, leaving open the possibility that The Big Bad Wolf and Santa
Claus may reside in some other domain.3 Thomasson contends that [S]tatements
like there is no Lear, or there are no unicorns are quite naturally interpretable as
claims that ... there is no (real) person who is Lear or there are no (real) animals
that are unicorns, implicitly limiting the relevant domain to real persons and real
animals, and allowing that Lear may refer to something outside that domain and
that some other domain may contain unicorns.4

1a
[Not all suspect or controversial entities are challenged by negative existentials in the way fic-
tional objects are. Realist accounts of numbers, properties, and propositions dont have to contend
with, explain away, ordinary apparently serious and literal denials that there are such things, that
they exist. Amie Thomasson (1999, 2003)likens fictional entities to laws, governments, and literary
works, all of which she takes to be abstract artifacts. But Tolstoys War and Peace and the government
of Italy do not face the challenge of negative existentials. There is no War and Peace and The gov-
ernment of Italy does not exist are not, intuitively, straightforwardly and literally true, as (1)(4)
are. Santa Claus and the Big Bad Wolf are, in this respect, like the present King of France, phlogiston
and ether, and unlike War and Peace and the government of Italy.]
2
Van Inwagen 1977, 308 n.l 1.Thomasson 1999, 111112.
3
Terence Parsons invoked the restricted quantification strategy in Parsons 1980, 7, and espe-
cially Parsons 1982, 365366. Van Inwagens footnote hints at it (Van Inwagen 1977, 308 n.11), and
Thomasson 1999, 112113, endorses it.
4
Thomasson 1999, 112. In Thomasson 2003, she abandons the restricted quantification strat-
egy (possibly because of the considerations Iadduce below). Instead she proposes to follow Keith
Donnellan in taking Holmes doesnt exist to be true just in case the historical chain of the use
of the name ... [leads back] to what Keith Donnellan calls a block. (Thomasson 2003, 141)But
Donnellan introduced the notion of a block precisely to account for failures of reference; a name
86 In Other Shoes

This is not a satisfactory response. Although domain restrictions on quanti-

fied there is no and there are no constructions are perfectly common and
ordinary, predicative statements concerning existence do not admit of corre-
spondingly restricted readings. It is only with considerable strain, if at all, that
Ican hear Cheetahs dont exist as an assertion merely that none of the ani-
mals in the Taronga Zoo are cheetahs, as elliptical for Cheetahs dont exist in
the Taronga Zoo. Cheetahs dont exist is almost inevitably understood as a
denial that there are cheetahs anywhere, even if the zoo and its holdings are the
topic of conversation. Suppose Charlie is the name of a particular Cheetah.
Charlie doesnt exist is hardly a comfortable way of saying just that Charlie is
absent from the Taronga Zoo, that he is somewhere else, if anywhere. This utter-
ance would more likely serve as an announcement of Charlies death. (Notice
also that the quantified statement There exist no cheetahs, unlike There are no
cheetahs, is also not a natural way of saying merely that no cheetahs are in the
Taronga Zoo. This suggests that the use of exists prevents the quantification
from having a contextually determined domain restriction.)
So, although taking (1)and (2), when true, to say only that The Big Bad Wolf
and Santa Claus are not to be found among real animals or real people, rather
than that they are simply nothing at all, may appear to be a viable option, it is
far less plausible that this is what (3)and (4)mean. And of course parents, in
reassuring frightened children or fessing up about Santa Claus, are just as likely
to assert (3)or (4)as (1)or (2). It is no less natural or reasonable to deny the
existence of fictions than to deny that there are such, and no less plausible that
in doing so one is speaking the literal truth. If (3)and (4)are literally true, it is
hard to avoid concluding that those fictitious things are nothing at all, that they
simply are not, period.
As always, there are artificial, ad hoc fixes open to the determined realist. One
would be to declare that claims of existence and non-existence admit selectively
of limited construals, that exist(s) can be, in effect, elliptical for exist(s) in
reality (understanding reality not to comprise everything) but not for exist(s)
in the Zoo. This is not an attractive move.
The anti-realist conclusion can be avoided if we are willing to distinguish
between being and existence, a distinction attributed to Meinong and defended
in an important study by Terence Parsons.5 Rather than taking the extension of
exist(s) to vary with the context of utterance, the Meinongian understands it
to be fixed but to exclude some entities. There are such things as The Big Bad

whose historical chain ends in a block is one that does not refer to anything (Donnellan 1974,
2230). Thomassons realism forces her to insist that, despite the block, Holmes refers to a character.
Parsons 1980. Although Russell is famous for rejecting this distinction, he introduced it himself
5

in Russell 1903, 427. Cf also Edward Zaltas comments in Zalta 1988, 103.
 Negative Ex istential s and Fi c ti on 87

Wolf and Santa Claus, according to Parsons, but they dont exist. So Parsons will
accept (3) and (4), while denying (1) and (2). Parsons can then explain the
truth of statements like (1) and (2), when they are true, by claimingas he
doesthat they concern a domain limited to existing things, allowing that The
Big Bad Wolf and Santa Claus are present in other domains.
Many find the distinction between being and existence and the claim that there
are nonexistent things unappealing, to say the least. It would seem blatantly ad
hoc, Isuspect, were it not for its historical pedigree. (Actually, it may be true
to say, of an existing thing, that there is no such thing, when the domain of quan-
tification is restricted. There is no Charlie can be true when Charlie exists
somewhere other than in the Taronga Zoo.) Most recent realists holdwhat
they take to be a more common sense positionthat everything there is (in any
domain) exists, and that among existing things are fictional entities like The Big
Bad Wolf and Santa Claus.6 For these realists the problem remains.
The reasonable conviction that there is no ontological difference between exis-
tence and being, that to exist is no different from having being (in some domain
or other), has obscured the fact that these linguistic expressions operate differ-
ently. Noticing that the phenomenon of domain restrictions on quantifiers could
in principle explain how the literal truth of (1)and (2)may be compatible with
realism about fictional entities, realists have apparently taken for granted that a
similar explanation is available for the compatibility of (3)and (4)with realism.
This is a mistake. The evident literal truth of (3)and (4)remains a severe prob-
lem for realist theories.6a, 7

References
Donnellan, K. 1974, Speaking of Nothing, Philosophical Review 83:1, pp. 331.
Everett, A. 2013, The Nonexistent, Oxford:Oxford University Press.
Parsons, T. 1980, Non-Existent Objects, New Haven:Yale University Press.
Parsons, T. 1982, Are There Nonexistent Objects? American Philosophical Quarterly 19:4, pp.
365371.
Russell, B. 1903, The Principles of Mathematics, London:George Allen & Unwin Ltd.
Sainsbury, M. 2010, Fiction and Fictionalism, London:Routledge.

6
I shall defend the thesis that there are things Ishall call creatures of fiction, and that every sin-
gle one of them exists. I do not see any important difference between there is and there exists...
(Van Inwagen, 1977, 299, 300. Cf. p.302.). Thomasson 1999 insists repeatedly that fictional entities
are existing things of a perfectly ordinary variety.
6a
[Amie Thomasson (2009) allows that the greatest difficulty for artifactual views [which she
favors] arises in handling ... non-existence claims (p.15). In Thomasson (2007, ch. 6)and else-
where she does propose a new and interesting account of negative existentials. Ishare worries about
it that have been raised by Mark Sainsbury (2010, 108111) and Anthony Everett (2013, 154163).]
7
Thanks to Stacie Friend, Thomas Hofweber, Ian Proops, and Alberto Voltolini.
88 In Other Shoes

Thomasson, A.L. 1999, Fiction and Metaphysics, Cambridge:Cambridge University Press.

Thomasson, A.L. 2003, Fictional Characters and Literary Practices, British Journal of Aesthetics
43:2, pp. 138157.
Thomasson, A. 2007, Ordinary Objects, Oxford:Oxford University Press.
Thomasson, A. 2009, Fictional Entities in J. Kim, E. Sosa, and G. Rosenkrantz, eds, A Companion
to Metaphysics, 2nd ed., Oxford:Blackwell Publishing Ltd.:1011.
Van Inwagen, P. 1977, Creatures of Fiction, American Philosophical Quarterly 14, pp. 229308.
Walton, K. L. 1990, Mimesis as Make-Believe: On the Foundations of the Representational Arts,
Cambridge, MA:Harvard University Press.
Walton, K.L. 2000, Existence as Metaphor? in A. Everett and T. Hofweber, eds, Empty Names,
Fiction and the Puzzles of Non-Existence, Stanford: Stanford University Press: 6994.
Reprinted in this volume.
Zalta, E. N. 1988, Intensional Logic and the Metaphysics of Intentionality, Cambridge, MA: MIT
Press.
 7

Existence as Metaphor?

A person who asserts:

Deposed government officials are sent to Siberia and written out of his-
tory; they no longer exist.
or
She has a life; she really exists.
or
You know what Ive come to think, Jake? Ithink you dont exist at all.
There is too many of you ... youre different all the way through, every
time. You cancel yourself out ... Youre nothing ... 1
is likely to be using the predicate exist(s) metaphorically. Iwill not be talking
about statements like these. Ishall be concerned, instead, with garden variety
singular existential statements suchas:
Neptune (the planet) exists.
and
Sir John Falstaff does not exist.
uttered in ordinary garden variety contexts in which they are understood to be
(probably true) characterizations of the real world. It would seem bizarre to sug-
gest that these are metaphors. Statements about fiction suchas
Falstaff is a buffoon.
and
Narnia is inhabited by rabbits, dogs, and unicorns.

1
See Barth (1967), page67.

89
90 In Other Shoes

may seem, if not metaphorical, at least less (more?) than literal in some manner.
The same can be said of existential statements suchas:
Godot is only a figment of Vladimirs and Estragons imaginations; he
doesnt exist.
understood as a (possibly true) statement about the world of Becketts Waiting
for Godot. But how could it even occur to anyone to imagine that ordinary occur-
rences of Falstaff does not exist and Neptune exists are metaphorical?
I will argue for the astonishing thesis that indeed they are not metaphorical
anyway, my conclusion will be more or less this. Ihope there will be some inter-
est in the more or less, and in my reasons for raising the question in the first place,
and for not being entirely confident about the answer.
Here is how the question occurred to me. A few years ago I found myself
thinking about metaphor in terms of what Icalled prop oriented make-believe.
Prop oriented make-believe was central to my earlier account of existential
claims, although Ididnt call it that then.2 This raised the question of what con-
nection there might be between metaphor and ordinary existential statements.
More recently, Fred Kroon has argued that some close relatives of ordinary
claims of nonexistence are metaphorical.3

I. Overview of the Make-Believe Approach

I begin by sketching, with blinding speed, the outlines of my make-believe
theory and its treatment of existential statementsthe best Ican do in twenty
five words or less, give or take a few.4 My tempo will be just slow enough to
bore readers who are familiar with my views, Iam afraid, but sprightly enough
to mystify those who arent. (I wish Icould believe in the existence of the aver-
age reader. It would be nice if there were someone who will be neither bored nor
mystified. But at least Ican pretend.)
The central idea of the make-believe approach is, of course, that what seem to
be commitments, by speakers or theories, to nonexistent entities are to be under-
stood in a spirit of pretense or make-believe. They are, in one sense, not meant
seriously and not to be taken seriously. When, having exposed our parents lies

See Walton (1993), pp.3959, and Walton (1990), Part4.

2

See Kroon (1996). Yablo (2000) proposes regarding talk about various philosophically con-
3

troversial (purported) entities as metaphorical, although he does not focus on existence claims.
Crimmins (1998) proposes a pretense account of existence claims.
4
In this section Isketch some of the main ideas of Chapter1 of Walton (1990). The notion of
prop oriented make-believe, which Iintroduce in the following section, is developed more fully in
Walton (1993).
 Ex istence a s Metaphor ? 91

about Christmas, we go on to talk about Santa Claus anyway, and when we

talk about Sherlock Holmes, the Fountain of Youth, Vulcan, and ether, while
accepting that there arent really any such things, that they dont exist, we are
recognizing and utilizing fictions to the effect that they do exist, or better, fictions
to the effect that in using names and referring expressions like Santa Claus,
Sherlock Holmes, the Fountain of Youth, Vulcan, and ether, we success-
fully pick out (existing) objects.
Pretense and make-believe have a reputation for frivolity, like computer
games played while taking a break from serious number crunching or word
processing (when the boss isnt looking). But the pretense Ispeak of is serious
business, even if it doesnt involve seriously supposing that we actually refer to
what we pretend to refer to. We engage in make-believe in order to think and talk
about features of the real worldoften ones that matter, and sometimes ones
that are not easy to think or talk about in any other way.
The main theoretical resources needed for my account are, first, the notion
of a kind of mental activity or mental attitude which Icall imagining. (Others
speak of pretending or making-believe.) This often goes with pretense behavior,
actions of pretending, as when children flail around with sticks pretending to
have a sword fight, imagining that their sticks are swords and their flailing fenc-
ing. Secondly, Ineed the notion of something being true in a fiction, or true
within a pretense, or fictional.
These resources are ones that we must have on hand quite apart from a
theory of existence and of nonexistent entities. We need them to understand
the nature of stories, pictures, and theatrical performances, our interest in such
things and our emotional responses to themnot to mention childrens games.
Icant imagine trying to make sense of the institution of fiction or games of dolls
without giving a prominent place to some kind of imagining or making believe,
and to the notion of propositions being fictionaltrue in a story or true in
a fiction. So the make-believe account of existential statements gets the main
resources it needs pretty much for free.
There are differences of degree to be attended to. And perhaps some exten-
sions, as well as refinements, of ordinary notions of pretending and of fictions
are needed. Imaginings can be more or less vivid, conscious, explicit. We can be
caught up emotionally in a fiction or pretense or make-believe activity, or we can
stand back and observe it coldly, intellectually, objectively. Afiction or a game
of make-believe may be large or small, recurrent or fleeting, it may be engaged in
deliberately, or more or less automatically. Icount it an advantage of my theory
that it allows us to understand thinking and speaking about things that dont
exist, even in contexts remote from obvious works of fiction, as continuous
with the understanding and appreciation of ghost stories, historical novels, and
games of dolls or mud pies.
92 In Other Shoes

What is true in a fiction, or fictional, depends on real world facts. Children may
play a game in which bicycles are horses, and a garage is a corral. The real world
fact that a bicycle is in the garage makes it fictional, true in the make-believe, that
a horse is in the corral. Icall the bicycles and the garage props. Facts about them
generate fictional truths. The colors and shapes on the surface of a painting and
events occurring on stage in a theatrical production are props which generate
fictional truths, thereby establishing the world of the picture or the world of
the play.
Participants in the make-believe imagine what they recognize to be fictional,
they pretend that it is true. They imagine that there are horses in a corral when
they see bicycles in the garage. But it is fictional that there are horses in the corral,
if bicycles are in the garage, even if no one knows about the bicycles and so no
one imagines horses in the corral. What is fictional is something for participants
to discover, something they can be ignorant of or mistaken about. The fiction-
ality of a proposition consists in there being a prescription, in a given cultural
context, that participants imagine it to be true, whether or not anyone knows
about the prescription or actually imagines the proposition. All that is needed,
in the context of the game, for the prescription to imagine horses in a corral to
be in force, is the presence of bicycles in the garage.
Participants and their actions are often props themselves, generating fictional
truths about themselves. The fact that Jennifer is riding a bike makes it fictional
that she is riding a horse. This goes for verbal actions as well. If Jennifer says
Giddyup!, or The cattle rustlers were here last night, she makes it fictional
that she is urging her horse on, or saying that cattle rustlers were around the
night before.

II. Prop Oriented Make-Believe

Make-believe serves different purposes in different situations. Sometimes we
are interested primarily in what is fictional, in the content of the fictional or
make-believe world. On other occasions our interest is primarily in the props,
or the real world facts that generate the fictional truths. Ispeak of content ori-
ented and prop oriented make-believe.5 Props can be mere tools or vehicles for
establishing fictional worlds which we find exciting or interesting or poignant
or moving. But in some cases it is the props themselves that matter. Their role in
generating fictional truths in a game of make-believe may help us to understand
them, to think and talk about them, even if the fictional world itself is of no inter-
est at all.

5
See Walton (1993).
 Ex istence a s Metaphor ? 93

Fictions in the arts are often content oriented, and so are childrens games of
make-believe. We focus on what happens in the world of a story:on whether the
hero will arrive in time to rescue the heroine, why Hamlet was so wishy washy,
who done it, etc. Content orientation is especially evident when appreciators
are, as we say, caught up emotionally in the fiction. They may hang on a storytell-
ers every word, but only because of what the words reveal about the characters
and their fates.
Here is an extreme example of the opposite kind, an example of prop ori-
ented make-believe. Asmall fiction can help one learn how to tie a bowline.
First you make a loop. Then you say to yourself, The rabbit comes out of the
hole, goes under the log, and back into the hole, as you manipulate the rope
accordingly. This is hardly a gripping story. I certainly wasnt caught up in it
when Iwas taught how to tie the knot, as a child might be caught up in the story
of Peter Rabbit. And my purpose in manipulating the rope was not to make it
fictional that a rabbit scampers around in a certain way, as a child pushes a toy
truck around on the floor in order to create a fiction in which a truck tries to
outrun a pursuing police car, crashes into a barrier, or whatever. Iwasnt inter-
ested in the fictional scamperings of a rabbit. Ijust wanted to know how to tie
the knot correctly.
Consider this Not Very Tall Tale:
Once upon a time an organism was infected by a virus. The End.
This is no great shakes as a story. But if your computer tells it, i.e. if your computer
is such as to make it fictional that it suffers from a viral infection, this says some-
thing important about the computer. Describing the computer in this manner
saying that it has a virusis a simple and efficient way of conceptualizing and
communicating what has happened to your computer. (Of course the metaphor
works only for people who already know what viruses are.) We dont examine
the computer, as we do the words of a story, to discover the fictional events it
establishes; we look at what fictional events it establishes in order to understand
the state that it is in.
Many metaphors involve prop oriented make-believe. Computer virus is
one. Here are two others:

The sky is a snotty handkerchief.6

We add his letter to the large delta of unanswered mail on his desk.
[The letter will be there until] it has been gently silted over by about
twenty other pleasantly postponed manuscripts.7

6
Marge Piercy, quoted in Sommer and Weiss (1996), p.391.
7
Christopher Morely, quoted in Sommer and Weiss (1996), p.2.
94 In Other Shoes

A bronze or stone equestrian statue generates fictional truths about the

shapes of a horse and its rider, hence about the carriage and demeanor of both
and possibly their characters or personalities as well. Our interest, typically, is
in these fictional truths. We think of a mountain as a sculpture also, when we
speak of the shoulder or a saddle of the mountain, and when we see a head or a
face in it. This make-believe might be content oriented, but it is likely to be prop
oriented. We may want to know about the saddle of the mountain in order to
decide how best to cross over to the next valley, and we may have no interest in
the fictional saddle for its own sake.
I do not want to leave the impression that make-believe in the arts and in
childrens games is exclusively or even mainly content oriented, or that the
make-believe associated with metaphors is always exclusively or mainly prop ori-
ented. And of course there are combinations of prop- and content-orientation,
and intermediate cases. Also, we may have various different kinds of interest in
either content or props. Afew examples will illustrate some of the complexities:
When the children are playing their bicycle-horse game, Mom might tell
Johnny not to ride the fence even if some of his cattle are wandering away,
because she doesnt want him to be playing near a busy street. What for the
children is probably a largely content oriented make-believe, is for her a prop
oriented one. In this case, she could just as easily talk simply about the props,
without invoking any make-believe. She could just tell Johnny not to go near the
busy street. But given that the game is up and running, she can play along with it
as a way saying what she wants to say.
Even the children may, sometimes, be more concerned with the props than
the fiction. A child who challenges a friend, Ill bet my horse is faster than
yours, might be interested in nothing more than whether he can beat the friend
in a bicycle race. He may not care at all about fictional truths concerning horses.
Representational works of art are often of interest partly for the manner in
which they generate fictional truths as much as, or more than, for the fictional
truths they generate. Consider Picassos Bull, constructed from the seat and han-
dlebars of a bicycle. The fictional bull is no big deal, and neither are the bicycle
parts. What is interesting is the relation between them, the generating of a sur-
prisingly convincing bull by such simple and unexpected means. The bull is
noteworthy only for what it is made of, and the bicycle parts only for what they
make.8
The sky is a dirty handkerchief is not just a way of pointing out certain
physical features of the atmosphere. It also expresses an attitude about the
weather. It does so by virtue of the content of the make-believe, the content
of the make-believe in which the condition of the atmosphere is thought of as

8
See my discussion of ornamental representation in Walton (1990), pp.274289.
 Ex istence a s Metaphor ? 95

a prop. This remark is very different from describing the sky as, for instance, a
warm bubble bath.
Even if a metaphor is intended to introduce an almost entirely prop oriented
make-believe, the content often has a way of sneaking back in. If I describe a
particularly brutal ax murderer as an animal or a beast, you might reply that Iam
being unfair to the animal kingdom, to the likes of Buddy, Checkers, and every-
ones Fido. Maybe what Iasserted was just about the ax murderer. But there was
a Gricean implication about animals, due to the metaphor Ichose to use in mak-
ing my point. On the other hand, if Idescribe the ax murderer as, say, a vicious
rabid dog, and you respond that this is unfair to vicious rabid dogs, you may be,
not defending the honor and reputation of vicious rabid dogs, but rather empha-
sizing, even more, the brutality of the ax murderer.

[Rex Morgan has] all the personal pizazz of decaffeinated jello.9

is a metaphor masquerading as a device to describe Rex Morgan, as though

he is the focus of interest. But the content of the make-believe steals the show.
Probably we dont really care whether Rex Morgan (the cartoon character) really
is boring enough to justify the metaphor. The speaker has in effect created his
own intriguing little absurdist fiction, disguised as a metaphorical description
of Rex.

III.Assertion
Acts of verbal participation can be real assertions. In pretending to say one
thing, one may actually be saying, asserting, something else. Iwill make do with
a fuzzy, intuitive notion of assertion.10 All Iwant to do now is to outline several
different patterns that may obtain, on most any reasonable account of assertion,
several ways in which what is asserted may be related to what is pretended. (I will
say nothing about the related issue of whether, when a speaker utters a sentence
in pretense, thereby asserting something other than what it literally means, the
sentence has a special secondary meaning, e.g., a metaphorical meaning.)11
When the make-believe is content oriented, the assertion is likely to be about
the content of the make-believe. The reader of Gullivers Travels who says, in pre-
tense, There is an island southwest of Sumatra where the horses are 4.5inches

9
National Public Radio, 3 March 1995.
10
What is asserted is one kind of speakers meaning. It might be understood in Gricean terms as
an instance of what a speaker means by an utterance.
11
See Davidson (1984).
96 In Other Shoes

tall is likely to be asserting that it is fictional in the novel that there is such an
island.
When the make-believe is prop oriented the speaker is (naturally enough)
likely to be saying something about the prop, not about the fiction or
make-believe. Johnny may say to his mom that his horse is in the corral, as a
way of telling her that his bicycle is safely stowed in the garage for the night.
He is pointing out the real world circumstance which makes it fictional that his
horse is in the corral. There is nothing about make-believe or fiction or pretense
in the content of the assertion. Make-believe simply explains how the sentence,
My horse is in the corral can serve as a way of saying that his bicycle is in the
garage. To say that a computer has a virus is probably to assert something about
the contents of the computers hard disk, not anything about a fictional virus,
or about fiction or pretense at all. Acomplex feature of the hard disk makes
it fictional that the computer has a virus, and in pretending to assert that the
computer has a virus, the speaker simply indicates that the hard disk is in this
condition.
In many instances of verbal participation, however, the words uttered, taken
literally, do not express a proposition, or at least it is arguable that they do not.
The speaker cannot be asserting that the proposition expressed by her words
taken literally is fictional, nor that the circumstances which make it fictional
obtain, if there is no such proposition. Iaccept the view that, since there is no
such thing as Lilliput, Lilliput is an island southwest of Sumatra does not
express a proposition. Ialso subscribe to the view that there is no such property
as being a unicorn, or a hobbit, or a snark, and hence that There once was a fam-
ily of hobbits fails to express a proposition.
Many metaphorical statements are not just false when taken literally (some of
them are true, of course), but are in one way or another incoherent, often com-
mitting egregious category mistakes. Some complex predicates with obvious
metaphorical uses are composed of bizarre combinations of adjectives or nouns;
the predicate as a whole, taken literally, refers to no property that anything could
conceivably possess, and perhapsdepending on what we are willing to count
as a propertyto no property at all. Here are some examples:

Hateful thoughts enwrap my soul in gloom,

[Radio voice] We have a cello concerto on tap, after this
announcement,
I warrant they would whip me with their fine wits till Iwere as crest-
fallen as a dried pear,12

12
Falstaff in Shakespeare, Merry Wives of Windsor, IV, 5.
 Ex istence a s Metaphor ? 97

When it comes to public relations, [Kenneth Starr] is a one-man

train wreck.13

Some metaphorical statements, or anyway some idiomatic expressions, are not

even grammatical.

I wont have any truck with that,

He took a shine to her,
She lost her cool.

We also speak of burning issues, food for thought, mountains of debt,

weighty arguments, chewing someone out, wrestling with a problem, and
being saddled with responsibility. It is arguable that attributing these predi-
cates to a subject is not to express, literally, a proposition.
A speaker may pretend that her words express a proposition, however, even if
they dont, and she may pretend to be asserting a proposition by means of them.
Moreover, it may be fictional, true in the pretense, that she asserts something
true, even if there is no proposition that, fictionally, she asserts. Whether this is
fictional will depend on real world circumstances. It is fictional that one speaks
truly when (in discussing Tolkiens The Hobbit) one says Bagins is a hobbit.
Fictionally, that sentence expresses a true proposition, although it doesnt actu-
ally express a proposition at all. Features of the text of the novel, together with
conventions about how to understand them, make this fictional. (I wont try to
spell out the details.) The novel is such that to say, Bagins is a hobbit is, fiction-
ally, to speak truly, and to say Bagins is a unicorn is, fictionally, to speak falsely.
What might speakers actually be asserting, when there is no actual proposition
that they are pretending to assert? In place of the proposition that ones words
(taken literally) express, we have the proposition that in uttering them, in pre-
tending to assert as one does, one is speaking the truth. When the make-believe
is content oriented, the speaker typically asserts that to pretend in this manner
is fictionally to speak truly. When the make-believe is prop oriented, the speaker
typically asserts that certain circumstances obtainthe circumstances that do
or would make it fictional that she speaks truly. In saying Bagins is a hobbit one
is likely to be asserting that to pretend in the way demonstrated is fictionally to
speak truly (in official games using The Hobbit as a prop). In describing some-
thing as food for thought, one is probably saying that it is interesting, puzzling,
and thought provoking, these properties being what makes it fictional that the
speaker is expressing a truth.

13
NewYork Times, editorial, 5 April 1998.
98 In Other Shoes

We must not suppose that it will always or even usually be possible to specify
the relevant circumstances without mentioning the make-believe. We may have
epistemological access to the facts of which we speak only by means of their role
in make-believe, and we may be able to refer to them only via the fictional truths
they generate, only as the circumstances that generate such and such fictional
truths. Or it may be difficult, at least, to conceptualize the underlying facts or to
think about them perspicuously without engaging in or alluding to make-believe.
Iand others have discussed several kinds of cases in which make-believe is essen-
tial or helpful, in one way or another, in our thinking about various kinds of
facts.14 Here is another example:
There is (I assume) no such thing as absolute motion and rest. One object
is in motion or at rest relative to another, but neither is in motion simpliciter, or
stationary simpliciter. Our perceptual experiences seem not to accord with these
facts, however. We see one object as fixed, and another as in motion. What we
see as fixed can change. When my train pulls out of the station Imay at first see it
(and myself) as stationary and a train on the next track as in motion; Imay then
switch to seeing the other train as fixed and mine as moving. Neither perspec-
tive is correct; neither train is in motion simpliciter, or stationary simpliciter. The
switch in perspective may be induced by the realization, or the perception, that
the other train is not moving relative to the earth, while my train is. The earth is
not fixed absolutely or in motion absolutely either, however, but only relative to
other things (the sun, the solar system).
We can think of this familiar phenomenon as one of perceiving in accordance
with a fiction, with what we know to be a fictionthe fiction that there is such a
thing as absolute motion and rest. The perceptual content of ones visual experi-
ence, at a particular moment, includes the fiction that one or another particular
object is fixed.
It is not impossible to describe events without relying on this fiction; we can
say that two objects are at rest, or in motion, relative to one another. Speaking
this way may be necessary to avoid ambiguity, when different people are think-
ing of different things as fixed. But descriptions of relative motion are awkward.
When ambiguity is not a danger, it is usually easier and more perspicuous to
speak and think as though the fiction is true, pretending that some things are
really stationary and others really in motioneven if everyone involved knows
that this is not so. We recognize a kind of make-believe in which facts about rela-
tive motion and rest generate fictional truths about absolute motion and rest. By
speaking in terms of this fiction, pretending to be making claims about absolute
motion and rest, we actually describe relations of relative motion and rest, i.e.

See, for example, Walton (1993), pp.4345, Crimmins (1998), pp.28, and Yablo (2000). See
14

also Hills (1997).

 Ex istence a s Metaphor ? 99

the circumstances that do or would make it fictional that, in pretending as we

do, we speak truly.
Perceiving in terms of the fiction seems not to be optional, even if speaking
and thinking in terms of it is. Try as Imight, Icannot perceive things as Ithink
they really are; Icannot see two objects as moving relative to one another with-
out seeing one of them, or a third thing, as fixed absolutely. (I wonder if it would
be possible to train myself to do this.) The make-believe seems to be indispens-
able as far as perception is concerned; only by means of it do Ihave perceptual
access to the actual facts, facts about relative motion and rest. The convenience
and perspicuity of descriptions of things in terms of absolute motion and rest is,
no doubt, a result of this feature of perception.
It is arguable that notions of absolute motion and rest are not merely mis-
taken as applied to events in the real world, but that they dont even make sense.
It may be best to say that there are no coherent propositions, perhaps no propo-
sitions at all, which we pretend to assert when we say, e.g., that a car came to a
complete stop at a stop sign. The difference between perceiving my train to be
(absolutely) stationary and another one (absolutely) in motion, and perceiv-
ing the reverse to be the case, may not amount to perceiving one or another
(coherent?) proposition to be true. It does involve perceiving the trains in a
way that makes it natural for English speakers to apply the predicates, moving
and stationary differently. (More needs to be said about what this difference
amounts to.)

IV. Establishing and Extending Fictions;

Conniving Pretense
No elaborate ceremony is needed to initiate or introduce a new game
of make-believe, in many cases, even an unfamiliar one. (This is already
obvious, given that metaphors, fresh as well as established ones, involve
make-believe.) Imight just say, I bet my horse can run faster than yours,
as we are riding our bikes, and you get the idea. If Irefer to a person who
happens to be an elevator operator as an indoor aviator, it may be obvious,
without my spelling it out, that I am introducing a make-believe practice
in which to operate an elevator is, fictionally, to pilot an aircraft. In these
and many other instances, the game is introduced by making what is in the
context easily construed as a move in it, by participating verbally. Ipretend
to say something about horses, for example, thereby launching a fiction in
which bicycles count as horses.
Once a game is begun, there are usually natural ways of extending it, ones
that do not have to be explained or made explicit. The make-believe introduced
100 In Other Shoes

by remarking that we are all in the same boat is naturally extended by the sug-
gestion that, since we are all in the same boat, it behooves us to row in the same
direction.
Engagement in make-believe tends to be infectious. One speaker may extend
anothers metaphor, continuing the pretense begun by the first. Astoryteller pre-
tending to report on a house break-in by three bears may inspire his listeners
to inquire, in pretense, whether the bears meant any harm, or to observe that
they didnt. Alistener may, in pretense, report to a third person that three bears
did break into a house, the residence of a certain Goldilox and her family. What
Icalled work worlds thus spawn larger game worlds in which the work is a prop,
and in which appreciators participate.15
Stories and other representational works of art are designed to initiate
games of make-believe of certain kinds, to introduce pretense with which
appreciators play along (implicitly if not explicitly). But sometimes we take
liberties, combining or extending or altering fictions in ways that contra-
vene the established or official limits of the original. Many such unofficial
games (as Icall them) are intuitively natural, however, and recognizing them
may serve our purposes, purposes which go beyond simply appreciating
the work. Manguel and Guadalupis The Dictionary of Imaginary Places is a
travel guide describing, in alphabetical order, exotic places like Brobdingnag,
Middle-earth, Shangri-la, Oz, and Ruritania, places which share the property
of being imaginary, the blurb on the back cover tells us, and also the prop-
erty of having been invented by writers like Jonathan Swift, J.R. R.Tolkien,
Edgar Allen Poe, L.Frank Baum, and Edgar Rice Burroughs. We can say that
Brobdingnag is larger than the Blessed Island. Brobdingnag (from Gullivers
Travels) is a 6000 mile long peninsula on the coast of California north of
Monterey, while the Blessed Island (from Lucian of Samosata, True History,
2nd c. A.D.), an island in the Atlantic Ocean whose bodiless inhabitants
wear purple spider webs, is a mere 500 miles long. This is a nice example
of an unofficial game of make-believeunofficial, that is, with respect to
Gullivers Travels and True History. It is made official by The Dictionary of
Imaginary Places.
The make-believe in this case is likely to be prop oriented (with respect to
the unofficial game one is participating in). Saying that Brobdingnag is larger
than the Blessed Island is likely to be a way of pointing out the circumstances
that make it fictional in the unofficial game that one speaks truly, namely the fact
that there is a size such that it is fictional in Gullivers Travels that Brobdingnag
is larger than it, and fictional in Samosatas True History that the Blessed Island

15
See Walton (1990).
 Ex istence a s Metaphor ? 101

is smaller. The content of the utterance concerns these other fictional worlds,
though not the world of the unofficial game the speaker is participating in.16
Sometimes we play along with discourse or activities that are not them-
selves pretense. We pretend to be serious in a way others really are. And in
so pretending we may be reporting on their serious discourse or activity. In
discussing the findings of the court at the Salem witch trials Imight say, Martha
Carrier is a witch, pretending to believe and to assert what the court decided
was true. One might declare that the planet Vulcan has such and such mass, as a
way of specifying the content of the Vulcan hypothesis.
Quoting someone directly can be understood as pretending to speak in a cer-
tain way in order to show how the quoted person spokemost obviously when
there is sarcasm or mimicry in the quoting persons tone of voice. An example
from Herb Clark and Richard Gerrig:

So her mother said [whinny voice] No, you cant go out before you
make your bed.17

Moments of pretense often creep into indirect quotation and other predomi-
nantly third person descriptions of other people. One mightsay,

John is absolutely determined to get to the summit of the mountain, by

golly. Life wont be worth living if he doesnt.

In using the words, by golly, the speaker pretends to express determination,

thereby indicating Johns determination to achieve the summit. The speaker
probably is not saying that her own life wont be worth living if John doesnt fin-
ish the climb; she is again pretending to express what she takes John to feel, what
he does or might seriously express. The speaker connives with, plays along

16
See my discussion of Robinson Crusoe was more resourceful that Gulliver, Napoleon was
more pompous that Caesar, and Sherlock Holmes is more famous than any other detective, in
Walton (1990), p.406 and pp.413414. [Postulating unofficial games in which two or more works
of fiction serve together as props avoids having to treat statements like Crusoe was more resource-
ful than Gulliver and statements like Crusoe was stranded on a desert island in entirely different
ways, understanding the former but not the latter to require that the names Gulliver and Crusoe
have referents.]
[Mark Sainsbury (2010), pp.123124, takes a similar line. Anna Karenina was more intelligent
than Emma Bovary can be understood as agglomerating the two fictions, he says. Anna Karenina
and Madame Bovary together may make it fictional that this proposition is true.]
17
See Clark and Gerrig (1990), p.776. Apparently this is not an actual case, Clark and Gerrig say,
but it is characteristic of ones we have heard from children six to ten years old (p.775). See also
Clark and Gerrig (1984).
102 In Other Shoes

with, what she takes to be Johns attitudes and the ways he might seriously
express his attitudes.18

V. Existential Statements
To say that Neptune or Falstaff exists, Ihave suggested, is to say that attempts
to refer of a certain kind are successful. (Lets call these Neptunian or
Falstaffian referring attempts.) To say that Neptune or Falstaff doesnt exist
is to say that such referring attempts do not succeed. This sounds strange. We
seem to be talking about Neptune and Falstaff, not about kinds of attempted
reference (notwithstanding the fact that there is no Falstaff to talk about). After
all, the sentences we use consist in the name, Neptune or Falstaff, with a
predicate attached. The impression that Falstaff and Neptune are what we speak
of is explained by the fact that we are pretending that this is soor rather, we
pretend to refer successfully by means of the names and to attribute properties
to the referents.
In pretending to refer by means of the names, the speaker displays, shows,
demonstrates, the kind of attempted reference she is talking about. (In the case
of Neptune exists this is not mere pretense; the speaker actually makes an
attempt of the kind she is talking about. Her attempt succeeds if what she says
is true.)19 In attaching the predicate, exists or does not exist, she declares the
kind of attempted reference indicated by the use of the name to be successful or
unsuccessful; she avows or disavows attempts to refer of that kind. We can think
of the existential claims like this. To say Neptune exists is tosay:
Neptune:That was successful.
To say Falstaff doesnt exist is tosay:
Falstaff:That didnt work.
In both cases the demonstrative that refers to the kind of attempted reference
illustrated by the utterance of the name.
Other predicates besides exists and does not exist can serve the purposes
of avowal and disavowal. Is real and is actual are used to attribute success to
referring attempts. Predicates that may be used to say that referring attempts fail
include:is fictitious, is a (merely) fictional character, is a mythical beast,

18
This is an instance of what some literary scholars call free indirect discourse. See Walton
(1990), p.375 ff., especially pp.379380.
19
Isay more about how producing or indicating an instance of a kind serves to specify the kind
in Walton (1990), pp.425427.
 Ex istence a s Metaphor ? 103

is feigned, is counterfeit, is a phantom (limb), is a (mere) toy, is stuffed,

is a product of an overactive imagination, is a trick of light (Evans), is an
artifact of the experimental setup or ... an artifact of the researchers biases,
is a cultural artifact, is a failed posit, was invented by Walt Disney, and is an
empty concept. Some of these predicates have other uses as well. Some actual
things are toys; to call a childs top or a Barbie doll a toy is simply to point out
that it belongs to this category. But to say that Barbie is a (mere) toy (or only a
doll), may be to say that a certain kind of referring attempt (attempts to refer to
a person) fail. Walt Disney did invent some things, certain animation techniques,
for instance, and one can point out that a certain technique was invented by him.
Committees, governments, and laws are cultural artifacts, and can be described
as such.20 But to say that Donald Duck was invented by Walt Disney or that
he is a cultural artifact is probably to say that there is no such thing, i.e. that
Donald Duck-ish referring attempts fail. If Donald Duck is anything he is a duck
(a talking duck); not an invention or a cultural artifact. Some concepts may be
empty; the concept of Donald Duck (if there is such a thing) is one of them. But
Donald Duck himself is not a concept.
In pretending to refer by means of a name such as Falstaff or Santa Claus
or Oedipus, speakers play along with, connive with, the pretense of the rel-
evant fiction or myth. To attach the predicate, does not exist, or is a fictional
character, or is a mythical beast, is to continue the pretense further, pretend-
ing to attribute a property to the thing supposedly referred to, even while one
actually declares the reference unsuccessful. This is a little bit like mom playing
along with the bicycle game, pretending to advise Johnny not to ride the fence,
while actually bringing him back to reality, warning about the dangerous traffic.
So it is part of the implied, unofficial fiction that some things exist and some
things do notrather, that exist(s) expresses a property which some things
possess and others lack. Also, the fiction has it that, among nonexistent things,
some are mythical, others are imaginary, fictional, etc. and that some bears
are made of flesh and blood, others are stuffed, and still others are painted or
sculpted or drawn with a very fine camel-hair brush. In El Idioma de John
Wilkins, Jorge Luis Borges makes something like this unofficial game official. It
is said that according to a certain Chinese encyclopedia animals are divided
into the following categories: (a) belonging to the Emperor, (b) embalmed,
(c)tame, (d)sucking pigs, (e)sirens, (f)fabulous, (g)stray dogs, (h)included
in the present classification, (i)frenzied, (j)innumerable, (k)drawn with a very fine
camel-hair brush, (1)etcetera, (m)having just broken the water pitcher, (n)that

20
Whether there are such things as techniques, committees, governments, laws, concepts, etc., is
not at issue here. Iwrite as though there are, confident that in doing so Iam saying something under-
standable, whether or not it is something about such entities. See Yablo (2000).
104 In Other Shoes

from a long way off look like flies.21 Ernst Gombrich can be understood to be recog-
nizing a similar unofficial game when he speaks of a snowman as a member of the
species man, subspecies snowman, and when he suggests that the museum turns
images into art by establishing [a]new category, a new principle of classification
that creates a different mental set. Take any object from a museum, say Riccios Box
in the Shape of a Crab. ... On the desk ... this object would belong to the species
crab, subspecies bronze crab.22
It is fictional that one speaks truly when one says, N is mythical, if (roughly)
N-ish attempts to refer go back to a myth. This is the real world circumstance that
makes this fictional. It is fictional that Ispeak truly in saying Pooh is a stuffed bear
if the indicated referring attempts go back to a stuffed toy. It is fictional that Ispeak
truly in saying N doesnt exist if N-ish attempts to refer fail, for whatever reason.

VI. Comparisons with Metaphor

Existence claims are like many metaphorical statements in that the speaker
asserts what she does by engaging in pretense. She pretends to make an asser-
tion of a particular kind, thereby actually asserting something of a different kind.
She actually asserts that certain real world circumstances obtain, viz. those that
would make it fictional in the implied make-believe that she speaks truly.
As in the case of some metaphors, there is no proposition which one pretends
to assert, in making an existence claim. When Isay Neptune exists or Falstaff
doesnt exist, Ipretend to assert a proposition having the property of existence
(as well as Neptune or Falstaff) as a constituent. But there is no such property,
Iclaimed in Walton (1990), and hence no such proposition. There is, however,
the proposition that in speaking as I do I speak the truth. This proposition is
fictional if Neptunian referring attempts succeed and Falstaffian ones fail, i.e. if
it is fictional that Neptune (as used by the speaker on that occasion) refers to
something which possesses a property denoted by exists, and Falstaff refers
to something which lacks it. What Iactually assert is that Neptunian referring
attempts do succeed, or that Falstaffian ones fail.
Metaphorical uses of language are usually or always guidedsometimes
in mysterious ways, sometimes in obvious onesby the literal meanings of
the words used.23 It is because we know and understand the literal meaning of
being infected by a virus that we are able to use and understand the meta-
phorical description of a computer as infected by a virus. There is a semantic

21
Obras Completas, p.708. Thanks to David Hills.
22
See Gombrich (1961), p.100 and p.114.
23
See, for example, Goodman (1968) and Hills (1997).
 Ex istence a s Metaphor ? 105

connection between the content of the speakers pretended assertion and what,
in so pretending, she actually asserts. This is so even when the content of what
the speaker pretends to say consists in something less than a proposition. If there
is no such property as being saddled with responsibility, and no proposition
which Ipretend to assert when Idescribe someone as saddled with responsibil-
ity, Iam nevertheless pretending to assert a proposition having the property of
being saddled and also the property of responsibility as constituents; Ipretend
that there is such a proposition and pretend to be asserting it. Ipretend to be
indicating a manner in which the person is saddled, and to be describing a rela-
tion that she bears to responsibility. It is by recognizing this pretense that hearers
understand what Iam actually asserting.
Is anything like this true in the case of existence claims? Does exist (s) have a
literal meaning which guides its use in characterizing referring attempts? Ones
first reaction is likely to be this:the literal meaning of exist(s) just is its possess-
ing the function of characterizing attempts to refer. Could there be another literal
meaning that guides this one? There could be, but Idoubt that there is.
Fred Kroon has argued that, since Ideny that there are genuine properties
which phrases like is a (merely) fictional character, is a failed posit, and is
a mythical beast, fictionally stand for, Icannot make sense of the connection
between what a sentence says in the scope of the pretense and what speakers
assert when uttering the sentence, Icannot, he thinks, explain how we succeed
in ascribing properties to attempted acts of reference through only pretending
to ascribe properties to the referents of Falstaff and Santa Claus.24 He offers
a useful example (which Imodify in order to avoid possible confusions) to illus-
trate his worry.
Consider a game, inspired by Lewis Carrolls poem about the Jabberwock,
involving an assortment of toy animals. Lets say that, as in the poem, it is fictional
that some animals are toves and that others are sneetches, i.e. it is fictional in
the game that each of these words picks out a property that some animals pos-
sess. But, as in the poem, there are no actual properties which, fictionally, they
pick out. Suppose we adopt the arbitrary convention that animal figures made
of wood represent toves, and that figures made of plastic represent sneetches.
Now we know that Smokey is correctly called a tove, in the fiction, if Smokey is
represented by a wooden figure; in that case to say Smokey is a tove is, fiction-
ally, to speak truly. There still is no property which the speaker is pretending to
attribute in saying this, however; we merely pretend that there is one.
According to Kroon (I paraphrase his conclusion to accommodate my modi-
fication of his example), the following ought to be clear:if we ... say Smokey is
a tove there is no senseat any rate, no interesting sensein which we thereby

24
See Kroon (1996), pp.179180.
106 In Other Shoes

declare that a particular toy is made of wood. Ianswer that this could very well
be what the speaker is asserting. (It is too much to ask that the sense in which
this is true be an interesting one; a mundane sense will do.) Suppose the game is
employed in a prop oriented manner. The plastic animal figures are suspected of
being poisonous and must be removed from the nursery. Itell you that Smokey
is a tove, meaning that the figure in question is made of wood and can safely
remain. There doesnt have to be a semantic connection between what a speaker
pretends to assert and what she actually asserts by engaging in the pretense.
In this case there is no content at all to the speakers pretended assertion. The
speaker exploits a stipulated rule of make-believe governing the words used.
Singular existential statements may be something like this. There may be sim-
ply a brute convention to the effect that to attach the predicate exists or does
not exist to a subject is to declare the indicated kind of referring attempt suc-
cessful or unsuccessful. There does not have to be a property expressed by the
predicate, a property one is pretending to attribute, which helps to determine
what one actually asserts. A grammatical form is being used in a nonstandard
way; the speaker is pretending to say something about the (pretended) referent
of the subject expression. But we dont have to assume that she is using exist(s)
or does not exist in a nonstandard or alternative or secondary way semanti-
cally, that she is pretending to use it with one meaning as a way of actually saying
something different. The only meaning of the predicate that is involved may be
the one consisting in its suitability for characterizing referring attempts as suc-
cessful or unsuccessful.
What about predicates like is a (merely) fictional character, is a failed
posit, is a trick of light, is a mythical beast? As Kroon points out, these
phrases serve not just to declare that the kind of attempted referrings indicated
by the subject expression fail; they say something about how or why they fail.
To describe something as mythical is different from describing it as imaginary,
or as a failed posit. These predicates are composite, and it is clear that their com-
ponents (fictional, mythical, posit) have meanings which help to guide the
use of the whole, which help to determine what is being said about the kinds of
referring attempts the speaker is talking about. It does not follow that the predi-
cate as a whole expresses a property which the speaker is pretending to attri-
bute to something referred to by the subject expression. (We might allow that
the predicate as a whole possesses a meaning, without insisting that its meaning
consists in its expression of a property.)
Consider idiomatic expressions suchas:

Ill fix your wagon,

He chewed her out,
She has a checkered past,
 Ex istence a s Metaphor ? 107

I am not prepared for the lecture; Ill have to wing it,

She made no bones about it,
He went bananas,
She chickened out,
He made a bee line for the refreshments,
She bent over backwards to be fair,
She blew her stack,
He is in the swim of things,
The computer went haywire,
He would have no truck with that proposal,
This behavior is beyond the pale,
She was beside herself with anger,
She stole his thunder.

These are familiar ways of saying things that, otherwise, have little, if any, salient
connection with the standard literal meanings of the words used.25 One would
be hard pressed to predict the idiomatic uses of these sentences simply from
the literal meanings and the context of utterance. Speakers must simply learn
that pretending to assert that someone fixed someones wagon, or pretending to
assert this using these words, is, in English, a way of saying that the first person
did the second one in, and that to make no bones about something is to be
forthright about it. (This may be a reason not to classify such idioms as meta-
phors.) The literal meanings of the words probably do provide some guidance,
however, in at least some of these cases. Knowing what checkered, chew,
wing, cool, and thunder, mean in other contexts surely helps one to learn
and to remember the idiomatic uses of checkered past, chew him out, wing
it, lose her cool, and stole his thunder. But this is no reason to assume that
the complex predicate as a whole expresses a property when its constituents are
understood in the standard literal ways. Some of these predicates do, but others
arguably do not. What would it be for a person, literally, to have a checkered past,
or to wing a lecture, or to lose her cool, or to chew someone out?
That the standard literal meanings of the words in many of these phrases are
operative in their idiomatic uses is clear from the fact that we can understand
certain transformations of them on first hearing, for example:

Make only a few bones about it,

I wouldnt touch it with an 11 foot pole,
Ill repair your cart,

25
Many of my observations about idioms here are discussed much more thoroughly in Nunberg
etal. (1994), pp.491538, and also Davies (1983).
108 In Other Shoes

He has a wasp in his fedora,

She found her cool,
This gadget boasts a deafening cacophony of bells and whistles,
[Geoffrey] Fieger has this tree-trunk of anger on his shoulder, which
resonates among people who feel disenfranchised by the system.26

Understanding the above requires recovering the specific words of the relevant
familiar idioms. In the last case, one must call to mind the phrase, a chip on his
shoulder. But what enables us to recover these words is a semantic link between,
for example, the literal meaning of chip and the literal meaning of tree-trunk.
It might be that exist(s) expresses a property, in some or all of its uses, even
though the supposition that it does is not needed to account for its use in char-
acterizing referring attempts. And it is possible that the sense it has in expressing
this property guides its use in characterizing referring attempts in the way the
literal meanings of metaphors guide their metaphorical uses. Gareth Evans and
(I believe) Fred Kroon hold that existence is a universal property, one which
everything possesses by default. Iclaimed otherwise in Walton (1990). (This
is not an issue for me to stomp my foot about, however; none of the central
features of my account of singular existential statements depends on it.) Those
who think there are nonexistent objects, of course, take it to be a discriminating
property, one that some things possess and some things lack.
If existence is a property necessarily possessed by everything, it is not the
only one; self-identity is another. To say that existence is necessarily universal
is not yet to specify what property it is. But suppose that existence is a universal
propertynever mind which one. How might it figure in a pretense account of
singular existential statements?
The account would have it that in saying Falstaff does not exist the speaker
pretends to describe the referent of the name as lacking this property, existence,
and that in so pretending she asserts that Falstaff-ish referring attempts fail. The
relevant pretense is one in which exist(s) expresses a discriminating property.
So we will be pretending, of what we know to be a universal property, that it is
actually a discriminating one; we will be pretending that there are things that
lack it.
Something like this picture of the working of singular existential statements
may fit some uses of other predicates. To assert:
Falstaff is not self-identical. (or Falstaff is not himself)
could be a metaphorical way of saying that Falstaff doesnt exist, that Falstaff-ish
referring attempts fail. The speaker pretends to be claiming that an entity which

26
NewYork Times, 3 August 1998.
 Ex istence a s Metaphor ? 109

she refers to as Falstaff lacks the property of self-identity; she pretends that
some things are self-identical and some things are not. By engaging in this pre-
tense, she genuinely asserts that Falstaff-ish referring attempts are unsuccessful.
(The connection here is this:since, as we all know, nothing lacks self-identity,
attempts to refer to things lacking self-identity are doomed to failure.) Acorre-
sponding metaphorical way of saying that Neptune exists is to assert:
Neptune is self-identical.
Are Falstaff doesnt exist and Neptune exists like these metaphorical claims?
Iam not sure, but Iam sceptical.
A rather feeble reason for scepticism is the fact that Falstaff doesnt exist
doesnt feel like Falstaff is not self-identical. Falstaff doesnt exist doesnt
sound to my ear like an indirect, or figurative, or metaphorical way of saying
what is said. This impression might be explained by the hypothesis that Falstaff
doesnt exist is a dead metaphor, one that is so familiar that, like chair leg,
mouth of a river, and keep your eyes peeled, it has lost its aura of metaphori-
cality. But then Iwonder if the metaphor was ever alive. Maybe it was dead on
arrival. If so, was it ever a metaphor at all? (I dont rule out this possibility.)
A somewhat less feeble reason for scepticism is this: I can recapture the
metaphorical character of chair leg, mouth of the river, and keep your eyes
peeled, by attending to their obvious literal meanings. (In the case of mouth
of the river Isuppose that this involves thinking of the opening of the river as
portraying or representing a mouth. It is harder to say what it amounts to in the
other cases.) But it isnt at all clear that Ican recaptureor capture for the first
timea way of hearing existence claims as metaphors. Presumably reviving the
metaphor would require keeping in mind supposedly literal uses of exists, cases
in which it is used to express the universal property of existence. But are there
any? Straightforward singular negative existentials are not such. Positive singular
existentials, which serve the supposedly metaphorical function of characterizing
referring attempts, could conceivably be functioning at the same time to attribute
the universal property of existence to the subject, but it isnt at all clear that they
are. Gareth Evans seems to think that exists has this literal sense in quantifi-
cational and modal contexts. Iam not convinced.27 In any case, this literal sense

27
See Evans (1982), pp.345348. Ithink that exist(s) in such contexts can be understood con-
vincingly without supposing that it has this sense. One who says This might not have existed, for
instance, utilizes the pretense that exists expresses a (discriminating) property and that there is a
proposition to the effect that the object she refers to as this possesses this property. In saying what
she does she pretends to be claiming that this proposition is possibly false. What makes it fictional
that she speaks truly (if it is fictional that she does), and what she actually asserts to be the case, is
something like the fact that it is possible that referring attempts of the kind exemplified by her utter-
ance of this should have been successful.
110 In Other Shoes

(if it exists) doesnt seem to be salient in ordinary existence claims. And Ihave
argued that we dont need to suppose that it is operative in order to understand
how by pretending to predicate exists or doesnt exist to (pretended) things,
we manage to characterize referring attempts as successful or unsuccessful.
By contrast, x is identical to y has obvious uses when the success or failure
of referring attempts is not in question. And so does x is self-identical, at least
in philosophical contexts. And it is clear that the sense identity statements have
in these cases is being exploited when Isay Falstaff is not self-identical, mean-
ing, metaphorically, that the guy isnt.28
I am not eager to offer a definition of metaphor or to pronounce on the
limits of the category. But it seems reasonable to regard the dependence of
what is asserted on distinct literal meanings of the words used, and the guid-
ing function of the content of what the speaker (merely) pretends to sayas
a necessary condition for an (assertive) utterance to be metaphorical. If this is
a necessary condition, it looks as though we can escape the apparently bizarre
thesis that exist(s), as it occurs in garden variety singular existential statements,
is metaphorical.
Nevertheless, the fact that making an existential claim, like speaking meta-
phorically, is to engage in make-believe, the fact that the speaker utters her words
in pretense, means that we do not have to take her apparent ontological com-
mitments seriously. To claim that Falstaff doesnt exist is not to claim, seriously,
that there is something, Falstaff, which lacks existence, any more than asserting
what one does in saying We are all in the same boat is to claim that there really
is a boat that we are all in, or the use of phrases like raising hackles, on pins
and needles, in seventh heaven, beyond the pale, and hold your horses
commits the speaker to the existence of (things called) hackles, pins and needles,
seventh heaven, the pale, or horses.

VII. Appendix:Logical Form

(ReplytoPetervanInwagen)
Paraphrases must preserve the logical form of the statements they paraphrase
and the entailment relations obtaining among them.29 Peter van Inwagen has

28
Is identical to, understood literally, may not be what it seems to bea predicate expressing
a relational property which everything bears to itself and to nothing else. See Crimmins (1998) and
Kroon (2001). Nevertheless, the literal sense of Falstaff is not self-identical, whatever it amounts to,
guides the metaphorical use of this sentence in asserting that Falstaff does not exist.
29
Logical form reprinted by permission of the publisher from Mimesis as Make Believe: On
the Foundations of the Representational Arts by Kendall L. Walton, pp. 416419, Cambridge,
Mass.:Harvard University Press, Copyright 1990 by the President and Fellows of Harvard.
 Ex istence a s Metaphor ? 111

expressed scepticism about whether paraphrases like those Ipropose for state-
ments concerning fiction satisfy this requirement.30 Van Inwagen observesthat

(1) There is a fictional character who, for every novel, either appears in that
novel or is a model for a character who does.

appears to have a certain complex quantificational structure, a structure that

validates the inference from (1)to:

(2) If no character appears in every novel, then some character is modeled on

another character.

But none of the paraphrases of (1) and (2) that I recommended in Walton
(1990) has these logical forms. On our primary model suggested there, both
will be paraphrased by something of thisform:

To engage in pretense of kind K is fictionally to speak truly in a game of

such-and-such a sort.

(the relevant kind of pretense being different in the two cases). An alternative
is to regard (1)and (2)as attesting to circumstances which, if present, would
make it fictional of one who pretends in the relevant manner that he speaks truly
(circumstances concerning the corpus of extant novels). Obviously we cannot
expect any such paraphrases to mirror the quantificational structures exhibited
by (1)and (2). Paraphrases of neither sort show (1)to entail (2)by virtue of
logical form.
Not only can we live with this result; we will thrive on it. Recall, first, that
what our paraphrases seek to capture is what speakers say in uttering the sen-
tences cited, not what the sentences themselves mean or what propositions
they express, if any. What speakers say simply does not have the logical forms
indicated by the sentences they use. To assume otherwise would be question
begging. But we do need to explain why people use sentences displaying logical
forms different from those of what they assert in uttering them. And it certainly
seems as though what is said by means of (2)follows deductively from what is
said by means of (1); it is not easy to envisage accepting (1)while dissenting
from (2). If the quantificational structures of what is said do not guarantee this
entailment, what does? If the entailment does not hold, why does it seem to?
Explanations are easily provided, but first lets say a little more about what speak-
ers might assert by means of (1)and (2).

30
See Van Inwagen (1985).
112 In Other Shoes

Utterances of (1)and (2)can be understood to involve unofficial games of

a rather ordinary sort, ones with some approximation of the following features:

(a) All novels are props in them and most of what is fictional in any novel is
fictional in them; the unofficial games combine the games authorized for
each individual novel in a way familiar from Walton (1990), section 10:4.
(b) It is fictional in these combination games that the universe is divided
into realms corresponding to the various novels.31 To say that a charac-
ter appears in a certain novel is, fictionally, to locate a person in a certain
realm.
(c) To write a novel of a certain sort is to make it fictional of oneself, in games
of the implied sort, that one creates people (characters) and endows
them with certain properties. (Compare: Jane Austen created Emma
Woodhouse.)
(d) When, as we say, an author models a character on some preexisting charac-
ter, it is fictional that he creates someone to be like some other person, that
he makes someone in the image of someone else. (In speaking of charac-
ters rather than people the speaker betrays his pretense, but this does not
affect the content of the assertion.)32

It is fictional in unofficial games of the kind implied, no doubt, that (1)and

(2)the sentences themselvesexpress propositions that have the logical
structures they appear to have, ones by virtue of which (1) entails (2). And
it is fictional of Ellen, if she utters (1)assertively while participating in such a
game, that what she asserts entails, by virtue of logical form alone, what would
be asserted by (2). This partly explains the impression that what Ellen actually
asserts by means of (1)thus entails what would be asserted by (2).
What does Ellen genuinely assert by means of (1)? In the simplest and pri-
mary case (unlikely though it may be) she is participating in an unofficial game
of the sort described and asserting that to pretend as she does is fictionally to
speak truly. Her pretense is a kind of pretending to assert something of the form
displayed by that sentence; it is fictional that she asserts a proposition of that
form. So naturally she uses that sentence, even though what she actually asserts
about this kind of pretense has a very different form.
If Ellen speaks of this kind of pretense without engaging in it, she still
refers to it. If, rather than speaking about it she is pointing out the presence of

31
This is not an uncommon feature of unofficial games that combine other games, though there
may be very little to be said about what, fictionally, realms are or what it is, fictionally, for things to
belong to the same or different ones. See Walton (1990), section 11.1.
32
See Walton (1990), section 11.1.
 Ex istence a s Metaphor ? 113

circumstances by virtue of which to so pretend is fictionally to speak truly, it is by

indicating the kind of pretense that she calls attention to the circumstances she
claims to obtain. In either case the sentence she uses, (1), displaying the logical
form it does, suits her purpose, since to pretend in the manner she refers to or
indicates is fictionally to assert something of that form.
But it is only fictional, not true, that (1) has a quantificational structure
such as to entail (2). And it is at most fictional, not true, that what is asserted
by means of (1)entails by virtue of logical form alone what is asserted by (2).
Nevertheless, one may be speaking truly in asserting

(3) (1)entails (2)by virtue of logical form alone.

Assertions of (3)can themselves be understood as we understand other state-

ments concerning fiction. The speaker, in the primary case, indicates a kind of
pretensethe pretense of asserting that (1)entails (2)by virtue of logical form
aloneand claims that to so pretend in an unofficial game of an implied sort
(the kind implied by (1)and (2)) is fictionally to speak truly. This claim is true,
even though (1)does not entail (2)by virtue of logical form alone; (3)taken
literally is false. The moral of the story, again, is that we must take care to dis-
tinguish between fictionality and truth, between what is the case and what is
merely pretended to be the case.
The likelihood of confusion is enhanced by the fact that what is asserted by
(1)probably does entail what is asserted by (2), though not by virtue of logical
form alone. The principles constituting the implied games of make-believe are
likely to be such that it cannot be fictional that to assert (1)is to speak truly
unless it is fictional that to assert (2)is to speak truly. The unofficial games would
have to be rather exotic logical fantasies for this not to be so. And the principles
are likely to be such that the circumstances required to make it fictional that
one speaks truly in asserting (1)are ones that make it fictional that one speaks
truly in asserting (2). No wonder it is difficult to conceive of (1)being true and
(2)false.
 Postscript to Chapter7:Existence as Metaphor?

FictionalismDavid Lewis and Simon Blackburn

Fictionalism (what some call hermeneutic fictionalism) is a strategy or family of
strategies for making sense of discourse (ostensibly) about ontologically dubi-
ous entities of one kind or anothere.g., fictional characters, or numbers, or
moral properties, or colors (or about otherwise questionable states of affairs).1
My treatment of talk about fictional entities and its extension to existence claims,
based on the notion of prop oriented make-believe, is fictionalist.2
David Lewis (2005) takes Simon Blackburns quasi-realist account of moral
discourse to be a variety of moral fictionalism (319). Blackburn (2005) dis-
agrees, claiming not only that his theory is not fictionalist, but that fictionalism
suffers from grave defects which his quasi-realism avoids. Although Blackburn
is interested mainly in moral or evaluative language, his primary worry, if valid,
applies to fictionalist treatments of other realms of discourse as well. But fiction-
alists neednt fret. Blackburns problem is a problem only for fictionalisms of a
very inadequate variety, albeit one suggested by Lewiss own account of fiction.
The kind of fictionalism Ipoint to in Existence as Metaphor? and elsewhere
bypasses it cleanly and completely. (I wont try to decide whether Blackburns
quasi-realism should count as a fictionalist theory, or whether, fictionalist or not,
it is viable or better than its rivals.)
There are no such things as fictional characters, or numbers, or moral proper-
ties, or colors, some say, not really; they are not constituents of the universe. Yet
we seem frequently to refer to them (or in other ways imply that they exist), and
much of what we say about them seems to be true. Fictionalists propose some-
thing like the following:That there are things of the kind in question is a fiction,

1
Cf. Eklund (2011) and the essays in Kalderon (2005).
2
Cf. Walton (1990, 1993)as well as Existence as Metaphor?

114
 Ex istence a s Metaphor ? 115

a fiction analogous in some ways to those of novels and stories. We pretendwe

only pretendto refer to them and to say (true or false) things about them. But
the fiction serves a purpose; there is a point in pretending as we do. Doing so may,
for instance, be a way of asserting something other than what the speakers words
literally express.
John Locke, Blackburn says, might be understood as holding that we talk as if
there are colours, although there are none really (324). Afictionalist about colors,
according to Blackburn, would take it to be true in a fiction, fictional, that there are
colors, that things are colored,3 and would claim that in saying roses are red or
canaries are yellow we are pretending to assert that roses really are red or canar-
ies really are yellow. But what would it be for there actually to be colors, Blackburn
asks rhetorically? How does our world differ from some other [possible] world
in which there are colours, really? (324). He points out that Lewiss account of
truth in fiction requires us to understand the idea of the fiction being told as known
fact (325). Is it true in the colour fiction, that canaries are yellow? To answer,
Blackburn claims, we need to understand what it is for canaries are yellow to be
fact, and to be known as such, although in our world it is not. This is a tall order. If it
is neither a fact nor known to be such in our world, what is different in those worlds
in which it is? (325). Likewise for moral fictionalism:I say that it is bad to neglect
the needs of children. According to this version of moral fictionalism, Iam taken to
be saying that in the moral fiction, it is bad to neglect the needs of children, although
it is not bad really. So:what would it be for it to be fact and to be known and told
as such, that it is bad to neglect the needs of children? It is not so in this world, evi-
dently, so what is different about worlds in which it is? (325). No answer is forth-
coming, Blackburn implies; fictionalist theories simply cannot get off the ground. 4
The fictionalist I envisage has a ready response. There is no need to sup-
pose that the speakers words (Canaries are yellow or Neglecting children is
wrong), taken literally, express a proposition, one that is true in some possible
world.5 Fictionalism requires only that it be fictional in the implied game that

3
Blackburn has Locke asking whether there are colors. The question is better put as whether
objects are colored. Anominalist whose ontology excludes properties in general will deny that there
are colors, color properties, but might allow that objects are colored. She might accept that Canaries
are yellow, but not that Yellow is the color of canaries; she need not understand the state of affairs, canar-
ies being yellow, to have as a constituent the property of yellowness. Likewise, nominalists might
agree that neglecting children is wrong, while denying that being wrong is a property possessed by
acts of child neglect.
4
Blackburn (2005) eventually allows that the premise he has relied on, that when we talk of fic-
tions, we know the contrast with fact, is doubtful (335), perhaps explaining the implication in the
above quotation that his is only one version of moral fictionalism. He seems to be leaving room for
a more radical variant than Ihave suggested, however, one that departs more fundamentally from
ordinary practices of storytelling and perhaps should not be called fictionalism at all.
5
Cf. Existence as Metaphor? pp.9697 and VI; also Walton (1990), pp.391, 396.
116 In Other Shoes

these words express a proposition, whether or not they actually do, and the fic-
tionalist can have the speaker pretending to be asserting a proposition even if
there is no actual proposition that she pretends to assert.6 Thus can fictionalism
about colors or moral discourse get off the ground.
I am not endorsing either of these fictionalisms, just claiming that Blackburns
worry is no reason to reject them. My agenda is mainly, of course, to protect my
own fictionalist account of discourse about fictional entities from similar objec-
tions. Gulliver was captured by the Lilliputians and Hamlet is a fictional char-
acter created by Shakespeare taken literally, like many metaphorical sentences,
do not express propositions, propositions that could be fictional or that speakers
could pretend to assert as true.7 Nevertheless it may be fictional in an implied
(official or unofficial) game of make-believe that the speaker speaks truly.
The reader will recall also, from Existence as Metaphor? (VI) that even if
the sentence a speaker utters in introducing a prop oriented game of make-believe
does not express a proposition, the semantic values of its individual words are
bound to play a role in determining what game of make-believe is implied, what
principles of generation are operative. (Syntactic features of the utterance will
contribute also, and of course the context is crucial.)
The fictions fictionalists postulate are supposed to be useful. Iclaimed that in
the fictional entity case the speaker actually asserts somethingnot a proposi-
tion her words express taken literally, but something that is true just in case it is
fictional in the relevant game that she speaks truly. Iwould expect fictionalists
about colors to say the same. Amoral fictionalist might also, but it is open to
her to locate the usefulness elsewhere, to understand speakers to be expressing
attitudes or recommending behavior, for instance, when they pretend to assert
as they do. The fictionalist might understand these actions, whatever they are,
to be appropriate or proper just in case it is fictional in the implied game that she
speaks truly.

References
Barth, J. 1967. The End of the Road. NewYork:Doubleday and Company.
Blackburn, S. 2005. Quasi-Realism no Fictionalism. In Mark Eli Kalderon (Ed.), Fictionalism in
Metaphysics (pp. 322338). Oxford:Oxford University Press. All page references are to this
article.
Clark, H., and R. Gerrig. 1990. Quotations as Demonstrations. Language 66:764805.

6
The fictionalist neednt claim that the speaker actually engages in pretense. Her purposes may be
well served simply by introducing, calling attention to, a certain prop oriented game of make-believe,
without participating in it. Walton (1990), pp.395396, and Walton (1993), pp.4243, 46, 49.
7
Walton (1990), pp.391, 396.
 Ex istence a s Metaphor ? 117

Crimmins, M. 1998. Hesperus and Phosphorus: Sense, Pretense, and Reference. Philosophical
Review 107:148.
Davidson, D. 1984. What Metaphors Mean. In Inquiries into Truth and Interpretation (Chap.17).
Oxford:Oxford University Press.
Davies, M. 1983. Idiom and Metaphor. Proceedings of the Aristotelian Society 83:6783.
Eklund, M. 2011. Fictionalism, The Stanford Encyclopedia of Philosophy (Fall 2011 Edition),
Edward N. Zalta (ed.), URL = <http://plato.stanford.edu/archives/fall2011/entries/
fictionalism/>.
Evans, G. 1982. The Varieties of Reference. Oxford:Oxford University Press.
Gombrich, E. 1961. Art and Illusion. NewYork:Pantheon Books.
Goodman, N. 1968. Languages of Art. Indianapolis:Bobbs-Merrill.
Hills, D. 1997. Aptness and Truth in Verbal Metaphor. Philosophical Topics 25:11753.
Kalderon, M.E. 2005. Fictionalist Approaches to Metaphysics. Oxford:Oxford University Press.
Kroon, F. 1996. Characterizing Nonexistents. Grazer Philosophische Studien 51:16393.
Kroon, F. 2001. Fictionalism and the Informativeness of Identity. Philosophical Studies
106:197225.
Lewis, D. 2005. Quasi Realism is Fictionalism. In Mark Eli Kalderon (Ed.), Fictionalism in
Metaphysics. Oxford:Oxford University Press, 314321.
Nunberg, G., I. Sag, and T. Wasow. 1994. Idioms. Language 70:491538.
Sainsbury, M. 2010. Fiction and Fictionalism. London:Routledge.
Sommer, E., and D. Weiss. 1996. Metaphors Dictionary. Detroit:Visible Ink Press.
van Inwagen, P. 1985. Pretence and Paraphrase. In Peter J.McCormick (ed.), The Reasons of Art/
LArt a ses raisons. Ottawa:University of Ottawa Press, 414422.
Walton, K. 1990. Mimesis as Make Believe. Cambridge, MA:Harvard University Press.
Walton, K. 1993. Metaphor and Prop Oriented Make-Believe. The European Journal of Philosophy
1:3957 and reprinted in this volume.
Yablo, S. 2000. A Paradox of Existence. In A. Everett and T. Hofweber (Eds.), Empty Names,
Fiction, and the Puzzles of Non-Existence. Stanford, CA:CSLI Press, 197228.
 8

Projectivism, Empathy, and

MusicalTension
Beauty is an emotional element, a pleasure of ours, which nevertheless
we regard as a quality of things, ... [in short] pleasure objectified.
George Santayana, The Sense of Beauty, 1896

This is a challenging idea, to say the least.1 Nelson Goodman warned of the
temptation ... to indulge in mumbo-jumbo about objectification in attempting
to explain the nature of aesthetic experiences and aesthetic objects, and derided
the sophisticated theory that what counts is ... pleasure objectified, pleasure
read into the object as a property thereof. What can this mean, he asked,
apart from images of some grotesque process of transfusion?2
I will not exactly aim to make sense of Santayanas claim; Iwont have anything
to say about the concept of beauty. But Ishall propose a way of understanding
some aesthetic properties that bears some analogy to what Santayana says about
beauty. The properties of tension and relaxation (or release) in music will be my
main examples, although it will be obvious that if what Isay about musical ten-
sion and relaxation is right, it applies to some other aesthetic properties as well.
I shall also consider how we go about ascertaining the psychological states of
other people, how we discover what others are thinking and feeling. An attrac-
tive recent theory has it that, in many instances, we simulate in ourselves the psy-
chological processes occurring in the other person and attribute to her what we
find ourselves thinking or feeling (actually or in imagination) as a result. Ionce
considered myself a good friend of the simulation theory, but have since become

1
This paper grew out of a talk presented at the University of Virginia in 1995, and on several other
occasions, including the conference on Self, Mind and Knowledge in honor of Sydney Shoemaker at
Cornell University. Thanks to the various audiences, and to Marion Guck, Fred Maus, Andrew Mead,
Susan Pratt Walton, and especially David Hills. Thanks also to Sydney Shoemaker, whose perceptive
and patient mentoring many years ago has guided my philosophical reflections ever since.
2
Nelson Goodman, Languages of Art (Indianapolis:Bobbs-Merrill, 1968), 243.

118
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 119

partly disillusioned.3 Ibelieve that we engage in mental simulation, as simula-

tion theorists usually characterize it, far less often than they suppose. But Iam
not eager to endorse the simulation theorys main rival, the theory, according
to which we come to understand others by applying, implicitly, a folk psycho-
logical theory. Too much attention has been paid to this rivalry. The smoke of
the battle has obscured important features of what Iwill call empathetic under-
standing of others. Empathy may sometimes amount to simulation as it is tradi-
tionally understood, but sometimes it does not. And it may or may not involve
the utilization of a folk psychological theory.
These two topics are linked. Listeners experiences of tension and relax-
ation in music bear some analogy to empathetic understanding of other people.
Acareful look at empathetic understanding will illuminate the experiences of
musical tension and relaxation, and will help to clarify the nature of these prop-
erties. The analogy would appear to support the idea, put forward by various
authors in various forms, that listeners think of or experience music as though it
is a person, or as containing or presenting one or many (fictitious) personae.4 My
conclusions, however, will tend mostly in the opposite direction.

I. Musical Tension and Relaxation

Tension and release or relaxation are, by virtually all accounts, of absolutely fun-
damental importance in music. They are so pervasive that it is hardly necessary
to cite examples, although Iwill mention several as we go along. Roger Sessions
remarked that The principle of tension and relaxation is perhaps the most
important single principle of musical rhythm, and its bearing on all questions
of musical expression cannot be overestimated.5 Victor Zuckerkandl refers to
force rather than tension: In music there would hardly be anything left to
describe if force had to be excluded from the discussion. Force is as real as music

3
The present paper revises some of my claims in Spelunking, Simulation, and Slime:On Being
Moved by Fiction, in Emotion and the Arts, ed. Mette Hjort and Sue Laver (New York: Oxford
University Press, 1997), 3749 and reprinted in this volume. What I call participation in
make-believe is not, in general, a form of mental simulation, if mental simulation is understood as
process simulation (ibid.).
4
See, for example, Carolyn Abbate, Unsung Voices:Opera and Musical Narrative in the Nineteenth
Century (Princeton: Princeton University Press, 1991); Edward T. Cone, The Composers Voice
(Berkeley:University of California Press, 1974); Stephen Davies, Musical Meaning and Expression
(Ithaca, N.Y.:Cornell University Press, 1994), 36769; Jerrold Levinson, Musical Expressiveness,
in The Pleasures of Aesthetics (Ithaca, N.Y.: Cornell University Press, 1996), 90125; Fred Maus,
Music As Drama, Music Theory Spectrum 10 (1988):5673; Aaron Ridley, Music, Value and the
Passions (Ithaca, N.Y.:Cornell University Press, 1995).
5
Roger Sessions, Harmonic Practice (NewYork:Harcourt, Brace, and World, 1951), 84.
120 In Other Shoes

itself.6 Fred Lerdahl and Ray Jackendoff speak of the tensing and relaxing, inher-
ent in the motion of pitch-events, the incessant breathing in and out of music in
response to the juxtaposition of pitch and rhythmic factors.7
The musical scholars quoted above are concerned exclusively with Western tonal
music. What about music of other kinds? Progressions from relaxation to tension
and from tension to relaxation, the buildup of tension and its subsequent release or
resolution, may play a less significant role in some non-Western music, if they are
present at all, than they do in the music of Bach, Mozart, and Beethoven. But much
non-Western musicmuch Javanese gamelan music, for instancedoes present a
relatively steady state of tension, or a steady state of relaxation (not release), and
presenting these states is a centrally important function of the music.8 One Javanese
musician, however, describes progressions of tension and release in gamelan music
like those we find in Western tonal music:The melodic flow of a gongan [a phrase
or section of a piece ending in a stroke of the large gong] consists in an initial state-
ment, tension, and resolution.9
The sources of tension and relaxation, the musical characteristics respon-
sible for these qualities, are many and various. Some music theorists have
concentrated on harmonic parameters, on the tension (often) inherent in dis-
sonances, and the relaxation that attends their resolution into consonances.
But characteristics of melody, rhythm, meter, dynamics, tempo, texture, tim-
bre, and form are obviously important also.10 The unpredictable syncopations

6
Victor Zuckerkandl, Sound and Symbol:Music and the External World, trans. Willard R.Trask
(Princeton:Princeton University Press, 1956), 372.
7
Fred Lerdahl and Ray Jackendoff, A Generative Theory of Tonal Music (Cambridge, Mass.:MIT
Press, 1983), 179, 285. In a psychological study, Carol Krumhansl asked subjects listening to a
Mozart sonata to indicate the amount of tension heard at each point throughout the piecepre-
supposing that there is, at every point in the music, a degree of tension or relaxation (Krumhansl, A
Perceptual Analysis of Mozarts Piano Sonata K.282:Segmentation, Tension, and Musical Ideas,
Music Perception 13 [3] [1996]:411).
8
Cf. A.Anderson Sutton, Variations in Central Javanese Gamelan Music:Dynamics of a Steady State
(De Kalb, III.: Northern Illinois University, Center for Southeast Asian Studies, 1993), esp. 37,
16465, 19495, 199201.
9
Sumarsam, Internet posting, June 9, 1998. He continues:
Even in kendhangan [drumming patterns], say the kendhangan kalih style of ladrang com-
position, you can feel the tension gradually rises, starting in the beginning of the third
kenongan [the third of four phrases in the ladrang form]; [the] drumming pattern toward
the last kempul [a medium sized gong which sounds at the middle of the fourth phrase] is
the peak of the tension, and then resolution toward GONG. Kethuk salahan [an especially
elaborate pattern played by a small gong, the kethuk], if it is played, helps to call the atten-
tion of the ensemble that the GONG is about to come. (Thanks to Susan Pratt Walton.)
10
Wallace Berry accepts the premise that no change distinguishing contiguous sound events can
be neutral with respect to intensity. Thus, pitch change, however slight, is suggestive of modification
in the degree of intensity, as are any and all changes in tonal reference, harmonic content, rhythmic
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 121

in Stravinskys Rite of Spring produce almost unbearable tension. Steady,

relentless, entirely predictable, driving rhythms, characteristic of Beethoven,
often do so as well.11 Rapidly repeating sixteenth note accompaniment figures
common in baroque string music may stir up a frenzy.12 The rising melodic
motion of the opening theme of Bartoks Music for Strings, Percussion and
Celesta involves a small but definite increase of tension, which subsides as the
melodic line descends. The pattern of tension and relaxation in subsequent
inversions of the theme is clearly not the same. Various musical devices for
creating tension are often used in combination, of course. Obvious examples
are everywhereincluding in Arnold Schoenbergs Verklaerte Nacht and in
Samuel Barbers Adagio for Strings.
It is one thing to identify the sources of tension in music, and quite another to
say what musical tension is. What is it that harmonic dissonances in tonal music
or Stravinskys syncopations or Beethovens driving rhythms are responsible for?
This would seem to be an especially philosophical question, but philosophers of
music have been strangely silent about these properties.
There are obviously different kinds of tension and relaxation to be distin-
guished. And in some cases one might prefer terms other than theseenergy
or force or intensity or stress or restlessness, for instance, in place of ten-
sion; stability or calmness or repose in place of relaxation. Relaxation
following tension is sometimes aptly characterized as a resolution of the ten-
sion, but sometimes it is not. Sometimes ones impression is of a conflict being
resolved or a difficulty overcome or a question answered. But in other instances
musical tension merely dissipates. The tension may be of a kind that doesnt call
for resolution or solution. This is true, it seems to me, of the tension attending
Vivaldis rapidly repeating sixteenth note accompaniment figures. Iresist charac-
terizing tension in general as a demand for resolution, or defining it in terms of
expectations or the frustration of expectations.
I will not be concerned with these differences now. My present interest is in
what many varieties of musical tension and relaxation, and many of their rela-
tives, have in common.
Lets begin with this question:Are tension and relaxation in the music or
in the listener? The best quick answer is both.

activity, textural complexity and quantity, metric structure, and coloration (Berry, Structural Functions
in Music [Englewood Cliffs, N.J.:Prentice-Hall, Inc., 1976], 910).
11
Cf. the final movement of Beethovens Fourth Symphony, for instance. Also, the relentless driv-
ing rhythm of Sampak, Javanese gamelan music used to accompany fight scenes in shadow puppet
performances.
12
In the Presto movement of the Summer concerto of Vivaldis The Four Seasons, for example.
Contrast the sustained accompaniment in the Adagio Molto of Autumn.
122 In Other Shoes

II. Tension in the Music

We hear tension in music. And we hear resolutions or relaxations of tension.
These would seem to be perceptual properties of music if any aesthetic proper-
ties are. Perhaps they and properties such as gracefulness, elegance, and garish-
ness are secondary ones, like colors, and are in the object only in whatever sense
colors are. Secondary properties are said by some to be powers or dispositions to
affect observers in certain ways, or properties things possess by virtue of bearing
relations of one sort or another to (potential or ideal) observers. Even if some
such view is right, it seems to me that we experience colors as neither dispositions
nor relational properties, but as intrinsic, monadic features of physical objects.
Iam following the lead of Mark Johnston and Sydney Shoemaker on this point.
Johnston writes that Steady colors, as opposed say to highlights, do not appear
to be relational properties and do not appear to be dispositions to look colored,
even if they actually are such. A course of experience as of the steady colors is a
course of experience as of light-independent and observer-independent proper-
ties, properties simply made evident to appropriately placed perceivers by ade-
quate lighting. By contrast, he says, a course of experience as of the highlights
reveals their relational nature. They change as the observer changes position
relative to the light source, etc. They wear their light- and observer-dependent
natures on their face. Thus there is some truth in the oft-made suggestion that
(steady) colors dont look like dispositions. ... 13 Shoemaker thinks that various
relational properties possessed by external objects, those of producing or being
apt to produce experiences of certain kinds (e.g., color experiences) in particular
observers or in observers with certain sensory constitutions, are not represented
in the observers experience as relational; we are not aware of them as relational
properties (prior to philosophical reflection).14
Whatever the real ontological character of musical tension and relaxation,
it seems clear that listeners experience them as intrinsic properties of passages
of music, not as relational properties or as dispositions or powers to affect us.
One way to see that this is so is to contrast musical tension and relaxation with
the properties of being surprising, or amusing, or disgusting. The latter wear their
observer-dependent natures on their face; we experience them as what they
arecapacities to surprise or amuse or disgust.
Is the tension that we hear in music actual tension, tension literally so called?
Probably not, if this implies that the sounds themselves are literally tense. There
is more than one kind of tension, and perhaps more than one (literal) sense of

13
Mark Johnston, How to Speak of the Colors, Readings on Color:The Philosophy of Color, ed.
Alex Byrne and David R.Hilbert (Cambridge, Mass.:MIT Press, 1997), 141. See also 175.
14
Sydney Shoemaker, Phenomenal Character, 23440.
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 123

tense. But it is not at all clear that sounds are among the kinds of things that can
be tense in any literal sense of the term. Maybe we should say that music repre-
sents or expresses or suggests or portrays (actual) tension, that what we hear in it is
fictive tension, or metaphorical tension, or an impression or illusion of (actual) ten-
sion. One music theorist remarked that music embodies or represents tension and
relaxation in successions of sound, as a painting embodies, represents, a tree in a
two-dimensional array of pigment.15 There is no actual tree in the picture, but there
is a representation of one, a picture-tree, and we see it when we see the picture.
Is hearing tension in music like seeing a tree in a picture? The analogy is strained
at best. Although represent is the right word for the picture of a tree, embody,
fuzzy though it is, seems closer to the mark in the case of musical tension. Some
prefer to describe music as expressive of tension.16 This may go better with the idea
that music embodies tension; to say that a work of art expresses melancholy or
anguish or joy seems approximately equivalent to saying that it is melancholy or
anguished or joyful. But suspicions may be aroused by the fact that, notwithstand-
ing their obvious importance and pervasiveness, musical tension and relaxation
are not among the commonly cited paradigms of objects of expression. We shall
note shortly one respect in which describing tense music as expressing tension is
apt to mislead. Nevertheless, listeners enjoy an experience that can reasonably be
characterized as hearing tension in (a passage of) music. And music properly so
heard surely has something to do with some variety of real tension.
Patterns of musical tension and relaxation are, no doubt, partly responsible
for whatever expressive properties music may possess, for its expressing gaiety
or exuberance or melancholy or determination, whether or not it also expresses
tension and relaxation. Tension and relaxation may also have something to do
with what music represents, when it is representational. In section X we will see
how similar properties of a painting may serve representational ends.
What kind of real tension is it that music represents or portrays or expresses
or embodies or gives an impression of or is experienced as possessing? Is it phys-
ical tension, like that of a coiled spring or a taut cable or flexed muscles? Is it a
psychological property, a tense feeling or sensation or mood? These questions
are complicated by others:Is a tense feeling a feeling as of somethings being
physically tense, or just a feeling with a certain phenomenal character? If tense
applied to music is a metaphor, we need to ask whether the non-metaphorical
sense of the term underlying this metaphorical use is a physical or a psychological

15
Anthony Newcomb, contribution to American Musicological Society symposium entitled
Listening with Imagination, NewYork, 1995.
16
This seems to be Jenefer Robinsons position. See her The Expression and Arousal of Emotion
in Music, Journal of Aesthetics and Art Criticism 52 (1) (1994):19. Some speak of tension in music as
a metaphor. See, for example, Krumhansl, A Perceptual Analysis, 427, cf. 429. To embody tension
might be understood to mean being tense metaphorically.
124 In Other Shoes

one. Maybe tense predicated of psychological states is itself metaphorical, par-

asitic on its application to physical objects or states. If so, does this metaphorical
sense play a role in the metaphorical application of the term to music? Or is
there a single unitary concept of tension, of which physical and psychological
tension are species?
Lerdahl and Jackendoff remark that the opening tonic [of Mozarts A-Major
Piano Sonata, K331] seems to stretch like a rubber band, which in measure 4
belatedly springs loose.17 They and others use a variety of physical analogies in
characterizing patterns of tension and relaxation in music:attractional fields or
forces, pulling events in certain directions, gravitation, inertia.18 Those who regard
musical tension as an expressive property probably have a psychological state
in mind, for expression is almost always understood to take only psychologi-
cal states (emotions, moods, feelings, attitudes, perhaps character or personality
traits) as objects.
I dont find myself much tempted by either alternative. Neither seems unrea-
sonable, but neither is compelling. And Idont think it makes much difference
which examples of musical tension we consider. Most any tense instrumental
music can be used effectively to illustrate either physical or psychological ten-
sion. Tense passages of Schoenbergs Verklaerte Nacht might serve well to illus-
trate a story or cartoon in which a slingshot is stretched to its limit, or tectonic
plates confront one another at a seismic fault. But the same music would be
equally effective illustrating a story in which a character faces a difficult decision
with momentous implications. The music alone, apart from a story or a text or a
title, seems neutral. It does not seem that, in order faithfully to characterize my
experience of tense music, Imust choose.
This is surprising. Physical tension and feelings of tension are very different
sorts of things, and we are not likely to confuse them in other contexts (even if
they are species of a single genus). One might expect that we would have sorted
out the nature of musical tension as well. Idont think we have. Ithink the ques-
tion is unlikely even to arise in the minds of most listeners and critics, even when
they explicitly think of and describe passages of music as being tense.
Certain other predicates commonly applied to music (and to works of art of
other kinds) seem clearly to be meant in a psychological sense, predicates such
as exuberant, melancholy, confident, timid, angry, and the old favorites
happy and sad. Other predicates are more likely to suggest physical states or
events:turbulent, flowing, rigid, jerky, rising, falling, etc. Ifind myself
curiously uncertain which of these groups tense belongs with. (Again, there

17
Lerdahl and Jackendoff, A Generative Theory, 123.
18
See Fred Lerdahl, Calculating Tonal Tension, Music Perception 13 (3) (1996):361, and his
references to S.Larson.
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 125

are complications. The predicates of the second, physical category, when they
are applied to music, might suggest something mental, a feeling of turbulence or
rigidity, an experience of falling, etc. And psychological predicates may suggest
physical behavior by means of which the psychological states are expressed.)19
I believe that musical tension is indeterminate in this unusual respect.
(Certain other aesthetic properties may be similarly indeterminate, including
calmness, severity, being disturbed, or in turmoil.) We have here a reason to be
skittish about saying that music expresses tension:To do so would be to resolve,
artificially, the ambiguity or indeterminacy in favor of psychological tension.

III. Tension in the Listener

Tension is not just in the music. Ithink it is clear that one typically feels tension,
or something like tension, when one listens to tense music, and that one relaxes
when the tension in the music is released or resolved or replaced by relaxation.
Often, my muscles actually do tighten in response to tension in music and then
relax when the music calms down. Sometimes Ihave a sense of the tendency of
the music to make me tense, even if it doesnt actually do so; Ifeel its influence.
Perhaps one feels tension in imagination, or imagines feeling tension. Lets let
any or all of this qualify as tension or something like it.20 Notice that, although
my experience is of course a mental state, my body is sometimes in a physical
state of tension, and Iexperience it as such. Whether the tension in the music
is psychological or physical, or something else, is a different question.
Tense music is thus (with qualifications) tension-evoking. Musical tension is
something like the property of being surprising or exciting or disgusting, i.e.,
surprise- or excitement- or disgust-inducing (even though, as Imentioned, we
experience tension and these -ing properties differently). It is unlike a tree in a
picture in this regard. Idont feel anything like a tree when Irecognize a tree in
a picture, nor do Iperceive the pictures capacity to induce tree-ish feelings; my
experience is more like one of perceiving a tree. The tree in the picture is out
there. Musical tensionsome of it, anywayis (typically) in me, the listener.
One can sometimes recognize tension in music without feeling tense. Ajaded
piano teacher whose senses are numbed by long years of service may recognize
tension when he hears it. He might do so without feeling even a tendency of

19
Richard Wollheim mentions turbulence as something that may be an object of expression. Yet
he concur[s] with the traditional requirement that what is expressed is invariably a mental or psy-
chological phenomenon (Painting As an Art [Princeton:Princeton University Press, 1987], 80).
20
The output of a simulation may be either actual or pretend states. See Walton, Spelunking,
Simulation, and Slime, 3749, and in this volume, pp.271286.
126 In Other Shoes

the music to make him tense. Ibelieve that jaded piano teachers can be said not
only to hear that there is tension in music, but also to hear the tension itself. But
in hearing the tension, they are hearing, if not feeling, its capacity to excite ten-
sion.21 One may also be able to recognize that something is surprising or disgust-
ing or amusing even if, for one reason or another, one is not actually surprised
or disgusted or amused. And Isee no objection to allowing that one can per-
ceive a things surprisingness or disgustingness or amusingness, without actually
experiencing anything like surprise or disgust or amusement. Being surprising
is, nonetheless, a capacity to surprise. And musical tension is, in part, a capacity
to induce tension in listeners.
But (to reiterate) there is also tension in the music, or anyway we experience
tension as being in the music. We attribute to the music what we feel, or what the
music has a tendency to make us feel. Something that is surprising or disgusting,
by contrast, is usually not itself in any sense surprised or disgusted, nor do we
experience it as such. The surprise or disgust is only in the observer.22
Not all tension-inducing music is tense itself. Music sometimes produces
tension in us which we do not read back into it. Harmonic or formal anoma-
lies may create uncertainties about where the music is going and leave the lis-
tener tense as a result. But the music itself might (as it were) sail blithely along,
unconcerned and unbothered. Music can be irritating or maddeningbecause
it strikes us as ostentatious or arrogant or self-indulgent or sappy, or because
it threatens to put us into a stupor we would rather not be in, or because it is
blatantly designed to manipulate us. Along with the listeners irritation comes
tension. But the music itself need not be tense or perceived as being tense, any
more than it is or is perceived as being irritated or angry. Some music makes
me tense by boring me to frustration. But what bores and frustrates me may
be precisely a lack of tension in the music. Iset aside, for now, music which is

21
And those thin clouds above ... / Those stars ... / Yon crescent moon ... / Isee them all so
excellently fair, / Isee, not feel, how beautiful they are! (Samuel Taylor Coleridge, Dejection:An
Ode).
22
Jenefer Robinson rightly emphasizes the tendency of music to make listeners feel tense or
relaxed, and its capacity to disturb, unsettle, startle, excite, calm, and soothe:The expression of a
feeling by music can sometimes be explained straightforwardly in terms of the arousal of that feeling
(Robinson, Expression and Arousal of Emotion, 19). Stephen Davies objects to her suggestion
that the tension, and so forth, of music consists in its power to arouse a corresponding automatic
response in the listener, claiming instead that the relevant properties are of the music, that they are
possessed not as causal powers but intrinsically (Davies, Musical Meaning, 1045). Each of them is,
in my view, partly right and partly wrong. Musical tension consists partly in a power to evoke tension
in listeners, but only partly; it is also an intrinsic property of the music. Aaron Ridley gives an account
of expression in music which is more like my account of musical tension. To perceive melismatic
qualities of music as expressive is to respond sympathetically to them. See Ridley, Music, Value and
the Passions, 138.
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 127

Figure1 Claude Monet (French, 18401926), The Break-up of the Ice (La Dbcle or Les
Glaons), 1880.Oil on canvas, 60.3cm x 99.9cm (233/4 in. x 39-15/16 in.). University of
Michigan Museum of Art. Acquired through the generosity of Russell B.Stearns (LS&A,
1916), and his wife Andree B.Stearns, Dedham, Massachusetts. 1976/2.134.

merely tension arousing. Ishall be interested in cases in which we hear tension

in music, in addition to feeling tense ourselves or recognizing the musics capac-
ity to arouse tension.
To clarify how the tension induced in the listener and the tension embodied
in the music are related, lets look at a contrasting example. Consider Claude
Monets La Dbcle (fig. 1). There is coldness in the picturein a relatively clear
sense in this case:La Dbcle represents coldness, or it represents things (the
river, the weather) as being cold. But Idont just recognize the frigid scene when
I look at the painting. I also feel cold myself, or something like coldness
not literally, perhaps, but Imay actually shiver, or feel as though Iam going to.
Perhaps the best way of describing my state is to say that Ivividly imagine feeling
cold or that Ifeel cold in imagination. (This imaginative experience is, of course,
not just an intellectual contemplation of the proposition that Ifeel cold.) In any
case, Ihave feelings which are appropriately, if not literally, characterized as feel-
ings of coldness.23
So the tension of a passage of music and the coldness of La Dbcle are both,
somehow, in the work and both are also, somehow, felt by the appreciator. The
two cases seem to me to be significantly disanalogous, however, although the
difference may not be as sharp as it appears at first to be. Briefly, viewers feel

23
La Dbcle might have the opposite effect on viewers from northern climes.
128 In Other Shoes

cold when they look at the painting because of the (represented) coldness in
the painting. But it is because a passage of music makes listeners feel tense that
there is tension in it. Irecognize the cold scene, the ice on the water, etc., in La
Dbcle. As a result, Ishiver; Ifeel cold. But it is by virtue of the feelings of ten-
sion which certain passages of Verklaerte Nacht arouse in me that Iattribute ten-
sion to them; Ido not first recognize a portrayal of tension, to which Irespond
with tension of my own. There might easily be a painting representing a cold
scene, which Irecognize as such, even if it has no tendency at all to make me feel
cold (the small reproduction of the Monet displayed above, for instance). But if
a passage of music has no tendency at all to elicit in me anything like a feeling of
tension, if Idont hear it as something that might, under favorable circumstances,
make me feel tense, Iwould not, Ithink, be inclined to call it tense (in anything
like the sense of tense that Iam considering).
I said that the difference between these two examples may not be as sharp as
it seems. It is possible that the colors of the painting, apart from what they repre-
sent, have some tendency to make me feel cold. And Idont rule out that some
suggestion of (represented) tension in the music is responsible for my feeling
tense. To some extent, Isuspect, the explanations go in both directions in both
instances. Nevertheless, it is largely the feelings of tension the music produces
in me that encourage and justify my attribution of tension to the music. And it
is largely the other way around in the case of the cold painting. Imight add that
there is no reason to suppose that appreciators who feel cold or tense, and who
attribute coldness or tension to the work, must be aware of which comes first,
their feeling or their attribution, or aware that the explanation does or does not
go in both directions.

IV.Projectivism
Santayanas definition of beauty as pleasure objectified recalls David Humes
observation that the mind has a great propensity to spread itself on external
objects, and his characterization of taste as gilding or staining ... natural objects
with the colors borrowed from internal sentiment.24 Contemporary incarna-
tions of such views travel under the banner of projection or projectivism.
Some have suggested that colors and other secondary qualities, or moral prop-
erties, are really just projections of features of our experiences onto the world.25

24
David Hume, A Treatise of Human Nature, ed. L. A. Selby-Bigge, 2d ed. (Oxford: Oxford
University Press, 1888/1978), bk. I, pt. III, section XIV.
25
Paul A. Boghossian and J. David Velleman, Colour as a Secondary Quality, Mind 98
(1989):81103.
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 129

Richard Wollheim gives an account of expressive properties of art in terms of a

kind of projection.26
Sydney Shoemaker confessed to being in the uncomfortable position of
finding [projectivism concerning secondary properties] both plausible and
unintelligible.27 The impression of unintelligibility derives, Ibelieve, from the
presumption that the properties of our experiences which we supposedly proj-
ect onto external objects are ones that external objects could not possibly pos-
sess.28 There is something profoundly unattractive about the view that there is
something like a category mistake, the attribution to things of features they could
not have, involved in the content of every sense experience.29
I do not find projectivism with respect to colors especially plausible. But
Idont find it exactly unintelligible either, at least not simply because it postu-
lates systematic category mistakes in the intentional content of our experiences
of the world. Imentioned Shoemakers suggestion that certain properties which,
on reflection, we know to be relational (dyadic) are nevertheless experienced as
being monadic. He also suggests that the relation being to the right of is sometimes
experienced as being dyadic and is naturally thought of as dyadic, although we
realize on reflection that it is at least triadic, (x is to the right of y relative to z).30
This is a pervasive and systematic error in the content of our experiences, and
the error is akin to a category mistake. There is no dyadic relation of being to the
right of; hence, no pair of objects could exemplify this relation.
We have begun to see that listeners do experience music as being tense, even
if the attribution (taken literally) is a category mistake. This seems rather differ-
ent from the case of colors, however. For we arguably are perfectly aware of the
mistake when we experience music as tense; we dont fool ourselves. If pushed
we are likely to retreat to the idea that the music portrays or expresses or repre-
sents or suggests tension, or that the attribution is metaphorical. No such retreat
is readily available in the case of colors. Perhaps we do fool ourselves into think-
ing that objects are colored, if projectivism about colors is right.
Richard Wollheim speaks of projecting ones sadness or melancholy onto
the world or onto a work of art. Thiswhat he calls complex projectionis not

26
Wollheim, Painting as an Art, and Wollheim, Correspondence, Projective Properties, and
Expression in the Arts, The Mind and Its Depths (Cambridge, Mass.: Harvard University Press,
1993), 14458. Roger Scruton toys with (but essentially rejects) the idea that listeners transfer their
own experiences of movement and passion to the music. Understanding Music, in Scruton, Art and
Imagination (London:Methuen, 1974), 9499.
27
Sydney Shoemaker, The First-Person Perspective and Other Essays (Cambridge: Cambridge
University Press, 1996), 139 n.
28
Ibid., 250.
29
Cf. ibid., 102.
30
Sydney Shoemaker, Phenomenal Character, 234.
130 In Other Shoes

experiencing the object as being sad or melancholy in the sense that persons are
sad or melancholy, however, so Wollheim avoids saddling experiences with cat-
egory mistakes, or even with systematic error. It does involve experiencing the
world as of a piece with ones emotion. Wollheims notion of projection is very
different from the one Iwill be developing. And the examples his is designed
to deal with seem very unlike those Iam interested in. For one thing, Iwill be
concerned with properties of works of art other than expressive ones, where
expression is assumed to be only of psychological states. Wollheim understands
what he calls projection to be motivated by anxiety, a desire to rid oneself of the
emotion, in the case of sadness or melancholy, or to sustain it, in the case of love.
And he thinks that projecting ones melancholy or sadness helps to relieve one of
it.31 None of this seems to me to be true, ordinarily, in the cases Ihave in mind.
Finally, his examples seem not to involve anything like imaginatively occupying
shoes other than ones own.

V. Mental Simulation
I switch gears now and consider, not our experiences of works of art, but our
experiences of other people and the judgments we make about their mental
lives. The hypothesis that people are targets of projection escapes the worries
Ihave just been considering. Properties of my experience can be attributed to
other subjects of experience without committing category mistakes. And such
attributions may well be true.
It is well known that our own emotional or affective states play a vital role in
our judgments about the mental states of others. Rather than simply perceiv-
ing that another person has certain thoughts or feelings, or inferring this from
what Iperceive, Igo partly by my affective responses. Irecognize someones
arrogance or pompousness partly by noticing the resentment or irritation
Ifeel toward him. Apersons aggressive manner shows itself in the defensive-
ness it arouses in me or in my feelings of intimidation. Icome to understand
how helpless someone is in part by noticing my tendency to nurture him or
her. It would be a mistake, Ibelieve, in the first two examples, to regard my
resentment or irritation as nothing but a symptom of the persons arrogance or
pompousness, as a clue which leads me to suspect arrogance or pompousness,
which Ithen verify more directly by considering the specifics of his behavior.
Arrogance and pompousness are response-dependent properties consisting
(partly) in a propensity to produce irritation in others, just as being disgusting

31
Wollheim, Painting as an Art, 82, 84; Wollheim, Correspondence, Projective Properties, and
Expression, 151.
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 131

consists (perhaps entirely) in a propensity to elicit disgust. To be arrogant is,

in part, to be irritating.
These are not instances of the objectification of ones feelings, or projec-
tion. It is not what one feels oneselfirritationthat one attributes to the
other person, but something differentpompousness or arrogance. When we
engage in what has been called mental simulation, however, we attribute psycho-
logical states to others on the basis of our experience of the same psychological
states, or at least similar or corresponding or analogous ones. Iput myself in
your shoes, in imagination, and judge from how Ifeel or what Iam inclined to
do, as a result, that that is how you feel or what you are inclined to do. Ishall be
especially interested in emotional contagion, which is sometimes described
perhaps not aptlyas a primitive kind of mental simulation. Imay catch a
mood or feeling from another person, and thereby have reason to attribute it
to her.
Simulation theory is controversial. One point of contention concerns whether
the simulation theory collapses into its supposed rival, the theory theory. Some
have claimed that to engage in simulation is itself to deploy a psychological the-
ory, to utilize (tacit) knowledge of correlations between the observable circum-
stances and/or behavior of other people and their thoughts and feelings. For
now, Iwill say just that if this is so, it does not make simulation superfluous. For
simulation can then be understood as a special, distinctive way of accessing and
utilizing ones psychological theory, and it behooves us to explore what is special
about it.
Suppose Iwant to predict how someone will feel or what she is likely to do if
she misses her plane, or suddenly comes across an intruder in her basement, or
goes on a caving expedition for the first time, or is condemned to death by firing
squad. According to the simulation hypothesis, Iimagine myself in similar cir-
cumstances; Iimagine missing my plane or discovering an intruder in my base-
ment, etc. Imay also imagine having certain beliefs or desires or attitudes which
Itake the other person to have, if they are different from my own. Ithen find
myself deciding on a course of action or experiencing certain feelings, in imagi-
nation, and I conclude that the other person will actually choose that course
of action or experience those feelings. Ithus project my (imagined) experience
onto the other person.
Why should we have confidence in this procedure? What reason is there
to suppose that the state Ifind myself in as a result of imaginatively occupying
anothers shoes is likely to be similar to, or a pretend version of, the state she is
actually in? The usual answer is that my imaginative experiment activates psy-
chological mechanisms in me which are fundamentally similar to those at work
in the other person. These mechanisms operate on actual and pretend states in
similar ways, so they can be expected to yield similar or corresponding results.
132 In Other Shoes

In short (as some but not all simulation theorists put it), the process occurring
in the simulator models the process occurring in the simulatee.32
A (successful) simulation, as Iwill use the term, is one in which not only the
inputs and outputs of the simulator correspond to those of the person being
simulated, but the processes linking them do as well. (I wont try to say in what
respects or to what extent the two processes must correspond.) This is what is
meant by simulation in much of the literature33 (although some have called
this process or process driven simulation, and distinguished it from other kinds of
simulation in which the processes do not correspond).34
The result of my imaginative experiment is usually not exactly the same
as the state the target person is in, of course. My cognitive system operates
off line or the output is taken off lineit is disconnected from its usual
role in guiding behavior. If in simulating someone who confronts a burglar
Ifind myself deciding to call 911, Idont actually pick up a telephone. And
Isuppose that the output of my simulation doesnt count as a genuine deci-
sion; it is merely in imagination that I decide to call 911.35 But I attribute
to the other person an actual decision and an actual action. Some outputs
of simulationstypically ones with no essential connection to deliberate
behaviorare not merely imagined or pretend states, however. On imagin-
ing confronting a burglar Imight actually experience a sudden panicky feel-
ing and break out in a cold sweatand Imay attribute this state to the target
of my simulation.
If the shoes I actually occupy are enough like those of a person I want to
understand, imagining may not be necessary. I might predict how others will
experience a roller-coaster ride by actually trying it out myself. My simulation,
in this case, is not an imaginative experiment; neither the inputs nor the outputs
are pretend states.

32
We call upon our similarity to other people, in particular the similar functioning of our minds
(Gary Fuller, Simulation and Psychological Concepts, in Mental Simulation, ed. Martin Davies and
Tony Stone [Oxford:Blackwell,1995], 21). Sec Jane Heal, Simulation and Cognitive Penetrability,
Mind and Language 11 (1)(1996):47; and Alvin Goldman, Empathy, Mind, and Morals, in Mental
Simulation, ed. Davies and Stone, 189. Robert Gordon denies that the simulator uses himself as a
model. See his The Simulation Theory: Objections and Misconceptions, Folk Psychology, ed.
Martin Davies and Tony Stone (Oxford:Blackwell, 1995), 11420.
33
Gregory Currie, Imagination and Simulation:Aesthetics Meets Cognitive Science, in Mental
Simulation, ed. Davies and Stone, 15960; Susan L Feagin, Reading with Feeling (Ithaca, N.Y.:Cornell
University Press, 1996), 8487; Ian Ravenscroft, What Is It Like to Be Someone Else? Ratio 2 (11)
(1998):177, 183.
34
Alvin Goldman, Interpretation Psychologized, in Folk Psychology, ed. Davies and Stone, 85.
35
The output state should be viewed as a pretend or surrogate state, since presumably a simulator
doesnt feel the very same affect or emotion as a real agent would (Goldman, Empathy, Mind, and
Morals, 189).
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 133

VI. Expressive Behavior

Simulation purports to account for our ability to ascertain another persons
psychological states on the basis of the conditions that give rise to them. In the
examples mentioned so far, the simulator begins by projecting himself into the
other persons circumstances. But what about cases in which we judge anothers
mental states from her behavior? Do we engage in simulation in these instances?
We do not merely observe the behavior, in many such cases; frequently we
behave in similar ways ourselves, or imagine doing so. The nearly universal ten-
dency to mimic, more or less automatically, the expressive behavior of other
people, their facial expressions, gestures, and bodily attitudes, has been well doc-
umented. Motor mimicryresponding to smiles with smiles and to frowns with
frownsis evident even in very young infants.36 On seeing someone leaping
with joy or strutting with pride, we may feel our muscles contracting as though
we are leaping or strutting ourselves, even if we dont actually do so.
Reproducing in ourselves, either actually or imaginatively, expressive
behavior that we observe in others surely plays an important role in coming
to understand what they think and feel. Robert Gordon speaks of simulation
in such cases. [M]uch of the work of simulation appears to be carried out by
rather superficial imitative mechanisms. An example is the automatic and often
subliminal muscular mimicry of the bodily postures and especially the facial
expressions of others, beginning with the infants smiling response to smiles.
Feedback from such motor mimicry appears to be an important factor in the
recognition of emotions in others.37 It is plausible that imagined or actual mim-
icry leads one to feel something like what the other person feels. To ascertain
what feelings a person expresses by frolicking, prancing, strutting, swagger-
ing, cavorting, or romping in a certain manner, Ifrolic, prance, strut, swagger,
cavort, or romp in a similar manner (actually or in imagination), and note what
Ithen feel.

36
A. N.Meltzoff and A. K.Moore, Imitation of Facial and Manual Gestures by Human Neonates,
Science 198 (1977):7578; A. N.Meltzoff and A. K.Moore, Newborn Infants Imitate Adult Facial
Gestures, Child Development 541 (983):7029; Janet Beavin Bavalas, Alex Black, Charles R.Lemery,
and Jennifer Mullett, Motor Mimicry as Primitive Empathy, in Empathy and Its Development, ed.
N.Eisenberg and J. Strayer (Cambridge:Cambridge University Press. 1987), 31738.
37
Robert Gordon, The Simulation Theory, 113. Gordon continues:Other imitative mecha-
nisms would seem to play an important role in ascribing content to the others expressive behavior.
One such mechanism is mimicry of perceptual orientation, especially gaze mimicry. ... So, by mim-
icking anothers gaze, Ilearn what the content of her perceptual experience is. This is ascertaining
the others mental state on the basis of its cause, the circumstanceslooking in a certain direction in
a certain situationwhich give rise to it. It is thus importantly different from the mimicry of a per-
sons bodily postures, etc., which figures in our attribution to her of the mental states those postures
express.
134 In Other Shoes

When mimicry leading to feelings or moods or emotions like those of the

person mimicked is sufficiently automatic and spontaneous, we speak of the
feelings or moods or emotions being contagious or infectious. Nellie is fidgeting
nervously. I find myself fidgeting and feeling nervous also, when in her com-
pany. Attending a party in which the mood is one of gay abandon, Iam caught
up in the spirit of the revelers. My own mood shifts to match theirs, and Itoo
behave with gay abandon. My fidgeting is unwitting mimicry, motor mimicry, of
Nellies fidgeting, which results in my feeling nervous (actually feeling nervous,
probably, in this case). Imay then judge Nellie to be nervous. Iam aware of her
nervousness because she makes me nervous; her nervousness is evident to me
by virtue of infecting me. In like manner, Imay discover that someone is calm
or excited or gleeful or down in the dumps because Ifind myself, when in her
company, infected with these feelings. The notion of emotional contagion will
be important when we return to the topic of musical tension.
In my recent examples, as in the instances of simulation outlined earlier,
Ioccupy the other persons shoes, or shoes like hers, and attribute to her what
Ithen find myself feeling. The difference is that the shoes Ioccupy are not in
a certain location, a certain situation, but rather are moving or behaving in certain
ways, or configured in a certain way.
This is a big difference. How can simulation of this kind (if that is what it is)
work if, as common sense has it, expressive behavior is caused by the mental
states being expressed? If my psychological mechanisms are relevantly like yours,
Imight expect the same or corresponding inputs, causes, to lead to the same or
corresponding outputs, effects. But there is no reason, in general, to expect that
my psychological mechanisms will run backwards, taking as inputs the outputs
of yours, and yielding as outputs the inputs of yours. So how can (actually or
imaginatively) mimicking someones expressive behavior be expected to repro-
duce in me feelings or emotions like those which cause his expressive behavior?38
Some well-known psychological studies establish that, to some extent at
least, behavioral manifestations of emotional states are causes rather than results
of those states.39 Iam sure that there is some truth in the Jamesian idea that we
are sad because we cry and afraid because we flee, rather than the other way
around. (Bodily movements and facial expressions are the causes at least of the
sensations of ones body moving or being fixed in the ways in question.) Insofar

38
If the simulators and simulatees causal processes do run in opposite directions, we wont have
an instance of (process) simulation.
39
Cf. Paul Ekman, Facial Expressions: New Findings, New Questions, Psychological Science
3 (1992): 3438; Paul Ekman and R. J. Davidson, Voluntary Smiling Changes Regional Brain
Activity, Psychological Science 4 (1993):4245; and Pamela K.Adelman and R. B.Zajonc, Facial
Efference and the Experience of Emotion, Annual Review of Psychology 40 (1989): 24980. See
Stephen Darwall, Empathy, Sympathy, Care, Philosophical Studies 89 (1989):265.
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 135

as the mental states of the people we aim to understand are results of their
behavior, there is no special problem for the simulation theory. Ido not think
this is the whole story, however. Surely there are causal relations running in the
other direction as well, from emotional experiences to expressive behavior. And
Iwould be surprised if the mechanisms involved do not figure in something like
simulations, or empathetic understanding, of others mental lives.
Here is one way they might do so. The simulator might observe the behavior
of the simulatee, and then try out various feelings or emotions, imagining expe-
riencing each of them to see which one leads him to behave or be inclined to
behave or imagine behaving in the manner the simulatee does.40 Iwould expect
the simulator to utilize something like a theory in choosing which hypotheses
to test by simulation. Often, probably, simulation simply serves to confirm judg-
ments made on the basis of ones theory. The confirmation is of course limited
insofar as more than one kind of experience might result in the behavior in
question.

VII.Memory
I believe that coming to understand others by imagining behaving as they do
sometimes involves a very different mechanism. Consider this case:An experi-
enced but intuitive table tennis player is asked which way one must slice ones
paddle in hitting the ball, in order to make it curve to the left, or to the right. To
answer, she might imagine moving her paddle to the right, and then find herself
imagining seeing the ball curve to the left. So, she concludes, a slice to the right
produces a curve to the left. Alternatively, she might start with the effect, imagin-
ing the ball curving to the left after it is hit, and then find herself imagining hav-
ing sliced her paddle to the right when she hit it. The point of these imagining

40
Both Gordon and Goldman have suggested approximately this: Gordon, Folk Psychology
as Simulation, in Folk Psychology, ed. Davies and Stone, 6466; Goldman, Interpretation
Psychologized, 82. See also Stephen Stich and Shaun Nichols, Folk Psychology:Simulation or Tacit
Theory, in Folk Psychology, ed. Davies and Stone, 12932.
[In Simulating Minds:The Philosophy, Psychology, and Neuroscience of Mindreading (Oxford:Oxford
University Press, 2006), 6.3.1 and 7.12, Goldman again proposes that this generate-and-test pro-
cedure may account for retrodictive mindreading. Reverse Simulationcasual processes operat-
ing in reverse order in the simulatoris not an option in the case of most cognitive processes, he
claims. In VII below, however, Isuggest that, when memory traces come into play, causes and effects
may indeed be reversed in the (imaginative) experience of the attributor, as compared to that of the
target (although this is not process simulation). We should expect that imagining or perceiving a per-
son engaging in certain verbal or nonverbal behavior might, with the help of memory traces, induce
one to imagine having had certain feelings or attitudes, obviating the need for a generate-and-text
strategy.]
136 In Other Shoes

exercises may be to ascertain something about physical events. But they also
enable the imaginer to make predictions about her or others mental lives. In the
first variant of the experiment, the imaginer predicts that if one hits the ball with
a rightward slice, she will then see it curving left.
It is not by means of simulation (process simulation) that she arrives at this
conclusion. The imaginer does not rely on the (off line) operation of the causal
mechanisms which are at work in the actual case. Actually slicing ones paddle
to the right, in a real game of table tennis, initiates a complex series of events
involving air flow around the ball, friction, differential pressure, etc., resulting
in the balls curving left and in the players seeing it curve left. Nothing at all like
or analogous to this physical process occurs in the imaginers mind. The causal
mechanism whereby imagining hitting the ball leads to imagining seeing it curve
involves, Iwould suppose, the activation of memory traces of the persons actual
experiences playing table tennis on previous occasions. (She need not explic-
itly recall any particular experiences, of course.) Such memory traces come into
play also, no doubt, when the imaginative experiment is run backwards, when
the person first imagines the ball curving left, and then finds herself imagining
having sliced it to the right. Memory traces have no place at all in the causal
story whereby actually hitting a table tennis ball with a rightward slice produces
a left-curving shot and a perception of a left-curving shot.41
Only someone who has played table tennisonly someone who has, in real
life, sliced the ball to the right and seen it curve to the left and so possesses the
required memory tracescan successfully carry out this imaginative experi-
ment. (Well, having watched table tennis games, or conceivably having been told
of them, might suffice.) But simulation can work, in principle, even if one has
never experienced or heard of or thought about situations of the relevant kind. If
you have never been spelunking and have never been in an elevator or closet or
other closed space, and so have never experienced the slightest twinge of claus-
trophobia, you might nonetheless test yourself for claustrophobia by simulating
an adventure in a cave. You imagine squirming on your belly in a dark twisting
passageway, your pack tied to a foot and dragging behind you because there isnt
room to wear it on your back. You can expect to find yourself feeling uneasy, if
you are susceptible to claustrophobia; you might even sweat profusely. At least
Ipresume that you can expect this. You can if the imaginative experiment works
by activating the off line operation of a causal mechanism, your propensity
to feel claustrophobic, which is a permanent feature of your psyche and which
would operate on line should you actually go on a spelunking expedition.

41
The table tennis example is much like that of ascertaining the number of windows in ones
house by counting them in imagination. As Stich and Nichols observe, the latter is not an instance of
off-line (process) simulation. See Stich and Nichols, Folk Psychology, 140.
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 137

Although the table tennis players imaginative experiment is not simulation as

Iam understanding it (process simulation), it mimics simulation in its outward
form. The input for the experiment is an imagined experience, the experience
of hitting the ball, which yields as output another imagined experience, that of
seeing the ball curve. This sequence convinces the experimenter that an actual
instance of the first experience would result in an actual instance of the second.
The difference is that the causal process whereby the imagining of the one expe-
rience results in the imagining of the other is entirely unlike the causal process
whereby the first experience would actually result in the second. Is the result of
this experiment to be trusted? Are the two processes, the one involving memory
traces and the very different real world one, likely to yield similar or correspond-
ing results? It would seem so. But Ihave said nothing about why this should be
so, or why we might expect it to be so. Some will hold that the imaginer relies on
a tacit theory embodied in his memory traces. This depends, of course, on what
is meant by tacitly possessing and using a theory. But the idea that the imagina-
tive experiment is theory driven is less persuasive than it seems, for reasons
Iwill sketch shortly.
It may well be that some or many of the standardly cited instances of (pro-
cess) simulation are not such. Even if the relevant real world causal mechanism
is entirely psychologicalthe mechanism whereby the experience of missing
ones plane makes one upset, for instancethe imaginative experiment, which
simulationists suppose to work by virtue of the off line operation of similar
mechanisms, may crucially involve something entirely different, the activa-
tion of memory traces, in whatever way the table tennis example does. In many
cases it will be difficult to decide. There need be no phenomenological differ-
ence between the two kinds of experiments. If you imagine missing your plane
and then find yourself being tremendously upset in imagination, there may be
nothing in your experience to indicate whether memory traces which would not
come into play should you really miss your plane play a role in your imagina-
tive experiment. There will be mixed cases, no doubt, cases in which parts of
the processes correspond and parts of them do not, or cases in which genuine
simulation runs alongside and is reinforced by a different causal process. This is
a matter for empirical investigation, of course. My guess is that memory traces
of some sort play an important role in just about all imaginative experiments of
the kind we are considering and that there are few if any pure instances of mental
(process) simulation. But Iwould guess, also, that there are many impure or par-
tial instances, that the causal processes linking the inputs and outputs of imagi-
native experiments often correspond in significant measure to those linking the
relevant real world experiences.
Others have recognized roles memory may play in simulations or empathetic
experiences. Ian Ravenscroft thinks that memory may make simulation more
138 In Other Shoes

likely by facilitating the causal process whereby the input of ones simulation
leads to an empathetic response.42 The table tennis example shows that the role
of memory may be to replace much of the causal process that occurs in the real
situation with a very different one, hence disqualifying it as an instance of simu-
lation (in his sense and mine). Shaun Nichols etal. note that memories triggered
by beliefs about another persons situation may contribute to ones empathetic
emotional responses, and observe that unlike off-line simulation, such informa-
tion based accounts [of empathy] dont appeal to pretend or deviant inputs.43
Iam urging that pretend inputs, imagining oneself in the others shoes, may be
what triggers the memories and leads to the emotion; merely believing or know-
ing what situation the other is in may not do the trick.
The role of memory traces in the experiment doesnt necessarily mean that
the two processes are disanalogous. Memory traces might play a corresponding
role in real life cases. When Iactually miss a plane and am upset, memory traces
of previous instances in which Ior someone else missed planesor trains or
appointmentsmight be partly responsible.
We do not now have to be surprised by the fact that we sometimes come to
understand others by imagining ourselves engaging in behavior or putting on
facial expressions that result from their mental states. If these are not instances
of simulation, they may be imaginative experiments informed by memory traces
(either traces of memories of ones own expressions of mental states or of others).
If a persons kicking up her heels in a certain manner is an effect and not a cause
of exuberance or glee, imagining kicking up ones heels in that manner, or actually
doing so, may still (help to) cause one to experience something similar. And this
may give observers reason to conclude that the person they observe is gleeful or
exuberant. It doesnt matter for my purposes whether this is genuine (process)
simulation. Lets call it an instance of empathy, whether or not the causal process
in the observer corresponds to the one occurring in the other person.
Readers will have noticed an affinity between what Isaid about cases in which
memory traces are involved and Humes account of inferences, based on the rela-
tion of cause and effect, to things not present to the senses or memory. When
an object appears, that resembles any cause in very considerable circumstances,
the imagination naturally carries us to a lively conception of the usual effect.
This is so because of a custom or habit acquired from past experience.

Now as we call every thing CUSTOM, which proceeds from a past

repetition, without any new reasoning or conclusion, we may establish

42
Ravenscroft, What Is It Like?
43
Shaun Nichols etal., Varieties of Off-Line Simulation, in Theories of Theories of Mind, ed. Peter
Carruthers and Peter K.Smith (Cambridge:Cambridge University Press, 1996), 61.
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 139

it as a certain truth, that all the belief which follows upon any present
impression, is derivd solely from that origin. When we are accustomd
to see two impressions conjoind together, the appearance or idea of the
one immediately carries us to the idea of the other.

Hume thinks we infer causes from effects as well as effects from causes in this
manner.44

VIII.Content
As we have seen, there is a considerable variety of kinds of cases in which we
learn about others by imaginatively or actually occupying their shoes. But there
are important similarities among them. In particular, what we learn contains a
demonstrative element.45 This is so whether or not the imaginative process cor-
responds to the real process, whether or not memory traces play a role in either
process, whether we judge the other persons psychological state from circum-
stances that give rise to it or from behavior that results from it, and whether we
engage in the imaginative experiment automatically or deliberately.
When Ilearn about Nellies mental state by being infected by it, the content
of my judgment is probably, in the first instance, something of the form:She
feels like (or something like) this, where this refers to an aspect of my own
state of mind. We might put this by saying that Iuse my nervousness to represent
Nellies.46 Imay alsoor Imay notfind a predicate to characterize her state;
Imight describe her as nervous. But even if Ido come up with a predicate, it is
likely to be less specific than what Irepresent to myself about her using my own
mental state; the predicate is likely to express only part of the intentional content
of my judgment.
Of course, Ihave to pick out the aspect of my mental state which Iuse to
represent the other persons. Isee no reason to suppose that Imust make use of
a predicate in doing this; the respect(s) in which Ijudge her state to be like mine
will come out in my dispositions to judge third persons, or myself at different
times, to be in the same state. Apredicate may be helpful, however. Imight say
to myself, She has this kind of nervousness or She is upset in this manner.

44
Hume, A Treatise of Human Nature, bk. I, pt. III, section XIII; bk. I, pt. III, section VIII; bk. I,
pt. III.section II, section VI.
45
Iam here developing a suggestion made by Jane Heal. See Jane Heal, How to Think about
Thinking, in Mental Simulation, ed. Davies and Stone, 3352.
46
David Hills suggested putting it this way. If Iexperience only in imagination the feeling Iattri-
bute to her, then this refers not to what Iactually feel but to what Iimagine feeling, what Ifeel in
imagination.
140 In Other Shoes

Even so, the reference to my own state of mind is essential to the full specificity
of the state Iattribute to her. Imay sometimes consciously modify my feelings
representation of the other person. Knowing that Ihave a greater tendency to
panic than she does, or that she is more likely to enjoy uncertainty, Imay say that
she feels like this, minus the panic, or that she feels like this, except that she enjoys
it. In any case, using my mental state to represent hers, Iattribute to her a state
which Iprobably cannot fully characterize by means of psychological predicates
(or predicate-like expressions of a language of thought). And even if Ido have a
predicate that is exactly right, it need not figure in my representation of her state.
My knowledge about her is, in this sense, nonconceptual.
This point holds generally for judgments of others based on simulation or
empathy. If, imagining myself in the shoes of a person who misses his plane,
Ijudge him to be upset, the content of my judgment, in the first instance, is that
he feels (something) like this. Imay or may not also declare him to be upset, using
that or another predicate. Likewise for the table tennis case. One judges, in the
first instance, that a ball hit with a slice to the right will travel in a way that looks
like this. In all of these cases, Iuse my own mental state to represent anothers.
The content of judgments of others based on an affective response that does not
match their mental states may contain a demonstrative element of a different kind.
If Idetect a persons arrogance partly by means of my resentment toward him, my
judgment is not:This is how he feels, where this refers to an aspect of my state of
mind. But it may amount to:His feeling (attitude) is such as to elicit this.
The fact that, in my judgment about the other person, my state of mind serves
to represent hers, seems to me to lessen considerably the attractiveness of the
idea that in arriving at the judgment Imust be utilizing a psychological theory
the idea that the simulation is (as some say) theory driven. Can it be said that
Ipossess a concept of the property Iattribute to her if Ihave nothing like a predi-
cate for that property and no way of expressing it except via my own instantia-
tion of it? Iam not sure. But Isee no reason to suppose that Imust possess the
concept before Iexperience the mental state in question. And if Ilack the con-
cept, how can Ipossess and utilize a theory involving it in the process by which
Icome to experience it? What Ihave, rather than a theory, is simply a capacity to
be affected in a certain way, should Iimagine being in shoes of a certain sort.46a

We are now in a position to make several further observations about emotional

infection which will be important when we return to musical tension.

46a
[So Idisagree with Alvin Goldmans claim that simulation based mindreading requires prior
possession of mental state concepts. (Simulating Minds, 188).]
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 141

Emotional infection, automatic empathy, can occur without my making any

judgment at all about the other person. One can empathize with someone with-
out empathetically understanding him or her. When Nellie makes me nervous,
Ido not necessarily use my nervousness to represent her state of mind. Imight
feel nervous and be aware that Ido, without having any idea what caused my
nervousness. In that case Iwont attribute nervousness to her.
It is also possible, however, that Ishould sense that Nellie is nervous without
realizing how Isense that she is. Imay not be aware that Iam nervous, letalone
that Icaught my nervousness from her. (What is the content of my judgment
about her, in that case? Try this:Iam aware of what is in fact my nervousness, but
Ithink of it as merely imagined nervousness, or as something like an image or a
representation of nervousness. My nervousness, though Ithink of it in this way,
is the referent of the this when Ijudge that Nellie feels like this.)
When Iam infected by Nellies nervousness, and my state of mind serves as
a representation of hers, Iprobably not only make a judgment about her state
of mind, I probably also experience her as being nervous. More specifically,
I probably experience her as feeling (something) like this. Suppose, however,
that although I am nervous and aware of my nervousness, I dont experience
Nellie as responsible for my state. Imight then learn in some other way that she
is responsiblea perceptive friend whose judgment Itrust might tell me, for
instance. And Imay infer thatjudge thatNellie feels like this. Iam now using
my state of nervousness to represent hers, but Idont experience her as feeling like
this, or even as being nervous.
I would expect projectivists about color to make points analogous to at least
some of what Ihave just said about the empathetic detection of Nellies nervous-
ness. If in seeing a stop sign as red one is projecting the color of ones experience
onto the stop sign, the intentional content of the experience will be, in the first
instance:The sign is this color, where this refers to the color of ones experi-
ence. If Ihave what seems to me to be an after image, but which is actually a
perception of a stop sign, my experience will not have the content That stop sign
has this color. But if my eye doctor sets me straight, Imay then judge that the
stop sign looks like this. There remains the fact that, according to the projectivist,
my experience as of the stop sign being red is in errorit does not possess that
property. My experience as of Nellies being nervous may, of course, be correct.

IX. Empathizing with Things

I have been discussing simulation of and empathy with other people, other sen-
tient beings. Lets look at responses to inanimate things, like works of art. Some
works of art contain characters, fictional sentient beings, and we often respond
142 In Other Shoes

to them, empathize with or simulate them, in much the way we do actual people.
But what about our responses to the works themselves?
Mental simulation is not always simulation with someone, and empathy does
not always have a sentient being as its object.47 My bones ache, sometimes, at
least in imagination, when Iobserve an especially uncomfortable looking chair.
I expect that this is because I imagine sitting in the chair, and imagining this
makes me feel something like the discomfort that actually sitting in it would
involve. (The imagining probably occurs spontaneously, and Imay be unaware
of it.) But there is no person whom I simulate or empathize with. I imagine
myself in a situation Iam not actually in, but not one Iunderstand anyone else
to be in. There is no need to suppose even that Iimagine another person sitting
in the chair or recognize a fictitious person there and feel with him. The shoes
that Iimaginatively occupy are not those of anyone else. (This example is much
like that of Monets winter scene.)
In other cases Iimagine being in a situation that is occupied by something,
but not by a person. Imay empathize with a column supporting a building, imag-
ine being in its shoes. Ifeel or imagine feeling the weight of the building on
my shoulders, experiencing intense kinaesthetic sensations as Istrain under the
load. But Iam not responding to what Irecognize as another sentient or feeling
person. Ican, if Iset my mind to it, personify the column, think of it as sentient
and imagine its feeling the weight of the building, as Ifeel that weight myself in
imagination. But Idont have to regard the column with an anthropomorphic
eye in order to enjoy the empathetic experience Idescribed. My imagining being
in the situation a column is in suffices to explain my (imagined) kinaesthetic
feelings; there is no need to imagine the column experiencing such feelings.
Simulation or empathy based on something like expressive behavior, an expe-
rience that begins with my imagining (or actually engaging in) expressive behav-
ior, need not be a response to a person or to something Ithink of or imagine as a
person. Certainly infants who mimic smiles at the tender age of several hours are
not responding to what they understand to be other sentient individuals. Imay
find myself swaying with a tree blowing in the breeze, or leaning to one side as
Iobserve the Tower of Pisa. There is something like motor mimicry going on
in these cases, and Imay feel unsettled or unstable, or diagonally inclined. But
I probably am not entertaining anything like the thought that the tree or the
tower experiences feelings or sensations.

47
Empathy, like the German Einfhlung, which it translates, was originally used to charac-
terize experiences of inanimate objects, as well as experiences of other people. See Theodor Lipps,
Einfhlung, Inner Nachahmung, und Organempfindungen, Archiv fr die Gesamte Psychologie 2
(1903):185204; Vernon Lee, Empathy, in Lee, The Beautiful (Cambridge:Cambridge University
Press, 1913).
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 143

These examples are obviously much like instances of infectious emo-

tions or emotional contagion. But these terms would be misnomers, since
the emotion (or feeling or sensation) Iexperience does not derive from any-
one elses experience of a similar one. And when we are infected by the actual
emotions of others, being infected does not necessarily involve being aware
of the people whose emotions infect usbeing aware of them as people.
Perception of the motion of their bodies, thought of as mere physical events,
may initiate motor mimicry on my part and infect me with the feelings in
question.
Our empathetic responses to inanimate objects are often grounds for attrib-
uting properties to them. And there may be something like projecting features
of ones experience onto inanimate objects. Imagining myself in the shoes
of a column may help me to appreciate the forces acting on it, the stresses
and strains it undergoes. That a building is stable, or off balance, or likely to
topple over, may become evident to me when Irespond empathetically to it.
Iimagine being in the position or attitude or situation the building is in. Ithen
find myself imagining experiencing kinaesthetic sensations indicating that my
body is (for instance) about to topple over. As a result, Ihave the impression
that the building I am looking at is about to topple over; I experience it as
being in that condition. This is a rough and dirty method of understanding
our physical environment, one that can easily give mistaken results. But it is
often useful in making the condition of our fellow material objects vivid and
obvious to us.48
If we speak of projection, here, we must do so with care. It is not features of
my experience which Iproject onto the object, but rather properties my expe-
rience represents my body as possessing. My experience is as of the buildings
being (not feeling) like this, where this refers not to the way Ifeel but to what
state Ifeel my body to be in.
There is something more like the projection of features of experience, psy-
chological properties, onto physical objects or inanimate environments than
this, however. To explain what Imean, and to get us back to a consideration of
musical tension, it will be helpful to consider ways in which a picture may depict
a persons mental states, and the ways viewers ascertain what the person in the
picture is feeling or thinking.

48
This point is nicely illustrated in a childrens book, Forrest Wilson, What It Feels Like to Be
a Building (Washington, D.C.:Preservation Press, 1988). Everyone can understand buildings. ...
Buildings feel the same stresses and strains that people do. For this reason you can put yourself in a
buildings place. When you feel what it feels like to be a building, you can talk to buildings and they
will talk to you in building body language. ... PUSH or PULL or SQUASH SQUEEZE DROOP
TUG BEND or BRACE, thats what it feels like to be a building!
144 In Other Shoes

X. Empathy in the Service of Representation

The obvious way to depict a persons inner life is by means of his behavior or
posture or facial expressions. Apicture may portray someone as feeling tense or
nervous by portraying him as fidgeting nervously. Viewers might simply infer
from the fidgeting, using a bit of psychological theory, that the character is ner-
vous (that this is so in the world of the picture). But as with actual people, the
nervousness may be infectious. The characters (depicted) fidgeting may make
viewers nervous, and thereby make it seem to them that he is nervous. They
may experience him as experiencing feelings like their own, as being nervous.
Experiencing him this way does not, of course, require that the viewer notice
what it is about his depicted manner or demeanor that reveals his nervousness.
What produces the affect in a spectator, what makes her nervous, need not be
the characters fidgeting or any other properties he is represented as possessing.
Features of the work itself, considered nonrepresentationally, may do the job.
Ibelieve that this is true of one of Van Goghs self-portraits (fig. 2). Something
about this painting makes me feel, if not nervous, at least somewhat uneasy or
tense. On reflection, it seems to me that it is not, primarily, aspects of the sitters
(depicted) facial expression or bearing or anything else about him that has this
effect, but rather features of the paint on the canvas:the busy brushwork in the
background and on the jacket, the choppiness of the strokes on the face and
beard. But Ithink of my affective response as an empathetic response to the per-
son in the pictureand Ibelieve this is the expected and proper way to think of
it. Ithink of Van Gogh, the sitter, as infecting me with uneasiness or nervousness,
and so Iascribe uneasiness or nervousness to him; Ijudge that (in the world of
the picture) Van Gogh is uneasy or nervous.
The painting makes it fictional that Van Gogh is uneasy, but not by generating
fictional truths about his appearance or facial expressions or demeanor. Features
of the picture that seem not to generate any fictional truths directly, and certainly
none about the sitters appearance, serve indirectly to generate fictional truths
about his state of mind. They do so by virtue of their infecting the viewer with
uneasiness. (I take it to be fictional, true in the picture-world, that Van Gogh is
manifesting his uneasiness, somehow, in his manner or demeanor, although it is
indeterminate how exactly he is doing this. And it is fictional in my game world,
it is part of what Iimagine, that something or other in his manner or demeanor
is responsible for my uneasiness but that Idont notice what it is.)
Similar examples are perhaps more obvious in the case of film music. It is well
known that background music can affect how we understand the characters and
events depicted on the screen. The music can color our perception of the states
of mind and personalities of characters, encouraging us to regard them as sinister
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 145

Figure2 Vincent Van Gogh, Portrait de lartiste, 1889 (RF 1949 17). Paris, Muse
dOrsay, don de Paul et Marguerite Gachet pour le muse du Jeu de Paume en 1949.

or vapid or terrified or determined or blas or panicked.49 The mechanism by

which this happens, Ibelieve, often parallels exactly the story Ihave just told
about the Van Gogh self-portrait. The music arouses in us certain feelings or
sensations. Imagining ourselves to have been infected by the characters we see
on the screen, we ascribe similar feelings to them.
Now consider brushwork or qualities of line, shape, and color like those in the
Van Gogh self-portrait, but in a non-figurative painting, a painting that does not
depict a person or anything else. Such a work may set in motion a process like
that of simulation or empathy, activating, Iwould guess, the same psychological
mechanisms that we habitually use in understanding other people, although in
this case there is no person (actual or fictional) to be the target.

49
See Kendall L. Walton, Mimesis as Make-Believe: On the Foundations of the Representational
Arts (Cambridge, Mass.: Harvard University Press, 1990), 172, 334. Jerrold Levinson sketches
some examples which arc nicely understood in this manner. See Levinson, Film Music and
Narrative Agency, in Post-Theory:Reconstructing Film Studies, ed. David Bordwell and Nol Carroll
(Madison:University of Wisconsin Press, 1996), 24882, at 26063.
146 In Other Shoes

The same may be true of film music without the film, or music that never
was attached to a film. Consider what we might call nervous music, used as
background in a film. And suppose that this music induces us, in the manner
Ijust described, to attribute nervousness to a character on the screen. We
feel nervous and have the impression of being infected by the characters
nervousness. We thus experience him as feeling as we do, and we conclude
that (in the world of the film) he is nervous. If Ihear the music alone, with-
out the film, it is likely to produce much the same affect in me, something
like nervousness, and Iam likely to experience my nervousness as an infec-
tion from outside. Since I dont experience or identify any sentient being
as the source of my infection, my experience is not as of so-and-sos being
nervous, or as of that persons being nervous.50 But Ido, all the same, feel as
though Iam in the company of one or more nervous people, even though
Idont (even in imagination) pick out or identify any such person. Isuggest
describing this experience as an experience as of there being someone or
other who is nervous.
Where, according to my experience, is there a nervous person or persons?
Well, Isuppose the answer is in the musicsomewhere in there. For it is the
music that Iexperience as infecting me with nervousness. Let this count as a
kind of projection of my nervousness onto the music. It explains, in any case, the
naturalness of describing the music as nervous, and it makes recognizing ner-
vousness in music very much like recognizing nervousness in people.
The projection is technically in error; the music is not literally nervous, as
Iam, and to say that it is would be a category mistake. But of course in describing
it as nervous, we dont mean, literally, that it possesses the mental state of being
nervous. The property we do attribute seriously to the music is something like
the property of being apt to produce in listeners an experience as of there being
someone or something who is infectiously nervous.
The listeners experience does not amount to recognizing a fictional charac-
ter or characters in the music, infectiously nervous one(s), nor as personifying
some of the sounds one hears, thinking of them as persons. Icant, in my imagi-
native experience, identify who it is who infected me, not even as the person
who makes me feel nervous. For the source of an emotional infection in real
life can be a group of people, as well as a single person. Anervous crowd may
make me nervous, and it may do so even if no individual member of it alone
does. Indeed, a crowd may be nervous even if none of its members are. If a crowd
is responsible, there will be no such thing as the person who infected me with
nervousness. And there is no way to tell from the infection itself, simply from

50
To avoid quantification over fictional characters, put it this way:Idont, in imagination, identify
a person and experience him/her as being nervous.
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 147

my feelings of nervousness together with the impression that it is an infection

from without, whether a crowd or an individual is responsible. So the infection
doesnt enable me to identify any particular person or group of people. Unless
Imake such an identification by sight or by some other means, my experience
will not be as of so-and-sos being nervous, or as of that persons being nervous,
but merely as of there being someone or other, one or more person or persons
or groups of persons, who is/are nervous.
Music induces emotional infection, often, without any indication of an other-
wise identifiable person. So Ihave an (imaginative) experience as of there merely
being one or more nervous Nellies or nervous crowds around. Since Iuse my
own nervousness to represent the state Iexperience myself being infected by,
Ican characterize my experience as one of there being one or more persons (or
groups) around, who are in this state, the state Iam in.

XI. Musical Tension Again

The reader will have anticipated that what Ihave said about nervous music is
close to what Iwant to say about musical tension. Isuppose that nervousness in
music is one kind of musical tension. But several wrinkles remain.
The tension Ifeel when, in real life, Iempathize with a tense person may be
indistinguishable from the tension Ifeel when Iempathize with a bent bow, a
stretched cable, or a compressed column. Imagining myself in the situation the
person or inanimate object is in, or mimicking its behavior or bearing or
demeanor, Ifeel tense. Ordinarily Ialso see what is infecting me and identify
it as sentient or as inanimate. But the difference is probably not evident from
the infection itself, from my feeling of tension. Tense music infects me with ten-
sion, but the music is likely not to suggest, at all clearly, either a tense person as
opposed to a tense inanimate object or the reverse. So my experience is simply
as of there being tension, there being one or more things or persons or groups
of persons which or who, is or are, tense. Or rather, it is an experience as of there
being something or someone or other, or several such, that/who is/are in this
statethe state Iam in. Musical tension is the property of being apt to elicit an
experience of this kind.
Enjoying the experience Idescribed is a far cry from recognizing a persona,
a fictitious tense person, in the music. Nevertheless, tense music will have to be
significantly humanlike, one might suppose, if musical tension is to be under-
stood along the lines I am suggesting. For it may seem that music can infect
human listeners with tension only if it resembles the behavior or bearing or ges-
tures typical of tense people, even if listeners need not think of or experience it as
being sentient or as containing a person. Emotional contagion or infection may
148 In Other Shoes

appear to occur primarily as a result of motor mimicry.51 Mimicking an infec-

tiously exuberant or anxious person, Ibehave or imagine behaving as though
I am exuberant or anxious, and this tends to make me exuberant or anxious.
Only if music behaves as tense humans do will our mimicking it amount to
behaving as though we are tense.
Some tense music does behave more or less as tense people do; rhythmi-
cally tense music often does. Perhaps most nervous music is recognizably like
nervous humans. (Many recent accounts of musical expression rely heavily on
supposed resemblances between music and human expressive behavior.)52 But
it is not easy to see how dissonant harmonies or dense textures or tense-making
formal and tonal characteristics can be nontrivially similar to the body language
of tense people. And it hardly seems that such features of music provoke any-
thing like motor mimicry in listeners. Iwouldnt know how to go about mim-
icking, with my body, the overlapping of durational segmentation and tonal
segmentation which Carl Schachter holds responsible for tension in the opening
eight measures of Mozarts Piano Sonata K.457.53
Nevertheless, it seems obvious to me that musical tension deriving from dis-
sonant harmonies or dense textures or Schachters overlapping segmentations
is infectious; it tends to make listeners tense. Are there subtle resemblances of
some sort, even if we cant put our finger on them, and something like motor
mimicry? Perhaps in some cases, but probably not in all. Ido not think that emo-
tional infection or contagion must involve motor mimicry or that it must rely on
behavioral similarities.
Consider our experiences of friends in the animal kingdom, and judgments we
make about their mental lives. (The shameless armchair psychology to follow is
meant to engage the readers intuitions, not to be conclusive.) Cats purr and dogs
wag their tails. In so doing, they express or exhibit attitudes or feelings or moods
at least we think they do. Purring cats are comfortable and contented; tail-wagging
dogs are excited, joyful, eager to please. These are familiar if rather unspecific
human states of mind, ones we often experience ourselves and recognize in other
people. But purring and tail wagging are anything but human ways of expressing
these states of mind; people do nothing of the sort when they are contented or
excited or joyful. Yet observers familiar with the ways of dogs and cats would seem
to recognize contentment, excitement, and joy in them just as immediately and
automatically, and just as certainly, as they do in fellow human beings.

51
How, then, does emotional contagion work? A central mechanism seems to be mimicry
(Darwall, Empathy, Sympathy, Care, 265).
52
Cf. Davies, Musical Meaning; Peter Kivy, The Corded Shell: Reflections on Musical Expression
(Princeton:Princeton University Press, 1980); Ridley, Music, Value and the Passions.
53
Carl Schachter, Rhythm and Linear Analysis:APreliminary Study, in The Music Forum, Vol.
IV, ed. Felix Salzer (NewYork:Columbia University Press, 1976), 33034.
 P rojec tiv ism , Empathy, and Mus i cal Te ns i on 149

The theory theorist will say that this familiarity amounts to possession of a
rudimentary theory of feline or canine psychology, which we call upon in order
to identify Fidos or Kittys mental states. Certainly it is implausible to suppose
that we simulate Fido when he shows us his excitement by wagging his tail.
(Simulation might be involved when we judge Fido to be excited because he
hasnt seen his beloved master all day.) But it seems to me that our understand-
ing of dogs and cats may be much like empathetic apprehension of the men-
tal states of other people. We respond affectively to dogs and cats, obviously.
And we often share their feelings, feel with them. We experience (something
like) contentment or excitement, or anxiety or fear, when they do. It is possible
that we first identify the mental states of our feline or canine friends and then
respond empathetically with a like state. But Iam inclined to think that, as in the
case of humans, the empathetic response often comes first; we identify or recog-
nize Fidos excitement or Kittys contentment because we experience something
like it ourselves, because it infects us. And we use our affective states to represent
theirs; we judge Fido or Kitty to feel like this. Emotions or moods are sometimes
contagious across species. Our susceptibility to infection from animals, as well
as from people, is a form of sensitivity to their states of mind.
Emotional contagion triggered by purring or tail wagging does not involve
motor mimicry; the infected person surely does not reproduce the purring or
tail-wagging in herself, not even imaginatively. Memorymemory traces of
experiences with cats and dogs whose purring or tail wagging coincided with
contentment or excitement as evidenced by other, more humanlike behaviors
has a role in the process, no doubt. This shouldnt be a surprise if memory plays
the roles Ithink it does in some of our previous examples.
If Iam right about cats and dogs, there is no need to suppose that infectiously
tense music must resemble people. Listeners may be infected without recogniz-
ing any approximation of human behavior in the music, and without engaging in
motor mimicry. The listener may respond physically to the tension in the music
nonetheless, exhibiting her tension in her own characteristic waysby tapping
her fingers and clenching her fists, for instance, not by doing anything like mod-
ulating to a distant key or becoming harmonically dissonant or acquiring a dense
texture.
Dogs and cats can lead us astray. It would be misleading to say that, just as
there are distinctively canine or feline ways of expressing or exhibiting excite-
ment or contentment, there are distinctively musical ways of expressing or
exhibiting tension. Listeners do not experience tense music as a bizarre creature
expressing feelings of tension by means of dissonant harmonies, dense musical
textures, and modulations to new keys, or as presenting or representing such a
creature. Neither do listeners hear in the music peculiar inanimate objects that
exhibit physical tension by these means, or things possessing indeterminately
150 In Other Shoes

psychological-or-physical tension exhibited or expressed thus. Dissonant har-

monies, dense musical textures, and modulations to new keys neednt be under-
stood to be or to represent behavioral or physical manifestations of tension at all.
Like the brushwork in Van Goghs self-portrait, they work simply by producing
in appreciators an experience of tension felt as an infection from without.54
Unlike Van Goghs self-portrait, tense music need not represent an identifi-
able (fictional) possessor of tension, letalone a person who feels tense. Tense
music can easily be made to represent some such thing or person, however.
Atext or title, or just an appropriate context or tradition in a musical culture may
do the trick. Some kinds of tense music are easily taken to represent a persons
expressing feelings of tension in specific (human) ways, or somethings exhibit-
ing physical tension in a particular identifiable manner, although in many cases it
will be indeterminate how, in the musical fiction, the tension is manifested. But
any such representation requires an imaginative understanding of the music that
goes beyond perceiving, and feeling, its musical tension. Musical tension and
relaxationtheir many varieties and many relativescontribute enormously
to the representational functions music often serves. They have a lot to do with
musics expression of emotions as well, whatever this expression consists in. But
these are optional layers of meaning built on top of such fundamental proper-
ties as those of musical tension and relaxation. It is important to realize how
much there is to music, how much there is of what is moving about it, in several
senses, quite apart from any of the usually recognized kinds of representational
or expressive qualities.

54
R.K. Elliott proposed that we sometimes hear music as if it is an expression of emotion, but
a mode of expression sui generis (R. K.Elliott, Aesthetic Theory and the Experience of Art, in
Aesthetics, ed. Harold Osborne [Oxford: Oxford University Press, 1972], 152. Cf. also Malcolm
Budd, Music and the Emotions [Boston:Routledge and Kegan Paul, 1985], 134; and Jerrold Levinson,
Film Music.) This awkward suggestion is attractive only on the assumption, which Ireject, that it
must be by presenting behavioral manifestations of emotion that music expresses them.
 9

Listening with Imagination

Is Music Representational?

Plato characterized the music of the flute and lyre as mimetic, assimilating it to
painting and poetry.* This attitude contrasts with the modern tendency to distin-
guish music sharply from the representational arts. Eduard Hanslick and others
insist that music is just sound or sound structure, that its interest lies in the notes
themselves, not in stories that they tell or anything that they mean. Peter Kivy
calls music an art of pure sonic design.1 There is, to be sure, explicit program
music. And music sometimes combines with words or images to form a repre-
sentational whole, as in song, opera, film, and dance. But some will set aside
the combinations as impure instances of music, mixtures of music with other
things. And purists dismiss program music as of little intrinsic interest or even
as only marginal examples of music. Music itself, pure, absolute instrumental
music such as Bachs Brandenburg Concertos, Brahmss symphonies, and Anton
Weberns Five Pieces for Orchestra, appears to be quite a different animal from
the standard mimetic or representational arts, such as (figurative) painting and
literature.
Given the strength of purist intuitions, it is disconcerting to discover how
quickly qualms arise. Distinguishing absolute music from program music is not
nearly as easy as one might have expected. There is no sharp line between explicit
and subtle program music, or between subtle program music and music that is
as unprogrammatic as it gets, and one can be puzzled about the location even of
the fuzzy lines. When musicwhat taken by itself would seem to be absolute

*
This is a revision of the third of three Carl G.Hempel Lectures given at Princeton University in
May 1991. It develops a suggestion Ibroached in What Is Abstract about the Art of Music? Journal
of Aesthetics and Art Criticism 46 (Spring 1988):35960 and reprinted in this volume (pp.224227),
and mentioned in Mimesis as Make-Believe:On the Foundations of the Representational Arts (Cambridge,
Mass.: Harvard University Press, 1991), pp. 33536. I am indebted to Karol Berger, David Hills,
Marion Guck, Anthony Newcomb, and Alicyn Warren for helpful discussions and comments.
1
Peter Kivy, Is Music an Art? Journal of Philosophy 88 (October 1991):553.

151
152 In Other Shoes

musicteams up with words or images, the music often makes definite repre-
sentational contributions to the whole, rather than merely accompanying other
representational elements. Opera orchestras and music on the soundtracks of
films frequently serve to describe the characters and the action, reinforcing or
supplementing or qualifying the words or images. Mere titles often suffice to
make music patently representational; indeed Icannot imagine music which an
appropriate title could not render representational. Music stands ready to take
on an explicit representational function at the slightest provocation.
If music can be nudged so easily into obvious representationality, can we be
confident that without the nudge it is not representational at all? Most, if not all,
music is expressive in one way or another, and its expressiveness surely has a lot to
do with its susceptibility to being made explicitly representational. To be expres-
sive is to bear a significant relation to human emotions or feelings or whatever it
is that is expressed. Why doesnt this amount to possessing extramusical mean-
ings, and why shouldnt expressiveness count as a species of representation?2
What is to stop us from saying that exuberant or anguished music represents
exuberance or anguish, or instances thereof?
One possible answer is that music is expressive by virtue of its capacity to
elicit feelings in listeners, and that possessing this capacity doesnt amount to
representation. Arousal theories of expression have not been popular recently.
Theorists typically prefer to locate the feelings expressed in the music rather
than in the appreciator, and so must face the question why the feelings in the
music, those the music expresses, arent represented. Is exuberance in the
music in the way that a train is in a picture of a train? Arousal theories have
obvious difficulties, but there is more to them than is usually acknowledged.
Further considerations which I will adduce shortly do more than raise
qualms; they put the burden of proof squarely on those who would resist Platos
assimilation of music to poetry and painting. They may not, however, cure the
inclination to resist, or rid us of the initial intuition that there is a gaping chasm
of some sort between (absolute) music on the one hand and painting and lit-
erature on the other. Iprefer to understand representation in such a way that
virtually all music qualifies. But we should not be satisfied until we can accom-
modate and explain the contrary inclinations.
No sophisticated theory of representation will be needed to see why music
might reasonably count as representational, however unattractive that conclu-
sion may appear initially. The hard part will be recapturing a sense of musics
purity, understanding how fundamentally music differs from the paradigmati-
cally representational arts, whether or not we count it as representational.

2
This possibility alone shows Kivys peremptory declaration in Is Music an Art? that music is
not representational to be seriously premature.
 Liste ning w ith Imag inati on 153

I. Imagination in Musical Experience

Literary and pictorial representations establish fictional worlds. There is the world
of a story and the world of a picture. Does music have fictional worlds? We might
be tempted to speak of the world of the music when we listen to a Brandenburg
Concerto, for instance, but this may be a world of a very different sort. Story worlds
contain (fictional) ghosts and goblins, or murderers and detectives, or jealous lov-
ers, or tragic heroes. There are people in the world of Bruegels The Peasant Dance
and unicorns in that of the Unicorn Tapestries. But if a Brandenburg Concerto has a
world, it may seem to be one that contains nothing but notes, harmonies, melo-
dies, rhythmic motives, developments, etc.the material of the music itselfnot
fictional characters and fictional events represented by the music. (This neednt
mean that the world is not a fictional one.) One can always construct a world of
the usual sort for a piece of music. One can, if one wants to, make up programs for
the Brandenburg Concertos, tell stories to go with themstories about ghosts and
goblins or murderers and detectives, or whatever one allows the music to suggest.
But such stories seem irrelevant at best to an understanding of the music, and
they are more likely to hinder than to enhance appreciation.
So speak purist intuitions. But lets look further. Fictional worlds are imagi-
nary worlds. Visual and literary representations establish fictional worlds by
virtue of their role in our imaginative lives. The Garden of Earthly Delights gets
us to imagine monsters and freaks. On reading Franz Kafkas story, A Hunger
Artist, one imagines a man who fasts for the delight of spectators. It is by pre-
scribing such imaginings that these works establish their fictional worlds. The
propositions we are to imagine are those that are true in the fictional world, or
fictional. Pictures and stories are representational by virtue of the fact that they
call for such imaginings.3 Music also induces imaginings. If we look carefully,
especially if we are willing to look under the surface, we stand to find more than a
little imagining in our experience of music, even of fugues and sonatas, and many
of our imaginings would seem to be called for by the music. Why doesnt the
content of these imaginings constitute fictional worlds, the worlds of the music?
And doesnt this make the music representational, as literature and painting are?
In what ways does music engage our imaginations? Alarge and diverse range
of cases needs to be considered, although what is to be said about many of them
will depend on how broadly one construes the notion of imagining. Imagining
as Iunderstand it can be spontaneous, nondeliberate, a passive experience rather
than something one does. Dreaming is one kind of imagining. Ialso favor under-
standing imagining in a way that allows for implicit or unacknowledged or

3
This is the central feature of the account of representation Idevelop in Mimesis as Make-Believe.
154 In Other Shoes

nonconscious or subliminal imaginings; one may imagine something without

noticing that one does. To insist that a person imagines only if the thought of
what is imagined occurs to her would be far too restrictive. (This doesnt mean
that it will be easy to ascertain whether a person does engage in a given imagin-
ing when the thought does not occur to her.)
Some imaginings that listeners plausibly engage in are about elements of the music
itself, about tones and harmonies and melodies. Beethovens String Quartet opus 59 #3
opens with a diminished seventh chord, giving the impression that what we hear first
is not its beginning, that it started before we heard it. Perhaps we imagine that it did.
On hearing the arpeggiated Prelude to Bachs G-Major cello suite, perhaps we imag-
ine the intermittently sounding pedal tone to be sounding continuously. (Heinrich
Schenker speaks of imagining in cases like this.4) We may imagine events of a piece to
be causally related in various ways. We speak of one musical idea or event growing out
of another, of one interrupting or interfering with another, of one preparing the way
for another. In many instances we probably imagine that there is a nomological con-
nection of some sort between events without imagining what specifically is the cause
of what. This is enough to explain our expectation that a tonic harmony will succeed
a dominant seventh, for instance, even if, having heard the piece many times before,
we know that the cadence is deceptive.5 We imagine (subliminally anyway) that causal
principles are operating by virtue of which the occurrence of the dominant seventh
makes it likely that a tonic will follow, and on hearing the dominant we imaginatively
expect the tonic, whether or not we actually expect it.6 If, or to the extent that, these
various imaginings are prescribed, we have fictionality.
In our examples so far, the characters in the fictions are elements of the
music. It is fictional that the initial diminished seventh chord of Beethovens
opus 59 #3 was preceded by sounds we did not hear, that the pedal tones of the
Bach Prelude sound continuously, that certain musical events are nomologically
connected in such a way that the occurrence of some makes the occurrence of
certain others likely. But the world of the music is a fictional one, not just part of
the real world, even if it is populated by immigrants from the real world; they
behave, in the musical world, in ways they do not in the real one. (Fictional
worlds contain plenty of real world immigrants. There are novels, stories, yarns,
about Julius Caesar, Napoleon, Richard Nixon, the Civil War, etc.)
Music, in the cases I have described, is like nonfigurative paintings which
present fictional worlds populated by features of the paintings themselves, as

4
Heinrich Schenker, Counterpoint:ATranslation of Kontrapunkt, trans. John Rothgeb and Jrgen
Thym (NewYork:Schirmer Books, 1987), vol. 2, book 2, chap.2, 2, pp. 5659.
5
See Leonard Meyer, Emotion and Meaning in Music (Chicago: University of Chicago Press,
1956).
6
See Edward Cone, Three Ways of Reading a Detective Storyor a Brahms Intermezzo, in
Music:AView from Delft, ed. Robert P.Morgan (Chicago:University of Chicago Press, 1989), p.87.
 Liste ning w ith Imag inati on 155

when it is fictional that one rectangular shape lies in front of another.7 If, as
Irecommend, we count such nonfigurative paintings as representational, much
music will qualify as well. This result need not distress musical purists. They may
be willing to call music representational in a sense that applies also to paintings
of Piet Mondrian and Kasimir Malevich and Frank Stella, provided that they can
still find a way to distinguish both sharply enough from literature and figurative
painting. This may not be easy, however, especially in the case of music. It is not
easy to deny that music often has fictional worlds containing characters that are
not themselves features of the music, as we shall see.
There is a lot of mimicry in music. Instrumental music sometimes mimics vocal
music. Keyboard instruments, percussive though they are, sometimes play can-
tabile.8 Vocal music occasionally mimics instrumental music. Dance rhythms are
used in pieces that are not dances. Stravinsky mimics a baroque musical style in his
Pulcinella Suite and so does Ernst Bloch in his Concerto Grosso. Mere similarities
do not necessarily induce imagining or constitute make-believe. But it is surely not
out of the question that one is to imagine the melodic line of the Adagio Cantabile
movement of Beethovens Pathetique Sonata as being sung, and that it is best played
in such a way as to encourage this imagining.9 Here, as in the previous examples,
the actual sounds of the music belong to its fictional world, but so does a (fictive)
person. It is fictional that a person is singing them.
This is one of many cases in which one has a sense of performers actions by which
they produced the sounds or composers thoughts as they wrote the score. We may
not care what the performer or composer actually did or thought or what feelings she
might actually have been expressing thereby. The impression the music gives of hav-
ing been produced in a certain manner or as being the expression of certain feelings
or emotions may be what we are interested in.10 Joseph Kerman suggests several such
characteristics in the Heiliger Dankgesang of Beethovens String Quartet opus132:

The mystic aura is furthered by the unnaturally slow tempo and the
scoring or, rather, by what seems to be an unnaturally slow tempo on

7
Cf. Mimesis as Make-Believe, 1.8. Ifollow Richard Wollheim in understanding representational-
ity in a way that covers much nonfigurative painting.
8
The left hand begins to sing like a cello. David Lewin, Auf dem Flusse:Image and Background
in a Schubert Song, 19th-Century Music 6, no. 1 (1982):53.
9
The slow section of the third movement of Beethovens Ab-Major Piano Sonata, opus 110, is
titled Recitativo, and obviously mimics a vocal recitative. (Alicyn Warren pointed out this example
to me.)
10
The notion of composers personae comes under this heading. See Edward T.Cone, Persona,
Protagonist, and Characters, in The Composers Voice (Berkeley: University of California Press,
1974), pp. 2526; Jerrold Levinson Hope in the Hebrides, in Music, Art, and Metaphysics
(Ithaca:Cornell University Press, 1990); Fred Maus, Music as Drama, Music Theory Spectrum 10
(1988); and Bruce Vermazen, Expression as Expression, Pacific Philosophical Quarterly 67 (1986).
156 In Other Shoes

account of the scoring. The image is orchestral:forty strings could sus-

tain the hymn at this speed with comfort, but four can bear it only with
a sense of strain, tenuousness, and a certain gaucherie. This Beethoven
certainly wanted. ... Again one thinks of the Great Fugue, another work
in which the instruments are made to outdo themselves, and in which
their unhappy striving is incorporated into the essential aesthetic.11

I take the notation, run amok, in the score of William Krafts Momentum
(1966) to be advice that the performer make the music sound as though he has
run amok.
It is not a large step to regarding music that gives an impression of the com-
posers or performers actions or feelings or thoughts to be representing itself as
the product of a composers or performers acting or feeling or thinking in cer-
tain ways, to be mandating that listeners imagine this to be so.
Expressive music, some say, is music that suggests or portrays or somehow
recalls expressive human behavior, behavior by means of which human beings
express exuberance or anguish or gaiety or agitation or serenity or anger or timid-
ity or boldness or aggressiveness.12 This will include music that represents itself
as resulting from such expressive behavior, but there is no reason to suppose that
music cannot simply portray expressive behavior without portraying itself as the
product of such behavior. In any case, there can be no doubt that some expressive
music is expressive by virtue of connections with human behavior. There is little
strain in thinking of some musical passages as representing, as inducing us to imag-
ine, exuberant or agitated or bold behavior. Vocal music portrays expressive verbal
behavior, including not just the utterance of certain words but the manner of their
utterance, a tone of voice. The expressive quality of an utterance, the tone of voice,
remains in much instrumental music even without the words. Some music has
more or less obvious connections with nonverbal behavior, with physical move-
ment. This is evident in the case of marches and dances, but listeners tendencies
to tap their feet or move in response to rhythmic features of other music as well
suggests that they understand music of many kinds to have some important con-
nection with agitated or calm or determined or lackadaisical behavior.
Where there is behavior there is a behaver. If music represents an instance
of behaving calmly or nervously or with determination, it represents, at least

Also, my Style and the Products and Processes of Art, in my Marvelous Images:On Values and the
Arts (NewYork:Oxford University Press, 2008).
11
Joseph Kerman, The Beethoven Quartets (NewYork:W. W.Norton, 1966), p.256.
12
See, e.g., Stephen Davies, The Expression of Emotion in Music, Mind 89 (1980); R. K.Elliott,
Aesthetic Theory and the Experience of Art, in H. Osborne, Aesthetics (Oxford:Oxford University
Press, 1972); Peter Kivy, The Corded Shell:Reflections on Musical Expression (Princeton:Princeton
University Press, 1980); Levinson, Hope in the Hebrides; Vermazen, Expression as Expression.
 Liste ning w ith Imag inati on 157

indirectly, someone so behaving. So the fictional world contains human

beings, anonymous fictive agents, whether or not the sounds themselves are
characters in it.
The prevalence and variety of imaginings in our experience of music, including
many of the examples Ihave mentioned and others as well, is reflected in the preva-
lence and variety of metaphors we use to describe it. We call passages of music exu-
berant, agitated, serene, timid, calm, determined, nervous. We speak of rising and
falling melodies, of wistful melodies and hurried rhythms, of motion and rest, of
leaps, skips, and stepwise progression, of statements and answering phrases,13 ten-
sion and release, resignation and resolve, struggle, uncertainty, and arrival. Music can
be impetuous, powerful, delicate, sprightly, witty, majestic, tender, arrogant, peevish,
spirited, yearning, chilly.14 Ido not think that metaphorical descriptions always indi-
cate imaginative experiences (even subliminal ones) on the part of listeners, but in
many instances they do. We imagine agitation or nervousness, conflict and resolu-
tion. Sometimes we imagine (somethings) rising or descending. (Or we can easily
get ourselves to imagine thus as we listen to the music; an awareness of this possibil-
ity may color our hearing even when we dont actually engage in the imagining.) The
metaphors purists are least able to avoid, those of tension and release, motion and
rest, seem to me to involve imagination. To appreciate music one must feel tension
and release; one must allow oneself to imagine motion and rest.15
What we have noticed so far seems suspiciously like the beginnings of story
fragments in music, the beginnings of programs. Iconclude this collection of
examples with one which is a little less fragmentary. Consider the opening of the
Adagio movement of Mozarts A-Major Piano Concerto, K.488:16

13
Lewin speaks of the mimesis of a giant question mark. (Auf dem Flusse, p.54.)
14
These last examples are Hanslicks. (On the Musically Beautiful, trans. Geoffrey Payzant
[Indianapolis:Hackett, 1986], pp. 9, 10, 32.)
15
For a discussion of the ways in which metaphors do and do not involve imagination and
make-believe, see my Metaphor and Prop-Oriented Make-Believe, The European Journal of
Philosophy 1, no. 1 (1993). Reprinted in this volume.
16
Iam indebted to Marion Guck for introducing me to this passage and pointing out several of
its interesting features. She discusses it in Taking Notice:AResponse to Kendall Walton, Journal of
Musicology 11, no. 1 (Winter 1993):4551.
158 In Other Shoes

The last half of bar 7 is in the dominant, heading for the tonic, F# minor. But it
doesnt get there for a whilenot until the cadence in bars 11 and 12. Instead, the
dominant goes first to a D-major triad (the submediant), in bar 8.This is a decep-
tive cadence, an instance of Meyers thwarted expectation, but a very special one.
The left hand does go immediately to the tonic, on the first beat of bar8.The right
hand gets there too, but not until the second beat. And by then the bass has moved
down to D.That gives usaccidentally, as it werethe D-major triad instead of
the tonic. The D major is understood later as the dominant of a Neapolitan 6th,
which resolves eventually to the dominant and then to the tonic.
The upper voice is late coming to the Ain bar 8.There are precedents for this
tardiness earlier in the passage. The upper voice was late getting to the A(andF#)
at the beginning of bar 3; in bar 4 it participates in a suspension; in bar 6 it is late
getting to the C#. In the first two cases the bass waits patiently for the soprano
to arrive. But in the second phrase, the bass cant wait. It is locked into a (near)
sequence, which allows no delay. In bar 6, as in bar 8, the bass has moved on,
changing the harmony, by the time the soprano arrives.
One could tell a story to go withthis:

A character, call her Dalia, is going to catch a train. She has a habit of
being late. And in bar 7 she dalliesshes off chasing butterflies. She
dallies so long that she misses the train (the bass), which is on a fixed
schedule and cant wait. But missing the train sets up a fortuitous meet-
ing (D major), which leads to unexpected new adventures (G major).

This is sillylike musical renditions of thunderstorms or locomotives or

feline serenades. The music certainly doesnt tell this story. We arent supposed
to think of the story as we listen. And it is likely to be distracting. But the lateness
of the upper voice, and its dallying quality, the rigidity of the basss progression,
the fortuitousness or accidentalness of the D-major triad, the movement to some-
thing new, are in the music. To miss these is, arguably, to fail fully to understand
or appreciate the music. And talking about the train, the butterfly chase, etc., is
one way of bringing out the lateness, the dallying, the fortuitousness.17 Some of
this at least is a matter of imagining. We imagine somethings being late, prob-
ably without imagining what sort of thing it is. And we imagine a fortuitous or
accidental occurrence. (The imaginings neednt involve saying to oneself that

17
Susanne Langer calls a program for pure music a crutch (Philosophy in a New Key, 3d ed.
[Cambridge, Mass.:Harvard University Press, 1942], pp. 24243). Crutches are sometimes helpful
and sometimes get in the way. This one does no harm as long as the make-believe the story intro-
duces is understood to be prop-oriented in the sense Iexplicate in Metaphor and Prop-Oriented
Make-Believe.
 Liste ning w ith Imag inati on 159

something is late, fortuitous, etc.) If the music told the story, it would certainly
be representational. But why shouldnt it count as representational anyway, as
representing instances of lateness, fortuitousness, etc.? (It would be inadequate
to think of the music as merely indicating or expressing the property of lateness;
it portrays a particular [fictitious] instance of somethings being late, even if
nothing much can be said about what it is that is late. Listeners imagine some-
things being late on a particular occasion; they do not merely contemplate the
quality of lateness.)
Some deceptive cadences consist in a clunking, unprepared-for progression
from V to VI, and give the impression of the composer playing a trick on us; we
can hear him saying, afterwards, Ha, ha; Ifooled you! But in the Mozart exam-
ple, my sense is rather that things just happen, in the natural course of events,
to turn out as they do; the VI chord results from occurrences earlier in the pas-
sage, including the top voices dallying and the basss rigid schedule. Idont think
Iwould have this sense if Iwerent engaged actively, if subliminally, in imaginings
like those Ihave described.
Does the dallying cause or explain the lateness? Try a less dallying melodic
line in bar 7:

If the lateness now seems to you less expected, less inevitable, more in need of
explanation, than it did in the original, this is evidence that in Mozarts version you
implicitly imagined dallying, and that you imagined it to be responsible for the late-
ness. (How could she have missed the train? She was right there when it arrived!)
There is room for disagreement, concerning some of the above examples,
about whether normal or appropriate musical experiences involve imagining or
make-believe in the ways Ihave described. But it is clear that we cannot simply
dismiss out of hand the idea of musical works having fictional worlds. It looks as
though they may have worlds teeming with life, just under the surface at least
like swamp water seen through a microscope. If we follow through on our purist
inclinations to reject stories or images or meanings attached to music as unmusi-
cal, if not childish or silly, we must begin to wonder how much of what we love
about music will be left. Yet if we accept pervasive make-believe in music, the
question of how to account for the evident contrast between (absolute) music,
and literary and pictorial representations, becomes pressing. Our experiences of
music seem shot through with imaginings, yet I, at least, continue to resist the
idea that Bachs Brandenburg Concertos and Brahmss symphonies have fictional
worlds, as Crime and Punishment and Hamlet do.
160 In Other Shoes

II.Differences
If musical works do have worlds, and if they involve very much of the make-believe
Ihave suggested they might, they are zoosfull of life, but discrete bits of life,
each in its own separate cagenot a working ecological system. It is not easy
to make sense of the fictional world of a fugue or a sonata as a coherent whole,
to see what the various diverse bits of make-believe have to do with each other.
It will be fictional that there are instances of upward and downward move-
ment, statements and answers, causes and effects, singing, unperceived sounds,
determined or aggressive or timid behavior; all of these fictional truths jumbled
together with few coherent links among them. There will rarely be a plot line
for the listener to follow, even as brief a one as Imanaged to find inor impose
onthe Mozart passage.
Musical worlds will be radically indeterminate with respect to the identity
and individuation of agents.18 Is it fictional that the agent who behaves aggres-
sively in one phrase is the same one who behaves placidly in the next? Do we
imagine that a single person behaves first aggressively and then placidly, or that
different agents engage in the aggressive and the placid behavior? In an answer-
ing musical phrase, is the character giving the answer different from the one
who made the statement or asked the question, or does the original fictive
speaker reply to herself? Sometimes we may have some sense of how to answer
such questions; often we will not.19
It may seem that the various bits of make-believe do not even belong to the
same fictional world, that the musical work has multiple worlds. But there is no
good way of deciding where one world stops and another begins.
If we think of a musical work as a prop in a game of make-believe, the picture
seems to be that of a succession of momentary skit fragments, unrelated to one
another. This picture contrasts starkly with the profound sense we often have
of the unity and coherence of musical works, the sense that their parts belong
together, that one phrase leads naturally to the next, and that any surprising
sequences ultimately seem to have been justified.20 Perhaps the unity is to be
18
See Fred Maus, Music as Drama.
19
Uncertainty about identity and individuation may be no accident in an aural representational
art. Hearing, in real life, is typically less important than sight in the acquisition of knowledge de re,
knowledge about particular things. On hearing the thunder of a team of galloping horses, Imay have
little notion which clops are made by the same horse and which by different ones. But Ican do much
better, when Isee the horses, in identifying which seen bits of horse belong to the same horse and
which do not.
20
Anthony Newcomb emphasized to me the difference between a high degree of connectedness
among the parts of a whole, and the whole having a coherent or intelligible shape; some theme and
variation movements possess the former and lack the latter. My remarks in this paragraph apply to
both of these varieties of unity.
 Liste ning w ith Imag inati on 161

explained in purely musical terms, even if the elements that are unified include
ones with significant representational roles.21 Compare a wallpaper design con-
taining depictions of objects of many different sortsa truck here, a dinosaur
there, an ice cream cone over therewith no very salient connections among
them. We may be expected to notice various individual depictions, but not to
think about how they are related within the fictional world, nor perhaps even
to think of them as part of the same fictional world. The overall pattern may
still be a highly unified one however, even if its unity does not consist in a uni-
fied fictional world. The depictions may all be in the same representational style,
and the overall design may be formally coherent. Likewise, perhaps, with music.
Musical coherence may consist more in coherence of sound patterns than in
unity of representational content. There may still be representational content, of
course, and it may be important.
If the elements of fictionality in a musical work do not cohere well and the
works unity is based on something else, some may be inclined to deny that these
elements constitute a fictional world. If, rather than telling the Dalia story, the
Mozart passage presents more or less disassociated instances of things being
late, somethings being on a fixed schedule, a fortuitous incident, and a change
to something new, it may seem artificial to attribute all this to a single fictional
world and presumptuous to speak of multiple fictional worlds. Even if a listener
does imagine certain connections among the incidents, these imaginings may
strike one as optional, as not mandated especially by the music itself, and so not
contributing to a fictional world of the musical work. (They may belong to the
world of the listeners imagination, however.)
There is a more important reason to hesitate attributing fictional worlds to
musical works, even while recognizing the rich imaginative component in lis-
teners experiences that Ihave described. Explaining it will require further stage
setting.
Paintings and novels are what Icall props in games of make-believe, hav-
ing much in common with dolls and hobby horses. All of these props provoke
imaginings. The child playing with a doll imagines a baby, as the spectator of a
picture of a dragon imagines a dragon. The imaginings children engage in when
they participate in make-believe are not just about babies and horses, however;
the children imagine about themselves as well. Achild imagines (himself or her-
self) putting a baby to bed or riding a horse. The child belongs to the world of
make-believe; it is fictional, in that world, that he or she puts a baby to bed or
rides a horse. Spectators of pictures and readers of stories also imagine about

21
See Peter Kivy, A New Music Criticism? Monist 63 (1990): 26067, and Auditors
Emotions: Contention, Concession and Compromise, Journal of Aesthetics and Art Criticism 51,
no.1 (1993):911. Iowe the wallpaper example to David Hills.
162 In Other Shoes

themselves. On viewing a picture of a dragon, I imagine (myself) seeing a

dragon. On reading A Hunger Artist I imagine being told (by the narrator)
about a professional faster, or at least Iimagine knowing about such a person.
The appreciator does not belong to the world of the story or the world of the
picture. But the appreciator uses the picture or story in a game of make-believe
which has its own world, one to which the appreciator does belong. In the world
of the picture, the work world, there is only a dragon. But in the world of the
viewers game with the picture, he sees a dragon.
(Absolute) music and the paradigmatically representational arts induce in
appreciators significantly different imaginings about themselves. One differ-
ence is evident in the ways in which music and painting portray space. Pictures
represent spatial properties of things. So, arguably, does musicwhen melo-
dies rise and fall, when there is movement from one key to another, arriv-
als and departures, dense textures, open fifths, etc., or at least in some such
instances. (It is less clear in music than in painting what sorts of things fic-
tionally possess the spatial properties.) But musical space, unlike pictorial
space, is usually presented in a perspective-less manner. It is fictional in my
game with a painting that Ihave a spatial perspective in the fictional world,
that I bear certain spatial relations to the objects in the painting. Fictionally
Isee a mountain towering above me; Iimagine seeing it from below. Or fic-
tionally Isee a ship in the distance sailing toward me. The make-believe spati-
ality of music seems not to give rise to similar imaginings de se. When a rising
melody makes it fictional that something rises, do Iimagine something rising
up toward me from below, or something above me rising away from me? At
what pitch am I? When there is movement from one key to another, where
fictionally am Iin relation to it? Am Ito imagine movement toward or away
from me, or from my left to my right, or what? Answers are hard to come by.
Listeners seem not to have spatial perspectives, even when musical worlds
are spatial. The music appears to have its own separate space, one unrelated
to the listeners space.
Occasional exceptions underscore this point dramatically. The middle sec-
tion of Debussys Nocturne, Fetes, mimics a band approaching the listener
getting louder and louder. We imagine the musicians approaching us. We have a
spatial position in the fictional world of our game relative to them, a spatial per-
spective. The contrast with more usual musical portrayals of space is striking. (It
is curious that crescendos and diminuendos do not very often give the impres-
sion of something drawing closer to us or getting farther away, but they dont.)
Looking ahead, Ishould mention another possibility:that I am what moves, that
Ifeel as though Iam rising, or imagine myself rising, when Ihear a melody as ris-
ing, and that the listener has the impression of moving from one place to another
when the music moves from one key to another.
 Liste ning w ith Imag inati on 163

The independence of musical space from us is linked to a more general

respect in which fictional worlds of musical works, if there are such, appear
often to be isolated from us. The experience of perceiving pictures is, at the most
basic level, an experience of imagining seeing; on looking at a picture of a tree
one imagines seeing a treefrom a certain angle and distance.22 But the experi-
ence of listening to music, even obviously representational music, does not in
general involve imagining hearing (or imagining perceiving of any kind). When
music represents soundssounds of trains, babbling brooks, bird callslisten-
ers imagine hearing these sounds. But much representational music does not
represent sounds at all. Arising melody portraying the ascension of a saint into
heaven doesnt portray the sound of the ascension; Ihave no idea what the ascent
of a saint to heaven sounds like. Music probably does not portray sounds when
it portrays nonvocal behavioral expressions of emotion (aggressive rhythms).
Tension and release in music, resignation and resolve, motion and rest, are likely
not to involve representations of sounds. When no sounds are portrayed, lis-
teners do not imagine hearing things as they listen to the music. They do not
have the kind of perceptual access to the fictional world that spectators have to
the worlds of paintings. One imagines instances of tension and release, but one
probably does not imagine perceiving them. So one does not imagine having a
particular perspective on them.
Things have taken a strange turn. We seem headed toward a conception of
music that Ifind very unattractive, one very contrary to my experience of music.
Ihave emphasized the importance of participation in our games of make-believe
with paintings and novels, the importance of imagining doing and experienc-
ing things in connection with what is represented, e.g., imagining seeing things.
What we seem headed toward is a conception of music involving much less par-
ticipation, much less imagining de se, than there is in painting and literature. It
may be fictional in our games that we know about instances of tension and relax-
ation, and perhaps about behavioral expressions of such states; we may imag-
ine knowing about them. But we dont imagine perceiving them. This is a little
like our response to descriptions, in a literary work, of characters doing various
things and experiencing various feelings and emotions. But many literary works
have narrators, and it is fictional that the narrator tells us about the characters.
Even this is lacking in music. The listener doesnt imagine being told about
someones being tense or relaxed or euphoric or agitated. It would seem to be
indeterminate (in many instances) how fictionally we know. We imagine know-
ing about them without, it seems, imagining anything concerning how we came
to know. We dont imagine having a particular perspective on the fictional world.

Sometimes the distance can be specified only approximately, and sometimes the angle is
22

ambiguous, e.g., in some Cubist works.

164 In Other Shoes

To say there is not much participation is to say that there is not much of a
game world, a world to which we ourselves belong. We would seem then to be
distanced from the events of the work world, the fictional struggles and agita-
tions and tension and release being most prominently a part of a world to which
we do not belong.
My experience of music is not at all like this. My impression is the opposite
of being distanced from the world of the music (if we can call it a world). Ifeel
intimate with the music, more intimate, even, than I feel with the world of a
painting. The world of a painting (as opposed to the world of my game with the
painting) is out there, something Iobserve from an external perspective. But it is
as though Iam inside the music, or it is inside me. Rather than having an objec-
tive, aperspectival relation to the musical world, Iseem to relate to it in a most
personal and subjective manner.
Some will say that, yes, Iam intimate with the music, with the auditory phe-
nomena, but not with a fictional world that it creates. I am sympathetic, but
things are not this easy. My intimacy is not just with sounds; it is with tension
and relaxation in the music, with exuberance and wistfulness and aggressiveness
and uncertainty and resolution. I share the purists skepticism as to whether
these add up to a fictional world like that of a picture or a novel, but we have to
admit that they are part of the stuff that such fictional worlds are made of. And
we should wonder how involvement with mere sounds could be gripping in any-
thing like the way involvement with monsters and dragons, innocent damsels,
evil villains, and tragic heroeseven fictional onescan be. Why should we
care what happens to a four note motive consisting of three eighth notes on a
given pitch followed by a half note a third below? We do not follow the fortunes
of musical motives in quite the way we follow the fortunes of Romeo and Juliet
or Anna Karenina, wishing them well or ill and worrying about what might hap-
pen to them.
I mentioned the idea that it is by portraying vocal or other behavioral expres-
sions of feelings that music portrays the feelings. This is how feelings are, most
obviously, presented in theater and painting. Schopenhauer and others have
claimed that music gets at feelings more directly.23 If music does bypass behavior
and portray feelings directly, the listener will in one sense be closer to the por-
trayed feelings. She will not have to go through the portrayal of the behavior in
order to ascertain what feelings are portrayed. But this does not provide for the
intimacy that Ithink Ihave with the feelings portrayed in music. We dont expe-
rience this intimacy when we read descriptions in a literary work of characters

23
Schopenhauer objects strenuously to musical imitations of phenomena of the world of per-
ception. The World as Will and Representation, vol. 1, trans. E. F.J. Payne (NewYork:Dover, 1969),
p.264. See also pp.257, 25960, 26264.
 Liste ning w ith Imag inati on 165

emotional states (rather than their behavior). Consider a novel that tells us
straight out that a character experiences a warm sense of security.24 The contrast
between the relevant words of the novel and a musical passage expressive of a
warm sense of securityin Brahms, for instancecould hardly be greater.

III. Imaginative Feeling

Lets try something different. Isuggest that music sometimes gets us to imagine
feeling or experiencing exuberance or tension ourselvesor relaxation or deter-
mination or confidence or anguish or wistfulness. This accords with the idea that
music sometimes portrays anguish, not by portraying behavioral expressions of
anguish but more directly, and also with the thought that our (fictional) access to
what is portrayed is not perceptualwe imagine introspecting or simply expe-
riencing the feelings, rather than perceiving someones expressing them. And it
goes a long way toward explaining the intimacy Isaid Ifelt with the anguish in
the music. On reading that a character in a novel experiences a warm sense of
security, one will imagine knowing about an instance of a warm sense of security,
someone elses, but one probably will not imagine feeling this way oneself.25
More needs to be said. Listeners are intimate with the feelings the music
expresses, with exuberance or anguish, for instance, if they imagine experiencing
exuberance or anguish. But this doesnt explain their intimacy with the music, with
the musics exuberance or anguish. Musical experiences are not just experiences
caused by music; they are experiences of music.26 We dont merely hear the music
and enjoy certain experiences as a result of hearing it. One possibility is that music
stimulates imaginings that are in part imaginings about the sounds themselves.
This is almost right but not quite. Listeners imaginings are, in many instances,
about their experience of hearing the sounds rather than the sounds themselves.
The point here is analogous to one concerning pictures. Apicture of a dragon
induces the spectator to imagine seeing a dragon. But a vivid verbal description
in a story about a dragon might do this much. The picture induces the spectator

24
Many literary works indicate characters feelings by means of a narrators description of their
feelings. The following portrayal of a characters warm sense of security seems not to be indirect
even in this way:Chance circumstances which facilitated [Panchos] intentions:the approach of a
ferocious-looking stray dog who frightened Fanny and gave rise to an unmistakable show of courage
on the part of Pancho, which awakened in Fanny a warm sense of security (Mauel Puig, Heartbreak
Tango, trans. Suzanne Jill Levine [NewYork:Random House, 1969], p.85).
25
Not anyway when reading something like the above Puig passage. In other cases one might
empathize with a characters feeling of warm security, imagining feeling this way oneself. Iwill discuss
such instances shortly.
26
See Malcolm Budd, Music and the Emotions (NewYork:Routledge, 1985), p.123.
166 In Other Shoes

also to imagine of her actual visual experience of the picture that it is her visual
experience of a dragon. Ones seeing of the picture is not just a stimulus but part
of the content of ones imaginative experience. Anguished or agitated or exu-
berant music not only induces one to imagine feeling anguished or agitated or
exuberant, it also induces one to imagine of ones auditory experience that it is
an experience of anguish or agitation or exuberance.
R. K. Elliott, in a perceptive and suggestive essay, describes what he calls
experiencing music from withinexperiencing it as if it were our own expres-
sion and feeling the expressed emotion non-primordially.27 Elliotts character-
ization of this experience is sketchy, but it clearly has much in common with the
experience Ihave described. He has the listener enjoying something akin to an
experience of the emotion expressed, not just (somehow) observing the emo-
tion in the music. Arousal theories, which in their crudest and least plausible
form say that music expressive of exuberance or grief is simply music that makes
listeners exuberant or stricken with grief, at least recognize that to appreciate
expressive music is to feel something. Neither Elliott nor Ithink the apprecia-
tor, in general, simply and straightforwardly experiences the emotion the music
expresses. Malcolm Budd encourages Elliott in my direction by suggesting that
feeling an emotion non-primordially be explained in terms of make-believe.28
(Elliott himself speaks of imaginatively enriched perception.) But in other
respects Budd develops Elliotts account in ways I do not find plausible (and
at least some of which Elliott need not accept)before finally dismissing it.
To experience music [either from within or from without] as if it were the
expression of emotion it would be necessary . . . to imagine someone giving
voice to the sounds of the music and in doing so to express his emotion, Budd
says. If Iexperience [a piece of music] M from within then Imake-believe that
Ifeel [an emotion] E and that Iam expressing my E in the sounds of M:these
sounds are issuing from me as a consequence of my feeling E and they bear the
imprint of E.29 Budd rightly observes that we do not very often hear expressive
music in this way. The kind of experience Ihave described doesnt involve imag-
ining oneself producing the sounds of the music, or imagining the emotion to be
expressed in sound-emitting behavior at all.30 One neednt imagine expressive

27
Elliott, Aesthetic Theory and the Experience of Art, p.152.
28
Budd, Music and the Emotions, pp. 127131. [It will be clear from my Thoughtwritingin
Poetry and Music (this volume) that Ido now find an approximation of this idea plausible, provided
that by the sounds of M we mean not the particular sound (tokens) that one hears, but sounds of
the same type. Indeed, Ithink listeners very often think of themselves as expressing feelings or emo-
tions by means of sounds of the kind they hear.]
29
Budd, Music and the Emotions, pp.131, 135.
30
Elliott makes it clear that expression of an emotion is, for him, not limited to behavior (Aesthetic
Theory and the Experience of Art, p.146).
 Liste ning w ith Imag inati on 167

behavior of any kind, nor anything at all about the sounds one hears. One imag-
ines experiencing the emotion, and one imagines ones experience of the sounds
to be ones experience of it.
Music sometimes induces actual feelings and sensations in listeners, not just
imaginings of such, and it sometimes affects our actual moods, if not our emo-
tions. There are tricky questions about how best to describe the various effects
music has on listeners. Alot will depend on what intentional content one takes
the various psychological states to involve. It is more plausible to say that music
makes listeners tense or relaxed or exuberant or agitated, in ordinary instances,
than that it arouses in them genuine, as opposed to imagined, anguish or deter-
mination or confidence or pride or grief (although the experience of vividly
imagining feeling anguish or determination or grief is likely itself be an emo-
tional one). Aperson who is deeply depressed might only imagine being exu-
berant when she listens to an exuberant fugue, whereas an originally cheerful
listener might be made genuinely exuberant. Idont doubt that even someone
who is depressed may become genuinely less depressed as a result of imagining
being exuberant in response to the music. Rather than trying to sort out merely
imagined feelings from genuine ones, Iwill understand the notion of imagined
feelings to include genuine ones as well. Music that induces me actually to feel
exuberant, thereby induces me to imagine feeling thus, and music might induce
me merely to imagine feeling anguish when Idont really.31
That listeners experiences of music include such imagined feelings (whether
they are actual or not) fits nicely with the tendency of music to elicit behavioral
responses like foot tapping, dancing, or swaying with the music. Some people
sing along with music; some are inclined to sing along but know better. When
music swells, one may swell with it. These are the beginnings of behavioral
expressions of feelingsfeelings of exuberance, agitation, gaiety, anguish, pride.
They may manifest, if not actual feelings of these kinds, the vivid imagining of
experiencing them. Compare the filmgoer who suddenly tenses, and perhaps
screams involuntarily, as he imagines being attacked by a slime and being ter-
rified. In this case also, vividly imagining experiencing certain feelings is inti-
mately tied up with behavior expressive of the feelings; the imagining elicits the
expressive behavior, and the person imagines this behavior to express actual feel-
ings of the kind in question.
I tried earlier to explain how music might portray feelings directly, rather
than via their behavioral expressions. Now we see that behaviorthe listeners

31
The affections of the will itself, and hence actual pain and actual pleasure, must not be excited,
but only their substitutes, that which is in conformity with the intellect as a picture or image of the
wills satisfaction, and that which more or less opposes it as a picture or image of greater or lesser pain
(Schopenhauer, The World as Will and Representation, 2:451. Italics in original).
168 In Other Shoes

expression of the feelings she imagines experiencinghas come back into the
picture. But this does not mean that the music portrays an instance of the feel-
ing by portraying someones behavioral expressions of it. It may be that some
music which can be taken to portray a person behaving in certain ways and
thereby expressing certain feelings, might alternatively be understood to involve
the listeners behavior. (It might be understood in both ways, as we shall see.)
But it doesnt portray the listeners behaviorcertainly not in anything like the
way a film about me portrays my behavior if it shows me expressing my feel-
ings. Behavior is likely to come into the picture because in listening I, or any of
us, may be induced to imagine not only experiencing certain feelings but also
expressing these feelings behaviorally.
Wallace Stevens describes the intimate connection between hearing music
and imagining feeling (or actually feeling) as follows. Notice that he doesnt
indicate anything about observing, even to empathize with, a consciousness dis-
tinct from the listener, presented or represented or portrayed or suggested by
themusic.

Just as my fingers on these keys

Make music, so the selfsame sounds
On my spirit make a music, too.
Music is feeling, then, not sound;
And thus it is that what Ifeel,
Here in this room, desiring you,
Thinking of your blue-shadowed silk,
Is music. It is like the strain
Waked in the elders by Susanna.
Of a green evening, clear and warm,
She bathed in her still garden, while
The red-eyed elders watching, felt
The basses of their beings throb
In witching chords, and their thin blood
Pulse pizzicati of Hosanna.32

Sounds are curiously unusual in their tendency to elicit responses like foot
tapping and singing along. Visual designs in motion have no comparable ten-
dency. Ihave no inclination to tap my feet or dance or even sway back and forth
to abstract motion pictures, even ones with a beat, a regular persistent rhythm.
It is hard to imagine jiving with a blinking traffic light, or even a battery of traffic

32
Wallace Stevens, Peter Quince at the Clavier, in Collected Poems (NewYork:Knopf, 1951),
stanza I.
 Liste ning w ith Imag inati on 169

lights operated by a jazz percussionist. People who dance at sound and light
shows dance to the sound, not the light.
There may be reasons why sounds are better suited than sights to play the
role Ihave claimed sounds often do play; reasons to expect that, if there is an
introspective art of the kind Ihave described, as opposed to a perceptual one,
it is more likely to be a sound art than a visual one. Aural experiences may be
better suited than visual ones to count as, fictionally, experiences of feelings or
emotions; experiences of sounds, as we construe them, may be more naturally
imagined to be experiences of feelings or emotions than experiences of sights
are. How might this be?
Here is an easy first point:Hearing is something we cannot easily turn off; we
cant close our ears as we can close our eyes. The same is true of our introspective
sense. We cant simply turn off at will feelings of agitation or serenity or anguish
or a sense of foreboding or of well-beingor our access to such feelings. In this
and other ways also, seeing is a more active sense than hearing is. When our
eyes are open we choose what to look at. Short of moving to a different location,
we cant choose what sounds we hear, although we can to some extent ignore
certain sounds and pay attention to others. Likewise, we dont have much direct
control over what we feel, short of changing our situation or circumstances. We
can concentrate on some feelings and ignore others, to an extent. But there is
nothing in feeling, any more than in hearing, much like looking at one thing
rather than another.
Other analogies between hearing and feeling are related to what I call the
Cavell-Calvino observation about sounds. Stanley Cavell remarked that we think
of sounds as independent entities separate from their sources, in a way we do
not think of sights; we reify or objectify sounds.33 We speak of clatters, bangings,
whinnys, murmurs, echoes, creaks, clangs, rustles, grumbles, gurgles. Sounds
like smellsfill rooms and cross streets; sights dont do that. Italo Calvino puts
the point well in A King Listens:

The music comes and goes, in gusts, it oscillates, down in the rumbling
groove of the streets, or it rises high with the wind that spins the vanes
of the chimneys.
And when in the darkness a womans voice is released in singing,
invisible at the sill of an unlighted windowWhat is it? Not that song,
which you must have heard all too many times, not that woman, whom
you have never seen: you are attracted by that voice as a voice, as it
offers itself in song.

Stanley Cavell, The World Viewed, enlarged ed. (Cambridge, Mass.:Harvard University Press,
33

1979), chap.2.
170 In Other Shoes

That voice comes certainly from a person; a voice, however, is not a

person, it is something suspended in the air, detached from the solidity
of things.34

We reify or objectify feelings and sensations, as we do sounds, and we

conceptualize them and our relations to them in similar ways. We think of
feelings of exuberance or anguish as entities distinct from their sources, and
sometimes as leaving their sources and surrounding or entering us. A feel-
ing, like a sound, may come over me. It may permeate my consciousness.
Both feelings and sounds wax and wane, independently of changes in their
sources. They get more intense, or diminish and then disappear; they can
be overwhelming or hard to detect. The same loss may cause grief of more
or less intensity in different people, just as the same train causes louder or
softer sounds at different distances. Sounds and feelings are both individu-
ated sometimes by their sources, and sometimes by their perceivers. And we
are of two minds about whether either can exist unperceived. We reify feel-
ings and sounds complete with their intentional properties (while thinking
of them as distinct from their intentional objects). The sound of a bell may
waft through the house. Anguish about a particular event may eat at me for
months afterwards; determination to succeed as a pianist may permeate ones
life; pangs of jealousy seep into ones consciousness. (There is, Iexpect, a lot
more to say on this score.34a)35
We do sometimes objectify sights distinct from their sourcesglimmers,
reflections, flashes of light, sheens, etc. But it is only in fairly special instances
that we think of things like these as the objects of our vision; usually we think
of ourselves as seeing physical objectstrees, houses, people, mountains. And
a glimmer or a sparkle, if not identical to the thing that glimmers or sparkles, is
thought of as attached to it. Glimmers and sparkles dont come to us from things
that glimmer and sparkle, as sounds do from the sounding object.

34
Italo Calvino, The King Listens, in Under the Jaguar Sun, trans. William Weaver (San
Diego:Harcourt Brace Jovanovich, 1988), pp. 50, 5354.
34a
[Somatic experiences of sounds are feeliings, physical ones. See my Two Kinds of Physicality in
Electronic and Traditional Music, in Bodily Expression in Electronic Music:Perspectives on a Reclaimed
Performativity, ed. Deniz Peters, Gerhard Eckel, and Andreas Dorschel (New York: Routledge,
2012)], II. Reprinted in this volume.]
35
Susanne Langer claims that particular feelings and particular sounds have similar logical struc-
tures. Philosophy in a New Key, pp. 226, 228, 244; Feeling and Form (NewYork:Scribners, 1953),
p.27. But my point is that feelings in general and sounds in general are conceived in analogous ways.
It is possible that conventions of some sort have a place in the explanation of how particular sounds
get associated with particular feelings. And it is possible that the natural affinity Langer finds between
certain sounds and certain feelings is a result of, rather than the source of, the tendency of the sounds
to get us to imagine experiencing those feelings.
 Liste ning w ith Imag inati on 171

It shouldnt be surprising that auditory rather than visual experiences are

imagined to be experiences of feelings of exuberance or anguish or foreboding
or well-being.36
There is no incompatibility between a works inducing an appreciator to
imagine someone else feeling exuberance or anguish, and its inducing the appre-
ciator to imagine feeling this herself. It may do both at once. Visual depictions
often do both when they elicit empathy for a character.
Empathizing with someone, I assume, involves imagining feeling the way
one takes the person to feel. Isee someone slice her hand with a kitchen knife.
I wince and jerk my hand back.37 I am not in pain. But I imagine (spontane-
ously, unreflectively, perhaps subliminally) the knife slicing through my flesh as
I watch it cut her, and I imagine feeling pain. My imagining this explains my
wincing and my jerking my hand back.
What stimulates my empathetic reaction in this case is what Isee happen-
ing to my friend, not, or not directly, how she feels or how she responds. But
other instances are different. If Iam watching my friends face when she is cut,
I might respond empathetically to the pain I see on her face. I may not even
see or know about the accident. One may simply observe a persons facial or
behavioral expressions and find oneself behaving empathetically, spontaneously
mimicking his expressions, contorting ones face as he contorts his. This reaction
is empathetic, and Itake it to consist in imagining feeling what one observes him
feeling. (Maybe one actually does feel as he does, in some respects; some aspects
of the feeling, as well as the expression, may be contagious.) One may empathize
with a dancer or an athlete, feeling ones muscles tense and relax in sympathy.
(When a visual object is or represents a person, it can evoke responses akin to
foot tapping.)
We respond similarly to fictional characters, imagining feeling as we perceive
them (fictionally) to feel. When Ilook at a portrait Imay contort my face in sym-
pathy with the portrayed characters facial expressions, just as Icontort my face
in sympathy with a real person. So a picture may induce me to imagine someone
else feeling anguish or elation and also, at the same time, to imagine (myself)

36
As tone is itself inwardness and subjectivity, it speaks to the inner soul (G. W.F. Hegel, Hegels
Philosophy of Nature, Vol. 2, trans. Michael John Perry [London:George Allen & Unwin Ltd., 1970],
p.71).
[Careful attention to the phenomenology of auditory experience makes the idea that we imag-
ine auditory experiences to be experiences of feelings much more plausible than it might seem at
first. See Bryan J. Parkhursts excellent The First-Person Feeling Theory of Musical Expression,
Postgraduate Journal of Aesthetics 9, no. 2 (2012).]
37
The example is borrowed from Richard Moran.
37a
[My Thoughtwritingin Poetry and Music (this volume) offers an explanation of how music
does this.]
172 In Other Shoes

feeling this way. Iimagine her feeling anguish or elation, and Iimagine feeling
the same in response to her.
When music (vocal music for instance) portrays a more or less definite char-
acter, we are likely to have this kind of empathetic response. But what needs to
be emphasized is that even when there is no definite character in the music, it
can get us to imagine feeling in certain ways. The music swells and Iswell with
it. I imagine feeling anguish or ecstasy as these qualities are expressed in the
music; Iimagine experiencing a sense of foreboding, as the music changes sud-
denly from a major key to the parallel minor. In such cases Iprobably do not
think of the music as portraying a person (distinct from me) who swells or feels
anguish or foreboding. My experience, phenomenologically, does have some
affinity with that of one who watches another persons facial expressions and
responds empathetically. But I may not have much of a sense of empathizing
with someone at all. No doubt this is partly because Ido not imagine perceiving
anyone when Ilisten, and because music is so fuzzy about the individuation of
particulars. The difference between imaginatively recognizing and feeling with
another person and merely imaginatively experiencing certain feelings oneself as
one listens to the music can be very subtle, especially given that the imaginings
are often implicit or subliminal. Iam sure that sometimes there is no fact of the
matter as to which is the case. But it would be a serious distortion of listeners
experiences to suppose that whenever music gets listeners imaginatively to feel,
it must be doing so by eliciting imaginative empathy with a person portrayed in
the music.38
There are other instances in which something gets us to imagine feeling a cer-
tain way without getting us to empathize with anyone. Isee my friends knife slip
and strike the cutting board. Isee a guillotine operate without a victim. In both
cases Iwince and draw back. But Idont observe someone else being hurt. Ido
imagine being cut or guillotined. Afilm of a hurtling roller coaster empty of pas-
sengers may nevertheless get us to imagine riding in it. Just looking at a comfort-
able rocking chair may induce one to imagine sitting in it.
Music differs from these last cases in that it often gets us to imagine hav-
ing certain experiences not by showing us circumstances that would produce
them, but by doing something more like showing or indicating either behav-
ioral expressions of the experiences in question or (somehow) the experiences
themselvesbut still (often) without in any very definite way portraying some-
one (distinct from the listener) exhibiting the expressive behavior or having the
experiences. It is as though the music provides the smile without the cata
smile for the listener to wear. How music manages this trick is a good question.

38
Elliott sometimes characterizes experiencing music, or a poem, from within as putting oneself
in the shoes of another person, e.g., the poet.
 Liste ning w ith Imag inati on 173

But the trick itself, the result, is not mysterious. Music gets us to imagine expe-
riencing a certain feeling, and possibly expressing it or being inclined to express
it in a certain manner. It often does this without getting us to imagine knowing
about (letalone perceiving) someone else having that experience or expressing
it in that manner.

IV. Is Expression Representation?:Game Worlds

without Work Worlds
I have a lot of imagining going on in the appreciation of music, a lot of imagin-
ing de se. There is a game world full of life. But what has happened to the work
world, the fictional world of the music itself? If there were a cat, it would be in
the work world. But if there is only a smile and the smile is mine, the listeners,
maybe there isnt even a place for a cat; maybe there is only a game world. The
work world is supposed to contain fictional truths generated by the music alone.
But the only fictional truths there are may be ones generated by the listeners
experience with the music, ones that belong only to the game world.
In the case of a picture of a dragon, there is a fictional world, the world of the
work, even when no one is looking at the picture. This is a large part of what it
means to say that the picture is representational. The picture, standing alone,
establishes a fictional world. But in music, when appreciation involves imagining
experiencing feelings in the way Idescribed, there are game worlds but no work
worlds. This gives us an important sense in which music is not representational.
Dragons exist independently of anyones perceiving or knowing about them,
if they exist at all. If there is a dragon in my garden, it is there regardless of
whether or not Isee it. When it is fictional that a dragon exists, it is fictional that
it exists independently of its being perceived. So it isnt surprising that we should
understand something that actually does exist independently of perceivers to
make it fictional that a dragon exists. Apicture fills the bill. It is there whether
or not anyone sees it, and it, by itself, makes it fictional that a dragon is there,
whether or not it is fictional that anyone sees the dragon. Work worlds comprise
fictional truths generated by the work alone. But feelings (of agitation, forebod-
ing, ecstasy) do not exist independently of people who feel them. (Even if they
can exist unperceived or unnoticed, they dont exist unowned.) So there is no
pressure to regard the music itself as establishing a fictional world in which there
are feelings (unless it is fictional that there is a person distinct from the listener
whose feelings they are). It is the listeners auditory experiences, which, like feel-
ings, cannot exist apart from being experienced, that make it fictional that there
are feelings. When the listener imagines experiencing agitation herself, there is
no reason to think of the music as making anything fictional. It is the listeners
174 In Other Shoes

hearing of the music that makes it fictional that she feels agitated. The only fic-
tional world is the world of her game, of her experience.
The absence of a work world goes a considerable way toward recovering
the abstractness or nonrepresentationality of music, toward explaining the
impression that music is not representational in the way painting and literature
are. Insofar as music is expressive in the manner Ihave described, it does not
have fictional worlds of the kind that (figurative) paintings and novels do.39 The
music itself is not a prop, as a painting or a novel is. What the music does is to
supply us with experiences when we listen to it, and we use these experiences as
props. It is the auditory experiences, not the music itself, that generate fictional
truths. Ican step outside of my game with a painting. When Ido, Isee the picture
and notice that it represents a dragon, that it calls for the imagining of a dragon
(even if Idont actually imagine this). But when Istep outside my game with
music and consider the music itself, all Isee is music, not a fictional world to
go with it. There are just the notes, and they themselves dont call for imagining
anything.40
The absence of a work world does not, however, prevent the listeners imagi-
nation from running wild, as she participates in her game of make-believe.

39
There is a work world insofar as music portrays characters behaving expressively. And as
Imentioned, most or all music has a fictional world and is representational in the way nonfigurative
painting usually is, though the fictional worlds in these cases differ importantly from the worlds of
figurative painting and literature.
40
It is misleading or worse to say, as Susanne Langer does, that the sounds are symbols for feelings,
even non-discursive symbols (whatever that means). If music has any significance, it is semantic ...
music is about feelings, it is their logical expression (Philosophy in a New Key, p.218). Idisagree.
Langers notion of presenting emotions for our contemplation, however, might be understood as get-
ting us to imagine experiencing them.
 10

Metaphor and Prop Oriented

Make-Believe

Dolls and hobby horses are valuable for their contributions to make-believe.1
The same is true of paintings and novels. These and other props stimulate our
imaginations and provide for exciting or pleasurable or interesting engagements
with fictional worlds. Adoll, in itself just a bundle of rags or a piece of moulded
plastic, comes alive in a game of make-believe, providing the participant with a
(fictional) baby. What in real life is a mere stick enables a child fictionally to ride
around on a horse, the better to chase bandits or stray cattle. Paint on canvas and
print on paper lead us into exciting worlds of mystery, romance, and adventure
and guide our travels through them.
But props are not always tools in the service of make-believe. Sometimes
make-believe is a means for understanding props. The props themselves may be
the focus of our attention, and the point of regarding them as props in (actual
or potential) games of make-believe may be to provide useful or illuminating
ways of describing or thinking about them. Participating in the game may not
be especially fun in itself and we may have little interest in the content of the
make-believe world or the subject matter of our imaginings. A game may be
cooked up simply to clarify or expose features of the props, simply so we can
observe their role in it. This is make-believe in the service of the cognition of
props. I call it prop oriented make-believe, and I contrast it to content oriented
make-believe, whose interest lies in the content of the make-believe, in the fic-
tional world. In Mimesis as Make-Believe Iemphasized the latter, exploring the
ways in which props of various kinds contribute to make-believe activities.2
Iwill focus now on prop oriented make-believe.

1
This is the second of three Carl G.Hampel Lectures presented at Princeton University in May,
1991. Igratefully acknowledge many helpful observations by the audience on that occasion, and by
David Hills and Gideon Rosen.
2
Walton (1990).

175
176 In Other Shoes

Paper airplanes, like hobby horses and toy trucks, serve as props in games of
make-believe. They make it fictional, i.e. true-in-the-world-of-make-believe, that
they are airplanes flying through the air, climbing, diving, landing on a runway,
crashing.3 But the fun of making and playing with paper airplanes does not derive
entirely, maybe not even primarily, from their role in make-believe. Children who
know nothing of actual airplanes and who think of what we call paper airplanes
merely as folded pieces of paper that behave interestingly when thrown, might
enjoy throwing them, watching them glide, experimenting with the effects of dif-
ferent folds on their flight, and so on. Ones interest may be in the paper con-
structions themselves, apart from any make-believe. Frisbees suggest a game in
which fictionally they are flying saucers. But most frisbee enthusiasts seem to be
interested in throwing, catching, and watching the plastic disks themselves, not
in fantasies about space travel.
There is nevertheless a point in calling the paper constructions airplanes and
the plastic disks flying saucers. These are convenient ways of indicating, for those
who know about airplanes and flying saucers, what these toys are and how they
work. The make-believe looks back toward the props themselves, rather than
forward to the fictional truths the props generate; it is prop oriented.
Paper airplanes and frisbees thus differ from such props as hobby horses,
non-flying airplane models (e.g. a model of the Wright brothers airplane), and
the kind of toy trucks that a child pushes around the floor. Merely manipulating
or looking at these things is likely not to be much fun. Ones interest is in the
make-believe to which they contribute, in fictionally riding a horse or observing
the Wright brothers airplane or driving a truck.4 In these cases make-believe
looks forward to the content of the make-believe; it is content oriented.
Where in Italy is the town of Crotone?, Iask. You explain that it is on the
arch of the Italian boot. See that thundercloud over therethe big angry face
near the horizon, you say; it is headed this way. Plumbers and electricians dis-
tinguish between male and female plumbing and electrical connections. We
speak of the saddle of a mountain and the shoulder of a highway.

3
To be fictional is to be (as we say) true-in-a-fictional-world, the world of a game of make-believe
or a representational work of art, for instance. Features of props are understood to make propositions
fictional, to generate fictional truths. It is because a folded piece of paper falls to the ground that it is
fictional that an airplane crashes. What is fictional in a game of make-believe is what participants in
the game are to imagine to be true. Propositions that are fictional can be true as well. It is both true
and fictional that something flies through the air, although it is only fictional, not true, that the flying
object is an airplane. Participants in the game with the paper airplanes are to imagine that an airplane
crashes, when the folded paper falls to the ground. See Walton (1990), Chapter1.
4
One fictionally rides a horse when one behaves so as to make it fictional, true-in-the-world-of-
make-believe, that one rides a horse, e.g. when one prances around the house straddling a hobby
horse, imagining oneself riding a (real) horse.
 Metaphor and P rop O r iented Mak e -B eli e ve 177

All of these cases are linked to make-believe. We think of Italy and the thun-
dercloud as something like pictures. Italy (or a map of Italy) depicts a boot. The
cloud is a prop which makes it fictional that there is an angry face. Male and
female plumbing or electrical connections are understood to be, fictionally, male
and female sexual organs. The saddle of a mountain is, fictionally, a horses sad-
dle. But our interest, in these instances, is not in the make-believe itself, and it is
not for the sake of games of make-believe that we regard these things as props.
Our participation is minimal at best.5 Imagining a boot, while seeing a map of
Italy or seeing it in my minds eye, may help me to understand your explanation
of the location of Crotone. But Idont contemplate the Italian boot in the way
one might contemplate Van Goghs Pair of Shoes or even Ren Magrittes The Red
Model II. Clouds can support extensive participation; one might, on a dreamy
summer afternoon, fictionally examine the furrows of an angry face, wonder
what it is angry about, and so on. One might be caught up emotionally in the
fictional world the clouds present. But such involvement is unnecessary if the
purpose is to identify which cloud you mean to point out. All this requires is
to recognize which cloud can best be understood to be an angry-face-picture.
To do that it may be helpful to have the experience of, fictionally, recognizing
an angry face, but no further participation is called for; there is no need to be
caught up emotionally in the fiction. The plumbing and electrical connections
invite scarcely any participation in the game in which they are understood to be
props, despite its sexual subject matter. The conscientious plumber does his job
without, fictionally, leering at the fixtures. (This plumbing terminology can be
vaguely titillating, however, and it might cause embarrassment, especially when
one comes across it for the first time. These reactions suggest that a certain per-
haps implicit participation in the game may be likely, perhaps even inevitable,
whether or not such participation helps the plumber to keep track of which fix-
tures can be connected to which others.) We may speak of saddles of moun-
tains and shoulders of highways without even thinking of make-believe, letalone
participating in it, although no doubt such thoughts were present when these
expressions were first introduced or learned.
Make-believerecognition of the possibility of make-believe, at leastis
useful in these cases, even if it is not exciting or pleasurable or edifying in ways
games of dolls and games with paintings and novels are. It is useful for articu-
lating, remembering, and communicating facts about the propsabout the
geography of Italy, or the identity of the storm cloud, or functional properties
of plumbing and electrical fixtures, or mountain topography. It is by thinking
of Italy or the thundercloud or plumbing connections as potential if not actual

5
Compare ornamental representations, which involve thinking about a game of make-believe
without participating in it. See Walton (1990), 7.6.
178 In Other Shoes

props that Iunderstand where Crotone is, which cloud is the one being talked
about, or whether one pipe can be connected to another. The purpose is cogni-
tive, but what Ilearn is not about boots, angry faces (or anger), or sex. The sub-
ject matter of the (potential) make-believe is merely useful.
There is nothing profound about the cognitive role of make-believe in these
examples. The facts it helps us to grasp and remember and communicate are
mundane, and the make-believe is dispensable, a mere convenience. There are
other ways of locating Crotone; we dont have to think of Italy as a boot. But
make-believe, we shall see later, plays a more essential and extensive role in our
understanding of props than is apparent from these examples.
Appreciation of visual and literary representations typically involves participa-
tion in content oriented games of make-believe, especially when the appreciation
includes the experience of being caught up in the story or emotionally involved
in the fictional world. But people sometimes find it convenient to devise ad hoc
prop oriented games, often modifications of the standard content oriented ones,
in describing the props themselves, the visual or literary representations, and
their surroundings.6 One might remark, for instance, that the author of a forth-
coming novel murdered several of his characters with a pencil; this may be a way
of pointing out that the author revised the novel so as to exclude those characters.
The remark indicates a (possible) game of make-believe in which revising a novel
in that manner makes it fictional that one kills characters with a pencil.
If the Metropolitan Museum borrows a portrait of Napoleon from the Louvre
for a special exhibit and has it shipped to NewYork on the Queen Mary, one
might observe that Napoleon is a passenger on the Queen Mary, thus invoking
a (possible) game in which the presence of a portrait on a ship makes it fictional
that the subject of the portrait is a passenger. Idont know whether anyone else
has thought of games like this, letalone participated in them. But there is noth-
ing exotic about them, and it takes only the remark that Napoleon is a passenger
on the Queen Mary, in a suitable context, to call the possibility of such games to
mind. There is no need for anyone to explain them.
Here are some other comments that can be taken in similar ways:
This statue isnt the original one. The Germans took the first Flaubert away
in 1941, along with the railings and door-knockers. Perhaps he was processed
into cap-badges.7
Christopher Robin had spent the morning indoors going to Africa and back
[i.e. reading about Africa], and he had just got off the boat and was wondering
what it was like outside, when who should come knocking at the door but Eyore.8

6
Ihave in mind what Icalled unofficial games of make-believe, in Walton (1990), 10.4.
7
Barnes (1984), p.l.
8
Milne (1928), p.9.
 Metaphor and P rop O r iented Mak e -B eli e ve 179

The chair behind the couch is not the stationary object it seems. Ihave
traveled all over the world on it, and back and forth in time. Without mov-
ing from my easy seat Ihave met important personages and witnessed great
events. But it remained for Kirk Allen to take me out of this world when he
transformed the couch in my consulting room into a space ship that roved
the galaxies.9
These examples illustrate the pervasiveness of make-believe in thought and
conversation, the prevalence of hints of, allusions to potential and often frag-
mentary games, in addition to sustained engagement with full fledged, estab-
lished games when we appreciate works of art. They also illustrate how little it
takes to introduce even rather novel games. The quotation from Lindner suffices
to introduce an unusual game in which a patients exotic tales of other worldly
events make it fictional that the psychiatrists chair is a space ship. We are con-
stantly inventing new games of make-believe and communicating them to each
other. This doesnt mean that we actively participate in these games. Many
of them are prop rather than content oriented; our interest being not in the
make-believe itself, but in the props. Thinking of the props as props in potential
games of make-believe is a device for understanding them.

Many remarks that serve to suggest or imply or introduce or call to mind games
of make-believe can themselves be moves in the implied games, acts of verbal
participation.10 In saying Napoleon is a passenger on the Queen Mary Imight
be pretending to assert that he is; it may be fictional, in the game my remark
introduces, that in saying this Iam claiming that Napoleon really is a passen-
ger on the Queen Mary. One may thus call attention to a game of make-believe
by engaging in it oneself. But the speaker need not actually participate in the
game in order to call attention to it. There are different degrees and kinds of
participation, and whether a speaker on a particular occasion does participate
will depend on how we choose to understand this notion. (One relevant consid-
eration will be whether, in saying Napoleon is a passenger on the Queen Mary
the speaker imagines herself to be asserting the literal truth of that sentence. But
it may be none too easy to decide whether she does imagine this.) What matters
is that to say Napoleon is a passenger on the Queen Mary is to say something
which obviously might be said in an act of verbal participation in a game of a
certain salient kind, and that in doing this one implies, suggests, introduces, calls
to mind, that kind of game.
The speaker is probably genuinely asserting something as well, whether or
not she is pretending to assert something. Saying Napoleon is a passenger on

9
Lindner (1954), p.223.
10
See Walton (1990), 10.2.
180 In Other Shoes

the Queen Mary is a colourful way of asserting that Napoleons portrait is stowed
aboard the ship. The colour consists in the utilization of make-believe as a device
for asserting this to be the case; the speaker asserts it by using a sentence that
might be used in pretending to assert that Napoleon is a passenger on the Queen
Mary, whether or not she actually so pretends. She at least alludes to a (possible)
act of pretended assertion, to an act of fictionally asserting, in the implied game
of make-believe, that Napoleon is a passenger on the Queen Mary. She is saying,
in effect, that fictionally to assert this would be, fictionally, to assert something
true, that circumstances are such that it is fictional that Napoleon is a passenger
on the Queen Mary. The circumstance that makes this fictional is the fact that
Napoleons portrait is aboard the ship.
So we have a way of describing the Queen Mary and Napoleons portrait which
depends on thinking of them as props in a game of make-believe of a certain sort.
Likewise, to say That pipe is male is a colourful way (a slightly off-colour way)
of saying that the pipe is designed to fit inside another pipe, that it is threaded on
the outside. The speaker implies a certain sort of game of make-believe in which
being threaded on the outside makes it fictional that a pipe is male. She goes
through the motions, at least, of fictionally asserting that the pipe in question is
male, and in doing so she, in effect, claims it to be fictional that the pipe is male,
i.e. she claims that it is threaded on the outside. The assertion amounts to the
claim that certain circumstances obtain, namely, the circumstances that would
make it fictional that she speaks truly if, fictionally, she asserts the literal truth of
what she says.
Notice that the content of the assertions in these instances, as given by these
glosses, includes no reference to make-believe. The speaker is simply describing
features of the prop or propsfeatures of things that are or would be props in
games of the implied kind. But it is by invoking make-believe that the speaker
says what she does about the props. Interest is focused on the props themselves;
the envisioned make-believe provides a way of describing them.
If, or to the extent that, statements alluding to make-believe can be para-
phrased in ways not involving make-believe, make-believe is not essential to
what is said. But make-believe sometimes has a more essential role in describing
and understanding props than it does in the examples Ihave given. Even so, the
make-believe may be of no particular interest in itself; it may serve merely to
clarify or illuminate the props. But it may be more or less indispensable for this
purpose. It may do more than simply add colour or provide conveniently memo-
rable or vivid ways of saying what could be said otherwise.
Mens restrooms are marked by stylized pictures of men on the doors;
womens rooms by pictures of women. Icons of people in wheelchairs indicate
facilities designed for the use of people with physical handicaps. These pictures
are used in visual games of make-believe, but ones that invite only minimal
 Metaphor and P rop O r iented Mak e -B eli e ve 181

participation.11 On seeing them, one imagines recognizing a woman, or a man,

or a person in a wheelchair; fictionally one does so. But thats about it. And this
minimal participation is not fun or pleasurable or satisfying or exciting in the
way that contemplation of pictures in art galleries is. The point, of course, is to
learn something about the picture itself, the propthat it is a man-picture or a
woman-picture, for instance, and hence an indicator of the mens room or the
womens room. (Our interest doesnt stop at the prop, in this case; our inter-
est in it is instrumental. Nevertheless it is an interest in the prop apart from the
make-believe world it contributes to, so the make-believe is prop oriented rather
than content oriented.)
Iconic signs are usually very stylized and standardized. But Iunderstand that
the restrooms at the Orson Wells Cinema in Cambridge Massachusetts were
marked by assorted stills of Katherine Hepburn and Cary Grant. Lets suppose
that any recognizable woman-picture or man-picture, no matter what its depic-
tive style and no matter what posture or attitude or environment the person is
depicted in, can be used for the purpose. Every womens room door sports a
different woman-picture. Some are in the style of Giotto; others mimic Rubens,
or Vermeer, or Degas, or Picasso. (I will suppose that these are bad Giottos,
Rubens, Picassos, etc., ones we would have little interest in contemplating. Our
objective is still simply to identify womens rooms by identifying signs on their
doors as woman-pictures.) Some womens room signs depict a seated woman
nursing a child; others a woman playing tennis, or giving a lecture, or bathing, or
climbing a mountain, or dancing, or descending a staircase. Some depict only a
womans face in closeup; others a silhouetted female figure in a vast landscape.
The variety of visual designs that serve to mark womens rooms is boggling. To
identify them and distinguish them from those marking mens rooms by char-
acteristics of line and shape would be hopeless. We succeed only if we use the
designs as props at least to the extent of fictionally recognizing a woman, or a
man, only if we see a woman or a man in them. Then we have no trouble. It is
not usually hard to identify woman-pictures in the style of Giotto or Rubens
or Degas as woman-pictures and to distinguish them from man-pictures, if we
engage in the appropriate make-believe. Make-believe is not merely a conve-
nience here, as it is in the case of male and female plumbing connections. Acer-
tain minimal participation in make-believe is essential; the mere thought of
potential games of make-believe doesnt suffice.
Perhaps make-believe is in principle dispensable even here. Perhaps there is
a complicated, disjunctive way of specifying members of the class of womens
room indicators in terms of lines and shapes. (To keep things simple, Iignore
the fact that as styles of depiction change new combinations of shapes may come

11
See Walton (1990), p.296.
182 In Other Shoes

to count as woman-pictures and womens room indicators.) But none of us can

expect to come close to spelling it out. It is hard enough even to say very exactly
what it is about a particular design that makes it a woman-picture; this requires
more artistic ability or a better eye than most of us have, and an ability to artic-
ulate what one intuitively knows which few artists possess. Moreover, even if
we had the relevant line and shape specification before us, the unity of the class
of woman-pictures, the similarity among them, would not be apparent in it. It
seems appropriate to classify these varied visual designs together only when we
see them as woman-pictures.12
It is arguable that the property of being a woman-picture is not identical
with that of possessing the complicated disjunctively specified design property.
Perhaps the former depends on, or is supervenient on, the latter. Some may
hold that being a woman-picture, like colors or moral properties or being funny,
is a response dependent property,13 the relevant response being that of seeing a
woman in the design. The design property is not response dependent, or anyway
it is not dependent on the same response.
By contrast, it would seem that saddle (of a mountain), male connector and
on the arch of the Italian boot are used to attribute properties having nothing
essentially to do with make-believe.14

It will have been evident that some of my examples are instances of metaphor.
Saddle applied to mountains and male applied to plumbing fixtures are meta-
phors in anyones book, dead ones anyway. My other examples may be less com-
fortably thought of as metaphors:Napoleon is a passenger on the Queen Mary,
Crotone is on the arch of the Italian boot, The ugly face in the sky is headed
this way, and There is a man said while pointing toward a mens room sign.
The ground of the distinction is unclear, however. To speak of the saddle of a
mountain is to think of the topography in question as though it is a representa-
tional sculpture, but one whose make-believe is oriented to the prop. It has been
Grand Central Station around here all day is a metaphor that involves thinking
of the household in question as a kind of unwitting theatrical portrayal of Grand
Central Station; one in which, again, the make-believe is prop oriented. The
cases of the Italian boot, the angry face in the sky, and the rest room icons consist
in regarding something as a representational picture whose make-believe is prop

12
Perhaps with enough practice we could learn to recognize woman pictures without either see-
ing women in them or explicitly identifying them by line and shape characteristics; perhaps we could
learn to recognize a shape gestalt which they share.
13
See Johnston (1989).
14
The dictionary defines saddle in the relevant sense as a ridge connecting two higher eleva-
tions, and male as designed for fitting into a corresponding hollow part. (Websters New Collegiate
Dictionary, 1979.)
 Metaphor and P rop O r iented Mak e -B eli e ve 183

oriented. If saddle and Grand Central Station in these contexts are metaphors,
why not also The ugly face in the sky is headed this way and There is a man said
while pointing toward a mens room sign?
I am not going to propose a theory of metaphor. This is because Iam very
unsure what to count as metaphors, and because Iam sceptical about whether
anything like the class of what people call metaphors is a unified one, whether
a single account will work for any reasonable refinement of that class. But Ido
want to explore the applicability of the notion of make-believe to some acknowl-
edged metaphors, and to sketch some advantages of understanding these meta-
phors, at least, in terms of make-believe.14a
Other metaphors that plausibly involve prop oriented make-believe are easy
to come by. Argument is war and the family of metaphors subsidiary to it,
including talk of claims being indefensible, criticisms being on target, winning
and losing arguments, shooting down arguments, attacking and defending posi-
tions, and so on,15 suggest a game in which what people say in the course of an
argument generates fictional truths about acts of war. The arguers or observers
of an argument participate in the game if they take argumentative behavior to
prescribe imagining acts of war, and imagine accordingly. But participation is
not necessary for using and understanding the metaphors; it is enough to rec-
ognize or be aware of the game. The metaphors can work even if no one has
ever participated in the game. The make-believe is prop oriented in that (or inso-
far as) it is the argument that one is interested in, and the make-believe war is
thought of as a device for describing or understanding the argument.
In this case a single game or kind of game crops up intermittently but persis-
tently in many different metaphorical utterances. Other metaphors of this sort
include those deriving from the thought that time is money,16 war metaphors
applied to sports, and sports metaphors applied to war. More localized meta-
phors which also might be thought of as involving prop oriented make-believe
include:Man is the cancer of the earth, Politician Jones started prairie fires on
his campaign trip in the midwest, an orgy of eating, and (at least before they
died) bottle neck, traffic jam, waves of immigrants, chair leg, and mouth of
a river. (Metaphors that strike me as less plausibly amenable to this treatment
include Time flies, Her spirits are rising, She always took the high road in busi-
ness dealings. He knows which side his bread is buttered on, and Happiness is
a warm puppy. Perhaps not all of these are metaphors?)
The general idea is this:The metaphorical statement (in its context) implies
or suggests or introduces or calls to mind a (possible) game of make-believe. The
14a
[David Hills (1997, 2007, 2011)has pursued an account of metaphor along similar lines much
further than Idid.]
15
Lakoff and Johnson (1980).
16
Lakoff and Johnson (1980).
184 In Other Shoes

utterance may be an act of verbal participation in the implied game, or it may be

merely the utterance of a sentence that could be used in participating in the game.
In saying what she does, the speaker describes things that are or would be props
in the implied game. It may be possible in favorable cases to paraphrase what she
says about them with reasonable fidelity. Typically, the paraphrase will specify
features of the props by virtue of which it would be fictional in the implied game
that the speaker speaks truly, if her utterance is an act of verbal participation in it.
There are many variations on the theme, and many differences among meta-
phors. The example of rest room signs suggests that some metaphorical utter-
ances are not paraphrasable, at least not in the way Imentioned, although they
may still amount to descriptions of the (potential) props. Some metaphorical
utterances may not be assertions at all, even if they are declarative in form. And
metaphorical sentences are not always ones that might be used in acts of verbal
participation in the implied games. Nevertheless, we are now in a position to
clarify and explain much that has been said about metaphor. Then we can look
at some of the variations.

Many have taken metaphor to involve the bringing together of two distinct cate-
gories or realms or domains. Nelson Goodman speaks of the (literal) use of predi-
cates in one realm guiding their (metaphorical) application in another.17 We can
think of the two realms as (a) that of the props and the generating facts, and (b)
that of the propositional content of the implied make-believe. The latter is the
home realm of the predicates that are used metaphorically, the realm in which
they have literal application (I. A.Richards vehicle). The former is the new or
target or foreign realm (Richards tenor).
Goodman says little about how the predicates from one realm organize
another. My suggestion is that (in the case of some metaphors anyway) the
mechanism involves our thinking of objects of the new realm as props, as gener-
ating the fictionality of propositions concerning the home realm. The predicates
male and female get applied to plumbing fixtures by means of our thinking of
plumbing fixtures as props which generate fictional truths about sexual identi-
ties. Male applies metaphorically to plumbing connections which make it fic-
tional, in the implied game, that they are male.
This gives some content to talk of seeing or thinking of one kind of thing in
terms of another, under the influence of metaphors, or of metaphors yoking
different kinds of things together. Richard Moran speaks of metaphors getting
us to adopt a perspective, to see one thing as framed by another.18 This framing
effect of metaphors is independent of and prior to the use metaphors sometimes

17
Goodman (1968), pp.7480.
18
See Lakoff and Johnson (1980), p.36; Davidson (1984); Moran (1989), pp.87112.
 Metaphor and P rop O r iented Mak e -B eli e ve 185

have in making assertions. It will be present even when the metaphor is embed-
ded in a context in which it is not asserted, when it is merely a question rather
than an assertion, and when it is denied or negated.
All of this is accounted for if we think of the new perspective, the framing
effect, as consisting in the metaphors implication or introduction or reminder
of a game of make-believe. The health of General Motors is improving implies
a game of make-believe; it gets us to think of corporations as props in a game
(even if we dont participate in the game). It also serves to assert something about
General Motors. But approximately the same game of make-believe is implied
equally by the following: If General Motors health is improving, unemploy-
ment will drop; I wonder if General Motors health is improving. Is the health
of General Motors improving?, General Motors health is not improving. All of
these statements have the same framing effect; all of them introduce essentially
the same game of make-believe. Probably Caterpiller is in robust good health
and Xerox has a slight cold do so as well.
This account of the framing effect of metaphors, of their capacity to get us to
see one kind of thing in terms of another, contrasts with two other tempting pro-
posals. One is that it is a matter of seeing similarities. Regarding things (or states
of affairs) of one realm as generating fictional truths, as prescribing imaginings,
concerning another realm, is not essentially a matter of seeing similarities. Some
principles of generation19 are more or less conventional, and to the extent that
they are, they are likely not to depend on similarities. (For example, halos on
figures in Christian art make it fictional that they are saints.) One might have
thought that metaphors based on conventions cannot be metaphors. Granted,
if there are simply conventions that slide means one thing in photographic
contexts and another in connection with childrens playground equipment, the
conventions merely define distinct literal meanings of the terms. But if there is
a convention to the effect that a ridge connecting two higher elevations makes
it fictional that there is a saddle, we still have a metaphor. Calling a topographic
feature a saddle is not simply to say that it is a ridge connecting two higher eleva-
tions. Calling it this implies the game of make-believe in which the conventional
principle of generation just mentioned holds. In this sense the speaker gets us
to see or think of such ridges as saddles. (Not for the first time, of course; the
convention is a familiar one. But the metaphor reminds us of the game.) The
freshest, most lively metaphors may be ones that introduce games, principles
of generation, that are new to us. But metaphors like saddle (of a mountain) are
not dead in a sense that ought to make us deny that they are metaphors, so long
as they invoke, remind us of, the game of make-believe, familiar though it is. So

19
Principles of generation are principles specifying what features of props make what proposi-
tions fictional (i.e., true-in-the-fictional-world).
186 In Other Shoes

long as they do this, their use as applied to mountains is parasitic on their origi-
nal literal senses, and it is their use in the home realm, their application to riding
equipment, that guides their application to mountain topography.
It seems unlikely that metaphors like high and low pitches, and rising
and falling melodies, are grounded in similarities between pitch relations and
spatial relations, although they may be not merely conventional but in some
way natural. Ispeculate that the association has a lot to do with the fact that
more energy is usually needed to produce higher pitched sounds than lower
pitched ones, just as upward movement requires more energy than downward
movement. To sound a higher note on a wind or string instrument one blows
harder or stretches the string tighter. But in order to understand metaphors
like rising melodies and low tones we neednt know how they came about,
how it happens that we associate pitches and spatial positions as we do. The
utterance is not an assertion of a similarity or natural connection, or a pointing
out of one. All that matters is that these metaphors do pick out for us a game
of a certain sort. (Notice, incidentally, that, if age and familiarity are any indi-
cations, these metaphors are dead as doornails. Yet they remain metaphors.
Their make-believe is activeindeed it is content as well as prop oriented, as
we shall see.)
Many metaphors are not reversible.20 Life is hell is very different from, Hell
is life. But similarity is presumably symmetrical. Life resembles hell in exactly
the respects that hell resembles life. This should make us wary of construing
metaphor in terms of similarity. My proposal explains this irreversibility nicely.
Generates fictional truths about is not symmetrical. Aridge between two higher
elevations makes it fictional that there is a saddle, but the reverse does not hold
(not in the same game anyway).
A second tempting account of what it is to see one kind of thing in terms of
another is that this is a matter of imagining things of the one kind to be of the
other kind.21 This is not my view. On my view it is a matter of taking things of
one kind to prescribe imaginings about things of another kind, not (in gen-
eral) imagining things of the first kind to be of the second. Understanding the
dotted lines of a balloon in a cartoon to prescribe imagining that the words
in the balloon are thought but not spoken, is not to imagine that the dotted
lines have anything to do with unspoken thought; it is not to imagine any-
thing of the dotted lines at all. The lines merely prompt and prescribe cer-
tain imaginings, imaginings about the character whose portrayal the balloons
stem points to.

20
As Richard Moran (1989) points out, p.93.
21
I.A. Richards (1936) speaks of imagining the tenor to be the vehicle, pp.100101 and else-
where. Richards seems to associate this view closely with the idea that metaphors involve resemblance.
 Metaphor and P rop O r iented Mak e -B eli e ve 187

Some props do prescribe and prompt imaginings about themselves, how-

ever. An actor playing Hamlet probably makes it fictional not only that a prince
of Denmark hesitates, but that he himself (the actor) is a hesitating prince of
Denmark. So we are to imagine something about the actor, the propthat he is
a prince of Denmark and hesitates. It may be that the props in the plumbing case
and in the case of the mountain saddle are also objects. Probably participants in
the game are not merely to imagine a saddle, this imagining being prescribed by
features of the mountain, but are to imagine of the ridge that it is a saddle.
It is less clear in other cases that props in games implied by metaphorical
utterances are also objects. Consider rising melody, broken chord, moving to
a new key, wistful melody, a mountain of debt, a healthy (or sick) corpora-
tion, and the sea is laughing. If one were to participate in the game implied by
moving to a new key, what would one imagine to be moving? The piece, the
musical work? Perhaps one would just imagine somethings moving, an instance
of something moving, as one listens to the modulation. It is not easy to see how
one might imagine a corporation to be (literally) healthy or sick. (Nevertheless,
the corporation is the object of interest. It is a prop if not an object of prop ori-
ented make-believe.)
I should mention, again, that understanding and appreciating a metaphor
need not involve any actual imagining in any case. It is enough to recognize the
implied game, to be aware of prescriptions to imagine in certain ways, without
actually so imagining.

The make-believe that metaphors involve is, I have suggested, prop oriented.
Our interest is focussed on the props, on the alien or target realm, the tenor.
The make-believe is a device to clarify or illuminate the props. This may be so
even if make-believe is essential for this purpose. But sometimes we have, even in
cases of metaphor, something more like an intrinsic interest in the make-believe
itself. The props may serve this make-believe, and metaphors may engage their
service. One might want to make prop oriented make-believe a requirement for
metaphor (or for the kind of metaphors that are based on make-believe). But
make-believe can look forward to the content and back to the prop at the same
time. Some metaphors that are said to be essential are Janus-like in this way.
Talk of broken chords in music usually involves simple prop oriented make-believe.
Apassage consisting of broken chords is one that can be understood to make it fic-
tional that chords (simultaneously sounding pitches of a single harmony) are broken
apart. The property by virtue of which the passage makes this fictional is the sequen-
tial sounding of individual pitches of a single harmony. To say that there is a broken
chord in the bass is to say that the bass sounds individual pitches of a given harmony
sequentially. The latter property is likely to be all we are interested in. The only point in
using the metaphor, in invoking the make-believe, is to indicate this feature of the bass.
188 In Other Shoes

Contrast metaphors like high and low notes, and rising and falling (or
descending) melodies. Roger Scruton calls these metaphors essential. We hear
melodies rise and fall, he says, and this is a crucial aspect of musical appreciation.
We dont just transfer the term; we transfer the movement.22
I have no doubt that we do hear at least some rising melodies as rising. And if we
didnt, or if we heard (what we call) rising melodies as falling, our musical experi-
ences would be very different. By contrast, we rarely if ever hear broken chords
as broken. There is no hint of violence in the gently flowing arpeggios, the broken
chords, of Bachs C Major Prelude from the Well Tempered Clavier.
What does hearing a melody as rising (or hearing a melody rise) amount to?
Areasonable first stab would be that it is hearing the melody in a way that involves
imagining an instance of somethings rising. One certainly does this when a rising
melody illustrates a vocal text describing the rising of someones soul into heaven,
or when, in the case of pure instrumental music, the listener tacks onto the music
a story about, lets say, the launching of a space ship. (One might close ones eyes
and visualize the launching, accompanied by the music.) But one can hear a
melody (as) rising without making up much of a story or visualizing something
moving upward. Isuggest that ones hearing of the melody may still involve imag-
ining (an instance of) somethings rising, although this imagining is probably
very inexplicit (the thought that something is moving upward doesnt go through
ones mind) and also indeterminate (there is no answer to the question what sort
of thing one imagines to be moving upward or, probably, how far or fast it moves,
or where it arrives). Could it be that one is just aware, vaguely, of how easy or
natural it would be for the melody to elicit ones imagining of somethings rising?
Imagining somethings rising can be construed as participation in a game
implied by the metaphor, the melody rises. And the listeners interest is in part
focussed on this make-believe. So the orientation of the make-believe underly-
ing this metaphor is to the content as well as to the prop.
Make-believe in this case is not essential in the way it is in the case of the
rest room signs. In the rest room example, minimal imaginative participation
was necessary to the prop oriented function of the make-believe; we cannot
recognize mens and womens room indicators as such without seeing men and
women in the designs. But the experience of hearing melodies rise, hearing them
in a way that involves imagining upward movement, is surely not necessary for
recognizing (what we call) rising melodies and distinguishing them from falling
ones.23 Lets say that pitches with higher frequencies are timper than those with
22
Scruton (1983), pp.9495.
23
It is curious that we have no convenient way of specifying the property of the melody, the prop,
which this metaphor picks out, without using some variant of the metaphor. We even speak of higher
and lower frequencies. But nonmetaphorical predicates can easily be introduced. Nor is there an easy
way of specifying legs of chairs without using that metaphor.
 Metaphor and P rop O r iented Mak e -B eli e ve 189

lower frequencies, and that the latter are tomper than the former. Some melodies,
rising ones, proceed in a timpish direction, or better timpishly; others proceed
tompishly.24
The point of the metaphors is not just to distinguish timper and tomper
pitches and to identify timpish and tompish melodies; the make-believe looks
forward to the content as well as back to the prop. The make-believe world in
which ascendings and descendings occur is of interest in its own right. Although
the metaphors are not essential to the prop oriented function their make-believe
serves, they are important in pointing out and eliciting participation in the
make-believe itself.
These metaphors do look back to the prop. We are interested in the props,
the melodies, independently of their role in make-believe. Important structural
features of musicbalance, contrast, etc.depend on timper and tomper rela-
tions of pitches and timpish and tompish qualities of melodies, apart from the
make-believe our metaphorical ways of describing these properties introduce.
There is the important difference between contrary and parallel motion in coun-
terpoint. There is the significant change when a succession of timpish melodic
fragments suddenly gives way to a strikingly tompish one.
It seems to me that metaphors indicating expressive qualities of music involve
make-believe which, even more obviously than rising melody, are content as
well as prop oriented. Consider wistful melodies, and cheerful, or anguished,
or angry, or calm music. We hear music as wistful or cheerful or angry, i.e. in
hearing it we imagine somethings being wistful or cheerful or angry, and thereby
participate in the implied game of make-believe. And this participation is itself
an important focus of interest. Such expressive properties are also important
to the formal structure of a piece;25 the make-believe is oriented to the prop as
well as to the content. But in these cases, like that of the rest room signs, one
must participate in the make-believe in order to use the metaphors in classify-
ing music or melodies. Icannot specify wistful melodies just by their formal or
acoustic properties, any more than Ican recognize man-pictures by their shape
properties. So the make-believe implied by wistful melody is essential in both
of the ways Imentioned. It is essential in the way the make-believe of rest room
signs is, and also in the way the make-believe implied by rising melody is.
Ordinarily, Ithink, talk of the shape of a sonata movement, where this refers
to its formal structure, is prop oriented only. Talk of the shape of a melody is con-
tent as well as prop oriented. And so is talk of moving from one key to another.

24
Timper and tomper do have a historical connection with spatial terms; Iused spatial terms in
the process of introducing them. But lets suppose that this historical connection is lost in the mists
of history.
25
See Kivy (1990).
190 In Other Shoes

Although descending melody, like rising melody, looks both ways, falling
melody is often oriented to the prop but not the content. We may describe a
melody as falling, although we hear it only as descending, not as falling.

In what sense does a person, on hearing a metaphor or any utterance implying a

prop oriented game of make-believe, become aware of the implied game? In what
sense does the metaphor introduce one to or remind one of a game of make-believe?
In the simplest cases one is made aware of and can articulate the games principles
of generation. Perhaps the game introduced by talk of mountain saddles consists
entirely in the single principle that ridges connecting higher elevations make it fic-
tional that they are saddles. Such talk may make us fully aware of this principle.
But we usually do not have such explicit knowledge even of the most standard
and familiar content oriented games. Inoted our inability to specify the prin-
ciples of our make-believe games involving pictures, our inability to say what
patterns of shape and color constitute pictures of women or pictures of men. In
this instance and in many others, we do not look to the principles for guidance
in our engagement in the game. We do not formulate them for ourselves, and
then use them to decide what fictional truths pictures generate, what partici-
pants are to imagine. The imagining comes first. On seeing a visual design, we
simply find ourselves imagining a frog (for instance) and imagining seeing a frog,
and because of this we take the design to make it fictional that there is a frog,
and to be a frog-picture.26 (The inference is defeasible, of course.) If one wants
to spell out the principles of generation, one would have to read them off from
our practice, noting what sorts of designs induce what imaginings (in normal or
appropriately idealized observers under normal or appropriately idealized cir-
cumstances), and generalizing.27
Although we do not bear in mind specifications of the principles concern-
ing what sorts of designs generate what fictional truths, we do have abilities and
dispositions appropriate to pictorial games. We are disposed to imagine in pre-
scribed ways on viewing designs of relevant sorts, and to recognize what fictional
truths they generate. It is in this sense that we are cognizant of the games.
Picture games (many of them anyway) need no introduction. No one needs
to imply or suggest them or remind me of them for me to be cognizant of them
in the above sense. It does not take someone pointing to a design in a portrait

26
There is more to our response than this. We find ourselves not only imagining seeing a frog,
but also imagining of our actual visual experience of the design that it is our perceiving of a frog. See
Walton (1990), Chapter9.
27
Alternatively, we might understand the fictional truths to be generated by pictures propensities
to elicit certain imaginings in qualified viewers, rather than by their design properties. The principles
of generation might be understood to specify what propensities make what propositions fictional.
(Cf. my discussion of the acceptance rule for dreams, Walton (1990), pp.4449.)
 Metaphor and P rop O r iented Mak e -B eli e ve 191

museum or on a mens room door and saying, that is a man, to activate my

disposition to imagine appropriately in response to pictures and to recognize
what fictional truths they generate. On seeing a picture I (usually) respond
automatically.28
We do need to be prodded to engage in or even to recognize many other
games, however. This is what metaphorical utterances do. Even very familiar
games may not automatically come to mind when Iexperience things that would
be props in them. On observing a ridge between higher elevations, I do not
always imagine a saddle, nor does it always occur to me that the ridge might be
understood to make it fictional that there is a saddleunless someone reminds
me of the game by saying That is a saddle. Imight come across an instance of a
weighty argument, an under the table payment, someones coming out of the
closet, a writing style with punch, or an unfolding melody without the game
of make-believe the metaphor implies occurring to me. Imay need someone to
remind me of the game by using the metaphor.
What metaphors do, in many cases, is to activate relevant dispositions or abil-
ities, rather than to make us aware of the principles of generation. When some-
one describes a writing style as having punch or a melody as unfolding, Icannot
say very well what characteristics of a writing style or a melody make these met-
aphorical attributions appropriate, which ones generate fictional truths about
punches or about somethings unfolding. But I may be prepared to recognize
writing styles or melodies as having punch or as unfolding.
The dispositions that metaphors activate are often far more extensive than
these, and may involve whole families of predicates, not just the one or ones
originally used metaphorically. Acomment that a computer remembers a phone
number may prepare me to think of computers behaving so as to make it fic-
tional that they forget things, that they calculate, make decisions, and even lose
patience or complain about their handlers or give up on a task. Your describing
your household as Grand Central Station might dispose me to describe mine
as Coney Island, or as a cathedral on a Wednesday at midnight. Once some-
one establishes the precedent of describing people as animals by calling Jones a
skunk, we may think of other people as, fictionally, being other animals (a tiger,
beaver, pig, mouse, dinosaur). The remark that we are all in the same boat eas-
ily leads to a recommendation that we all row in the same direction. Metaphors
often function something like the stipulative launching of a (content oriented)
game of make-believe, which then grows naturally beyond the original stipula-
tion. In suggesting Lets let stumps be bears, or pointing toward a stump and

28
This is not to deny that my propensity to imagine appropriately may depend on my having
experienced various pictures in the past. And learning how to read new kinds of picturescubist
ones, for instancemay require additional experience.
192 In Other Shoes

declaring, Watch out for the bear, a child may establish a game in which the
presence of the stump makes it fictional that a bear is there. But the game is
bound to be far richer than this. It may be understood, more or less automati-
cally, that larger stumps count as larger bears and smaller ones as smaller bears,
that an appropriately shaped stump makes it fictional that a bear is rearing on
its hind legs; seeing a stump through the undergrowth will make it fictional that
one sees a bear through the undergrowth, and children can behave in obvious
ways so as to make it fictional that they run away from a bear in terror, or face it
bravely, or offer it a blueberry ice cream cone. Such extensions of the game the
child introduced are more or less inevitable, but it took an introduction to get it
started.
Metaphorical utterances, like stipulated launches of games of make-believe,
enable us to go on in new ways, to apply the predicates used in the original meta-
phor to new cases, and to apply related predicates metaphorically. If possess-
ing a concept consists in such abilities or dispositions to go on, as some have
suggested, metaphorical utterances expand our repertoire of concepts. The new
concepts are concepts of properties we might describe as those of being meta-
phorically metaphorically unfolding, or metaphorically having punch, or
being metaphorically under the table.
In uttering a metaphor one may assert that some such concept applies in a
certain instance. But the introduction of the concept, the metaphors role in
enabling hearers to acquire it, is independent of the assertion. It is part of, or a
result of, Morans framing effect, which a given metaphor and its negation, as
well as the same metaphor in nonassertive contexts, may possess equally. Insofar
as we are unable to specify the features of props by virtue of which a predicate
applies metaphorically to them, insofar as we just go on, we are likely to consider
purported paraphrases of the assertions in terms of such features inadequate.29

Many metaphors, especially the more interesting ones, do not enable us to go

on with assurance. They leave us uncertain or perplexed or in disagreement
about applications of the original metaphorical predicate and others in its fam-
ily. It is very unclear what games are introduced by Juliet is the sun, or by the
description of a musical passage as a rainbow.30 Not only can we not specify
the principles of generation, we are not prepared to identify with any assurance

29
The metaphorical assertion that X is might, however, admit of a paraphrase of the following
form:X is such as to make it fictional in game G that something (possibly X itself) is . Jones is a
squirrel might be paraphrased as Jones has whatever it take to make it fictional in game G that he is
a squirrel. This paraphrase is literal, Ipresume. But it is not the kind of paraphrase people look for. It
does not get rid of the predicates that are metaphorical in the original.
30
Pablo Casals so described a passage of Beethovens AMajor sonata for cello and piano, during
a master class at Berkeley.
 Metaphor and P rop O r iented Mak e -B eli e ve 193

which people are metaphorically the sun and which are not (no matter how
well we know them), or what musical passages are rainbows. Here is another
example:

Art is dead. Its present moments are not at all indications of vitality;
they are not even the convulsions of agony prior to death; they are the
mechanical reflex actions of a corpse submitted to a galvanic force.31

What do moments of art have to be like to be (metaphorically) reflexes of

a corpse, as opposed to convulsions of a person not yet dead? We can neither
say with any confidence, nor can we very well recognize which description is
appropriate for the present moment of art, or for other moments of art in this or
another culture. To the extent that the concept a metaphor introduces is unclear,
it will be unclear what (if anything) the speaker is asserting. But that may not
be the point of the metaphor. Its point may be, in part, to provoke us to think
about what sorts of games along suggested lines might be reasonable or natural
or intriguing.
Even if the nature of the game implied by a metaphor is fairly definite and a
fairly definite assertion is made, the metaphors interest may lie neither in the
assertion, nor in the introduction of new concepts. Consider There was anger in
the rays of the sun.32 Perhaps the game this metaphor introduces is one in which
all sunlight contains anger, in which sunlight always makes it fictional that there
is anger (although one might choose not to participate in or think about the
game in which this is so). If this is right, the metaphor seems not to introduce
any interestingly new concept, any new way of classifying things of the sort that
might serve as props in the game. And the assertion (if there is one) is trivially
true. What is of interest is the game of make-believe itself, but not simply the
content of the make-believe, the fictional truth(s) generated by the suns rays
(roughly, the fact that fictionally there is anger). The make-believe may be con-
tent oriented, but it is prop oriented also. And the interest lies in the combina-
tion of the two views, in the sunlights role as a prop in the envisioned game
(not just a classification of sunlight that thinking of it as a prop might enable
us to make). The metaphor shows us a way of regarding sunlightas making it
fictional that there is anger.
Many other metaphors would appear to be like this one. The sea is laughing
seems likely to be more significant as an expression of a way of regarding the sea
or some manifestations of itregarding it as a prop in the implied gamethan

31
Marius de Zayas (1912). Quoted in Danto (1986), p.81.
32
Mishima (1966).
194 In Other Shoes

as introducing a way of classifying states of the sea or as asserting something

about the sea on the occasion of utterance.
Metaphors thus make such things as sunlight and the sea into something
like representational works of art. AJapanese brush painting of a flower may be
interesting not (or not merely) because of what it makes fictional, but because
of how it makes it fictional, because of the manner in which the brush strokes
work to generate the fictional truths. To see how they do is to regard them in a
special way, and regarding them in this special way is an important part of ones
aesthetic experience of the painting. It is the function of pictures such as the
Japanese painting to serve as props in games of make-believe. This is not in gen-
eral the function of sunlight and the sea. In particular social contexts metaphori-
cal utterances accord them this function. Sunlight and the sea are found objects.
Metaphors do the finding.32a

References
Barnes, J. (1984), Flauberts Parrot. NewYork:McGraw Hill.
[Camp, E. (2009), Two Varieties of Literary Imagination: Metaphor, Fiction, and Thought
Experiments, in Midwest Studies in Philosophy 33.]
Danto, A.C. (1986), The Philosophical Disenfranchisement of Art. NewYork:Columbia University
Press.
de Zayas, M. (1912), The Sun Has Set, in Camera Work 39:17.
Davidson, D. (1984), What Metaphors Mean, in Davidson, Inquiries into Truth and Interpretation.
Oxford:Clarendon Press.
Goodman, N. (1968), Languages of Art. Indianapolis:Bobbs Merrill.
[Hills, D. (1997), Aptness and Truth in Verbal Metaphor, in Philosophical Topics 25.]
[Hills, D. (2007), Problems of Paraphrase:Bottoms Dream, in The Baltic International Yearbook of
Cognition, Logic, and Communication 3.]
[Hills, D. (2011), Metaphor, in Edward N. Zalta (ed.), Stanford Encyclopedia of Philosophy.
Stanford, CA:Stanford University Press.]
Johnston, M. (1989), Dispositional Theories of Value, in the Proceedings of the Aristotelian Society,
Supplementary Volume 63.
Kivy, P. (1990), Music Alone: Philosophical Reflections on the Purely Musical Experience.
Ithaca:Cornell University Press.
Lakoff, G. and Johnson, M. (1980), Metaphors We Live By. Chicago:University of Chicago Press.
Lindner, R. (1954), The Jet-Propelled Couch, in Lindner, The Fifty Minute Hour:ACollection of
True Psychoanalytic Tales. NewYork:Holt, Rinehart, and Winston.
Milne, A.A. (1928), The House at Pooh Corner. NewYork:Dutton.

32a
[My primary focus in this essay is on prop oriented make-believe, emphasizing its role in the
functioning of metaphors. But prop oriented make-believe also grounds what Iconsider the most
plausible variety of fictionalist metaphysical theories. Cf. Walton (2000) and the new postscript to it
in this volume, and Walton (2013).
Elisabeth Camp (2009) misrepresents Hills and my suggestions about metaphor, and the notion
of prop oriented make-believe. The two kinds of imagination that she faults me for ignoring corre-
spond closely to the very distinction Iinsisted on, that between prop oriented and content oriented
make-believe. See Walton 2013.]
 Metaphor and P rop O r iented Mak e -B eli e ve 195

Mishima, Y. (1966), Death in Midsummer. NewYork:New Directions.

Moran, R. (1989), Seeing and Believing:Metaphor, Image and Force, in Critical Inquiry 16.
Richards, I.A. (1936), The Philosophy of Rhetoric. Oxford:Oxford University Press.
Scruton. R. (1983), Understanding Music, in Scruton, The Aesthetic Understanding:Essays in the
Philosophy of Art and Culture. London:Methuen.
Walton, K. L. (1990), Mimesis as Make-Believe: On the Foundations of the Representational
Arts. Cambridge, MA:Harvard University Press.
[Walton, K.L. (2000), Existence as Metaphor? in Anthony Everett and Thomas Hofweber (eds.),
Empty Names, Fiction and the Puzzles of Non-Existence. Stanford, CA: Stanford University
Press. Reprinted in this volume.]
[Walton, K.L. (2013), Metaphor, Fictionalism, Make-Believe. Available on www.oup.com/us/
inothershoes.]
 11

Understanding Humor and

Understanding Music
All of a sudden Iwonder why Ihave to tell this, but if one begins to
wonder why he does all he does do ... why when someone has told us
a good joke immediately there starts up something like a tickling in the
stomach and we are not at peace until weve gone into the office across
the hall and told the joke over again; then it feels good immediately,
one is fine, happy, and can get back to work.
Julio Cortzar, Blow-up

Anthony is an anthropologist doing field work on Mars.* He notices that

Martians laugh on occasion. They emit from time to time, out of the opening
in their bodies from which they speak, a kind of staccato melisma. (I wont
inquire into how Anthony identifies this behavior as laughter.) Anthony doesnt
understand why they laugh when they do; he doesnt see whats funny. So he sets
about trying to find out. He takes careful notes on the circumstances in which
they laugh and those in which they do not. When necessary, he arranges experi-
ments, setting up various situations to see which ones elicit amusement from
the natives. Eventually, Anthony is able to predict when Martians will laugh. He
knows, for instance, that if Martha, or any other mature Martian from the upper
socio-economic class, sees a yellow square shape moving horizontally from left
to right across a movie screen, and if she has not seen such a phenomenon previ-
ously, she will break out in peals of hysterical laughter. For the sake of argument,
let us make the enormously unrealistic assumption that Anthonys predictions
are perfect.
Not only does Anthony predict successfully when Martians will laugh, he
knows what makes them laugh. He knows, for instance, that part of what makes
Martha laugh is the fact that the movement of the square is from left to right; he
has experimented with squares moving from right to left and observed that they

*
Iam indebted to Marion Guck for many helpful suggestions.

196
 Und erstanding Humor and Mus i c 197

elicit nothing but bored stares from Martians like Martha. And he knows that
the size of the squares does not matter, because he has observed that varying
their size does not change Martians responses.
But there is a sense in which Anthony still doesnt understand why Martha
laughs. It is a mystery to him why she should find a yellow square moving from
left to right amusing. So Anthony renews his grant and goes back to work.
Perhaps he investigates Martian physiology or neurology, or the architecture
of their logic boards. He knows what changes the moving squares produce in
Martians sensory receptors, and can trace the electrical and chemical effects of
these changes in other parts of their bodies, and he knows how all of this leads
finally to the up and down bouncing of their laughter. Or perhaps Anthony
comes up with a psychological explanation of a more or less mechanistic sort.
He discovers somehow or other, after painstaking research, that it is because
members of the upper classes have a slight sense of guilt about their position in
society, along with feelings of superiority, that they laugh at yellow, right-moving
squaresthey wouldnt laugh if they didnt feel the guilt or the superiority
and that the experience of observing the squares relieves or lessens their guilt.
Or maybe he discovers that their amusement depends on their realization that
Mars is smaller than Venus. Now Anthony can explain in much more detail why
Martha laughs when she does. But he still doesnt understand. He doesnt see
what is funny about the yellow squares, even to her. It is still a mystery to him
why she laughs.
The point is not that Anthony is not amused himself, that he doesnt laugh.
One need not laugh oneself in order to understand, in the sense in which
Anthony does not, why others laugh. I may not be amused, when others are,
because Iam in a bad mood, or tired, or because Ihave heard the joke too often
previously, or because Iam offended by its racist or sexist undertones, or because
Ihave outgrown that kind of humor. Yet Imay understand perfectly well what is
funny about it, why other people find it funny.
One must get the joke, at least, in order to understand why others laugh.
It is possible to get a joke even if, for one reason or another, one is not amused
by it. But getting the joke, in one sense anyway, is not enough. It may consist of
being aware of a pun or an obscure allusion, or in possessing a crucial piece of
information. (For example, Duchamps Mona Lisa: LHOOQ.) Iam supposing
that Anthony has all of the relevant information of this sort; he knows what the
Martians know. But he is still mystified by their laughter.
Notice that it is possible to understand why someone laughs, in the sense
in which Anthony fails to understand why Martha does, even if one lacks
experimental data of the kind Anthony has. Imay not realize that a comedians
speeding up his rate of delivery would ruin his joke, for example, or that it
plays on my repressed Oedipal feelings, or gives me relief from guilt. Imay be
198 In Other Shoes

utterly unable to predict what sorts of things will cause me and others with
senses of humor like mine to laugh. Yet when we do laugh Iunderstand why
we do, in the sense in question. The kind of explanation Anthony is able to give
of Marthas amusement is no more necessary than it is sufficient for this kind
of understanding.
Anthonys observations and experiments are insensitive to a crucial distinc-
tionthe distinction between what makes Martians laugh, and what they laugh
at. Anthony knows what causes Marthas amusement but not which of the causes
are also objects of it. What makes me laugh when a comedian tells a joke may
be, in part, the timing of his delivery. But Idont laugh at his timing. Television
producers know that laughter on the sound track encourages viewers to laugh,
but it is not the sound track laughter that viewers find funny. It may be only after
Ihave had a beer, or only after Ihave successfully completed a difficult project
and so am in a relaxed mood, that Ilaugh in a certain situation. But it is not the
beer or my completion of the project or my relaxed mood that Iam amused by
or find funny. These are causes but not objects of my amusement. Amusement is
an intentional experience, an experience of something. It is not a mere twinge or
tickle in the stomach that one feels as a result of hearing a good joke.
The kind of understanding Anthony is unable to achieve involves an aware-
ness of what it is that Martha laughs at. This is a little like knowing a persons
reasons for doing something, not just what causes her to do what she does. It
may be misleading to say that Martha has reasons for laughing, however; this
may imply that her laughter is deliberate in the way in which many of our actions
are. Knowing what Martha laughs at is more like knowing why a person is angry
(or jealous, or feels guilty) in the sense of knowing what he is angry at or angry
about, not just what causes him to be angry. If we call these his reasons for being
angry, this need not imply that his anger is deliberate. To understand a persons
anger one must know his reasons for being angry in this sense. And to understand
Marthas amusement Anthony must know what she is laughing at or amused by,
what her reasons are, in this sense, for laughing.
Knowing this is not sufficient for the kind of understanding Anthony is after,
however, nor is knowing what a person is angry about sufficient for understand-
ing his anger. Suppose Martha is fully aware of the objects of her amusement and
tells Anthony what they are, and suppose that he knows she is not mistaken and
is not lying. It may still be a mystery to him why she laughs at what she laughs at.
Anthony needs to understand Marthas reasons for laughing; knowing what they
are is not enough.
What Anthony lacks is a kind of understanding some have called Verstehen,
one that involves an ability to empathize with Martha when she laughs. But
what is that? Ithink we can say at least this much:Understanding what it is like to
be amused as Martha is is related to an ability to imagine being amused oneself,
 Und erstanding Humor and Mus i c 199

to imagine the experience of being amused in the situation in question and by

whatever it is that amuses Martha. (I do not mean simply an ability to imagine
that one is amused, or to suppose that one is.) One can imagine being amused
without actually being amused, of course. But imagining this does involveand
here Iwill have to be rather vagueexercising ones own sense of humor. To
imagine being amused in the way that Martha is and thereby to understand her
amusement Anthony would have to have and make use of a sense of humor that
is in relevant respects like hers. Anthonys observations and experiments do not
provide him with this sense of humor, and they do not enable him to engage in
the relevant imagining. Nor does anything Martha might tell him about what
features of the situation are objects of her amusement explain to him how to
imagine being amused by those features.

I am sure you have guessed by now that Iwill be interested in what analogies
there might be between understanding a joke and understanding a musical
composition
To say that in analyzing a piece of music theorists attempt to understand how
and why it works is not to say very much, but it is fair enough, Isuppose, as
far as it goes. This characterization of analysis is not entirely noncommittal. An
interest in how a piece works is an interest in how it works on or for listeners.
One doesnt merely try to explain the events of the piece, how they happen to
have come about or what they are like in themselves, as one might try to explain,
for instance, the birth of a star or the formation of Yosemite valley. (Or anyway,
Iwill not be concerned with analyses that do only that, ones that do not also seek
to explain how compositions work on or for listeners.) But to be interested in
how a piece works is not to be interested just in listeners experiences, construed
merely as experiences they have as a result of listening to the piece. Their experi-
ences are experiences of the music and to understand them one must understand
what they are experiences ofthe music.
This, incidentally, argues against the identification of pieces of music with
listeners experiences, as some have proposed.1 It is in the nature of these experi-
ences to be experiences of the piece; the distinction between the experience and
the piece is given in the experience itself. (Part of ones experience is a sense of
a difference between what one finds in the piece, and what one imposes on it or
does with it.)2 In focusing on the experience one will have to recognize the piece
of which it is an experience.
1
John Rahn, Aspects of Musical Explanation, Perspectives of New Music XVII/2 (Spring-Summer,
1979), 205, 217, 218.
2
To insist on recognizing the piece as an entity distinct from ones experience of it is not to deny
that the piece is a cultural construct of some sort, rather than something existing independently of
people or human institutions.
200 In Other Shoes

What makes a composition work is (to quote Kerman) what general prin-
ciples and individual features assure the musics continuity, coherence, organiza-
tion or teleology.3 What makes a given composition work may include its key
structure, inversions and retrograde relationships, melodic augmentations and
diminutions, regularities and irregularities of meter and rhythm, relationships
among pitch class sets, qualities of timbre and texture and changes in them, and
so on. It is the job of music theorists, not philosophers, to decide what charac-
teristics are important in what instances, and Iwill not presume to intrude. My
interest now is not in the answers to such questions but in the methods by which
one might arrive at them.
How does one discover what makes a composition work? Shall we run a series
of experiments, altering characteristics of the piece one at a time and asking
competent and experienced listeners, after each change, whether it still works
(or whether it still works in a particular way)? Will we eventually, after many
such experiments, succeed in isolating the characteristics that make it work?
This will not be an easy task, for obvious reasons. There are an enormous
number of distinct variables in even a simple pieceprobably infinitely many
of them; in any case more than one could hope to check out even in a lifetime.
The experimenter will have to make a lot of guesses about which variables are
worth testing. And many of the variables are not independent of one another.
The melodic shape of the bass line cannot be changed much without altering the
piece harmonically. So how will we set up an experiment to ascertain whether it
is the melodys shape or the harmony that helps to make the piece work? This is
not an unfamiliar problem in the experimental sciences, and there are, in prin-
ciple, ways of getting around it. One would have to devise a much more sophis-
ticated and complex experimental procedure than the one Isketched, and the
data would be less direct evidence for the conclusions. One might start by modi-
fying melodies in ways that have only slight harmonic implications, in order to
establish generalizations about what kinds of effects melodies with certain sorts
of shapes tend to have. These generalizations might then ground inferences con-
cerning what part of the effect of melodic changes which do involve significant
harmonic changes is attributable to the former.
But is any procedure of this sort what we want, even assuming that it can be
carried out? Music theorists rarely undertake anything like the kind of experi-
mentation Ihave described (although psychologists of music do). Could it be
that although this experimental method would be the ideal one, it is simply not
feasible; the difficulties of carrying it out are overwhelming. So theorists work
informally, sloppily, unsystematically, in effect guessing at what the results of

3
Joseph Kerman, Analysis, Theory and New Music, in Contemplating Music (Cambridge:Harvard
University Press, 1985), p.61.
 Und erstanding Humor and Mus i c 201

such experiments would be? Idont think so. Even if this kind of experimenta-
tion were feasible, it would be unsatisfying in the way that Anthonys research on
the Martians sense of humor is unsatisfying. It would tell us what it is about the
music that makes it work, what causes it to work, but it would not, not by itself
anyway, enable anyone to understand, in the way that Ithink most analysts want
to understand, why or how the music works, why it moves us.4
An indication of this is the fact that the experimental procedure Ioutlined
both the simple one and more sophisticated variants of itdo not require the
experimenter to exercise her musical intuitions; indeed she neednt even have
any. A Martian anthropologist with no musical sense at all, or none relevant
to our music, could in principle conduct the investigation, using native listen-
ers as subjects, just as Anthony might investigate Martian humor in the man-
ner I sketched earlier without engaging his own sense of humor. The Marian
anthropologist may be at a practical disadvantage in one respect. One would
want to exercise ones musical intuitions in deciding which hypotheses about
what makes a piece work are worth testing. But maybe Martian anthropologists
live for thousands of years, or the Martian anthropologist might be lucky and
accidently pick the right hypotheses early on. In any case, one might in principle
succeed in carrying out this investigation and learn what it has to teach without
exercising any musical intuitions at all.
This point can be obscured by the fact that the subjects of the imagined experi-
ments do use their musical intuitions; they have to tell the experimenter whether
a piece or a variant of it works. And the experimenter might use herself as a sub-
ject, even as her only subject. The experimenter may think of her task primarily
as that of explaining what makes the piece work for her. In that case she is the
obvious subject to use. Ideally, thenif only it were possibleshe will sit down
at the piano and alter the piece bit by bit, asking herself at each step whether it
still works. Eventually she comes up with a list of the features that make it work
for her.
Now Ido think that engaging in this process might help the appreciator to
understand how the music works for her. But this understanding does not con-
sist simply in accepting the results of the experiments, in being able to specify
the relevant features. The experimenter could employ subjects other than herself
and, using ordinary experimental safeguards, extrapolate the results to herself
with at least some degree of confidence. (Compare testing the effects of a drug on
a limited population and extrapolating the results to others.) The experimenters

4
Wittgenstein argues for something very much like this point, in his Lectures and Conversations in
Aesthetics, Psychology, and Religions Belief, edited by Cyril Barrett (Berkeley:University of California
Press, n.d.), pp. 1921. Frank Jackson tells stories that are in some respects like my Martian story in
Epiphenomenal Qualia, Philosophical Quarterly, XXXII (April 1982), 12736.
202 In Other Shoes

own musical intuitions need play no role in this procedure, and it is clear that she
might accept the results, the list of features which make the composition work for
her, without achieving any particular understanding of how it works. There is still a
gap between knowing what features make the music work, and understanding how
and why it works.
But if the music does work on me, dont Ialready understand how it does? If
Ifind a joke funny myself dont Isee what is funny about it? Yes, in a way. Imay
not know the mere causes of my experience, of course, but that is not what we are
usually after anyway. What we are after is an understanding of the character of the
experience, not how it happened to have come about. Afull description of an expe-
rience need not mention its causes, but must specify its objects, its intentional con-
tent. Ido not fully understand what a person is feeling if Iknow merely that she is
angry, or jealous; Imust know what she is angry about or whom she is jealous of.
Ineed not know, however, that her feeling angry or jealous is (in part) a side-effect
of a drug she is taking, or (in part) a result of a chemical imbalance in her brain, or
that it results (in part) from such and such early childhood experience. The feeling
itself might have been the same had it been caused differently. But it would not have
been the same feeling if what the person is angry about were being passed over for a
promotion, for example, rather than an insult from a fellow worker.
When Iam moved by a musical composition I, presumably, have some sort of
awareness of the content of my experience, the particular features of the music that
it is an experience of. But Imay not be able to specify in much detail what they are.
To specify them, to articulate them, is to achieve a better understanding of how
the music works for me (as well as, possibly, to make it work better for me). This,
Ibelieve, is a large part of what many music theorists do when they analyze a com-
position:They spell out the intentional objects of their musical experiences (and
of the experiences of other listeners as well, to the extent that others experience
the music in similar waysor come to do so under the influence of the analysis).
The method is not the experimental one Idescribed, since that method does not
separate mere causes from objects of the experience. And in any case the theorist
has her own vague awareness of the content of her experience to start with.
The process of articulating the content of ones musical experiences is not
unlike that of becoming fully aware of the content of other intentional states
coming fully to realize what exactly one is angry about or worried about or artic-
ulating for oneself why one has qualms about pursuing a certain course of action,
as when a person is able to say, Now Iknow why Ihave always disliked himit
is his pretense of humility.5 Ido not have in mind so much what a person might

5
Arnold Isenberg, Critical Communication, in Aesthetics and the Theory of Criticism:Selected
Essays of Arnold Isenberg, edited by William Callaghan, etal. (Chicago:University of Chicago Press,
1973), p.171.
 Und erstanding Humor and Mus i c 203

do on the psychiatrists couch, as more ordinary and everyday reflections on the

content of ones intentional attitudes. (There would seem to be nothing corre-
sponding to Freuds notion of repression which must be overcome to achieve a
realization of the content of ones musical experiences.) We might also compare
the experience of having a word on the tip of ones tongue, and then trying to fig-
ure out what it is, and, of course, the process of figuring out why we laugh when
we donot just what makes us laugh, but what exactly we laugh at.
The primary means by which one uncovers the content of ones intentional
states is introspection. I am favorably disposed to Boretz observation that
introspection is the crucial testing-ground for the kind of personally inter-
ested musical experience [his] work has addressed and cultivated.6 I am not
taking introspective data as data on which an empirical scientific theory is
built, however (in accordance with Hempels deductive-nomological model, for
instance). My point is rather that achieving the introspective results is itself our
objective, or a large part of it. We seek to explain our musical experiences in the
sense of specifying, spelling out, what they are like, not in the sense of saying why
they are as they are, or how they came about. (If one wants to say that explaining
what something is like is actually a kind of description, Iwill have no objection.)
There is still room for explanations of other sorts, however. It certainly may
be of interest to know what kinds of sound events people, or a particular person,
hears in what kinds of wayswhich ones provoke experiences with what inten-
tional content. And to discover this would Isuppose constitute at least part of
an explanation of how it happens that we hear sound events in the ways we do. If
(as Iassume) objects are also causes, to discover by means of whatever empirical
investigation might be relevant the causes of my musical experience may at least
give me clues as to what the objects of the experience might be, and may thereby
contribute to the introspective task Ijust described.
The result of this introspective taskthe task of articulating what Iam angry
about, or what my musical experience is an experience ofis not a mere piece of
information. Ido not merely acquire the knowledge that what Iam angry about
is the subtly insulting tone of someones remark, or that what worries me about
so and sos heading a certain organization is (lets say) his tendency to ignore cer-
tain kinds of advice. Icould in principle acquire this information about myself by
being told by someone Itrusta psychologist or a friend Ibelieve to be percep-
tive; Imight take someones word for it. Or Imight read and believe a study that
concludes that when people like me are angry or worried in circumstances like
the one Iam in, that is what they are angry or worried about. The result of my
introspection is rather a recognition, an acknowledgment that that is indeed what
Iam worried or angry about. Now Isee that that is so. (It is not easy to say what

6
Benjamin Boretz, The Logic of What?, The Journal of Music Theory XXXIII/1 (1989), 113.
204 In Other Shoes

this means.) Of course, being told by a trusted friend might help me to see this,
to acknowledge it. But seeing it goes beyond merely concluding that what the
friend says it true. Recognizing, finally, what the word on the tip of my tongue
was is more than just knowing what it was.
Likewise, the introspection that may attend devising an analysis of a piece or
examining someone elses analysis, leads ideally to a recognition or acknowledg-
ment that such and such features of the music are included in the content of ones
musical experience, a recognition or acknowledgment that goes beyond acquir-
ing information about what features are part of the content of ones experience.
Imight acquire this information about myself simply by taking the word of a
theorist Iadmire and whose way of hearing Ihave reason to expect will be much
like my own, without engaging in any introspection at all. But upon introspec-
tion Isee for myself that my hearing of the piece corresponds to the analysis,
Icome to understand my experience as being an experience of the features in
question.7
The kind of understanding Anthony fails to achieve in his study of the
Martians sense of humor is the understanding of acknowledgment or recogni-
tion. We now see that it has two aspects. One might, in the first place, see the
humor of a situation. One does this when one acknowledges or recognizes ones
own amusement. (I am not sure one ever is amused without acknowledging that
one is.) It is harder to say what exactly one recognizes or acknowledges when
one sees the joke but is not actually amusedperhaps a blocked or inhibited or
potential reaction of amusement. But it is clear that the understanding does not
consist simply in possessing the information that one would laugh if one were
not tired, or jaded, or offended by the jokes racism, or whatever. The under-
standing consists in experiencing and acknowledging a certain response to the
joke, if not one of actual amusement.
It is a further step, beyond seeing the humor, to articulate just what it is about
the situation that is funny, to recognize or acknowledge the object of ones
amusement. When one does this one understands more deeply why the situa-
tion is funny, why those who laugh at it do so (assuming of course that they laugh
for the same reasons).
Music theorists investigating music of their own culture usually possess an
understanding of the first sort to start with, a recognition or acknowledgment
that the music works (not just acceptance of the fact that it does). Much of their
effort, Isuggest, goes to deepening this understanding in the way Idescribed,
recognizing or acknowledging, and articulating, the content of their musical
experiences. In many cases this makes the experience itself more satisfying. (My

7
See Nicholas Cook, Music Theory and Good Comparison: A Viennese Comparison, The
Journal of Music Theory XXXIII/1 (1989), 128.
 Und erstanding Humor and Mus i c 205

anger may be more satisfying after Ihave fully articulated what Iam angry about,
even if what Iam angry about has not changed.)
My suggestions so far will help clarify the issue of whether only what is or can
be heard in a piece of music is musically significant, and whether the analysis or
explanation of a composition ought to be restricted to specifying features of it
that listeners do or might hear.8
I urge that we avoid a narrow conception of what one hears. We are some-
times angry or jealous without being able to articulate fully what we are angry or
jealous about. We also, even more frequently perhaps, laugh without being able
to say in much detail what it is that we are laughing at. (It might be essential to
our amusement, in some cases, that we not be able to specify its objects very pre-
cisely.) So the content of a musical experience may include features of the music
the experiencer cannot specify. There may not always be a good way of ascertain-
ing what the content of any of these intentional states includes, of acquiring the
capacity to articulate them in their entirety. When possibilities are suggested, we
may recognize or acknowledge them as indeed among the objects of the experi-
ence (as having been so all along). But failure to make such an acknowledgement
cannot be taken as conclusive proof that they are not objects of the experience.
Recognition does not always come easily.
So analyses are probably, more often than one might have thought, specifica-
tions of what we hear. The possibility is open that even the Schenkerian deep
structure of a piece, or the fact that the foreground and middle ground are elabo-
rations of the deep structure, is in fact an unacknowledged part of the content of
musical experiences even of ordinary listeners. Listeners inability to specify the
tone row of a twelve-tone piece is not sufficient to establish that they do not hear
it. An analysis of a piece may amount to a speculation about what might be in the
unacknowledged content of the analysts or other listeners experiences. In arriv-
ing at such speculations, the analyst may depend more on general impressions of
what sorts of things are likely to be heard in what ways than on introspection of
her own musical experience, even if they are speculations about the character of
her own actual experience.

The objective of analysis is not always or only to explicate how listeners in fact
hear pieces, however. Many analyses are designed to explain and encourage new
ways of hearing. The obvious way of doing this is by specifying the intentional
content of the new hearing. (The information in the analysis does not by itself
enable one to hear the composition in the new way or even to understand that
kind of hearing in the sense Ihave been concerned with, but it may help.) But the
distinction between explicating a current way of hearing and pointing to a new

8
See Cook, pp.11741.
206 In Other Shoes

one is not clear-cut. Even an analysis consisting entirely in specifications of what

a listener already hears in a piece is likely to change the way she hears it.
We have already observed that coming to acknowledge the content of ones
musical experience makes for a change in the experience. Being angry and
fully aware of what one is angry about is not the same as being angry without
realizing exactly what it is that one is angry about. The experience of hear-
ing one melodic line as an inversion of another while being aware only of a
vague similarity between them differs from the experience of noticing that the
one melody is the inversion of the other. (To notice something is, approxi-
mately, to perceive it while acknowledging ones perceiving of it.) So an anal-
ysis which, by specifying what I hear in a piece, helps me to acknowledge
what Ihear in it, changes my experience at the same time that it enhances my
understanding of it.
There is much more to a musical experience than hearing certain features of
the music and noticing some of them. Musical experiences are multi-faceted
and multi-layered complexes of many intentional states, including expectation,
surprise, satisfaction, excitement, recognition, admiration, and even humor, as
well as hearing and noticing. (Some of the constituent intentional states take
others as objects, as when, in noticing features of the music, one acknowledges
the fact that one hears them.) Rather than comparing the experience of listen-
ing to music to that of laughing at jokes, we will do better to compare it to the
larger experience of appreciating a comedians routine, also a multi-faceted,
multi-dimensional experience involving not just humor but various other inten-
tional attitudes as well. Achieving an awareness of the cause of one element of
either a musical experience or an appreciation of a comedy routine often makes
the cause part of an object of another element of it, and in this way alters the
experience.
In addition to laughing at a joke, one may come to notice how the laugh was
achieved, how the timing of the comedians delivery, for example, contributed
to its effect. One may then admire the elegant or masterful means by which the
laugh was elicited, take delight in the artistry that went into it. The manner in
which the timing of the delivery helps to induce ones laughter may surprise or
impress the appreciator (or for that matter, disappoint or bore her). The timing
is just a cause, not an object of humor, but it becomes part of an object of the
larger appreciative experience.
Likewise in music. Amodulation may be surprising partly because of its tim-
ing, because of certain rhythmic features of the passage in question, even if it is
just the modulation, not the rhythmic features, that listeners are surprised by.
Accents or dynamic qualities may contribute to recognition of one passage as
an elaboration of another. If we come to notice either of these causal relation-
ships, we may relish or admire the elegant manner in which the surprise or the
 Und erstanding Humor and Mus i c 207

recognition is effected. And the relishing or admiration, as well as the surprise or

recognition, is part of the musical experience.7a
I assume that the causes are features we heard in the music all along but didnt
notice. Once we notice them, acknowledge our hearing of them, we are in a posi-
tion to notice how our surprise or recognition depends on them, and hence to
appreciate the way the surprise or recognition is brought about. One function
of an analysis specifying what we hear in a piece may be in this way to induce
awareness of causes of certain elements of our musical experience in order to
incorporate them in the content of other elements of it. The analysis at one and
the same time explains the listeners experience and extends it.
Analyzing musical compositions is not a matter of standing back from ones
musical experiences to investigate objectively the compositions or how they
work on us. Analysis is continuous with appreciation, and explaining or under-
standing how it is that one hears a piece is not to be separated from the experi-
ence of hearing it. Although an analysis might, coldly and matter of factly, inform
someonea musicologist from a different planet, for instanceof certain facts
about a way of hearing a piece, analyses are usually meant to induce the recogni-
tion or acknowledgement that constitutes understanding of that way of hearing,
understanding that goes beyond accepting what the analysis explicitly says and
involves exercising ones own musical intuitions. What one comes to understand
may be either the kind of musical experience one previously enjoyed, or a new
one. In either case the understanding is likely, in the ways Ihave described, to
deepen and enrich ones subsequent hearings of the piece.

[I have suggested that admiration is the central component of aesthetic pleasure, and judg-
7a

ments of aesthetic value. Kendall L.Walton, How Marvelous!:Toward a Theory of Aesthetic Value.
Reprinted in my Marvelous Images: On Values and the Arts (New York: Oxford University Press,
2008), pp. 320.]
 12

What Is Abstract About

theArtofMusic?
... music heard so deeply
That it is not heard at all, but you are the music
While the music lasts. ...
T. S.Eliot, The Dry Salvages

I
What is abstraction in any of the arts? Abstract works of art are sometimes
contrasted to representational (or figurative or objective) ones. Even if a
degree of abstraction is compatible with representation, as in the case of cub-
ism or Monets late work from Giverny, for instance, what is entirely abstract, in
at least one sense of the term, is nonrepresentational.1 Let us begin, then, with
the rough and ready commonsense distinction between those arts or works
of art that are said to be representational (or figurative or objective) and
those that are said not to be. In the first category we find most pre-twentieth
century painting and sculpture, virtually all literature and, except for a few
avant-garde experiments, all theater and film. Nonrepresentational, abstract
works includeprovisionally at leastmost architecture, twentieth-century
nonobjective painting and sculpture as well as much design and ornament
from throughout history, and of course musicpure or absolute music, that
is. What can be made of this distinction? Can it be made out at all?
There is a startling difference between music and the visual artspainting in
particularin their attitudes toward representation and abstraction. It can eas-
ily seem that music is naturally, normally abstract, whereas painting is naturally,

1
Richard Wollheim considers virtually all visual art representational, but distinguishes that
which is figurative from that which is not (Wollheim, On Drawing an Object, in On Art and the
Mind [Harvard University Press, 1974], 24, 25). Representational as understood here can be
taken to be an approximate equivalent of his figurative.

208
 W hat Is Abstrac t About the A r t of Mus i c ? 209

normally representational. Representational pictures are a part of nearly every cul-

ture. Even when graven images are forbidden, the very prohibition recognizes the
naturalness of pictorial representation, the power of the temptation to be resisted.
Abstract visual design, although no less universal, is often thought of as mere orna-
ment or decoration, implying both a lack of importance and a lack of centrality. What
is ornamental or decorative typically ornaments or decorates something else:often a
representational design, as when an ornamented frame surrounds a picture of a lady
or a still life; sometimes a utilitarian form, as in the case of an ornamented spoon.
Wallpaper designs and other decoration not subservient to something decorated
are in many instances not to be noticed especially or focused on. Some architectural
forms are centers of visual interest, of course, but even then the visual experience may
be thought of as subsidiary to, an enhancement of, the life that is led in that architec-
tural environment. We remember how the abstract painters of the early twentieth
century had to fight for the right to make abstract forms central. Painting has always
wanted to be real, Frank Stella observed recently, and by 1600 in Italy it had the
means to do it.2 By real Stella means illusionistically representational.
No one would say that music has always wanted to be real in this sense. In
music, abstraction is given the highest honors (even though musicinstrumental,
absolute musicprobably developed largely from speech, by way of poetry and
then song). Blatant program music is often considered silly or childish.3 Musical
depictions of trains, galloping horses, and the sounds of battle, though not uncom-
mon, have the status of experiments, oddities outside the mainstream of serious
music-makingthe more so the more realistic they arewhile the most serious
mainstream painters have worked avidly on developing and refining realistic per-
spective and modelling techniques. There are, to be sure, programmatic elements
in serious music (Haydns birdcalls, the tone-painting in Beethovens Pastoral
Symphony), but they are typically considered irrelevant to its value as music, inci-
dentals which the listener can safely ignore while concentrating on the musical
significance of the sounds.4
Music does have a respectable function in illustrating verbal texts and assist-
ing, reinforcing, representation in other media, as in song, opera, dance, and film.
Palestrina illustrates the words descendit de coelis with descending melodic
lines.5 Trumpets laugh in Bachs Cantata Der Himmel lacht, die Erde jubilieret. Flies

2
Frank Stella, Working Space (Harvard University Press, 1986), p.40.
3
Even by Tovey, who, as writers on music go, is hardly a purist. See Donald Francis Tovey,
Programme Music, in The Forms of Music (Cleveland, 1956), p.168.
4
Not a bar of the Pastoral Symphony would be otherwise if its program had never been
thought of (Tovey, Programme Music, p.168).
5
In the Credo of Missa Papae Marcelli. Iowe this and several of my other examples of represen-
tational music to Peter Kivys Sound and Semblance:Reflections on Musical Representation (Princeton
University Press, 1984).
210 In Other Shoes

buzz in Handels Israel and Egypt. But music with words or music in the service of
a story is considered by purists to be something less than the pinnacle to which
music can aspire, music in its highest form occurring in, for example, the classical
string quartet literature. Eduard Hanslick claims that the rigor with which music is
subordinated to words is generally in an inverse ratio to the independent beauty of
the former.6
Much abstract visual art is parasitic on the representational. Perhaps some
works are about representationality; at least their point sometimes consists
partly in their departure from the representational norm. Viewers are expected
to notice the absence of representational content. But absolute music is not
thus beholden to representational music. It stands on its own.
I suspect that this difference, the fact that in music abstraction is so often con-
sidered normal and representation requires justification, while in painting the
reverse is true, has something to do with two significant disanalogies between
vision and hearing: In the first place, vision is frequently more effective than
hearing as a means of identifying particulars, as a source of de re rather than mere
de dicto knowledge. (By listening, the pedestrian about to cross a street can tell
that one or more cars are coming, but he may not be able, without looking, to
identify any particular car, or even to determine whether or not there is more
than one. If he looks, he is unlikely not to notice at least one particular car.)7
Secondly, sounds are thought of as standing apart from their sources more easily
than sights are, as objects of perception on their own, independent of the bells or
trains or speech which might be heard by means of them. Asight is nearly always
a sight of something, in our experience; a sound can be just a sound. In any case,
since absolute music is so well established and highly regarded, and also tries
so hard to keep its distance from representation, we will do well to focus our
attention on it. Ultimately of course we would like to understand abstraction in
all of the arts, and indeed abstract elements in even the most representational
works.
What is abstract about (absolute) music, and how does it differ from the
obviously representational (figurative, objective) arts? I will explore three
lines of thought:(a)that music lacks meaning or semantic content, (b)that its
semantic content is more general than that of figurative painting, literature, etc.,
and (c) that music is somehow not perceptual, or is less so than painting and
literature are. Only part of the answer, if even that, is to be found in any of these
directions. But considerations raised in exploring all three will combine to sug-
gest a way of understanding music that not only clarifies its abstract character,

6
The Beautiful in Music, ed. and trans. Gustav Cohen (NewYork, 1957), p.40.
7
This accords nicely with the suggestion Iwill offer shortly that music has a tendency to express
propertiesuniversals rather than particulars.
 W hat Is Abstrac t About the A r t of Mus i c ? 211

but also promises to facilitate the daunting task of uncovering the secret of its
power.

II
Do music and other abstract arts lack a semantic dimensionmeanings, sub-
ject matter, propositional contentthat is to be found in the representational
arts? The works of Dickens, Vermeer, and Shakespeare refer to things outside of
themselves, they are of or about other things, they say things about the world,
they make statements, it would seem, whereas Bachs Art of the Fugue, the Taj
Mahal, and Mondrian compositions are just objectsthey just sit there. The
trouble with this negative characterization of abstraction, by itself anyway, is that
it exacerbates the mystery of the value of the abstract arts.
Questions about what abstraction is need to be approached with why ques-
tions at least in the backs of our minds:Why is there such a thing as abstract art?
Why and how do abstract works appeal to us? Why do we listen to music?
It is not easy to explain the interest and appeal of any of the arts. But the values
of abstraction seem particularly problematic, especially so if it is understood as the
absence of semantic properties. To be sure, there is a puzzle about why people seem
to care about Anna Karenina and Willy Loman, realizing that they are merely fic-
tional, and why people find the portrayal of such characters intriguing or entrancing
or moving. But it is obvious that Anna Karenina, Death of a Salesman, and many
other representational works are significantly concerned with topics of great inter-
est to us, even if they do not speak of particular real people and situations.8 They are,
in some sense or other, about love, or life, or war and peace, or success and failure,
or ambition, or defeat. This by itself shows where to look for a plausible explanation
of the values of representational art. But if the abstract arts are not about anything, if
they have no subject matter, it is hard to see even how to begin going about account-
ing for their power. If Bach fugues and Mondrian compositions are just things, pat-
terns of notes or shapes pointing to nothing beyond themselves, why in the world
should they interest us at all, letalone send shivers up our spines?
It is not surprising that some have tried to find semantic properties in music
and other (so-called) abstract art. Assimilating them to the obviously repre-
sentational arts might seem the only way of making explanations of their value
possible. These attempts have taken various forms. The horizontal and verti-
cal elements of Mondrians paintings have been associated with horizons and
cathedrals. Some have explicitly attributed statements to the most seemingly
abstractworks.

8
It is not obvious that depictions of ordinary still lives and mundane landscapes are, however.
212 In Other Shoes

Like all masterpieces of architecture, [the Rectors palace in Dubrovnik]

expresses an opinion about the activities which are going to be carried on
under its roof. Chartres is a speculation concerning the nature of God and
of holiness. The Belvedere in Vienna is a speculation concerning politi-
cal power. With its balanced treatment of its masses and the suggestion
of fecundity in its springing arches and proliferating capitals, the Rectors
palace puts forward an ideal of an ordered and creative society.9

Many have claimed abstract works to be informative or illuminating, often

assuming implicitly that the illumination is effected by virtue of meanings or
semantic content. Aldous Huxley has one of his characters in Point Counter Point
contend that the slow movement of Beethovens String Quartet opus 132, the
Heiliger Dankgesang, is a proof of the existence of God.
We may detect a grain of truth in such claims. If we were forced to choose
between considering the Heiliger Dankgesang a statement (lets not say a proof)
that God exists, or a statement that God does not exist, Iwould expect us to opt
unhesitatingly for the former. But many will reject the choice. Many of us find
many or most attributions of statements to abstract works of art crude and
gratuitous projections of ones own preoccupations or hangups onto works of art
that are unable to resist only because they are mute. And these attributions are
likely to strike us as having little to do with what is moving or satisfying or mar-
velous about the works. (Does one have to take the Heiliger Dankgesang to have
anything at all to do with theism, in order to appreciate it fully?) Claims that
the Belvedere is a speculation concerning political power or that the Heiliger
Dankgesang states that God exists are ripe targets for a good satire.10
There are less crude ways of finding semantic content in music, of course. Its
expressiveness is sometimes explained in semantic terms.11 What music expresses

9
Rebecca West, Black Lamb and Golden Falcon (London, 1967), p.235.
10
Such as this review of an Italian restaurant in the NewYorker:
... Ibegan my meal with an antipasto, which at first appeared aimless, but as Ifocussed
more on the anchovies the point of it became clearer. Was Spinelli [the chef] trying to
say that all life was represented here in this antipasto, with the black olives an unbearable
reminder of mortality? If so, where was the celery? ... At Jacobellis, the antipasto consists
solely of celery. But Jacobelli is an extremist. He wants to call our attention to the absur-
dity of life. Who can forget his scampi:four garlic-drenched shrimp arranged in a way that
says more about our involvement in Vietnam than countless books on the subject? ... For
desert, we had tortoni, and Iwas reminded of Leibnizs remarkable pronouncement:The
Monads have no windows. How apropos!
(Woody Allen, Fabrizios:Criticism and Response, The NewYorker, 5 February 1979.)
11
Deryck Cooke in The Language of Music (Oxford University Press, 1959) attempts to discover
a vocabulary of musical elements, each with its own expressive meaning. In Chapter2 of Languages
of Art (Indianapolis, 1968). Nelson Goodman holds that expressive works refer to properties
(or rather predicates).
 W hat Is Abstrac t About the A r t of Mus i c ? 213

is usually held to include human emotions, and human emotions certainly are
important to us. If music is, in some sense or other, about them, that people listen
to music should be no more mysterious or surprising than that they read novels
or look at figurative paintings, however much remains to be done to spell out the
nature of the interest.
It is easy to point to examples of expressive music. Much of the vocabulary
of human emotions is readily applied to music; we speak easily of musical pas-
sages being joyful, or tense, or anguished, or exuberant. What is meant when we
speak this way? Afamiliar first stab is to suggest that music is expressive by virtue
of mimicking the behavior by which people express their emotions. There are
Agitato movements, lilting melodies, driving rhythms, nervousness and calm-
ness, etc. But some qualities of music important to its expressive character have
no obvious connection with human expressive behavior. Although the major
mode is not invariably happy or the minor invariably sad, mode is by no means
expressively inert. Just try changing a melody from major to minor or vice versa
and see what happens. (There are also the more elaborate medieval church
modes, and Indian ragas.) If the major mode makes a particular melody cheerier
than it would be otherwise, it is not easy to argue that it does this by somehow
resembling or recalling cheery behavior. People dont change from major to
minor or from Phrygian to Lydian when their moods change.
Cats express contentment by purring and dogs show joy by wagging their tails,
though neither behavior bears any evident similarity to the ways you and Iexpress
these feelings. Just as there are specifically feline and canine expressions of feelings,
there may be specifically musical ones. It may take some experience with a musical
tradition to understand its manners of expression, as it takes familiarity with ani-
mals to detect theirs. But once the connection is learned, the detection of the emo-
tion, the reading of the expression is as immediate and automatic as it is in the case
of fellow humans.12
We might worry now that the assimilation of abstract art (insofar as it is expres-
sive, anyway) to representational art will be too successful. Does expressive music
simply represent occurrences of emotions, by representing behavioral expres-
sions? That is just what representational paintings and novels so often do. Idoubt
that we are prepared to obliterate completely the difference between representa-
tional works and abstract but expressive ones, or to take expression to be simply one
variety of representation.13

12
Amore radical suggestion is that there are specifically feline, canine, and musical emotions, not
just ways of expressing them, and that contentment in cats and cheerfulness in music are not the
same properties that these terms describe in people. But if this is so, why are we interested in musical
cheerfulness; what does it have to do with us?
13
The phrase abstract expressionism suggests that expressiveness (of one sort anyway) is
sharply distinct from representation. Beethoven spoke of expression in contrast to sound-painting
214 In Other Shoes

Some hope may be gleaned from the fact that the expressive range of music,
in contrast to literature and (figurative) painting, appears to be severely limited
in certain directions. Music can apparently express anguish or ecstasy, but it is
hard to imagine its expressing envy or guilt (without the help of text or title or
program notes). Music can be sad or joyful, but hardly embarrassed or jealous.

III
Attempts to find connections between music and the outside world have met
strong resistance from some quarters, even derision. That much music is expres-
sive (whatever exactly this means) may be undeniable. But how important its
expressiveness is, how much it has to do with the power of music and why we
listen to it, is quite another question. Hanslicks The Beautiful in Music is a tract
for the resistance.

In the pure act of listening we enjoy the music alone and do not think of
importing into it any extraneous matter.14
To the question:What is to be expressed with ... [euphony, rhythm,
melody, harmony, etc.]? the answer will be: Musical ideas. Now, a
musical idea ... is ... an end in itself, and not a means for representing
feelings and thoughts.15

That there is something to such purist attitudes is suggested by this compari-

son:Chess moves and chess games can be marvelous, beautiful, elegant. So can
proofs in logic and mathematics; so can scientific theories. But it seems strained,
to say the least, to attribute these aesthetic qualities to expressiveness, and
downright implausible to suppose that they consist in the expression of human
emotions. Achess game or a proof is not beautiful because it expresses joy or
anguish or determination or resignation, nor because it is in some sense about
love or ambition or the human condition, let alone because it represents, say
fate knocking on the door. If this is so there would seem to be a kind of aesthetic
value which is not to be explained in terms of subject matter, one which might be
important in music. When a surprising but elegant modulation or the entrance
of a fugue subject sends shivers up our spines, we neednt assume that the shivers
arise because joy or futility or whatever is somehow expressed, letalone because

(mehr Ausdruck der Empfindung als Malerei) in his note to the Pastoral Symphony. (Quoted in
Tovey, p.169.)
14
Hanslick, pp.1112.
15
Ibid, p.48. See also Edmund Gurney, The Power of Sound (London, 1880).
 W hat Is Abstrac t About the A r t of Mus i c ? 215

anything is represented. (For that matter, the appeal of Italian cuisine seems
hardly to lie in its expressiveness any more than in the statements it makes.
But this appeal would seem to be more like that of sounds appreciated in the
spirit of John Cage than that of Bachs The Art of the Fugue.)
Hanslicks purism is not pure. He allows metaphorical descriptions of music
(flight, reapproach, increasing and diminishing strength16), which arguably
point to important links between music and the outside world. Metaphors can
easily occupy a great deal of our discourse about music, if we let them. We speak
of ascending and descending motives, thick and thin textures, strain
and repose, conflict and concord, movement, return, destinations,
renewal, soaring and whispering melodies, throbbing rhythms, etc. This
will be of little consequence if the metaphors are no more than ways of speaking,
colorful means of describing musics formal or acoustic properties.17 But they
may well be essential,18 immortal. What is said when one speaks of ascent or
descent or movement or destinations in music may necessarily involve reference
to spatial phenomena. If it does, this fact will be welcomed by those who hope to
find a subject matter for music.
More than a few recent music theorists attempt to do Hanslick one better
and avoid even metaphors. Heinrich Schenker tended to shun words entirely in
his musical analyses, preferring diagrams that combine musical notation and his
own symbols.19 The most pure conception of music has it consisting of nothing
but sounds and relations among them:musical motives and their elaborations,
suspensions, inversions, strettos, modulations, recapitulations, Urstze, tone
rows, etc.all of this existing for its own sake and appreciated without refer-
ence to anything else.
Purists tend to have a deep reverence for music. Their feeling seems to be that
to attribute programs or emotional qualities or thoughts (except purely musi-
cal ones) to music is to trivialize it, to cheapen it, to insult it. Semantic content
doesnt do justice to the exceptional profundity of musical values. But in reject-
ing semantic content they are rejecting what can easily seem the most promis-
ing, even the only promising route to an explanation of musical value.
The deepest values in painting and even literature, as well as music, are some-
times held to be those which have least to do with representational content or

16
Hanslick, pp.47, 48.
17
Ibid, p.53.
18
Cf. Roger Scruton, Understanding Music, in The Aesthetic Understanding (London, 1983),
p.85.
19
Music as expounded by Schenker is ... concerned ... only with the internal relationship of
musical elements. Music is structure. Musical discourse must be purely musical ( Joseph Kerman,
Analysis, Theory, and New Music, in Contemplating Music [Harvard University Press, 1985], pp.
7475).
216 In Other Shoes

other semantic properties. This may be the point of Paters claim that all arts
aspire after the condition of music. Painting, though it may be more naturally
representational than music has had its purists:Clive Bell held that representa-
tional elements in the visual arts are irrelevant aesthetically, that only signifi-
cant form matters.20 Similar suggestions have been made even about literature,
which can scarcely avoid representationality.

[T]he genuine writer has nothing to say. . . . He has only a way of

speaking.21
What Ishould like to write is a book about nothing, a book depen-
dent upon nothing external, which would be held together by the inter-
nal strength of its style, just as the earth, suspended in the void, depends
upon nothing external for its support:a book which would have almost
no subject, or at least in which the subject would be invisible, if such a
thing is possible.22

It is as though the very remoteness from our lives of abstract elements in works
of art, which is what makes their power so hard to explain, is at the same time the
source of their power!23
It is too soon to conclude that music and other abstract arts lack significant
links to matters of nonmusical import, however. Music may connect with our
lives in ways which, though profoundly important, are less direct and less obvi-
ous than the examples of representational painting and literature and even
straightforwardly expressive works lead us to expect. The evident futility, even
the foolishness of attempts to say what a piece like The Art of the Fugue means
or what it is about does not close the door on the idea that it does have mean-
ings or a subject matter, and that they contribute significantly to its beauty.
Could it just be that there is no saying, that language is simply inadequate to
convey musics semantic properties or how it relates to our lives?24 This of course
is the familiar idea that music is ineffable.25

20
Clive Bell, Art (London, 1914). There is also, more recently, Clement Greenbergs objections
to representationality.
21
Alain Robbe-Grillet, quoted in John Barth, The Friday Book: Essays and Other Nonfiction
(NewYork, 1984), p.191. No source given.
22
Gustav Flaubert, in a letter to his mistress, Louise Collet. Quoted by Arthur Danto in The
Philosophical Disenfranchisement of Art (Columbia University Press, 1986), pp. 14849.
23
Hanslick seems to consider the appeal of music to be inexplicable. Cf. pp.15, 28, 50, 51, 52,
57, 67.
24
Iset aside the question of what sorts of connections between music and matters of nonmusical
interest count as semantic ones.
25
See for example Schopenhauer and, more explicitly, Suzanne Langer.
 W hat Is Abstrac t About the A r t of Mus i c ? 217

Substantiating this thought is a daunting task. How are we to establish that

music does have meanings (or whatever) if we cant specify what they are?
And why should we be unable to express them verbally? Satisfactory answers
to these questions are a long way off, but Iwill venture a start on the second
one. If we can understand how it might be that musical meanings are inex-
pressible in language we will have deflected the challenge to come up with
them. And we will have boosted the credibility of the idea that our halting
attempts to say, by the use of metaphors, for instance, what musical works
mean might sometimes point at least in the right general direction, notwith-
standing their evident inadequacy and the apparent impossibility of doing
better.
The ineffability of musical meanings, if it can be made out, will do much to
accommodate the intuition that there are none and to explain how music is
abstract, while the presence in music of meanings at all, however ineffable, takes
the edge off the mystery of our interest in it. But notions of abstraction other than
that of lacking or appearing to lack semantic content come into play as well, ones
which will turn out to be importantly related to ineffability. Lets look at what else
might be meant by calling music abstract.

IV
Abstraction, in one sense, is generality. Rather than lacking semantic content,
perhaps abstract works of art have content that is in some way especially general.
There are two possibilities here. It may be that musical works are like predicates
(house, building) that apply to many different things, whereas painting and
literature are analogous to proper names denoting individuals (or to statements
containing proper names). Or music may be merely more general than the obvi-
ously representational arts are, in the way that building is more general than
house. We will consider the second alternative first.
Hanslick urges that music cannot portray definite (specific) emotions, but
allows that it can express indefinite ones. His point, elaborated a little, seems to
be that emotions have two aspects. They have a cognitive component:thoughts
or judgments or evaluations or beliefs or attitudes. The feeling of hope [for
instance] is inseparable from the conception of a happier state which is to
come.26 They also involve a dynamic element, psychical motion,27 what we
might call feelings or sensations (understood not to have intentional objects).
Music is incapable of capturing the cognitive elements, but it can portray the

26
Hanslick, p.21.
27
Ibid, p.37.
218 In Other Shoes

psychical motion; it can express intensity, waxing and diminishing, hastening

and lingering.28
Different emotions can have the same psychical motions. Anger and fear
might (in one sense of the term) feel alike; what distinguishes them is their
cognitive content:anger involves a desire to harm something and perhaps the
thought that one has been wronged, whereas fear involves the thought that one
is in danger.29 So music cant distinguish between them. It cant express anger,
specifically, as contrasted to fear. But it might portray what fear and anger have
in common, a less specific state of feeling of which both are instances. Similarly,
music cant distinguish between love, longing, and religious fervor (not with-
out a title or a text or program notes), or between piety and eroticism,30 but
it may express more general ideas from which cognitive elements have been
abstracted.31
Music may be unable even to differentiate emotions from physical events.
There may be no telling whether a given musical passage expresses fury or a for-
est fire.32 Nonetheless, if it expresses (represents, portrays, depicts) something
of which both are instances, it does apparently have semantic content, a subject
mattera very general one at least.
The resistance many of us feel to claims that a piece of music is about this
or that weakens the more abstract, the more general the purported semantic
content is. It may be gratuitous, presumptuous, unmusical to suggest that a
sonata (apart from title or text) is about the Trojan war, or even about warfare.
But it seems less so to suggest that it is about struggle in general, struggle in a
sense that includes political as well as military struggles, legal battles and busi-
ness competition, personality clashes, struggles against poverty and for dignity,
struggles between ones own desires and ones better judgment, and so on as
far as the imagination can see. (Even if this judgment about the subject matter
of the sonata is not rejected as automatically as a more specific one would be,
however, it is hard to see how one might demonstrate that it is true or plausible,
and even how one might become convinced of it oneself. Part of the problem
may be that we have a very weak grasp of the relevant broad notion of struggle.)

28
Schopenhauer claims something like this also, but less clearly than Hanslick does. See The
World as Will and Representation I:26162; II:44950.
29
The constituent judgment is what transforms an indefinite feeling into a definite one
(Hanslick, p.21).
30
Ibid, pp.29, 35.
31
It is a peculiar fact that some musical forms seem to bear a sad and a happy interpretation
equally well. ... what music can actually reflect is only the morphology of feeling; and it is quite plausible
that some sad and some happy conditions may have a very similar morphology (Suzanne K.Langer,
Philosophy in a New Key, 3rd ed. [Cambridge, Massachusetts, 1971], p.238).
32
Cf. Scruton, p.66.
 W hat Is Abstrac t About the A r t of Mus i c ? 219

It may be silly to read into a recapitulation the thought of someones returning

to the home of his parents, but less silly to regard it as expressing a very general
notion of returning, of which not only returns to ones home but also returns to
health, to the scene of a crime, to ones former convictions, and who knows what
else are instances.
The point here is not that music and other abstract arts are concerned with
such generalities whereas the representational arts are not. Novels and figurative
paintings may ultimately treat notions of equal generality.33 The difference pre-
sumably is that they present them by illustrating them, by representing instances
of them. A novel may present the abstract notion of struggle by portraying
Napoleons struggles, whereas music may present that notion more directly.34
But in some cases music or a passage in music may present a very general con-
cept by being, not representing, an instance of it. Arecapitulation may itself be
one example of the general notion of returning. The recapitulation, a purely
musical event, may on the surface at least appear to be the main focus of the
listeners interest, as Napoleons invasion of Russia may appear the main focus
of War and Peace. This may be the element of truth in Hanslicks claim that the
ideas expressed by music are purely musical whereas the ideas expressed by
language are thoughts distinct from its medium.35
The generality of the semantic content of music and other arts might at first
seem a great virtue, allowing a work to speak to many different interests and con-
cerns. One person might be interested in Napoleons military struggles, another
who doesnt care beans about that may be immersed in family power struggles,
a third may be preoccupied with the conflict between his desire to succeed and
his natural laziness, someone else may have on his mind the tensions aroused by
someone who both attracts and repels him. Apiece of music which is about
struggle in general, which abstracts from what is unique to any particular strug-
gle, would seem to have something to say to each of them. One might expect
such a piece to get high marks on the test of time for aesthetic greatness. It ought
to appeal to different ages, different cultures, people with different interests and
preoccupations, to anyone who is concerned with struggles of any sort. It ought

33
A Tale of Two Cities is not about the French Revolution in the way that Carlyles History of the
French Revolution is. ... It is about love, loyalty, and self- sacrifice among human beings pungently
observed, ... [T]he more a novels main interest is in the time and place its about, the less likely it is
to be a significant work of literature in its own right ... (Barth, pp.188, 190).
34
Cf. Schopenhauer I, Book 3, 52. Iresist the temptation to proclaim Schopenhauer a precur-
sor of the proposals about music Iwill sketch, even though one can divine in a careful selection of
his remarks a certain similarity of spirit. The vagueness of the selected passages allows for construals
significantly incompatible with my suggestions, and I have little sympathy for the metaphysics in
which Schopenhauers views on music are embedded. Ican accept the support of the fact that his
views seem to derive from initial intuitions similar to those that motivate mine, but nothing more.
35
Hanslick, p.23. Cf. also p.67.
220 In Other Shoes

to outlive works dealing specifically with one kind of struggle, which will likely
be forgotten, when circumstances or interests change.35a Do we now have the
key to understanding the special aesthetic profundity some find in abstraction?
If we do, we are a long way from opening the lock with it. If struggles of a
certain sort are important to me, or a particular instance of returning is, it is
not clear why Ishould be interested in the notion of struggle or of returning in
general. Probably Ialready realize perfectly well that the object of my immedi-
ate interest is a struggle, or a return. What Iwant to understand, probably, is its
particular nature, the specific kind of struggle or return it is. Isnt generality just
vacuity?
Here is a story of great generality, one which abstracts from an enormous
number of specifics:

Once upon a time there was a person.

The End.

It is about personhood, Isuppose. All of us have a considerable interest in peo-

ple, in certain people and in certain particular aspects of them; no doubt this is
true of everyone in every culture and every age since the beginning of time. But
the story Ijust recited is notable for its excruciating lack of interest. It is vapid.
It doesnt speak to our specific concerns with particular people at all; it doesnt
connect with our lives. It is guaranteed to flunk the test of time (unless the
Guinness Book of Records immortalizes it as the dullest story ever told). We want
to hear more. We want details. Once upon a time there was a person who. ...
Or consider a theatrical event which abstracts from the cognitive aspects
of its characters emotions, their thoughts (broadly speaking), and represents
only the dynamic element, psychic motions; so no definite emotions are
portrayed. The actors laugh and cry, they scream, moan, and giggle, frown and
smile, stiffen and relax. But there is no indication of the characters thoughts, no
hint of their beliefs or desires or hopes or expectations. The audience can only
speculate about why they laugh and cry and frown and smile. Idont dare say that
such an event would be of no interest; if Idid someone would stage it and some-
how make it interesting (perhaps as an instance of theater of the absurd?). But
the idea is hardly promising. (And it is almost certain that such theater would
not be interesting or thrilling or satisfying in anything like the way Bachs Art of
the Fugue is.) Again, we want details; we want to see the emotions in context, to
understand what they are about and why they are felt.36

35a
[Cf. my The Test of Time, in Marvelous Images (NewYork, 2008), 2325.]
36
Hanslick doesnt think that the indefinite emotions that music expresses are responsible for
its power. These abstract notions ... are by no means the subject matter of the pictures or the musical
 W hat Is Abstrac t About the A r t of Mus i c ? 221

Is it even true that musical portrayals are especially general? Certainly music
and other nonrepresentational works have no monopoly on generality. Astick
figure of a person drawn in black ink probably abstracts from anorexia and obe-
sity, from color, race, sex, mood, type of clothing, etc. The person in the picture
may be sitting or standing, or running, but if the sketch is sufficiently stylized it
may abstract from any particular manner of running or sitting. It may be almost
as unspecific as the phrase, a sitting person. Much of the serious representa-
tional art we think of as (partially) abstract is general in this way, if not to this
extent. (A cubist portrait in which a rectangular shape represents the sitters
head doesnt thereby depict the sitter as having a rectangular head; it is simply
very unspecific about the shape of his head.)
I will suggest later that music does present cognitive elements of the emotions
it expresses (though in a manner very different from those in which painting and
literature most obviously do, and though the cognitive elements it presents may
be importantly different). But even if it doesnt, the difference between music
and the representational arts is hardly one of degree of generality. Music may
well express the dynamics of emotions with extreme specificity, in much more
detail than can be done easily or at all in painting or literature. Music may not
be able to distinguish between fury and fear, but it may portray very precisely the
nature of certain (nonintentional) feelings or sensations one might have when
one is either furious or afraid. Perhaps a particular recapitulation captures a spe-
cific manner of returning in great detail, one which might characterize a return
to Athens and a return to health and a return to earlier convictions.37 The dif-
ference between music and the representational arts may lie less in the degree
of generality of their semantic properties than in the respects in which they are
general and the respects in which they are specific.
It seems also that the properties which music is able to portray are often ones
for which we have no words. The English language groups all of the various sorts
of anger together, and separates them from cases of fear and from forest fires. It
less easily expresses sensations which some cases of fear and some cases of anger
may have in common, especially very specific ones (agitated or upset might
express unspecific ones), and less easily still captures what they share with for-
est fires. Struggle and return may strike one as puns when applied as broadly
as I have suggested. Music might serve to show us what certain instances of

compositions ... (p.23). The function of art consists in individualizing, in evolving the definite out
of the indefinite, the particular out of the general (p.38). Ido think that generality is important,
however, in ways which tie up with the preference for elegance and simplicity in scientific theories.
37
Mendelssohn claimed that music has more definite meanings than words do (Tovey, The
Meaning of Music, in The Main Stream of Music and Other Essays [Cleveland, 1959], p.397). Cf.
Schopenhauer I:259. Idisagree sharply here with Malcolm Budds claim that emotions can be fully
revealed by the use of language (Music and the Emotions [London, 1985], p.137).
222 In Other Shoes

returning from a trip, returning to health, returning to previous convictions, etc.,

have in common. (And of course struggle or return even in its extended sense
may correspond only approximately to the properties expressed by a piece or
passage of music.) One writer finds the apparent contradiction gay melancholy
the best he can do in describing the character of the alla tedesca movement of
Beethovens B-flat Quartet. Both elements are present, not as contrasting, but
strangely unified in one haunting phrase.38
So music organizes things differently from how language does; its categories
cut across those of English. This supports the claims of the ineffability of music,
of the impossibility of capturing what it expresses in words.39 But it is only a
small part of the story, as we will see.

V
Abstract things are sometimes contrasted to objects of perception; we see and
hear what is concrete, but we conceive or apprehend with the mind what is
abstract. Music and painting are usually thought of as perceptual arts, one aural
and the other visual, in contrast to literature (and, for that matter, to the concep-
tual art of the 1960s and 1970s). But there is a sense in which music is less an
aural art than (figurative) painting is a visual one. This comes out most obviously
in the fact that when music is representational or illustrative it often does not
represent sounds, whereas representational painting seems always to represent
sights.40 Patience, which Handel illustrates in Belshazzar by means of long notes,
is not an aural phenomenon, nor are instances of patience. Bach illustrates the
words I follow Christ with a canon in his Cantata 12, Weinen, Klagen, Sorgen

38
J. W.N. Sullivan, Beethoven:His Spiritual Development (NewYork, 1960), pp. 15152.
39
It also suggests another line of thought which Iwill not pursue now:that music treats not only
particular concepts, a particular way of organizing experience, but the very process of organizing and
reorganizing it, the process of adopting systems of classification and replacing one with another, of
reconceptualizing things, of adjusting ones conceptual scheme. (The comparison with the beauties
of chess can be understood to point in this direction.) There are in music constant reorganizations
and reclassifications and reconceptualizations of musical materials:thematic ideas, rhythmic motives,
harmonic progressions, and formal structures are combined, fragmented, recombined, made to look
like or unlike others, placed in new contexts, etc. (This occurs subtly and subliminally in even very
simple melodies, as well as explicitly in, for example, the development sections of complex sonata
forms.) Listening, then, is perhaps an exercise in the techniques by which we reconceptualize our
extramusical experiencewhether in the development of scientific theories or in everyday thought.
This may constitute a less direct connection with our lives than would the treatment in music of
notions of struggle or achievement or return, but surely not a less important one. (It may or may not
be understood in semantic terms.)
40
Kivy (p.40 ff.) points out that musical representation is not always of the sounds-like variety.
Cf. also Scruton, pp.6768. Neither seems to me to accord sufficient importance to this fact.
 W hat Is Abstrac t About the A r t of Mus i c ? 223

Zagen, one voice imitating another. Ascents and descents are often portrayed in
the obvious way.
Pictures can represent nonvisual phenomena, of course. Rudolf Arnheim
pointed out that in Sternbergs silent film, The Docks of NewYork, the report of a
revolver is indicated by the rising of a flock of birds.41 Odors can be represented
by depicting wrinkled noses. But it is by means of depicting sights in these cases,
the looks of things, that sounds and smells are portrayed. It is because we see the
birds suddenly take flight that we know the sound of a shot has occurred. But
musical portrayals neednt involve the portrayal of sounds at all. It is not by rep-
resenting any sounds that long notes indicate patience, or an ascending melodic
line someones rising into heaven.
I believe that this difference is tremendously important. When we look at
a picture we imagine ourselves taking in the sights that it depicts. On view-
ing a painting of a wheat field one imagines seeing a wheat field. One also
imagines, Iclaim, that looking at the picture, is an instance of seeing a wheat
field.42 Representational music frequently involves no such imagined perceiv-
ing. Idont imagine myself hearing patience or anyones being patient or any-
thing else, when Ihear the long notes of Belshazzar and understand them to
portray patience. Still less do Iimagine of my hearing of the music that it is a
case of hearing anything having to do with patience. To listen to music, even
representational music, is not, in general, to be perceptually involved with
what is represented. In paintings and other visual representations there are fic-
tional worlds to which we have perceptual access. Sometimes we are included
in the world in a special way, as when Caravaggios Bacchus offers the viewer
a glass of wine, or a character in a play asks the audiences advice. Insofar as
music is not perceptual, we dont have this access to a fictional world, and we
cant expect to be included in that special way. The listeners relation to what
is represented or portrayed would seem often to be fundamentally nonper-
ceptual, in this sense, and thus fundamentally different from the viewers rela-
tion to what a picture depicts. In fact, this nonperceptual character of musical
portrayals may connect with what is meant by those who describe music as
ethereal, incorporeal, disembodied (or unembodied?), and with sug-
gestions that music (music like Bachs Art of the Fugue, anyway) is introspec-
tive, cerebral, conceptual.
It goes nicely, also, with the idea that music involves generality in a way that paint-
ing does not, if we put it together with the suggestion by several recent philosophers

41
In Film as Art (University of California Press, 1966), p.34.
42
Cf. Kendall Walton, Pictures and Make-Believe, The Philosophical Review ( July 1973):283
319; and Walton, Looking at Pictures and Looking at Things, in Philosophy and the Visual Arts, ed.
Andrew Harrison (Reidel, 1987), pp.277300.
224 In Other Shoes

that all identification of individual things is at bottom demonstrative, indexical, and

hence(?) perceptual43the idea is not that music is more general than painting, but
that it is general in something like the way all predicates are, whereas painting, like
proper names, involve something more like reference to particulars.
Even the sketchiest stick figure drawing is in one sense a picture of a particu-
lar person, an individual.44 Viewers imagine themselves seeing and identifying
some one individual person, even if the picture indicates very little about how
he or she differs from other people. It may be that, when music does not induce
imaginative perception, it presents only concepts, properties, universals (even if
not very general ones), not particular things. Music may portray the notion of
patience, or ascent, or struggle, or return, but not particular instances of some-
ones being patient or ascending or struggling or returning.
It may seem puzzling, however, how a musical passage can get connected to
the notion of patience or struggle or whatever in the abstract, without somehow
portraying a particular (possibly fictitious) instance. Iwill suggest shortly how
this can be.

VI
I propose that, although music does not in general call for imaginative hear-
ing or imaginative perceiving, it often does call for imaginative introspecting.
We mentioned the possibility that music is expressive by virtue of imitating
behavioral expressions of feeling. Sometimes this is so, and sometimes a pas-
sage imitates or portrays vocal expressions of feelings. When it does, listeners
probably imagine (not necessarily consciously, and certainly not deliber-
ately) themselves hearing someones vocal expressions.45 But in other cases
43
Saul Kripke, John Perry, Hilary Putnam, etc.
44
This isnt to be taken literally. There may be no actual person whom the drawing depicts, and
I deny that there are any such things as purely fictional individuals, the man in the picture, for
instance. But it is to be taken seriously. Understood in an appropriate nonliteral manner, the stick
figure drawing is a picture of a particular person is true.
45
In Pictures and Make-Believe Iproposed that (for example) anguished music is music which
make-believedly, is an expression of someones anguish (p.299). Malcolm Budd examines this pro-
posal in Music and the Emotions, concentrating on the idea that expressive music is, make-believedly,
someones (perhaps the appreciators own) vocal expression of the emotions in question, and that
the listener imagines hearing these vocal sounds. My present suggestion makes use of the notions of
imagination and make-believe in a very different way, centering as it does on the idea of introspective
awareness of ones emotions rather than on that of someones expressing them, vocally or otherwise.
But Ido think that music is sometimes imagined to be someones expression of feelings. (Imagining
this is not incompatible with being imaginatively aware of the feelings in question.) Budd is quite
right to observe that the mere fact that a set of sounds is for someone make-believedly the vocal
expression of an emotion is ... not sufficient to endow it with artistic value for him (p.133; cf. also
 W hat Is Abstrac t About the A r t of Mus i c ? 225

they may, instead, imagine themselves introspecting, being aware of their

own feelings. Hearing sounds may differ too much from introspecting for us
comfortably to imagine of our hearing the music that it is an experience of
being aware of our states of mind. My suggestion is that we imagine this of
our actual introspective awareness of auditory sensations.46 If so, the music
probably can be said to portray particulars in the sense that figurative paint-
ings do, rather than simply properties or concepts. Presumably the listener
imagines experiencing and identifying particular stabs of pain, particular
feelings of ecstasy, particular sensations of well-being, etc., as in viewing a
painting one imagines seeing particular things. But introspection is different
enough from external perception to make music, understood in this way,
differ very substantially from painting. For a start, there is the important
fact that one introspects ones own psychological states, while one sees and
hears the expression of the psychological states of other people. There is also
the fact that the results of introspection seem immune from error in ways in
which the results of external perception are not. We have here a start toward
developing the thought, expressed aphoristically by Carroll Pratt, that music
sounds as feelings feel.47
We may also have stumbled upon some insight into evocation theories of
expression. There is a persistent tendency to regard expressive works of art
(musical or otherwise) as ones which arouse in appreciators feelings of the
kind expressed. Alot of argument has been directed against this thesis, but it
dies hard. On the present suggestion expressive works dont actually arouse
feelings but they do induce the appreciator to imagine herself experiencing
them. That is close enough to cause confusion, and to explain the appeal of
evocation theories. Notice that there is hardly any temptation to understand
the representation of emotions in terms of evocation. If anything counts as rep-
resenting anguish it is, Ipresume, representing a character (in a picture, or a
novel, or whatever) as being anguished. That involves inducing the apprecia-
tor to imagine that someone else feels anguish (perhaps, but not necessarily,
by inducing the appreciator to imagine seeing an anguished expression on
his face, or hearing him express his anguish verbally, or hearing a third per-
son describe his anguish). Imagining that can hardly be confused with feeling

p.141); much more does indeed need to be said. But as grounds for summary dismissal of this fact
as playing a significant role in musical value it is clearly inadequate. Sadness is important in our lives.
Asignificant connection between a musical work and sadness is a promising step toward an explana-
tion of why the work also is important to us. The same goes for the very different connection Iam
now proposing between expressive music and human emotions.
46
Ishould add that it is in a first-person manner, not a third-person one, that one imagines oneself
to be aware of his states of mind. The distinction is familiar, but Iwill not try to explicate it here.
47
Pratt, The Meaning of Music (NewYork, 1931), p.203.
226 In Other Shoes

anguish oneself.48 This gives us a sharp contrast between the representation

and the expression of emotions.
To the extent that expression consists in inducing imaginative awareness of
ones feelings, we also have an explanation of the fact that feelings (emotions,
moods, inner states of one sort or another) are so often taken to be the sole
or the primary objects of expression. Goodman is an exception. He speaks of
the expression of color, sounds, weight, fragility, movement, and being a glue
factory no less readily than of the expression of feelings; for him there is nothing
especially psychological about expression. Certainly he is right if expression is
what he says it ismetaphorical exemplification;49 virtually any property might
be metaphorically exemplified. But it seems to me that music does not express
color or fragility or angularity or weight in the same direct way that it expresses
anguish or joy or elation or despair (or anyway the dynamic elements of some
such feelings). (This point is obscured by predicates such as strength, strug-
gle, and heavy, which metaphorically characterize psychological states though
their primary application is elsewhere. The fact that a work can express strength
or heaviness need not conflict with the idea that expression is preeminently or
even exclusively of psychological states.)
Even if expression (in one sense) is a matter of imaginative self-awareness,
however, music may present, let us say (express indirectly, or in a different
sense?), nonpsychological properties or conceptsones akin to that of return-
ing or a certain manner of returning, or struggling, or power, or achievement.
What the listener imagines being introspectively aware of may be an impres-
sion of or a feeling about (the relevant variety of something like) returning or
struggling or power, a phenomenological response or reaction to it, a way of
experiencing or understanding it. She is thus imaginatively aware not just of
psychical motions, the dynamic elements of emotions, but of cognitive ele-
ments as well, objects toward which the feelings are directed. These objects
are not presented independently of the feelings about them, however. My sug-
gestion is not that the music portrays an objective event or circumstance, and
then induces the listener to imagine responding to it in a certain manner; it just
induces the listener to imagine the experience of responding to an object of a
certain sort. We might say that the music conveys subjective aspects of objec-
tive phenomena.
Does one imagine knowing about and responding to a specific (possibly ficti-
tious) instance of struggling or returning? Perhaps. If so, this need not involve
imagining anything about how one knows about it. The listener is likely not to

48
Although some might argue that to the extent that the appreciator empathizes with the charac-
ter, some of the anguish will rub off on him.
49
Goodman, Chap.2.
 W hat Is Abstrac t About the A r t of Mus i c ? 227

imagine hearing or perceiving a struggle, for example, or being told of one, or

having perceived or been told about one previously. The world of his imagina-
tion is likely to be indeterminate with respect to the means by which he is aware
of the (imaginary) struggle. The focus is on other aspects of it.
This means that musical portrayals can lack points of view of kinds common
in the representational arts. Paintings, typically if not inevitably, depict things as
perceived from a certain angle and distance or in certain conditions; novels pres-
ent events as reported or described in a certain manner by a person of a certain
sort (the narrator).50 But music may portray a struggle or a return or an ascen-
sion without portraying it in any such manner.
An even more basic point of view may be absent as well. It may be indetermi-
nate whether or not the imaginary instance is one in which one participates one-
self. Does the listener imagine herself returning or struggling or does she imagine
someone (or something) else doing so? I, for one, find it impossible to say, even
when it seems to me reasonably clear that a passage does portray a particular
instance of returning or struggling and that my experience involves imagining
such an instance. Iimagine someones struggle or return and my having certain
impressions of it or responding to it in a certain manner, but Iimagine neither
that Iam the one struggling or returning, nor that Iam not.51
Then again, what the listener imagines may not be a particular instance. She
may imagine simply having an impression or conception of a kind of returning
or struggling (or power or achievement).52 It is even more likely, Ithink, that
there will be no answer to the question whether what one imagines is a particular
instance or merely a kind. One may imagine having an impression of struggle
without imagining either that ones impression is of a particular struggle or that
it is not. Much the same feeling (a subjective sense of the dynamics of struggle)
might accompany either an awareness of a particular actual struggle or merely
the thought of (a certain sort of) struggle in general.53 The music may capture
this impression without being specific about its source.

50
Cf. my Points of View in Narrative and Depictive Representation, Nous (March 1976):4961.
51
R. K. Elliott distinguishes between experiencing music as ones own expression of an emo-
tion and experiencing it as that of another person, but he does not recognize the possibility of an
experience which is indeterminate between these alternatives. Such indeterminacy is as important
to theories linking music to the expression of emotions as the similar indeterminancy will turn out
to be to mine; it is an important part of what might be said to make music abstract, despite its
being, in one sense or another, expressive of human emotions (Cf. Elliott, Aesthetic Theory and
the Experience of Art, in Harold Osborne, Aesthetics [Oxford University Press, 1972], pp. 14557;
and Budd, Chapter7).
52
It is possible that she actually has such an impression or conception, as she listens. This must not
be considered incompatible with his imagining having one.
53
Ones impressions may be of a sort which are not even ostensibly impressions of particulars, in
the way that visual images seem to be.
228 In Other Shoes

VII
We have uncovered grounds for recognizing new and surprising dimensions of
generality in musical meanings.54 The musical portrayal of an awareness of an
individual thing may abstract from any specific mode of awareness, including
first- and third-person perspectives as well as perceptual and other means of
epistemological access. And music may fail to differentiate between properties
and their instances. These abstractions are, in varying degrees, awkward or dif-
ficult or impossible to capture in language, and they are foreign to much of what
we call thinking. When we have in mind a particular return or struggle we usu-
ally have some idea how we are aware of it. (One may have forgotten the source
of ones awareness of a struggle long past, but insofar as a musical passage can be
said to portray a particular struggle it would seem to portray a present instance,
a struggle occurring now.55) Although in English one can easily refer to things
without specifying how one knows about them, whether by perceptual experi-
ence or hearsay or what, Iunderstand that this is not so in some languages. It is
not easy in English to avoid having to choose between first- and third-person
constructions.56 One is very unlikely to be unsure whether a given struggler or
returnee is oneself or another (especially if the struggle is a presently occurring
one), and it is hard to conceive of thinking of someone in a way which is neither
first- nor third-person. (One can describe and think of oneself in a third-person
manner, even if one realizes that it is oneself.) There is no simple way to refer to
something as, indifferently, either a given property or an instance of it. (There
are ambiguities, but that is different.) Idont know what it would be like to be
aware of and to reflect introspectively on an impression of something which is
either a particular struggle or merely the thought or notion of struggle in gen-
eral, without knowing which it is.57

54
The mere fact that music elicits imaginings does not of course justify speaking of musical mean-
ings or even of semantic content, though what else is needed is a matter of dispute. What follows is to
be understood as an account of what sort of meanings music has, on the assumption that what it has
is meanings and not just capacities to induce imaginings.
55
If it does not, it is surely indeterminate whether what is portrayed is present or past or future,
and this is something we are unlikely to be unsure of.
56
One can avoid reference to a straggler or a returnee entirely by using the passive voice (A
struggle was engaged in). But if one does refer to the person in question one must use either a first-
or a third-person (or second- person) personal pronoun, or a name or description which indicates a
third-person perspective.
57
One may have a visual impression as of a particular thing, for instance, without knowing
whether or not it is veridical. But it purports to be of a particular in a way in which the impressions
of struggle that listeners may be imaginatively aware of may not be. What Isuspect we sometimes
do while listening to music, and not when we think verbally, is to imagine having an impression of
struggle, without imagining either that it does or that it does not purport to be of a particular.
 W hat Is Abstrac t About the A r t of Mus i c ? 229

We noted earlier that the properties which music expresses (or which figure
in one way or another in its meanings) are likely not to be easily accessible by
verbal means. The notion which a musical passage presents to listeners is prob-
ably not exactly that of struggle or of return, perhaps not even approximately
that, but rather something more general or less general or both. We now see that
the ineffability of musical meanings, the incommensurability of music and
(verbal) languages, may go much deeper than this. It is not surprising that any
(verbal) suggestions one might come up with about what music means seem,
so often, to be ludicrously inadequate. And if one cannot come even close to say-
ing what the semantic content of a musical passage is, we can surely understand
the powerful (even if mistaken) impression that it has none.
I have not claimed, as some have, that music has meanings of a sort no pos-
sible discursive language could express; establishing this would require a very
fancy argument. But the fact that important aspects of the semantic content of
music are in fact inexpressible in the actual languages familiar to listeners (or
even just that they are thus expressible only with great difficulty) promises to be
significant enough. Music may be a vehicle of thought in whatever sense (verbal)
languages are, but one which encourages a very different mode of thinking. To
the extent that what we think in music is unthinkable otherwise, the listener
feels in a different realm, a different worldone that is purely musical in the
sense that it is accessible to him only through music.
But only in this sense, if music does make reference to extra-musical reali-
tiesto (dare we try to say?) unnamable feelings or the dynamics of emo-
tions, or awarenesses of, indeterminately, ones own or anothers return, or
something vaguely like the property of being a conflict, or something which is
either that property or an instance of it. Such references, however indescribable,
may be part of the secret of the power music has over us. It may be impossible to
say which matters of interest a particular passage or piece is about and what it
says about them. Perhaps we cannot hope to explain specifically how and why
a particular passage or piece is powerful. But now that we understand better how
it might be that music treats of things that matter to us in ways that are beyond
description, the fact that music affects us deeply while seeming so remote from
our lives should be less a mystery. Much remains to be done in explaining musics
power, of course. But if these suggestions are on the right track we neednt be any
more astonished by its power than by that of the obviously representational arts.
We now know where to look for an explanation.
 13

The Presentation and Portrayal

ofSound Patterns

One of the most fundamental questions of musical aesthetics is this:Which is of

primary musical importance, musical works (symphonies, songs, sonatas, etc.)
or performances of musical works? Are works or performances the basic objects
of musical attention, musical appreciation, and musical judgment?1
Like many fundamental questions, this one is not often confronted directly,
or even clearly perceived. It is easy for the theorist to presuppose an answer to
it of one sort or another without being fully aware of either the answer or the
question. But work on other problems of musical aesthetics is likely to be col-
ored significantly by ones attitudes concerning this question, and for the sake of
clarity these attitudes need to be exposed and examined.
I will not address this fundamental issue immediately, however, but will delve
first into several more commonly discussed problems of how particular perfor-
mances are related to particular works. We can ask, of a particular performance,
(1)what work it is a performance of, (2)whether it is a correct performance of
that work, and (3)whether it is a good performance of it. Iwill focus on (1), but
the discussion will lead me to some observations concerning (2)and (3)as well.
It will be illuminating at various points to compare the relation between
musical works and their performances with relations between works of other
kinds and their instances. Iwill discuss specifically how literary works are related
to copies and (oral) readings of them, and how culinary dishes are related to
their instances, the food that people eat on particular occasions.
Eventually Iwill return to the fundamental issue of the relative primacy of
works and performances. I will sketch the position I propose to take on this

1
Ishould emphasize that this question, as Iconstrue it, is a question about our cultural insti-
tution of music. Iam asking what roles pieces and performances have in our institution, how they
are regarded and treated by participants in the institution:composers, performers, and appreciators.
Neither works nor performances are intrinsically primary, apart from their place in some cultural
institution.

230
 P re s entation and Por t rayal 231

issue, and show how my position opens the way for an illuminating theory of
how performances are related to the works of which they are performances.
A rather exotic example will serve to introduce the problem of what deter-
mines which work a given performance is a performance of. Suppose that musi-
cal scores on Mars specify very different sorts of properties from those that our
scores specify. Martian scores do not indicate what pitches a performer is to
play, or for what durations. Instead they give detailed instructions concerning
dynamics, tempos, articulations, vibrato, nuances of accent and timbre, etc.
instructions much more detailed than those provided by (traditional) scores in
our society. The performer of a Martian work is free to decide what pitches to
play and for what durations, but she is expected to play them with the dynam-
ics, articulations, timbres, etc., indicated by the composer. Different performers
playing from the same score will of course play different pitches and rhythms
(and hence different harmonies and harmonic rhythms) in executing the com-
posers instructions, just as on Earth different performers play the notes (pitches
with durations) specified by the composer with different dynamics, tempos, and
articulations.
Now let us imagine that a Martian composer, one Ludwig van Marthoven,
wrote a symphony, his sixth, and that the dynamics, tempos, articulations,
accents, etc., called for in his score happen to be precisely those that character-
ized a certain performance by the Chicago Symphony Orchestra of Beethovens
Sixth Symphony. Imagine, further, that a certain performance by the Martian
Philharmonic of Marthovens symphony has, by coincidence, the notes called
for in Beethovens score. Suppose that, in these and all other respects, the two
performances are acoustically indistinguishable.
This example might be treated in various ways. One way, which so far as Ican
see has little in its favor, is to say that there is but one musical work and that both
orchestras performed this one work. However, Marthovens score is radically dif-
ferent from Beethovens; many performances that comply perfectly with one of
them conflict drastically with the other. It would seem unreasonable not to rec-
ognize two different works corresponding to the different scores. Should we say
that each performance is a performance of both of these works at once, that is,
that the Chicago Orchestra inadvertently played Marthovens Sixth Symphony
while it was deliberately playing Beethovens, and that the Martian Philharmonic
inadvertently played Beethovens while playing Marthovens? This seems an
awkward and unintuitive way of construing the example.
Another alternative would be to hold that something is a performance of a
given work only relative to a musical system,2 and that relative to our musical

Amusical system may be understood as consisting, partly, of conventions concerning what per-
2

formance features are to be notated in scores.

232 In Other Shoes

system both performances are of Beethovens work and relative to the Martian
system both are of Marthovens. But this alternative still treats the two perfor-
mances alike. Iprefer to treat them differently.
The position Ifind most congenial is that the Chicago Orchestra performed
Beethovens symphony and not Marthovens, and that the Martian Philharmonic
played only Marthovens. This means that which work a performance is a per-
formance of depends at least partly on some non-acoustic properties of it,
since the two performances are acoustically identical. What makes the Chicago
Orchestras performance a performance of Beethovens symphony rather than
Marthovens is something about the context or setting in which it occurs:the
musical tradition it is part of, or the intentions of the performers, or the fact that
Beethovens score played a certain causal role in bringing it about, or some com-
bination of these circumstances.
Does it matter how we treat this case? Indeed, does it ever matter what work
we say something is a performance of? It is obvious from much more mun-
dane examples that our usual criteria for determining what is a performance of
what are not at all precise. How horrendously can a student orchestra clobber
Mozarts Jupiter Symphony and still be clobbering Mozarts Jupiter Symphony? If
someone whistles a garbled conglomeration of Carmen and a Josquin Des Prs
motet, is he thereby whistling either or both of these works, or parts of them?
Is the Dies Irae a separate musical work that is performed every time a larger
work (such as Berliozs Symphonic fantastique) in which the appropriate notes
are embedded is performed? How prominent must those notes be in the larger
work? Must they have been put there intentionally? Should a performance of a
four-hand piano transcription of a Brahms symphony count as a performance of
that symphony? These questions do not have clear answers. Is this because they
do not much matter? Is it because it makes little difference what is regarded as a
performance of what? We should not just assume that the performance-of rela-
tion is an important one, that it plays a significant role in the institution of music.
I contend that this relation is important, that what we take to be a perfor-
mance of what can make a profound difference to listeners experiences, that our
criteria for assigning performances to works are a crucial part of the institution
of music. But it will be clear why this is so only after Ihave given my account of
what makes something a performance of a given work.
Let us say, tentatively, that a musical work is a sound pattern. It is important to
avoid misunderstandings about my use of the term pattern. Apattern is some-
thing that particular objects or events may fit or fail to fit (although the pattern
itself is not a particular object or event). The American flag is a pattern of stars
and stripes, which the flag flying above the post office in Decatur, Illinois, fits, as
do many other similar objects. Something fits a given pattern by virtue of pos-
sessing certain properties. To fit the American flag pattern is to be a rectangular
 P re s entation and Por t rayal 233

object (or surface) with fifty stars and thirteen horizontal bands of the right col-
ors in the right relations. Apiece of material has (fits) a checkered pattern just in
case it consists of squares of alternating colors.
Whether something fits a given pattern is not a matter of how it is perceived,
of how it is structured, organized, or parsed, in the perception of an observer.
Acheckered tablecloth might be seen as having white squares on a black back-
ground, or as having black squares on a white background, or simply as having
alternate black and white squares. One might see in it the horizontal rows of
alternately colored squares, or the comparable vertical columns, or the diagonal
lines of similarly colored squares; the squares can be seen as diamonds rather
than as squares. But how the tablecloth is seen does not alter the color and shape
properties that it in fact possesses, and so does not affect what pattern or pat-
terns it fits by virtue of its possession of these properties. The checkered pattern
of the tablecloth may be structured in perception (or in thought) in any of many
ways, but that the tablecloth has this pattern depends only on its actual colors
and shapes.
There can be great differences among things that fit the same pattern. This is
because (ordinarily) only some of a things properties determine whether it fits
a given pattern. Checkered tablecloths can differ drastically in size, shape, mate-
rial, and in the size and colors of their squares, since these properties do not affect
whether something fits the checkered pattern.
Patterns are individuated by what must be true of particular things in order to
fit them. Pattern P and pattern P' are the same just in case the properties some-
thing must have to fit one are exactly the same as the properties something must
have to fit the other. So to specify a pattern is to specify what anything must
be like in order to fit it. Musical scores specify patterns. Ascore indicates vari-
ous performance features, thereby specifying the pattern such that whatever has
those features fits it.
A pattern is a sound pattern if the properties required for fitting it are acoustic
ones. A sound event (e.g., the succession of sounds produced by a particular
performer on a particular occasion) fits or conflicts with a given sound pattern
depending on qualities such as the pitches, durations, timbres, and volumes of
the sound. The patterns that musical scores specify are primarily sound patterns.3
Assuming that musical works are sound patterns, which sound pattern is
to be identified with a given work? Anatural first answer is that a work is the
pattern specified by its score. But some works do not have scores, and in oral

3
Many scores indicate not only what sorts of sounds are to be produced but also how they are to
be produced. If the latter indications are regarded as contributing to the specification of patterns, the
patterns specified are not pure sound patterns. Iwill ignore this complication for now but will return
to it in the final section.
234 In Other Shoes

musical traditions there are not even conventions for producing scores (con-
ventions as to which performance features are to be notated and which not).
Moreover, scores sometimes serve more than one purpose and their use for
some jobs may interfere with their specification of the appropriate pattern.
(Iwill explain this.)
I suggest, provisionally, that we think of a work as the pattern determined
by the sound properties a performance must have to be a correct or flawless
(not necessarily good) performance of it. (This suggestion will be modified
later.) Assuming that P is a performance of work W, what sound properties
must P have to be a correct performance of W? The answer enumerates the
properties anything must possess to fit the pattern constituting W; it specifies
what pattern W is.4
Scores in our musical tradition function primarily to lay down rules for cor-
rect performance. Discrepancies between a performance of a work and its score,
if it has one, are usually what count as mistakes. Insofar as this is so the score
specifies the pattern that constitutes the work. But some specifications in scores
are better construed as advice about how to perform the work well, rather than
as indications of what counts as a correct performance. Gyrgy Ligetis Requiem
(ca. 196567) contains instructions such as: change bow unobtrusively, as
though from afar, and stop suddenly as though torn off. The score for William
Krafts Momentum (1966) instructs several players to run amok at the climax.
Obviously composers can give advice about how to perform a work well, just as
anyone can, and there is no reason why their advice should not be written on
what we call scores. The mentioned instructions of Ligeti and Kraft are plausi-
bly regarded as this kind of advice. If a score is understood to go beyond indicat-
ing the conditions for correct performance, the pattern that it (taken as a whole)
specifies is not to be identified with the work. The pattern constituting the work
is (roughly) the one specified by the score minus whatever advice for good per-
formance it contains.
The fact that many scores are thought of as having the dual function that
Ihave described is, Ithink, very important. But for present purposes Iwill adopt
(for now) the simplifying assumption that scores merely formulate conditions
for correct performance, and hence that works that have scores are the patterns
specified by their scores.

4
This amounts to a view of the nature of musical works that Nicholas Wolterstorff considers and
rejects in Towards an Ontology of Art Works, Nos, 9 (1975):130, namely, his formula (2). Iwould
answer Wolterstorff s objection to (2)by denying that (in his terms) all performances of a work are
necessarily examples of the kind with which the work is identical; i.e., Iclaim that a performance of
a work need not fit perfectly the pattern that constitutes the work. Insofar as he endorses his formula
(4), Wolterstorff and Iagree, however, that the notion of a musical work is closely tied up with the
notion of what constitutes a correct performance of it.
 P re s entation and Por t rayal 235

The distinction between correct and incorrect performances is not limited to

works that have scores. We usually have a fairly clear idea of what counts as a cor-
rect performance of a work even if the conditions for correct performance have
not been notated. Our conception of the conditions for correct performance
amounts to our conception of the identity of the piece. Granted, not every ques-
tion about what sound properties are required for correct performance has a
definite answer. (This is true for scored as well as for unscored works.) Whatever
uncertainties there are about such questions translate into uncertainties about
what patterns works are.5
What is it for a performance to be a performance of a given work? One view
is that it is to fit perfectly the works pattern.6 The perfect-fit theory, as Iwill call
it, has the unsettling consequence that pieces cannot be performed incorrectly,
that the slightest mistake in an attempted performance of Mozarts D Minor
Piano Concerto, for example, aborts the attempt entirely; if the pianist slips and
plays an A-flat where the score calls for an Ahe has not succeeded in perform-
ing Mozarts D Minor Concerto at all, and hence cannot even be said to have
performed it incorrectly.
Moreover, Ifind the claim that fitting a works pattern is a sufficient condition
for being a performance of it just as questionable as the claim that it is a neces-
sary condition. Is every performance that fits perfectly the pattern of Mozarts D
Minor Concerto necessarily a performance of this work? If so, then we will have
to say that, in the example described earlier, the Chicago Symphony Orchestra
performed Marthovens Sixth Symphony while it was performing Beethovens
Sixth Symphony. This seems to me an infelicitous consequence of the perfect-fit
theory.
Developing a better theory requires consideration of the fundamental ques-
tion that Iraised at the outset.
Which are primary:works or performances; sound patterns or sound events?
It will be useful to see how one who holds to the primacy of performances might
spell out her position. This is a tempting view that needs to be understood clearly.
But Idont buy it. It does not describe accurately the roles that performances and
works actually have in our institution of music.
The ultimate objective of composers and performers, according to this posi-
tion, is the production of particular acoustic events that listeners are likely to
find pleasing or fascinating or satisfying or in some way worth listening to. These
sound events are the end products of the musical process. Ultimately they are
all that matter musically. The composers job is simply to give instructions to

5
Cf. Wolterstorff, pp.14041.
6
This is essentially the view espoused by Nelson Goodman in Languages of Art (Indianapolis,
Ind., 1968), pp. 11718, 18687.
236 In Other Shoes

performers, to provide recipes for producing performances. Success or failure

hinges entirely on how interesting or pleasing the resulting sound events are.7
We find it convenient to catalogue or classify musical sound events in many
different ways. We distinguish between sound events in E major and ones in the
Phrygian mode, between polyphonic and monophonic sound events, between
fugues and sonatas. We classify sound events according to how they were pro-
duced, whether by string instruments or by human voices or by electronic syn-
thesizers. We order them chronologically by the dates on which they occur
and by the dates on which scores that led to their occurrence were written. The
notion of musical works constitutes just another way of classifying sound events
(according to the view Iam sketching). Some sound events are performances of
one work, and others are performances of another one. The Martians Idescribed
can be regarded merely as classifying sound events differently than we do. (Their
classifications are useful for them because of the kinds of scores their composers
produce.) But such classifications are no more than conveniences. The sound
events themselves are what matter musically, regardless of how they are sorted.
We do describe musical works, not just performances, as good or bad,
delightful or profound, moving or sentimental. And we do speak of enjoying or
appreciating works. But this can be explained away. What appear to be musical
judgments of works are (one might contend) merely disguised generalizations
about performances. To call a Bartk string quartet delightful or profound, for
example, is simply to claim that all or most performances of that piece are delight-
ful or profound, or perhaps that all or most correct, or good, performances of it
are delightful or profound. An analogous case is the following:To assert that the
North American Bison is a herbivore is merely to advance the generalization
that all or most animals that are North American Bison are herbivores. It is not
to describe some abstract object called the North American Bison as a herbi-
vore. Also, what we call appreciation of a musical work is merely appreciation of
some performance of it plus anticipated appreciation of others. There is no such
thing as appreciation simply of the work itself.
Wracking our brains about which performances are of which works would
seem hardly worth our while, if performances are primary in the way described.
We can listen to sound events, and we can judge and appreciate them musically,

7
J.O. Urmson has suggested that, in all of the performing arts (including music, cooking, ballet,
etc.), what the creative artist does is to produce a recipe or set of performance instructions for the
performer. This is only part of the story, in my opinion, at least with regard to music. Cf. J. O.Urmson,
The Performing Arts, in H. D.Lewis, ed., Contemporary British Philosophy, Fourth Series (London,
1976), pp. 23952, and Literature, in G. Dickie and R. J. Sclafani, eds., Aesthetics: A Critical
Anthology (NewYork, 1977), pp. 33441. It may seem obvious that [m]usic is essentially sound;
the performer produces sounds in accordance with the instructions of the composer (Urmson,
Literature, p.388). But Ifind this claim misleading, for reasons that will become clear shortly.
 P re s entation and Por t rayal 237

without ever wondering what they are performances of. And if we are interested
we can investigate whether or not a given performance conforms to a given
score, or to a manuscript copy of a score, or to the score as the composer meant
it to be; we can investigate whether the performer was following, or meant to
be following, a given score, or whether the sounds he produced were the ones
he intended to produce; we can study the musical tradition to which the perfor-
mance belongswe can do all of this without worrying about what work the
performance is a performance of. And what more is there to ask? Once we have
collected as much of the above information about a sound event as we can, what
else could we want to discover in going on to ask what work it is a performance
of? The notion of musical works and that of sound events being performances
of musical works appear to be entirely inessential, ones we can easily do without.
The characterization sketched above of the roles of works and performances
in the institution of music is, Ibelieve, fundamentally mistaken, and Iwill shortly
offer a better one. But it is worth pointing out here, for the sake of contrast, that
culinary dishes and their instances are plausibly construed on the model of this
false picture of musical works and performances. What is of gastronomic inter-
est is simply the particular morsels of food that we eat, not the abstract culi-
nary dishes of which what we eat are instances. Our concepts of culinary dishes
merely amount to a convenient way of categorizing food. And judgments of a
dish are just generalizations about its instances. To say that Devastatingly Rich
Chocolate Cake is delicious is to say merely that all or most instances of it are
delicious (or that all or most accurately prepared, or well-prepared, instances of
it are delicious). What matters gastronomically is the taste of the food we eat,
not how it is classified. Notions of abstract entities such as Devastatingly Rich
Chocolate Cake are mere conveniences, ones we could do without and still say
everything that needs to be said.
I propose an alternative picture of works and performances. I suggest that
musical workssound patterns rather than sound eventsare objects of musi-
cal attention in their own right. Musical judgments of works are just that, musi-
cal judgments of works. To call the Bartk string quartet delightful or profound
is to describe the Bartk string quartet itself, not merely to generalize about its
performances. And appreciation of a quartet or cantata or sonata is appreciation
of that work, not just actual or anticipated appreciation of performances of it.
This is not to denigrate performances. Performances are vehicles for present-
ing or conveying musical works to us. Sound events provide access to sound pat-
terns. Aperformance is thus at least partly a means to an end rather than an end
in itself. But it is an important means to that end, usually an indispensable one.
Literature is analogous to music in important ways. A copy of a novel is a
vehicle for presenting a word pattern, which is literarily important. But there
is an obvious difference between copies of novels and performances of music.
238 In Other Shoes

Peculiarities of a copy of a novel such as the size, style, and color of the type, and
the quality of the paper are irrelevant in a way in which peculiarities of a musical
performance are not. One copy of a novel cannot be better or worse than another
in the way that one performance of a sonata may be better or worse than another
performance of it. Copies of a novel may of course vary in accuracy or legibility, but
that is another matter.
Words can be spoken as well as written, and word patterns can be conveyed to
us by vocal readings or recitations as well as by written copies. Copies are the usual
vehicles for presenting the word patterns of novels, whereas vocalizations are often
preferred in the case of poetry. Arecitation of a poem is a performance, and is very
much like a musical performance; nuances of intonation and inflection are impor-
tant far beyond the requirements of accuracy and intelligibility. But Iam interested
in the differences between performancesmusical or literaryand copies of liter-
ary works.
Part of the explanation for the fact that the details of a copy of a novel are irrel-
evant in a way those of a musical performance are not may be that whereas a copy
of a novel serves only to present a pattern, a musical performance not only does that
but is also an object of musical interest in its own right. The sound event and the
pattern it presents are both musically significant. (This dual function is not unheard
of in copies of literary works. The calligraphy of a manuscript may be admired in
addition to the pattern it presents. Copies of concrete poems obviously serve both
purposes.)
But performances are important in still another way, a way in which copies of nov-
els are not. Aperformance not only presents a pattern, but portrays it in a certain
light. It interprets, parses, organizes the pattern in some way or other, as well as indi-
cating what it is. And how a pattern is portrayed or interpreted, as well as what it is, is
musically significant. Aperformance may, for example, emphasize certain analogies
between parts of the pattern and obscure others. (One way to emphasize an analogy
is to play the analogous parts with precisely the same nuances of timbre, phrasing,
accent, etc.) One passage of a work may be portrayed as a restatement or variation or
development or elaboration of another. Aperformance may bring out features of the
pattern such as canons, strettos, melodic inversions, retrogrades, or augmentations; a
different performance of the same work may make these features more subtle. Aper-
former can choose to present an extended and complex section of the pattern as spun
from, growing out of, a simple but pregnant motive. Or she might render a simple
concluding passage as a distillation of the essence of the entire piece. Asection may
be portrayed as the beginning of a new idea, or alternatively as a continuation of an
old one. These are some of the ways in which performers interpret patterns.8

8
This helps to explain the appropriateness of saying that both performers and critics interpret
works of art. Cf. Richard Wollheim, Art and Its Objects (NewYork, 1968), p.73.
 P re s entation and Por t rayal 239

The organizing or interpreting of patterns by performances can be compared

to the ordering of data by theories based on them. And good performances often
have virtues analogous to those of good theories; they neatly accommodate all
parts of the pattern and provide an elegant, simple, illuminating, theoretically
satisfying view of them. The performance makes sense out of the pattern,
and makes it easily grasped by the listener. Every detail seems right; there is
a reason for everything. (I do not claim that these are the only virtues of good
performances.)
A copy of a novel, by contrast, merely presents a pattern; it does not also
provide a portrayal or interpretation of it. Different copies may differ greatly;
they may range from a handdrawn parchment to a modern pulp paperback. But
none of the copies provides an interpretation of the work that is more elegant or
revealing or interesting or satisfying or obvious or subtle than others do. Each
of the various copies (provided that it is legible and accurate) simply indicates
what the relevant pattern of words is.
Vocal readings of a poem, performances of a play, and tellings of a story,9 like
musical performances, serve to interpret patterns as well as to present them. This
is why their featuresintonations, inflections, costumes, stagings, the words
one chooses in telling a storymatter more than do features of a copy of a novel
that are not required for legibility and accuracy.
What about culinary dishes? We savor the most subtle tastes of our food.
Virtually every tinge or hint of flavor in a piece of Devastatingly Rich Chocolate
Cake is gastronomically significant. But this is not because the piece of cake
interprets or organizes a pattern, as does a musical performance or a recitation
of a poem. It is because the piece of cake does not serve to present a pattern at all;
our interest is in the taste of the piece of cake itself. Ido not claim that instances
of food could not be understood as presenting or portraying patterns; only that
they usually are not so understood. We rarely if ever think of what we eat as indi-
cating a pattern of taste features, and as organizing or interpreting it in a certain
way, stressing analogies between certain aspects of it, etc., a way different from
how another instance of the same dish might organize or interpret it.
The job, the purpose, of a performance is to present and portray a pattern;
but which pattern? Any sound event conforms perfectly to many different pat-
terns. The Chicago Symphony Orchestras performance discussed earlier fits the
pattern of pitches and durations that we call Beethovens Sixth Symphony. But
it also fits Marthovens Sixth Symphony, a pattern of articulations, dynamics,
accents, timbres; and it fits infinitely many other patterns as well. Asingle sound
event can convey different patterns to different listeners. We hear the Beethoven

9
Stories can be construed as patterns, not of words, but of plot features, which are presented and
portrayed by the specific words that the storyteller chooses.
240 In Other Shoes

pattern in the Chicago Orchestras performance. Our musical experience has

prepared us to discriminate and attend to patterns of notes rather than patterns
of articulations, dynamics, accents, and timbres. Presumably a Martian listener
would hear the Marthoven pattern in the same sound event, because of his dif-
ferent prior musical experience, and that is how he hears the performance by the
Martian Philharmonic. So the same sound event, or acoustically indistinguish-
able ones, convey or present different sound patterns to Earthling listeners and
to Martian listeners. On hearing indistinguishable performances the two audi-
ences recognize and appreciate different musical works.
Most Earthling listeners are not of course explicitly or intellectually aware of
the Beethoven pattern when listening to the performance; they are not capable
of reconstructing Beethovens score, an explicit specification of the pattern, from
what they hear. But most of them could demonstrate in other ways awareness
of the pattern. They may be able to sing parts of it, that is, produce vocal sound
events that fit the pattern at least approximately, or they may be able to recog-
nize other sound events as fitting the pattern, or parts of it. In fact, many lis-
teners could without much difficulty produce vocal sound events that fit parts
of the Beethoven pattern but differ drastically from the Chicago Orchestras
performance in other respects; they could sing melodies that they heard in the
performance but with very different dynamics, tempos, articulations, accents,
etc. Listeners who cannot produce such sound events may well be able at least
to recognize them as melodies that the Chicago Orchestra played, despite the
drastic differences between them and the sounds produced by the orchestra.
It would be enormously difficult for any of us to demonstrate awareness of the
Marthoven pattern in any of these ways, to sing the pattern of articulations,
dynamics, accents, etc., that characterizes the performance but with significantly
different pitches and durations. Nor could we readily recognize sound events as
fitting the Marthoven pattern if they differ substantially in other respects from
the performance.
I am supposing that Martian listeners have the opposite abilities, that after
hearing the performance they could produce or recognize other sound events
that fit (parts of) the same pattern of dynamics, tempos, articulations, accents,
but not ones that conform to the same note pattern. They hear the Marthoven
pattern in the performance but not the Beethoven one.
The members of the Chicago Orchestra had the Beethoven pattern in
mind, and it is the Beethoven pattern that they meant to convey to listeners.
Moreover, the musical culture in which they performed is one in which the stan-
dard purpose of performances is to convey note patterns rather than patterns of
dynamics, etc. Let us say that the function, the role, of the Chicago Orchestras
performance was to present the Beethoven pattern, and not the Marthoven pat-
tern. This, Isuggest, is what makes it a performance of Beethovens work and not
 P re s entation and Por t rayal 241

Marthovens. In general, Ipropose that we consider a sound event a performance

of a given work just in case its role, in the context in which it occurs, is to present
the sound pattern identified with that work. It should be obvious that this way
of understanding what it is to be a performance of a work makes the question of
which performances are of which works much more than a matter, to be decided
more or less arbitrarily, of how to catalogue or classify sound events.
Among the advantages of the theory Ijust sketched is that it allows perfor-
mances to be incorrect. Aperformance is incorrect if it doesnt perfectly fit the
pattern it has the job of presenting. If in a performance of Mozarts D Minor
Piano Concerto the pianists finger slips, hitting an A-flat instead of an A-natural,
the A-flat is a wrong note because the performance, in the context in which it
occurs, has the job of presenting a pattern that calls for an A-natural at that point.
(The pianist intended to play an A-natural; that is what the score he is playing
from calls for; and the audience can be expected to regard the A-flat as out of
place, ungrammatical, and hence to assume that it was unintended.) The func-
tion of the performance is still to present the pattern with the A-natural, the pat-
tern of the concerto, despite the slip. It is still a genuine performance of that
concerto.
Wrong notes and other mistakes are to be distinguished sharply from bad
(unsuccessful, ill-advised, unsatisfying) interpretations. These two sorts of per-
formance defects sound very different (to the practiced listener), and our theory
should explain why they do. To be unsuccessful interpretatively is not a matter
of failing to fit the relevant pattern, but a matter of portraying or interpreting
that pattern in an infelicitous way. An incorrect performance might be likened
to a scientific theory that conflicts with its data; a correct performance whose
interpretation is unsuccessful is more like a theory that is objectionable on other
groundsbecause it constitutes an inelegant or overly complex or unilluminat-
ing or otherwise unsatisfying ordering of the data.
We are now in a position to explain an intriguing difference between musi-
cal performances and copies of literary works. Typographical errors in a copy
of a novel correspond to wrong notes in a musical performance. Typographical
errors consist of a lack of correspondence between the copy and the word pat-
tern it is designed to convey. But they are not nearly as upsetting or disturbing
as wrong notes are. Atypographical error is of little consequence if we can tell,
somehow, how the text should have read; we simply take it as corrected and pro-
ceed. But wrong notes cannot be forgiven so easily. It is not enough merely to
ascertain what note should have been played. An obviously wrong A-flat both-
ers us even if we are perfectly aware that what should have been played is an
A-natural.
Why is this? At least part of the explanation lies in the fact that musical per-
formances, unlike copies of novels, serve not merely to present patterns but also
242 In Other Shoes

to interpret them. If the copy somehow gets across what the relevant pattern is,
it has done its job. But even if a sound pattern is decipherable from an incorrect
performance, the wrong notes are likely to disturb or destroy the performances
interpretation of the pattern. An A-flat where the pattern calls for an A-natural
may well prevent the performance from portraying the passage containing the
mistake as clearly parallel to or analogous to another passage with a correspond-
ing A-natural.
The picture Ihave drawn so far is too simple in several respects. In the first
place, a given musical performance may present to listeners not just one pattern
but several or many. Ido not mean merely that a performance may fit more than
one pattern, of course, but that listeners may hear them in the performance. And
it may be the function of a performance to present not just the pattern specified
approximately by the composers score but others as well. This means that, in
many cases at least, a musical work is better identified with a set of patterns, often
a hierarchically ordered series of them, than with a single pattern (though this
too will turn out to be an oversimplification). Richly complex relations among
the various patterns presented by a performance contribute importantly to the
interest and excitement of listeners experiences.
Exactly which patterns are to be heard in performances of a given piece is
a subtle and delicate question demanding the utmost in musical sensitivity.
Nor is it easy, even for the listener herself, to say which patterns she actually
does hear on a particular occasion. No doubt questions of both sorts frequently
lack definite answers. But we can see how things work by considering plausible
conjectures.
Consider a performance of a Bach Chorale. Isuggest that most listeners hear
in it not only the note pattern indicated by Bachs score, but also the more inclu-
sive pattern that a score in figured bass notation would indicate, the pattern that
any sound event with the bass line and harmonic structure of the Chorale will
fit, regardless of how the harmonies are realized in the upper voices. Evidence
that we do hear this figured bass pattern is provided, again, by the fact that it
is relatively easy for a person with a little musical training to sing or play vari-
ants of the Chorale in which the harmonies are realized very differently in the
upper voices, and it is relatively easy for the practiced listener to recognize such
variants. (Listeners may recognize the similarity between such a variant and a
performance of the Chorale as Bach scored it even if they cannot say what the
similarity consists in.) Notice that the figured bass pattern is more inclusive,
more general than the note pattern in the sense that anything that fits the latter
necessarily fits the former, but not vice versa.
I suggest, also, that listeners are likely to hear in the performance a third pat-
tern intermediate between the note and figured bass patterns, namely, a pattern
that would be indicated by figured bass notation with the upper parts written
 P re s entation and Por t rayal 243

out, but in which it is understood that the performer is free to embellish and
elaborate on the upper voices as she sees fit (within certain limits implicitly
agreed upon).10 To comply with this pattern a sound event must have the bass
line and harmonic structure of the Chorale and also upper voices that conform
in their general shape to what Bach indicated. But the upper voices neednt con-
tain the particular embellishments that Bach wrote out. Bachs score (under-
stood as requiring just the notes that are specifically indicated and allowing no
others) spells out one of many ways of realizing this pattern.
Patterns of the three kinds discussed above are, Ibelieve, to be found in much
of the music of the Common Practice Period.11 But many works of this cor-
pus, especially ones composed after Bach, probably involve patterns of still other
sorts as well. Aperformance of a Brahms song, for instance, may well present to
listeners and may have the function of presenting tothem:

1. a figured bass pattern;

2. a figured bass pattern with restrictions on the general shape of the upper
voices;
3. a note pattern; and in addition
4. a pattern such that to fit it a sound event must not only contain certain notes
and no others, but must also have certain dynamics, tempos, phrasings,
accents, etc., that is, approximately the pattern indicated by Brahmss actual
score.

Is there any limit to the number of patterns a performance might present?

Not in principle. No doubt there are limits to how many patterns human listen-
ers can hear at once in a given sound event, and even stricter limits on how many
they can attend to at once. But Iam inclined to think that listeners are commonly
aware at least vaguely of more than four patterns in a given performance. In any
case a performance may have the function of presenting more patterns than lis-
teners can hear or attend to simultaneously; listeners may hear different combi-
nations of the presented patterns on different occasions. This helps to account
for the remarkable endurance of much music, its ability to withstand enormous
repetition without losing listeners interest.
Patterns 14 constitute a hierarchy of progressively more specific, less inclu-
sive patterns. Any sound event that fits one pattern in the series will fit the preced-
ing ones as well, but may not fit those following it. Anote-perfect performance

10
For a fascinating collection of examples of permissible ways of elaborating various melodic
formulas in an earlier age, see Diego Ortiz, Tratado de glosas sobre clausulas y otros generos de puntos en
la musica de violones (Rome, 1553), ed. Max Schneider (Basel, 1936).
11
Not the same patterns; patterns of the same kinds.
244 In Other Shoes

of the Brahms song that also has dynamics, phrasings, etc., within the ranges
Brahms specified, that is, a performance fitting pattern 4, will automatically sat-
isfy the conditions for fitting patterns 13, but a performance with the bass line
and harmonic structure required by pattern 1 might not qualify as an instance of
any of the other patterns.
The patterns presented by performances of a given work need not be hierar-
chically ordered. Among those that listeners hear in a performance of the Bach
Chorale may be one instantiated by performances of different settings of the
same chorale melody, a pattern requiring the notes of the chorale melody in the
soprano but not much elseperhaps a certain harmonic structure or its out-
lines, perhaps not, and definitely not any particular melodic lines in the inner
voices. This pattern is neither more nor less specific than the intermediate one
consisting of the figured bass and its realization apart from embellishments, or
(probably) the figured bass pattern. Each is such that some sound events that fit
it would fail to fit the other.
But it is very common, it seems to me, for hierarchically ordered patterns to
be presented by a performance. Patterns so ordered constitute levels reminiscent
of Schenkerian theory.12 The relations between the levels can be understood in
terms of the relations Idescribed earlier between performances and the patterns
that Ipreviously identified with works. Each pattern in the series corresponds
to a particular way of portraying or interpreting its less specific neighbor. More
precisely, any performance that fits one of the patterns thereby not only fits the
patterns less specific neighbor in the series but portrays it in a certain manner.
To indicate what a given pattern in the series is, that is, to specify what a sound
event must be like to fit it, is at the same time to indicate a way of organizing or
interpreting the pattern on the next deeper level.
A performance of a work thus presents and portrays all of the patterns in the
series at once. In presenting one it presents and portrays the next; in present-
ing this pattern it presents and portrays a still more general one; and so on. In
fact, a performanceor rather, the very specific sound pattern that only a given
performance and acoustically identical sound events fitcan be regarded as the
shallowest level in the series, an extreme foreground, a surface structure lying in
front of the scored pattern. The other patterns in the series correspond to differ-
ent strata of deep structure, that is, to middleground, background, and interme-
diate levels.
It would seem arbitrary to identify a musical work with any single pattern, the
pattern indicated by its score, for example, if its performances have the function

12
Cf. Heinrich Schenker, Neue Musikalische Theorien und Phantasien, Universal Edition, vol. 1
(Vienna, 1906); vol. 2, pt. 1 (1910); vol. 2, pt. 2 (1922); and numerous secondary sources in recent
music theory.
 P re s entation and Por t rayal 245

of presenting other patterns as well.13 Lets say (for now) that the piece is the set
of patterns that it is the function of any performance of it to present. This will not
include the entire set of patterns a given performance of the piece serves to pres-
ent; it will exclude, for example, performance patterns, ones instantiated only
by sound events aurally indistinguishable from a given performance, for it is not
the function of all performances of the piece to present any particular perfor-
mance pattern. Works that have scores will be sets of patterns whose most spe-
cific member is, approximately, the pattern indicated by the score, though each
of its performances may present one or more other patterns in addition to these.
But it may seem rather arbitrary, also, to make the score the indication of the
cutoff point, or indeed to fix on any definite cutoff point at all. We should allow
some flexibility here, admitting more and less strict notions of piece identity.
Lets say that to be performances of the same piece, two performances must have
the function of presenting all of the same patterns down to a certain level of
specificity, leaving open what might be taken as an appropriate level of speci-
ficity. This will accommodate nicely our hesitations about deciding some ques-
tions of piece identity. In playing a simplified version of Mozarts D Minor Piano
Concerto, do a student pianist and a school orchestra perform the same piece
that Rudolf Serkin performs with the Vienna Philharmonic? Yes, on a loose
notion of piece identity; no, on a stricter notion requiring coincidence of pat-
terns to a lower level of specificity. We can happily leave the matter right there.
But we cannot rest with the identification of musical works and sets of sound
patterns.14
Most scores indicate to performers not only what sorts of sounds they are to
produce, but also the manner in which they are to produce them. Composers
usually specify the instruments to be used, and they sometimes give instructions
about how the instruments are to be manipulated (col legno, pizzicato, con sor-
dino, etc.).
If a musical work is a set of sound patterns and a correct performance is one
that fits those patterns, the sound-making techniques employed by a performer
will not affect the correctness of his performance, so long as they result in the

13
That suggestion would also have the counterintuitive consequence that different performances
that have the job of presenting the same scored patternif that is the anointed onebut very differ-
ent deeper-level patterns would count as performances of the same piece.
14
My discussion in this section is very sketchy. It relies heavily on more general considerations
developed in my Categories of Art, Philosophical Review, 79 (1970): 33467; reprinted in my
Marvelous Images:On Values and the Arts (NewYork, 2008). Jerrold Levinson applied these consid-
erations to music in ways that are similar in some respects to mine, in his What a Musical Work Is,
Journal of Philosophy, 77 (1980):528. The most important differences between his conclusions and
mine are, first, that he takes a single pattern rather than a set of them to be (partially) constitutive of
a work and, second, that he takes the relevant pattern to be a sound/performing-means pattern; no
pure sound pattern is involved.
246 In Other Shoes

right sounds. AChopin Prelude could be performed just as correctly on an elec-

tronic sound synthesizer capable of mimicking the piano perfectly as it can on
a piano. How then are we to understand Chopins stipulation that the Prelude is
for piano, and composers indications of performance-means generally? Perhaps
they serve merely to indicate, indirectly, the kinds of sounds required for correct
performance (in specifying that a piano is to be used the composer indicates that
sounds of the kinds pianos make are required for correct performance), and also
to advise performers as to how such sounds can be achieved.
This conclusion is not a comfortable one. One cause for unease lies in the fact
that the manner in which sounds are produced has a lot to do with their effects on
listeners. Arendition of the Chopin Prelude on the piano may be brilliant, flashy,
energetic, virtuosic; producing the same sounds electronically may constitute a
relatively dull performance. In listening to the pianist our impression may be of a
person and her instrument pushed to their limits, of enormous difficulties over-
come by deep concentration and heroic effort. The synthesizer, in contrast, may
seem merely a machine doing its jobmatter-of-factly, competently, effortlessly,
without emotion.15 (I am assuming that the listener realizes in each case how the
sounds are made.) In short, performances on the synthesizer and on the piano are
likely to sound very differenteven if they are aurally indistinguishable!
Since the Prelude was written for piano, we will want to describe it, not just
the pianists performance, as brilliant, energetic, heroic, virtuosic, etc. Surely
there is something seriously wrong, then, with the duller, distinctly unvirtuosic
electronic performance. Shall we say that at least some of the patterns that con-
stitute the piece are not pure sound patterns, but rather patterns such that to fit
them a sound event must have been produced in a certain manner in addition
to possessing certain acoustic properties?16 If so, we can declare the synthesizer
performance of the Chopin Prelude to be incorrect on the grounds that it fails to
fit one or more of the relevant patterns.
This proposal creates other difficulties. If the patterns constituting a piece are
not sound patterns they are not audible in performances in the way that sound
patterns are, and we will have to reconsider the sense in which performances can
be said to present them to listeners. Iwill shortly suggest a different way of giving
appropriate weight to the manner in which sounds are produced.
Musical works are not to be identified merely with certain patterns or sets
of patterns. We do need to recognize a set of patterns (sound patterns, Iclaim)

15
This difference will be less pronounced if the performance is heard in the category of piano
(-sounding) performances, as well as in that of electronic performances, and perhaps non-existent if
it is not heard also in some such category as that of electronic performances. But the difference will be
dramatic if it is heard as an electronic performance and not as a piano performance. (Depending on
the circumstances, this could be the appropriate way to hear it.) Cf. Categories of Art, (d), in sec. 3.
16
This is Levinsons suggestion, pp.1920.
 P re s entation and Por t rayal 247

corresponding to each work. But this set is best understood as only one compo-
nent of the work.
A modulation that is daring in Mozart may be mundane in a twentieth- or
late nineteenth-century composition. To indicate a given set of sound patterns
in the eighteenth century may be to compose an imaginative, surprising, exciting
piece of music; to specify the very same set of sound patterns in the nineteenth
or twentieth century may be to compose a dull, clichd work. The pieces are
distinct, since they have different aesthetic properties, but the patterns are the
same. So neither work is simply the set of patterns.
The difference between them lies in how their performances are to be heard,
and that is determined in part by the circumstances in which they were com-
posed. Very briefly:performances of the Mozart piece are to be heard as per-
formances of an eighteenth-century work.17 So heard, they sound exciting, and
since they sound exciting when heard appropriately, the piece is an exciting one.
Performances of the later work are to be heard as such. So heard they sound dull;
hence the piece is dull.
Lets identify the piece with the complex consisting of the set of sound
patterns and whatever circumstances (such as the date of composition or the
culture in which it was composed) go into determining how its performances
are to be heard.18 The means by which sounds are produced enters into the lat-
ter rather than the former component of musical works. Performances of the
Chopin Prelude are to be heard as piano performancessince Chopin wrote it
for piano. This, together with the fact that when so heard they sound virtuosic, is
(roughly) what makes the piece virtuosic. The use of a synthesizer in performing
the Prelude forces or demands or encourages listeners to hear the performance
otherwise. This is what is wrong with that performance. We neednt hold that
the sounds produced fail to fit one of the patterns (partially) constitutive of the
work, or one that its performances (qua performances of that piece) have the
job of presenting. The defect is of a different kind. The patterns constitutive of
the work are pure sound patterns, ones that are fully audible in its performances.
A musical work, Isuggested, is a set of sound patterns plus the circumstances
that go into determining how its performances are to be heard (in what cat-
egories they are to be heard). Among these circumstances are, in many cases,
composers designations of the instruments performers are to use and their
instructions about the manner in which they are to use them.

17
They are to be heard in the category of eighteenth-century works, where this category is to be
understood as a perceptually distinguishable one. Cf. Categories of Art, sec. 2.
18
Alternatively, we might identify the piece with the set of patterns plus the appropriate way or
ways of hearing its performances, or with the set of patterns plus the categories its performances are
to be heard in.
 Postscripts to Chapter13:The Presentation and
Portrayalof Sound Patterns

A. Heinrich Schenker:Levels
The Presentation and Portrayal of Sound Patterns is to be faulted for mention-
ing only barely the seminal work of Heinrich Schenker and Schenker-influenced
music theory.1 My primary focus was on certain aspects of the relation between
musical works and performances that Schenker discussed very little.2 But there
is an obvious affinity between his notion of structural levels in musical works
and my brief suggestions in the last several pages of Presentation and Portrayal
about hierarchically ordered patterns, an affinity that is most welcome from my
point of view.
There are differences also, of course. One is that there is nothing in my
scheme corresponding to Schenkers Ursatz (fundamental structure) which he
claims occurs in essentially the same form at the most fundamental level of every
composition (in the repertoire he is interested in). And whereas Schenker, in
the course of his perceptive analyses of particular works, has a lot to say about
what kinds of patterns tend to occur at different levels and what relations obtain
between them, my sketchy observations are meant simply to illustrate a very gen-
eral conception of how musical works might involve patterns and hierarchies of
patterns. Iput no restrictions on what (sonic) features might constitute patterns
at one level or another. There can be rhythmic patterns, patterns of dynamics,

1
Schenkers most influential work is Free Composition (Der Freie Satz), trans. and ed. E. Oster
(NewYork, 1979; orig. pub. 1935).
2
Schenker did offer advice to performers, arguing that performances should be informed by an
understanding of the structure of the work. The performance of a musical work of art can be based
only upon a perception of that works organic coherence. ... the concept of background, middle-
ground, and foreground is of decisive and practical importance for performance (Free Composition,
p.8). And he made specific recommendations about how to perform particular works.

248
 P re s entation and Por t rayal 249

timbres, textures, phrasing, articulation, registers, or formal characteristics, or

any combinations of these, as well as note patterns and patterns of melodic or
harmonic structures.
There is some advantage in stepping back, temporarily at least, from the spe-
cifics of a theory like Schenkers. Abare, unadorned picture of hierarchies of pat-
terns may apply to and illuminate a vastly larger range of music than Schenkers
more detailed picture does or is claimed to. We may find that the presentation
and portrayal of sound patterns is an important aspect of much non-Western
music, for instance (maybe even Martian music), not to mention Western music
other than the mostly Germanic masterpieces of the eighteenth and nineteenth
centuries that occupied Schenkers attentionif we do not expect similarities in
the kinds of patterns that are presented and portrayed.
I claim an insight going beyond anything that Schenker suggested, in my pro-
posal that relations between performances and pieces are to be understood, in
part, in terms of relations between patterns in a hierarchy (244). Aperformance
presents a very specific pattern, what Icalled a performance pattern, thereby
presenting and portraying the more general pattern or patterns that (partially)
constitute the piece. We can think of the performance as a musical surface lying
in front of the surface of the piece.
Another difference between my suggestions and Schenkers theory concerns
what we say about interpretation.3 Iunderstand the more specific patterns in a
hierarchy, the more superficial levels, to portray or interpret the more general,
deeper ones in a certain manner, bringing out (for example) certain relations of
similarity and difference among parts of the more general pattern and obscur-
ing others (238239, 244). Schenker is concerned (mostly) with interpretive
relations going in the opposite direction, and interpretations of a different kind.
Amusic theorist interprets surface level musical events by specifying their deep
structure (including the Ursatz), what they are elaborations of. These two ideas
are obviously not incompatible.

B. Lydia Goehr:Sound Patterns

andtheWorkConcept
I began Presentation and Portrayal by asking, Which is of primary musical
importance, musical works ... or performances of musical works? (230). This
question is misleadingly formulated. Early in my discussion, Iidentified musi-
cal works, tentatively, with sound patterns (232), and proceeded to argue that
works, as well as performances, are objects of musical interest. But I rejected

3
Thanks to Bryan Parkhurst.
250 In Other Shoes

this identification in the end, claiming that works merely correspond to and
are partially constituted by patterns or sets of patterns (246247). The force
of my argument is that patterns are objects of musical interest; listeners hear
them in performances, and appreciate them. Whether and in what ways works
themselves, not just the patterns they present, might be musically important or
objects of musical interest, are questions Idid not address.
This is important for clarifying how my proposals relate to Lydia Goehrs
view that the work-concept began to regulate musical practice at the end of the
eighteenth century.4 Iam skeptical of some of the details of her claim, includ-
ing the date. But the idea that the notion of musical works is a less than essen-
tial aspect of many musical cultures, including some Western ones, is true and
important, and insufficiently acknowledged. I am convinced that listeners in
some contexts and some musical traditions are not expected to hear musical per-
formances as instances of musical works, or to think of them as such. We need
not recognize or have a conception of musical works, however, in order to hear
patterns in music. Ihave no doubt that listeners did and were expected to hear
patterns in performances even in cultures in which works were not recognized,
and certainly in Western societies long before 1800. Hearing patterns neither
constitutes nor necessitates the recognition of musical works, but it must surely
have encouraged development of the work concept whenever that occurred,
and helped to shape it. No doubt the sociological and economic considerations
Goehr discusses also played a role.

C. What Are Musical Works?

Much recent discussion of works and performances, among philosophers of
music, has concentrated on metaphysical puzzles, on questions about what
ontological category musical works belong to, their identity conditions, and
what it is for a performance to be a performance of a given work. The results
have been mostly unexciting, partly because these metaphysical issues have been
treated largely in isolation from aesthetic concerns, from questions about the
value of music, what it does for listeners, how it contributes to our lives.5 An
account of the ontological nature of works and performances ought to mesh, in

4
Lydia Goehr, The Imaginary Museum of Musical Works: An Essay in the Philosophy of Music
(Oxford:Oxford University Press, 1992), p.111.
5
The debate about whether to compose a piece of music is to create something new or to discover
or pick out something (an abstract entity) that has always existed strikes me as particularly sterile.
That the composer exercises creativity, imagination, originality, is not at issue. One does not deni-
grate the creativity or originality of mathematicians if one speaks of their discovering proofs rather
than constructing them.
 P re s entation and Por t rayal 251

revealing ways, with insights about how and why music is important. It was with
this strategic consideration in mind that Ibegan my essay on an aesthetic note,
asking about the relative musical importance of works and performances (230).
My (tentative) thoughts about metaphysical matters grew out of my treatment
of this question.
Heinrich Schenkers interests are largely aesthetic rather than metaphysical;
he doesnt worry about what sort of things musical works are, or what it is for
a given performance to be a performance of a given work. So Iam pleased that
my discussion of pieces, patterns, and performances connects, eventually, with
some of his ideas about what makes the music he loves so wonderful.6

6
Thanks to Bryan Parkhurst for very helpful conversations.
 14

Fearing Fictions
[T]he plot [of a tragedy] must be structured ... that the one who is
hearing the events unroll shudders with fear and feels pity at what hap-
pens:which is what one would experience on hearing the plot of the
Oedipus.
Aristotle, Poetics*

I
Charles is watching a horror movie about a terrible green slime. He cringes in his
seat as the slime oozes slowly but relentlessly over the earth destroying every-
thing in its path. Soon a greasy head emerges from the undulating mass, and two
beady eyes roll around, finally fixing on the camera. The slime, picking up speed,
oozes on a new course straight toward the viewers. Charles emits a shriek and
clutches desperately at his chair. Afterwards, still shaken, Charles confesses that
he was terrified of the slime. Was he?
This question is part of the larger issue of how remote fictional worlds
are from the real world. There is a definite barrier against physical interactions
between fictional worlds and the real world. Spectators at a play are prevented
from rendering aid to a heroine in distress. There is no way that Charles can
dam up the slime, or take a sample for laboratory analysis.1 But, as Charless
case dramatically illustrates, this barrier appears to be psychologically transpar-
ent. It would seem that real people can, and frequently do, have psychological
attitudes toward merely fictional entities, despite the impossibility of physical

Work on this paper was supported by a grant from the American Council of Learned Societies.
Earlier versions were read at a number of universities in the United States, Canada, and Australia.
Iam grateful for the many helpful suggestions made on these occasions. Iam especially indebted to
Holly Smith, Robert Howell, and Brian Loar.
*
Chapter14. Translated by Gerald F.Else (Ann Arbor:The University of Michigan Press, 1967).
1
Iexamine this barrier in a companion piece to the present paper, How Remote Are Fictional
Worlds from the Real World?, Journal of Aesthetics and Art Criticism, XXXVII (Fall 1978):1123.

252
 Fear ing Fic tions 253

intervention. Readers or spectators detest Iago, worry about Tom Sawyer and
Becky lost in the cave, pity Willy Loman, envy Supermanand Charles fears
the slime.
But Iam skeptical. We do indeed get caught up in stories; we often become
emotionally involved when we read novels or watch plays or films. But to con-
strue this involvement as consisting of our having psychological attitudes toward
fictional entities is, Ithink, to tolerate mystery and court confusion. Ishall offer a
different and, in my opinion, a much more illuminating account of it.
This issue is of fundamental importance. It is crucially related to the basic
question of why and how fiction is important, why we find it valuable, why we
do not dismiss novels, films, and plays as mere fiction and hence unworthy of
serious attention. My conclusions in this paper will lead to some tentative sug-
gestions about this basic question.

II
Physical interaction is possible only with what actually exists. That is why
Charles cannot dam up the slime, and why in general real people cannot have
physical contact with mere fictions. But the nonexistence of the slime does
not prevent Charles from fearing it. One may fear a ghost or a burglar even
if there is none; one may be afraid of an earthquake that is destined never to
occur.
But a person who fears a nonexistent burglar believes that there is, or at least
might be, one. She believes that she is in danger, that there is a possibility of her
being harmed by a burglar. It is conceivable that Charles should believe himself to
be endangered by the green slime. He might take the film to be a live documen-
tary, a news flash. If he does, naturally he is afraid.
But the situation I have in mind is the more usual and more interesting
one in which Charles is not deceived in this straightforward way. Charles
knows perfectly well that the slime is not real and that he is in no danger. Is he
afraid even so? He says that he is afraid, and he is in a state which is undeni-
ably similar, in some respects, to that of a person who is frightened of a pend-
ing real-world disaster. His muscles are tensed, he clutches his chair, his pulse
quickens, his adrenalin flows. Let us call this physiological/psychological state
quasi-fear. Whether it is actual fear (or a component of actual fear) is the
question at issue.
Charless state is crucially different from that of a person with an ordinary
case of fear. The fact that Charles is fully aware that the slime is fictional is,
Ithink, good reason to deny that what he feels is fear. It seems a principle of
common sense, one which ought not to be abandoned if there is any reasonable
254 In Other Shoes

alternative, that fear 2 must be accompanied by, or must involve, a belief that one
is in danger. Charles does not believe that he is in danger; so he is not afraid.
Charles might try to convince us that he was afraid by shuddering and declar-
ing dramatically that he was really terrified. This emphasizes the intensity of his
experience. But we need not deny that he had an intense experience. The ques-
tion is whether his experience, however intense, was one of fear of the slime. The
fact that Charles, and others, call it fear is not conclusive, even if we grant that
in doing so they express a truth. For we need to know whether the statement that
Charles was afraid is to be taken literally or not.
More sophisticated defenders of the claim that Charles is afraid may argue
that Charles does believe that the green slime is real and is a real threat to him.
There are, to be sure, strong reasons for allowing that Charles realizes that the
slime is only fictional and poses no danger. If he didnt we should expect him
to flee the theater, call the police, warn his family. But perhaps it is also true that
Charles believes, in some way or on some level, that the slime is real and really
threatens him. It has been said that in cases like this one suspends ones disbe-
lief, or that part of a person believes something which another part of him
disbelieves, or that one finds oneself (almost?) believing something one never-
theless knows to be false. We must see what can be made of these notions.
One possibility is that Charles half believes that there is a real danger, and
that he is, literally, at least half afraid. To half believe something is to be not
quite sure that it is true, but also not quite sure that it is not true. But Charles
has no doubts about whether he is in the presence of an actual slime. If he half
believed, and were half afraid, we would expect him to have some inclination to
act on his fear in the normal ways. Even a hesitant belief, a mere suspicion, that
the slime is real would induce any normal person seriously to consider calling
the police and warning his family. Charles gives no thought whatever to such
courses of action. He is not uncertain whether the slime is real; he is perfectly
sure that it is not.
Moreover, the fear symptoms that Charles does exhibit are not symptoms
of a mere suspicion that the slime is real and a queasy feeling of half fear. They
are symptoms of the certainty of grave and immediate danger, and sheer terror.
Charless heart pounds violently, he gasps for breath, he grasps the chair until his
knuckles are white. This is not the behavior of a man who realizes basically that
he is safe but suffers flickers of doubt. If it indicates fear at all, it indicates acute
and overwhelming terror. Thus, to compromise on this issue, to say that Charles
half believes he is in danger and is half afraid, is not a reasonable alternative.

2
By fear Imean fear for oneself. Obviously a person can be afraid for someone else without
believing that he himself is in danger. One must believe that the person for whom one fears is in
danger.
 Fear ing Fic tions 255

One might claim that Charles believes he is in danger, but that this is not a
hesitant or weak or half belief, but rather a belief of a special kinda gut belief
as opposed to an intellectual one. Compare a person who hates flying. He real-
izes, in one sense, that airplanes are (relatively) safe. He says, honestly, that they
are, and can quote statistics to prove it. Nevertheless, he avoids traveling by air
whenever possible. He is brilliant at devising excuses. And if he must board a
plane he becomes nervous and upset. Igrant that this person believes at a gut
level that flying is dangerous, despite his intellectual belief to the contrary.
Igrant also that he is really afraid of flying.
But Charles is different. The air traveler performs deliberate actions that one
would expect of someone who thinks flying is dangerous, or at least he is strongly
inclined to perform such actions. If he does not actually decide against traveling
by air he has a strong inclination to do so. But Charles does not have even an
inclination to leave the theater or call the police. The only signs that he might
really believe he is endangered are his more or less automatic, nondeliberate,
reactions:his pulse rate, his sweaty palms, his knotted stomach, his spontaneous
shriek.3 This justifies us in treating the two cases differently.
Deliberate actions are done for reasons; they are done because of what the
agent wants and what he thinks will bring about what he wants. There is a pre-
sumption that such actions are reasonable in light of the agents beliefs and desires
(however unreasonable the beliefs and desires may be). So we postulate beliefs
or desires to make sense of them. People also have reasons for doing things that
they are inclined to do but, for other reasons, refrain from doing. If the air traveler
thinks that flying is dangerous, then, assuming that he wants to live, his actions
or tendencies thereto are reasonable. Otherwise, they probably are not. So we
legitimately infer that he does believe, at least on a gut level, that flying is dan-
gerous. But we dont have to make the same kind of sense of Charless automatic
responses. One doesnt have reasons for things one doesnt do, like sweating,
increasing ones pulse rate, knotting ones stomach (involuntarily). So there is no
need to attribute beliefs (or desires) to Charles which will render these responses
reasonable.4 Thus, we can justifiably infer the air passengers (gut) belief in the
danger of flying from his deliberate behavior or inclinations, and yet refuse to
infer from Charless automatic responses that he thinks he is in danger.
Someone might reply that at moments of special crisis during the movie
e.g., when the slime first spots CharlesCharles loses hold of reality and,

3
Charles might scream deliberately. But insofar as he does, it is probably clear that he is only pre-
tending to take the slime seriously. (See section v.)
4
Charless responses are caused partly by a belief, though not the belief that he is in danger. (See
section iv.) This belief is not a reason for responding as he does, and it doesnt make it reasonable, in
the relevant sense, to respond in those ways.
256 In Other Shoes

momentarily, takes the slime to be real and really fears it. These moments are too
short for Charles to think about doing anything; so (one might claim) it isnt
surprising that his belief and fear are not accompanied by the normal inclina-
tions to act.
This move is unconvincing. In the first place, Charless quasi-fear responses
are not merely momentary; he may have his heart in his throat throughout most
of the movie, yet without experiencing the slightest inclination to flee or call
the police. These long-term responses, and Charless propensity to describe
them afterwards in terms of fear, need to be understood even if it is allowed
that there are moments of real fear interspersed among them. Furthermore,
however tempting the momentary-fear idea might be, comparable views of
other psychological states are much less appealing. When we say that some-
one pitied Willy Loman or admired Superman, it is unlikely that we have
in mind special moments during her experience of the work when she forgot,
momentarily, that she was dealing with fiction and felt flashes of actual pity
or admiration. The persons sense of reality may well have been robust and
healthy throughout her experience of the work, uninterrupted by anything like
the special moments of crisis Charles experiences during the horror movie.
Moreover, it may be appropriate to say that someone pities Willy or admires
Superman even when she is not watching the play or reading the cartoon. The
momentary-fear theory, even if it were plausible, would not throw much light
on cases in which we apparently have other psychological attitudes toward
fictions.
Although Charles is not really afraid of the fictional slime depicted in the
movie, the movie might nevertheless produce real fear in him. It might cause
him to be afraid of something other than the slime it depicts. If Charles is a
child, the movie may make him wonder whether there might not be real slimes
or other exotic horrors like the one depicted in the movie, even if he fully real-
izes that the movie-slime itself is not real. Charles may well fear these suspected
actual dangers; he might have nightmares about them for days afterwards. (Jaws
caused a lot of people to fear sharks which they thought might really exist.
But whether they were afraid of the fictional sharks in the movie is another
question.)
If Charles is an older movie-goer with a heart condition, he may be afraid
of the movie itself. Perhaps he knows that any excitement could trigger a heart
attack, and fears that the movie will cause excitement, e.g., by depicting the slime
as being especially aggressive or threatening. This is real fear. But it is fear of the
depiction of the slime, not fear of the slime that is depicted.
Why is it so natural to describe Charles as afraid of the slime, if he is not,
and how is his experience to be characterized? In what follows Ishall develop a
theory to answer these questions.
 Fear ing Fic tions 257

III
Propositions that are, as we say, true in (the world of) a novel or painting or
film are fictional. Thus it is fictional that there is a society of tiny people called
Lilliputians. And in the example discussed above it is fictional that a terrible
green slime is on the loose. Other fictional propositions are associated not with
works of art but with games of make-believe, dreams, and imaginings. If it is
true in a game of make-believe that Johnnie is a pirate, then fictionally Johnnie
is a pirate. If someone dreams or imagines that he is a hero, then it is fictional
that he is a hero.
Fictional truths5 come in groups, and each of these groups constitutes a fic-
tional world. The fact that fictionally there was a society of tiny people and the
fact that fictionally a man named Gulliver was a ships physician belong to the
same fictional world. The fact that fictionally a green slime is on the loose belongs
to a different one. There is, roughly, a distinct fictional world corresponding to
each novel, painting, film, game of make-believe, dream, or daydream.
All fictional truths are in one way or another man-made. But there are two
importantly different ways of making them, and two corresponding kinds of fic-
tional truths. One way to make a proposition fictional is simply to imagine that it
is true. If it is fictional that a person is a hero because he imagines himself to be a
hero, then this fictional truth is an imaginary one. Imagining is not always a delib-
erate, self-conscious act. We sometimes find ourselves imagining things more or
less spontaneously, without having decided to do so. Thoughts pop into our heads
unbidden. Dreams can be understood as simply very spontaneous imaginings.
Fictional truths of the second kind are established in a less direct manner.
Participants in a game of mud pies may decide to recognize a principle to the
effect that whenever there is a glob of mud in a certain orange crate, it is true in
the game of make-believe, i.e., it is fictional, that there is a pie in the oven. This
fictional truth is a make-believe one. The principles in force in a given game of
make-believe are, of course, just those principles which participants in the game
recognize or accept, or understand to be in force.
It can be make-believe that there is a pie in the oven without anyones imagin-
ing that there is. This will be so if there is a glob in the crate which no one knows
about. (Later, after discovering the glob, a child might say, There was a pie in
the oven all along, but we didnt know it.) But propositions that are known to
be make-believe are usually imaginary as well. When kids playing mud pies do
know about a glob in the crate by virtue of which it is make-believe that a pie is
in the oven, they imagine that there is a pie in the oven.

5
Afictional truth is the fact that a certain proposition is fictional.
258 In Other Shoes

Principles of make-believe that are in force in a game need not have been for-
mulated explicitly or deliberately adopted. When children agree to let globs of
mud be pies they are in effect establishing a great many unstated principles
linking make-believe properties of pies to properties of globs. It is implicitly
understood that the size and shape of globs determine the make-believe size and
shape of pies; it is understood, for example, that make-believedly a pie is one
handspan across just in case that is the size of the appropriate glob. It is under-
stood also that if Johnnie throws a glob at Mary then make-believedly Johnnie
throws a pie at Mary. (It is not understood that if a glob is 40 per cent clay then
make-believedly a pie is 40 per cent clay.)
It is not always easy to say whether or not someone does accept, implic-
itly, a given principle of make-believe. But we should notice that much of the
plausibility of attributing to children implicit acceptance of a principle link-
ing the make-believe size and shape of pies to the size and shape of globs rests
on the dispositional fact that if the children should discover a glob to have a
certain size or shape they would imagine, more or less automatically, that a
pie has that size or shape. The children are disposed to imagine pies as having
whatever size and shape properties they think the relevant globs have. In gen-
eral, nondeliberate, spontaneous imagining, prompted in a systematic way by
beliefs about the real world, is an important indication of implicit acceptance
of principles of make-believe. Ido not claim that a person disposed to imag-
ine, nondeliberately, that p when she believes that q necessarily recognizes a
principle of make-believe whereby if q then it is make-believe that p. It must
be her understanding that whenever it is true that q, whether she knows it or
not, it will be fictional that p. It may be difficult to ascertain whether this is her
understanding, especially since her understanding may be entirely implicit.
But the spontaneity of a persons imagining that p on learning that q strongly
suggests that she thinks of p as having been fictional even before she realized
that q.
A game of make-believe and its constituent principles need not be shared
publicly. One might set up ones own personal game, adopting principles that no
one else recognizes. And at least some of the principles constituting a personal
game of make-believe may be implicit, principles which the person simply takes
for granted.
Representational works of art generate make-believe truths. Gullivers Travels
generates the truth that make-believedly there is a society of six-inch-tall people.
It is make-believe that a green slime is on the loose in virtue of the images on
the screen of Charless horror movie. These make-believe truths are generated
because the relevant principles of make-believe are understood to be in force.
But few such principles are ever formulated, and our recognition of most of
them is implicit. Some probably seem so natural that we assume them to be in
 Fear ing Fic tions 259

force almost automatically. Others we pick up easily through unreflective experi-

ence with the arts.6

IV

[The actor] on a stage plays at being another before a gathering of peo-

ple who play at taking him for that other person.
Jorge Luis Borges7

Compare Charles with a child playing an ordinary game of make-believe with

his father. The father, pretending to be a ferocious monster, cunningly stalks the
child and, at a crucial moment, lunges viciously at him. The child flees, scream-
ing, to the next room. The scream is more or less involuntary, and so is the flight.
But the child has a delighted grin on his face even while he runs, and he unhesi-
tatingly comes back for more. He is perfectly aware that his father is only play-
ing, that the whole thing is just a game, and that only make-believedly is there
a vicious monster after him. He is not really afraid.
The child obviously belongs to the fictional world of the game of make-believe.
It is make-believe that the monster lunges, not into thin air, but at the child.
Make-believedly the child is in grave and mortal danger. And when the child
screams and runs, make-believedly he knows he is in danger and is afraid. The
game is a sort of theatrical event in which the father is an actor portraying a mon-
ster and the child is an actor playing himself.
I propose to regard Charles similarly. When the slime raises its head, spies the
camera, and begins oozing toward it, it is make- believe that Charles is threat-
ened. And when as a result Charles gasps and grips his chair, make-believedly
he is afraid. Charles is playing a game of make-believe in which he uses the
images on the screen as props. He too is an actor impersonating himself. In
this section Ishall explain this proposal in detail. My main arguments for it will
come later.

6
I have developed the notion of make-believe truths and other ideas presented in this sec-
tion more fully elsewhere, especially in Pictures and Make-believe, Philosophical Review, LXXXI,
3 (July 1973): 283319. Cf. also Are Representations Symbols?, The Monist, LVIII, 2 (April
1974):236254. Ishould indicate that, in my view, there are no propositions about mere fictions, and
hence none that are make-believe. It is make-believe not that Gulliver visited Lilliput, but that a man
named Gulliver visited a place called Lilliput. Ishall occasionally ignore this point in the interest
of simplicity, for example, when Iwrite in section v as though the same slime resides in two different
fictional worlds. Compare How Remote Are Fictional Worlds from the Real World?, op. cit., note 22.
7
From Everything and Nothing, Borges, Labyrinths:Selected Stories and Other Writings, Donald
A.Yates and James E.Irby, eds. (NewYork:New Directions, 1962), p.218.
260 In Other Shoes

Charles differs in some important respects from an ordinary onstage,

self-portraying actor. One difference has to do with what makes it make-believe
that Charles is afraid. Facts about Charles generate (de re) make-believe truths
about him; in this respect he is like an actor portraying himself on stage. But the
sorts of facts about Charles that do the generating are different. Make-believe
truths about Charles are generated at least partly by what he thinks and feels,
not just by how he acts. It is partly the fact that Charles is in a state of quasi-fear,
the fact that he feels his heart pounding, his muscles tensed, etc., which makes
it make-believe that he is afraid. It would not be appropriate to describe him as
afraid if he were not in some such state.8
Charless quasi-fear is not responsible, by itself, for the fact that make-believedly
it is the slime he fears, nor even for the fact that make-believedly he is afraid rather
than angry or excited or merely upset. Here Charless (actual) beliefs come into
play. Charles believes (he knows) that make-believedly the green slime is bear-
ing down on him and he is in danger of being destroyed by it. His quasi-fear
results from this belief.9 What makes it make-believe that Charles is afraid rather
than angry or excited or upset is the fact that his quasi-fear is caused by the
belief that make-believedly he is in danger. And his belief that make-believedly
it is the slime that endangers him is what makes it make-believe that the slime
is the object of his fear. In short, my suggestion is this:the fact that Charles is
quasi-afraid as a result of realizing that make-believedly the slime threatens him
generates the truth that make-believedly he is afraid of the slime.10
An on-stage actor, by contrast, generates make-believe truths solely by his
acting, by his behavior. Whether it is make-believe that the character portrayed
is afraid or not depends just on what the actor says and does and how he contorts
his face, regardless of what he actually thinks or feels. It makes no difference

8
It is arguable that the purely physiological aspects of quasi-fear, such as the increase of adren-
alin in the blood, which Charles could ascertain only by clinical tests, are not part of what makes it
make-believe that he is afraid. Thus one might want to understand quasi-fear as referring only to the
more psychological aspects of Charless condition:the feelings or sensations that go with increased
adrenalin, faster pulse rate, muscular tension, etc.
9
One cant help wondering why Charless realization that make-believedly he is in danger pro-
duces quasi-fear in him, why it brings about a state similar to real fear, even though he knows he is
not really in danger. This question is important, but we need not speculate about it here. For now we
need only note that Charless belief does result in quasi-fear, however this fact is to be explained. [I
should have mentioned that this belief does not cause Charless quasi-fear by itself. Watching a bad
horror film, one might be aware that make-believedly one is in grave danger without experiencing any
quasi-fear; one might just laugh.]
10
This, Ithink, is at least approximately right. It is perhaps equally plausible, however, to say that
the fact that Charles believes his quasi-fear to be caused by his realization that the slime endangers him
is what makes it make-believe that his state is one of fear of the slime. There is no need to choose now
between my suggestion and this variant.
 Fear ing Fic tions 261

whether his actual emotional state is anything like fear. This is just as true when
the actor is playing himself as it is when he is portraying some other character. The
actor may find that putting himself into a certain frame of mind makes it easier to
act in the appropriate ways. Nevertheless, it is how he acts, not his state of mind, that
determines whether make-believedly he is afraid.
This is how our conventions for theater work, and it is entirely reasonable that
they should work this way. Audiences cannot be expected to have a clear idea of an
actors personal thoughts and feelings while he is performing. That would require
knowledge of his off-stage personality and of recent events that may have affected
his mood (e.g., an argument with his director or his wife). Moreover, acting involves
a certain amount of dissembling; actors hide some aspects of their mental states
from the audience. If make-believe truths depended on actors private thoughts and
feelings, it would be awkward and unreasonably difficult for spectators to ascertain
what is going on in the fictional world. It is not surprising that the make-believe
truths for which actors on stage are responsible are understood to be generated by
just what is visible from the galleries.
But Charles is not performing for an audience. It is not his job to get across to
anyone else what make-believedly is true of himself. Probably no one but he much
cares whether or not make-believedly he is afraid. So there is no reason why his actual
state of mind should not have a role in generating make-believe truths about himself.
It is not so clear in the monster game what makes it make-believe that the child
is afraid of a monster. The child might be performing for the benefit of an audience;
he might be showing someone, an onlooker, or just his father, that make-believedly
he is afraid. If so, perhaps he is like an on-stage actor. Perhaps we should regard his
observable behavior as responsible for the fact that make-believedly he is afraid.
But there is room for doubt here. The child experiences quasi-fear sensations as
Charles does. And his audience probably has much surer access to his mental state
than theater audiences have to those of actors. The audience may know him well,
and the child does not try so hard or so skillfully to hide his actual mental state
as actors do. It may be perfectly evident to the audience that the child has a case
of quasi-fear, and also that this is a result of his realization that make-believedly a
monster is after him. So it is not unreasonable to regard the childs mental state as
helping to generate make-believe truths.
A more definite account of the situation is possible if the child is participating
in the game solely for his own amusement, with no thought of an audience. In
this case the child himself, at least, almost certainly understands his make-believe
fear to depend on his mental state rather than (just) his behavior.11 In fact, let us

Observers might, at the same time, understand his behavior alone to be responsible for his
11

make-believe fear. The child and the observers might recognize somewhat different principles of
make-believe.
262 In Other Shoes

suppose that the child is an undemonstrative sort who does not scream or run
or betray his fear in any other especially overt way. His participation in the
game is purely passive. Nevertheless the child does experience quasi-fear when
make-believedly the monster attacks him, and he still would describe himself
as being afraid (although he knows that there is no danger and that his fear
isnt real). Certainly in this case it is (partly) his quasi-fear that generates the
make-believe truth he expresses when he says he is afraid.
My proposal is to construe Charles on the model of this undemonstrative
child. Charles may, of course, exhibit his fear in certain observable ways. But
his observable behavior is not meant to show anyone else that make-believedly
he is afraid. It is likely to go unnoticed by others, and even Charles himself may
be unaware of it. No one, least of all Charles, regards his observable behavior as
generating the truth that make-believedly he is afraid.

V
It is clear enough now what makes it make-believe that Charles fears the slime,
assuming that make-believedly he does fear the slime. But more needs to be said
in support of my claim that this is a make-believe truth. What needs to be estab-
lished is that the relevant principle of make-believe is accepted or recognized by
someone, that someone understands it to be in force. Icontend that Charles, at
least, does so understand it.
It is clear that Charles imagines himself to be afraid of the slime (though he
knows he is not). He thinks of himself as being afraid of it; he readily describes
his experience as one of fearonce he has a chance to catch his breath. So it is
at least imaginary (and hence fictional) that he fears the slime.
Charless act of imagining himself afraid of the slime is hardly a deliberate
or reflective act. It is triggered more or less automatically by his awareness of
his quasi-fear sensations. He is simply disposed to think of himself as fearing
the slime, without deciding to do so, when during the movie he feels his heart
racing, his muscles tensed, and so forth.11a It is just such a disposition as this,
we recall, that goes with implicit recognition of a principle of make-believe. If
a child is disposed to imagine a pie to be six inches across when she discovers
that that is the size of a glob of mud, this makes it reasonable to regard her as

11a
[R. M.Sainsbury objects to my theory of fiction primarily because of what he calls its highly
active picture of consuming fiction, in Fiction and Fictionalism (New York: Routledge, 2010),
pp.vivii, 19. This is a very unfair characterization of my view, given that imaginings like those of
Charles are automatically, non-voluntarily, provoked by works of fiction. Charles doesnt choose to
imagine being afraid, any more than he chooses to imagine the attacking slime. See also my Mimesis
as Make-Believe (Cambridge, MA:Harvard University Press, 1990), pp. 1416.]
 Fear ing Fic tions 263

recognizing a principle whereby the globs being that size makes it make-believe
that the pie is also. Similarly, Charless tendency to imagine himself afraid of the
slime when he finds himself in the relevant mental state constitutes persuasive
grounds for attributing to him acceptance of a principle whereby his experience
makes it make-believe that he is afraid.12
Several further considerations will increase the plausibility of this conclusion.
First, Ihave claimed only that Charles recognizes the principle of make-believe.
There is no particular reason why anyone else should recognize it, since ordi-
narily only Charles is in a position to apply it and only he is interested in the
make-believe truth that results. Others might know about it and realize how
important it is to Charles. But even so the principle clearly is in important respects
a personal one. It differs in this regard from the principles whereby an on-stage
actors behavior generates make-believe truths, and also from those whereby
images on the movie screen generate make-believe truths about the activities of
the green slime. These principles are fully public; they are clearly (even if implic-
itly) recognized by everyone watching the play or movie. Everyone in the audi-
ence applies them and is interested in the resulting make-believe truths.
This makes it reasonable to recognize two distinct games of make-believe
connected with the horror moviea public game and Charless personal
gameand two corresponding fictional worlds. The situation is analogous to
that of an illustrated edition of a novel. Consider an edition of Dostoyevskys
Crime and Punishment which includes a drawing of Raskolnikov. The text
of the novel, considered alone, establishes a fictional world comprising the
make-believe truths that it generates, e.g., the truth that make-believedly
a man named Raskolnikov killed an old lady. The illustration is normally
understood not as establishing its own separate fictional world, but as com-
bining with the novel to form a larger world. This larger world contains the
make-believe truths generated by the text alone, plus those generated by the
illustration (e.g., that make-believedly Raskolnikov has wavy hair and a reced-
ing chin), and also those generated by both together (e.g., that make-believedly
a man with wavy hair killed an old lady). So we have two fictional worlds,
one included within the other: the world of the novel and the world of the
novel-plus-illustration.

12
These grounds are not conclusive. But the question of whether Charles accepts this principle is
especially tricky, and there is reason to doubt that it can be settled conclusively. One would have to
determine whether it is Charless understanding that, if he were to have the quasi-fear sensations, etc.,
without realizing that he does and hence without imagining that he is afraid, it would still be fictional
that he is afraid. If so, the fictional truth depends not on his imagining but on his quasi-fear, etc. It is
hard to decide whether this is Charless understanding, mainly because it is hard to conceive of his
being ignorant of his quasi-fear sensations, etc. But insofar as Ican get a grip on the question Ithink
that the answer is affirmative.
264 In Other Shoes

Charless state of mind supplements the movie he is watching in the way an

illustration supplements what it illustrates. The movie considered alone estab-
lishes a fictional world consisting only of the make-believe truths that it gener-
ates (e.g., that make-believedly there is a green slime on the loose). But Charles
recognizes, in addition, a larger world in which these make-believe truths are
joined by truths generated by Charless experience as he watches the movie, and
also by truths generated by the images on the screen and Charless experience
together. It is only in this more inclusive world that make-believedly Charles
fears the slime. (And it is the larger world that occupies Charless attention when
he is caught up in the movie.)
The analogy between Charless case and the illustrated novel is not perfect.
The novel-plus-illustration world is publicly recognized, whereas the fictional
world established by the movie plus Charless experience of it probably is not.
Dolls provide an analogy which is better in this respect. Anyone who sees a doll
of a certain sort will recognize that it generates the truth that make-believedly
there is a blonde baby girl. The doll, regarded simply as a sculpture to be
observed from a distance, generates make-believe truths such as this. But a
child playing with the doll is playing a more personal game of make-believe,
one in which she herself is a self-portraying actor and the doll serves as a prop.
What she does with the doll generates make-believe truths, e.g., the truth that
make-believedly she is dressing the baby for a trip to town. Similarly, Charles
uses the screen images as props in a personal game of make-believe in which
he himself is a character. He plays his own game with the images. The screen
images, of course, do not lend themselves to bring dressed or manipulated in
all the ways that dolls do, and this limits the extent of Charless participation in
the game. But the relations and interactions between Charles and the images
do generate a number of important make-believe truths:that make-believedly
Charles notices the slime and stares apprehensively at it, that make-believedly
it turns toward him and attacks, and that make-believedly he is scared out of
his wits.13
One source of uneasiness about my claim that make-believedly Charles fears
the slime may have been the impression that this can be so only if Charles belongs
to the fictional world of the movie. (The movie itself doesnt depict Charles, nor
does it make any reference to him, so he doesnt belong to the movie-world.) My
two-worlds theory shows that this impression is mistaken and hence that the
uneasiness based on it is out of place.

13
One important difference between dolls and the screen images is that the dolls generate de re
make-believe truths about themselves and the images do not. The doll is such that make-believedly it
is a baby that is being dressed for a trip to town. But a screen image is not such that make-believedly
it (the image itself) is a green slime.
 Fear ing Fic tions 265

I have portrayed Charles so far as participating rather automatically in his

game of make-believe. But he might easily slip into participating deliberately.
The naturalness of his doing so gives added support to my claim that Charles
does recognize a make-believe world that he and the slime share, even when his
participation is not deliberate. Suppose that during the movie Charles exclaims,
deliberately, to a companion or to himself, Yikes, here it comes! Watch out!
How are we to understand this verbal action? Certainly Charles is not seri-
ously asserting that a slime is coming and warning himself or his companion
of it. Presumably he is asserting that it is make-believe that a slime is coming.
But the indexical, here, carries an implicit reference to the speaker. So Charless
exclamation shows that he takes it to be make-believe that the slime is headed
toward him; it shows that he regards himself as coexisting with the slime in a
make-believe world.
But this does not take us to the bottom of the matter. Yikes! and Watch
out! are not assertions, and so not assertions of what make-believedly is the case.
Moreover, if in saying, Here it comes, Charles were merely making an assertion
about what make-believedly is the case, he could well have made this explicit
and exclaimed instead, Make-believedly the slime is coming! or The slime is
coming, in the fictional world! But these variants lack the flavor of the original.
Charless exclamatory tone is absurdly out of place when the make-believe status
of the danger is made explicit. Compare how ridiculous it would be for an actor
playing Horatio in a performance of Hamlet to exclaim, when the ghost appears,
Look, my lord, it comes, in the fictional world of the play!
The comparison is apt. For Charles is doing just what actors do, pretending to
make an assertion. He is pretending to assert (seriously) that the slime is headed
his way. (Pretending to assert this is not incompatible with actually asserting
that make-believedly the slime is coming. Charles might be doing both at once.)
In my terms, Charles understands his utterance of Here it comes! to generate
the truth that make-believedly he asserts (seriously) that the slime is coming.
He is playing along with the fiction of the movie, incorporating it into a game
of make-believe of his own. This makes it obvious why it would not do to say,
Here it comes, in the fictional world! Saying that is simply not (normally) how
one would pretend to assert that a slime is (really) coming. The rest of Charless
verbal behavior is now easily explainable as well. In saying Yikes! and Watch
out! he is pretending to express amazement or terror and pretending to issue a
(serious) warning; make-believedly he is doing these things.
We have now arrived at the solution to a pair of puzzles. Why is it that in
everyday conversation we regularly omit phrases like in the fictional world and
in the novel, whereas we rarely omit other intensional operators such as It is
believed that, Jones wished that, Jones denies that? Why do we so naturally
say just Tom and Becky were lost in a cave rather than In the novel Tom and
266 In Other Shoes

Becky were lost in a cave, whereas it would be almost unheard of to shorten

Jones wishes that a golden mountain would appear on the horizon to simply
Agolden mountain will appear on the horizon (even if the context makes it
clear that Joness wishes are the subject of conversation)?
The explanation lies in our habit of playing along with fictions, of
make-believedly asserting, pretending to assert, what we know to be only
make-believedly the case. We mustnt be too quick to assume that an utter-
ance of p is merely an ellipsis for Make-believedly p (or for In the novel p).
This assumption is wrong if the speaker make-believedly is asserting that p,
rather than (or in addition to) asserting that make-believedly p. Charless fran-
tic, Yikes, here it comes! is an obvious case in point. Acase only slightly less
obvious is that of a person reading The Adventures of Tom Sawyer who remarks,
gravely and with an expression of deep concern, that Tom and Becky are lost
in a cave.
I do not suggest that the omission of in the novel is never a mere ellipsis.
Tom and Becky were lost in a cave uttered by a critic analyzing the novel could
easily have been expanded to In the novel Tom and Becky were lost in a cave
without altering the character of the remark. The critic probably is not pretend-
ing to assert that Tom and Becky were (actually) lost in a cave. But our habit of
dropping fictional operators persists even in sober criticism, and testifies to the
ease with which we can be induced to play along, deliberately, with a work of
fiction.
In German the indicative mood is used ordinarily only when the speaker is
committed to the truth of the sentence or clause in question. But fictional state-
ments constitute a striking exception to this generalization; the indicative is used
in fictional statements even though the speaker is not committed to their truth.
(One says, for example, Robinson Crusoe hat einen Schiffbruch berlebt, which
is indicative, even though one is not claiming that there actually was a person
named Robinson Crusoe who survived a shipwreck.) The explanation is that
speakers are often pretending to express their commitment to the truth of sen-
tences or clauses in fictional contexts. So naturally they use the indicative mood
in these cases; they speak as they would if they were not pretending. And the
habit of using the indicative persists even when there is little or no such pretense.

VI
The treatment of Charless fear of the slime suggested above can serve as a
model for understanding other psychological attitudes ostensibly directed
toward fictional things. When it is said that someone pities Willy Loman, or
worries about Tom and Becky, or detests Iago, or envies Superman, what is said
 Fear ing Fic tions 267

is probably not literally true.14 But the person is, actually, in a distinctive psycho-
logical (emotional?) state, even if that state is not pity or worry or hate or envy.
And his being in this state is a result of his awareness of certain make-believe
truths: that make-believedly Willy is an innocent victim of cruel circum-
stances, that make-believedly Tom and Becky might perish in the cave, that
make-believedly Iago deceived Othello about Desdemona, that make-believedly
Superman can do almost anything. The fact that the persons psychological state
is as it is, and is caused by such beliefs, makes it make-believe that he pities Willy,
worries about Tom and Becky, hates Iago, or envies Superman.
We have here a particularly intimate relation between the real world and fic-
tional worlds. Insofar as make-believe truths are generated by a spectators or
readers state of mind, he is no mere external observer of the fictional world.
Ascertaining what make-believedly is true of himself is to a large extent a matter
of introspection (or of whatever sort of privileged access one has to ones own
beliefs and sensations). In fact, when Charles watches the horror movie, for exam-
ple, introspection is involved in ascertaining not merely that make-believedly he
is afraid of the slime, but also make-believe truths about the nature and prog-
ress of his fear. If it is make-believe that his fear is overwhelming, or that it is
only momentary, this is so because his quasi-fear sensations are over-whelming,
or are only momentary. Make-believedly his fear grows more or less intense, or
becomes almost unbearable, or finally subsides, etc., as his quasi-fear feelings
change in these ways. So it is by attention to the nature if his own actual experi-
ence that Charles is aware of make-believe truths about the nature of his fear. He
follows the progress of his make-believe fear by introspection, much as one who
is literally afraid follows the progress of his actual fear.
It would not be too far wrong to say that Charles actually experiences his
make-believe fear. Idont mean that there is a special kind of fear, make-believe
fear, which Charles experiences. What he actually experiences, his quasi-fear
feelings, are not feelings of fear. But it is true of them that make-believedly they are
feelings of fear. They generate de re make-believe truths about themselves, and so
belong to the fictional world just as Charles himself does. What Charles actually
experiences is such that make-believedly it is (an experience of) fear.
Cases like that of Charles contrast strikingly with others in which an actual
person belongs to a fictional world. Consider a performance of William Luces

14
Assuming of course that the person realizes that he is dealing with a work of fiction. Even so,
arguments are needed to show that such statements are not literally true, and I shall not provide
them here. But it is plausible that pity, worry about, hate, and envy are such that one cannot have
them without believing that their objects exist, just as one cannot fear something without believ-
ing that it threatens one. Yet even if one can, and does, envy a character, for example, it may also be
make-believe that one does so, and this make-believe truth may be generated by facts of the sort my
theory indicates.
268 In Other Shoes

play about Emily Dickinson, The Belle of Amherst, in which Julie Harris plays
Emily Dickinson. Suppose that Emily Dickinson herself, with the help of a time
machine or a fortuitous reincarnation, is in the audience. In order to discover
make-believe truths about herself, including what make-believedly she thinks
and feels, Dickinson must observe Julie Harriss actions, just as any specta-
tor must. It is as though she is watching another person, despite the fact that
that person, the character, is herself. Dickinson has no special intimacy with
make-believe truths about her own mental state.15 The situation is basically the
same if Dickinson should replace Julie Harris in the lead role and act the part
herself. She still must judge from her external behavior, from what spectators
could observe, whether or not it is make-believe that she is afraid or worried or
whateverand she might easily be mistaken about how she looks to spectators.
It is still as though she considers herself from the outside, from the perspective
of another person.
This is clearly not true of Charles. It is not as though Charles were confront-
ing another person, a fictional version of himself, but rather as though he him-
self actually fears the slime. (Nevertheless, he does not.) Make-believe facts
about his fear, especially the fact that make-believedly it is his, are portrayed to
Charles in an extraordinarily realistic manner. And make-believe facts about our
pity for Willy, our dislike of Iago, and so forth, are similarly vivid to us. We and
Charles feel ourselves to be part of fictional worlds, to be intimately involved
with the slime, or Willy, or with whatever constituents of fictional worlds are,
make-believedly, objects of our feelings and attitudes.
We see, now, how fictional worlds can seem to us almost as real as the real
world is, even though we know perfectly well that they are not. We have begun
to understand what happens when we get emotionally involved in a novel or
play or film, when we are caught up in the story.
The theory Ihave presented is designed to capture intuitions lying behind
the traditional ideas that the normal or desired attitude toward fiction involves a
suspension of disbelief, or a decrease of distance. These phrases are unfortunate.
They strongly suggest that people do not (completely) disbelieve what they read
in novels and see on the stage or screen, that, e.g., we somehow accept it as fact
that a boy named Huckleberry Finn floated down the Mississippi Riverat
least while we are engrossed in the novel. The normal reader does not accept
this as fact, nor should he. Our disbelief is suspended only in the sense that

15
Ihave in mind those make-believe truths about her mental state that are generated by what hap-
pens on stage. Dickinson is not only a character in the play, but also a spectator. In the latter capacity
she is like Charles; her actual mental state generates make-believe truths about herself. Dickinson is
in a curiously ambiguous position. But it is not an uncommon one; people frequently have dreams in
which they watch themselves (from the outside) doing things.
 Fear ing Fic tions 269

it is, in some ways, set aside or ignored. We dont believe that there was a Huck
Finn, but what interests us is the fact that make-believedly there was one, and that
make-believedly he floated down the Mississippi and did various other things.
But this hardly accounts for the sense of decreased distance between us and fic-
tions. It still has us peering down on fictional worlds from reality above, however
fascinated we might be, for some mysterious reason, by what we see.
On my theory we accomplish the decrease of distance not by promoting
fictions to our level but by descending to theirs. (More accurately, we extend
ourselves to their level, since we do not stop actually existing when it becomes
fictional that we exist.) Make-believedly we do believe, we know, that Huck Finn
floated down the Mississippi. And make-believedly we have various feelings and
attitudes about him and his adventures. Rather than somehow fooling ourselves
into thinking fictions are real, we become fictional. So we end up on the same
level with fictions. And our presence there is accomplished in the extraordi-
narily realistic manner that Idescribed. This enables us to comprehend our sense
of closeness to fictions, without attributing to ourselves patently false beliefs.
We are now in a position to expect progress on the fundamental question of
why and how fiction is important. Why dont we dismiss novels, plays, and films
as mere fiction and hence unworthy of serious attention?
Much has been said about the value and importance of dreams, fantasy, and
childrens games of make-believe.16 It has been suggested, variously, that such
activities serve to clarify ones feelings, help one to work out conflicts, provide
an outlet for the expression of repressed or socially unacceptable feelings, pre-
pare one emotionally for possible future crises by providing practice in facing
imaginary crises. It is natural to presume that our experience of representational
works of art is valuable for similar reasons. But this presumption is not very plau-
sible, Ithink, unless something like the theory Ihave presented is correct.
It is my impression that people are usually, perhaps always, characters in
their own dreams and daydreams. We dream and fantasize about ourselves.
Sometimes ones role in ones dream-world or fantasy-world is limited to that
of observing other goings-on. But to have even this role is to belong to the fic-
tional world. (We must distinguish between being, in ones dream, an observer
of certain events, and merely observing, having, a dream about those events.)
Similarly, children are nearly always characters in their games of make-believe.
To play dolls or school, hobby horses or mud pies, is to be an actor portraying
oneself.
I suggest that much of the value of dreaming, fantasizing, and making-believe
depends crucially on ones thinking of oneself as belonging to a fictional

Agood source concerning make-believe games is Jerome L.Singer, etal., The Childs World of
16

Make-Believe (NewYork:Academic Press, 1973).

270 In Other Shoes

world. It is chiefly by fictionally facing certain situations, engaging in certain

activities, and having or expressing certain feelings, I think, that a dreamer,
fantasizer, or game player comes to terms with her actual feelingsthat she
discovers them, learns to accept them, purges himself of them, or whatever
exactly it is that she does.
If Iam right about this, people can be expected to derive similar benefits from
novels, plays, and films only if it is fictional that they themselves exist and par-
ticipate (if only as observers) in the events portrayed in the works, i.e., only if my
theory is on the right track.
I find encouragement for these speculations in the deliberate use of
role-playing in educational simulation games, and as a therapeutic technique in
certain kinds of psychotherapy (e.g., Gestalt therapy). Atherapist may ask her
patient to pretend that his mother is present, or that some inanimate object is
his mother, and to talk to her. He may then be asked to be the mother, and to
say how he feels (when he is the mother), how he acts, what he looks like, etc.
Iwill not venture an explanation of how such therapeutic techniques are effec-
tive, nor of why simulation games work. But whatever explanation is appropriate
will, Isuspect, go a long way toward explaining why we are as interested in works
of fiction as we are, and clarifying what we get from them. The important place
that novels, plays, and films have in our lives appears mysterious only on the sup-
position that we merely stand outside fictional worlds and look in, pressing our
noses against an inviolable barrier. Once our presence within fictional worlds is
recognized, suitable explanations seem within reach.

VII
A more immediate benefit of my theory is its capacity to handle puzzles. Icon-
clude with the resolution of two more. First, consider a playgoer who finds
happy endings asinine or dull, and hopes that the play he is watching will end
tragically. He wants the heroine to suffer a cruel fate, for only if she does, he
thinks, will the play be worth watching. But at the same time he is caught up
in the story and sympathizes with the heroine; he wants her to escape. It is
obvious that these two apparent desires may perfectly well coexist. Are we to
say that the spectator is torn between opposite interests, that he wants the hero-
ine to survive and also wants her not to? This does not ring true. Both of the
playgoers conflicting desires may be wholehearted. He may hope unreserv-
edly that the work will end with disaster for the heroine, and he may, with equal
singlemindedness, want her to escape such an undeserved fate. Moreover,
he may be entirely aware of both desires, and yet feel no particular conflict
between them.
 Fear ing Fic tions 271

My theory provides a neat explanation. It is merely make-believe that the

spectator sympathizes with the heroine and wants her to escape. And he (really)
wants it to be make-believe that she suffers a cruel end. He does not have con-
flicting desires. Nor, for that matter, is it make-believe that he does.
The second puzzle concerns why it is that works last as well as they do, how they
can survive multiple readings or viewings without losing their effectiveness.17
Suspense of one kind or another is an important ingredient in our experience
of most works:Will Jack, of Jack and the Beanstalk, succeed in ripping off the
giant without being caught? Will Tom and Becky find their way out of the cave?
Will Hamlet ever get around to avenging the murder of his father? What is in
store for Julius Caesar on the Ides of March? Will Godot come?
But how can there be suspense if we already know how things will turn out?
Why, for example, should Tom and Beckys plight concern or even interest a
reader who knows, from reading the novel previously, that eventually they will
escape from the cave? One might have supposed that, once we have experienced
a work often enough to learn thoroughly the relevant features of the plot, it
would lose its capacity to create suspense, and that future readings or viewings of
it would lack the excitement of the first one. But this frequently is not what hap-
pens. Some works, to be sure, fade quickly from exposure, and familiarity does
alter our experience in certain ways. But the power of many works is remark-
ably permanent, and the nature of their effectiveness remarkably consistent.
In particular, suspense may remain a crucial element in our response to a work
almost no matter how familiar we are with it. One may worry just as intensely
about Tom and Becky while rereading The Adventures of Tom Sawyer, despite
ones knowledge of the outcome, as would a person reading it for the first time.
Achild listening to Jack and the Beanstalk for the umpteenth time, long after she
has memorized it word for word, may feel much the same excitement when the
giant discovers Jack and goes after him, the same gripping suspense, that she felt
when she first heard the story. Children, far from being bored by familiar stories,
often beg to hear the same ones over and over again.
None of this is surprising on my theory. The child hearing Jack and the
Beanstalk knows that make-believedly Jack will escape, but make-believedly
she does not know that he willuntil the reading of the passage describing his
escape. She is engaged in her own game of make-believe during the reading, a
game in which make-believedly she learns for the first time about Jack and the
giant as she hears about them.18 It is her make-believe uncertainty (the fact that

17
David Lewis pointed out to me the relevance of my theory to this puzzle.
18
It is probably make-believe that someone (the narrator), whose word the child can trust, is giv-
ing her a serious report about a confrontation between a boy named Jack and a giant. Cf. my Points
of View in Narrative and Depictive Representation, Nos, x, 1 (March 1976):4961.
272 In Other Shoes

make-believedly she is uncertain), not any actual uncertainty, that is responsible

for the excitement and suspense that she feels. The point of hearing the story is
not, or not merely, to learn about Jacks confrontation with the giant, but to play
a game of make-believe. One cannot learn, each time one hears the story, what
make-believedly Jack and the giant do, unless one always forgets in between
times. But one can and does participate each time in a game of make-believe. The
point of hearing Jack and the Beanstalk is to have the experience of being such
that, make-believedly, one realizes with trepidation the danger Jack faces, waits
breathlessly to see whether the giant will awake, feels sudden terror when he
does awake, and finally learns with admiration and relief how Jack chops down
the beanstalk, killing the giant.
Why play the same game over and over? In the first place, the game may not
be exactly the same each time, even if the readings are the same. On one occasion
it may be make-believe that the child is paralyzed by fear for Jack, overwhelmed
by the gravity of the situation, and emotionally drained when Jack finally bests
the giant. On another occasion it may be make-believe that the child is not very
seriously concerned about Jacks safety and that her dominant feelings are admi-
ration for Jacks exploits, the thrill of adventure, and a sense of exhilaration at the
final outcome. But even if the game is much the same from reading to reading,
ones emotional needs may require the therapy of several or many repetitions.
 15

Spelunking, Simulation, and Slime

On Being Moved by Fiction

Works of fiction induce in appreciators thoughts about people, situations and

events; lets say that they induce appreciators to imagine them. The imagined
people, situations, and events are frequently ones that do not really exist or
occur, but we can (as we do) speak of them as constituting a fictional world,
the world of the novel or story or film.
This much is not controversial. But it is important to realize how little it
comes to, how much remains to be explained. Why should we be interested in
these nonexistents? Why should we bother thinking about or imagining them?
We havent yet distinguished novels and other fictions from a mere list of sen-
tencessentences used in a grammar lesson, for instance. When we read and
understand these sentences, they induce us to entertain the thoughts that they
express. But that is all. Fictional worlds seem so far to be worlds apart, worlds
having nothing to do with us, ones that we merely peer into from afar. As Iput it
in Mimesis as Make-Believe:

If to read a novel or contemplate a painting were merely to stand out-

side a fictional world pressing ones nose against the glass and peer in,
... our interest in novels and paintings would indeed be mysterious. We
might expect to have a certain clinical curiosity about fictional worlds
viewed from afar, but it is hard to see how that could account for the sig-
nificance of representations, their capacity to be deeply moving, some-
times even to change our lives.1

Consider childrens games of make-believe. Children do not peer into worlds

apart, nor do they merely engage in a clinical intellectual exercise, entertaining
thoughts about cops and robbers, or whatever. The children are in the thick of

1
Walton 1990:273.

273
274 In Other Shoes

things; they participate in the worlds of their games. We appreciators also par-
ticipate in games of make-believe, using works as props. Participation involves
imagining about ourselves as well as about the characters and situations of the
fictionbut not just imagining that such and such is true of ourselves. We imag-
ine doing things, experiencing things, feeling in certain ways. We bring much of our
actual selves, our real-life beliefs and attitudes and personalities, to our imagina-
tive experiences, and we stand to learn about ourselves in the process.2
There have been lively discussions recently, in philosophy of mind and cogni-
tive science, about what is called mental simulation. Fiction and the represen-
tational arts are rarely mentioned in them, but the notion of mental simulation
dovetails almost uncannily with my make-believe theory. Insights concerning
simulation reinforce and augment my theory of fiction. Indeed, the participation
in make-believe that Idescribed is itself a form of mental simulation.3
Many discussions of Mimesis as Make-Believe have concentrated on my nega-
tive claim that it is not literally true, in ordinary circumstances, that apprecia-
tors fear, fear for, pity, grieve for, or admire purely fictitious characters. Charles,
who fidgets and tenses and screams as he watches a horror movie, does not,
I argue, really fear the Slime portrayed on the screen. The reasons that have
been advanced against this claim are, in my opinion, very weak.3a Of even more
2
The preceding points are explained much more fully in Walton 1990.
3
For discussions of mental simulation, see Gordon 1987: chap. 7, and the essays collected in
Davies and Stone 1995a, and 1995b. Mental simulation is related to older notions of empathy, which
have figured prominently in the discussions of the arts, but it is their more recent incarnation in the
form of mental simulation that exhibits special affinity to my make-believe theory. Gregory Currie
links fiction to simulation theory in Imagination and Simulation: Aesthetics Meets Cognitive
Science, in Davies and Stone 1995b. See also Feagin 1996.
3a
[Several critics point to empirical research showing that fiction causes in appreciators affective
responses that are similar phenomenologically to ones experienced in corresponding real life cases.
This has never been in dispute. It leaves entirely open the question of whether appreciators are best
understood to be experiencing emotions having fictional entities as intentional objects. Iam happy
to agree with Paul Harris (2000, c hapter 4) that our emotional system is engaged when we are
absorbed in a fiction, which Itake to mean that similar physiological and neurological processes are
involved. This is what we should expect, since the results are phenomenologically similar.]
Derek Matravers (1991) takes it that one can be sad or afraid, without being sad about, or afraid
of, anything in particular. Then one is just feeling frightened or sad (324, his emphasis). He is cer-
tainly right about sadness. In Fearing Fictions (in this volume), Iunderstood fear always to have an
intentional object, and so took Charles not to be afraid (unless he is afraid of something other than
the Slime). If a feeling of fear need not have any object at all, Charles might well experience such a
feeling even if his fear of the slime is merely fictional. There remains the question of what is meant by
anything in particular. Perhaps a feeling of fear is necessarily a feeling that something or other bad
might happen. If so, genuinely feeling fear is still consistent with not literally fearing the slime. Iwont
speculate about whether, or how often, appreciators of fictions have feelings with non-particular
(propositional) objects of this sort. Note that Matravers does not count these feelings as emotions.
He may have in mind the common terminological recommendation that only object-directed states
be called emotions (as opposed to moods).]
 Spelunk ing , Simulation , and S lime 275

concern, however, is the undue emphasis that commentators have put on this
issue, at the expense of the positive aspects of my make-believe theory. This
presents me with the ticklish job of defending the negative claim while direct-
ing attention to other more important matters. Simulation theory will be help-
ful in both parts of this task. In particular, it will help to counter a surprisingly
prevalent assumption that imagining (and make-believe, which Iunderstand in
terms of imagining), or the kind of imagining central to my theory, can only be
a clinical, antiseptic, intellectual exercise, and so cannot have a central role in
explaining the genuinely emotional responses to fiction that appreciators (often)
experience.
It goes without saying that we are genuinely moved by novels and films and
plays, that we respond to works of fiction with real emotion. Some have miscon-
strued my make-believe theory as denying this. [T]he key objection to Waltons
theory, says Nol Carroll, is that it relegates our emotional responses to fic-
tion to the realm of make-believe.4 That would indeed be a mistake. In fact, our
responses to works of fiction are, not uncommonly, more highly charged emo-
tionally than our reactions to actual situations and people of the kinds the work
portrays. My make-believe theory was designed to help explain our emotional
responses to fiction, not to call their very existence into question. My negative
claim is only that our genuine emotional responses to works of fiction do not
involve, literally, fearing, grieving for, admiring fictional characters.
Lets begin with an experiment. Imagine going on a spelunking expedition.
You lower yourself into a hole in the ground and enter a dank, winding pas-
sageway. After a couple of bends there is absolute pitch darkness. You light the
carbide lamp on your helmet and continue. The passage narrows. You squeeze
between the walls. After a while you have to stoop, and then crawl on your hands
and knees. On and on, for hours, twisting and turning and descending. Your
companion, following behind you, began the trip with enthusiasm and confi-
dence; in fact she talked you into it. But you notice an increasingly nervous edge
in her voice. Eventually, the ceiling gets too low even for crawling; you wriggle
on your belly. Even so, there isnt room for the pack on your back. You slip it off,
reach back, and tie it to your foot; then continue, dragging the pack behind you.
The passage bends sharply to the left, as it descends further. You contort your
body, adjusting the angles of your shoulders and pelvis, and squeeze around and
down. Now your companion is really panicked. Your lamp flickers a few times,
then goes out. Absolute pitch darkness. You fumble with the mechanism ...
This experiment demonstrates the power of the imaginationthe power on
me of my imagination, anyway. Idid not for a moment, while Iwas composing
the preceding paragraph or reading it over, think Iactually was wriggling on my

4
Carroll 1990:7374.
276 In Other Shoes

belly in a cave, or really see the dank walls of the passageway close in on me in
the flickering light of a carbide lamp. Iimagined all of this, merely imagined it. Yet
my imaginative experience was genuinely distressing, upsettingloaded with
affect, as psychologists say. Even rereading the paragraph for the umpteenth
time gives me the shivers.
The results of the experiment may not be the same for everyone. You may
not find it distressing to imagine crawling in a cavemaybe you arent claus-
trophobic. In that case, a different experiment would probably demonstrate to
you the power of the imagination. Try imagining climbing a nearly vertical rock
face, looking down on a valley several thousand feet below, as the wind screams
around you. Or imagine being in an automobile accident, or discovering an
intruder in your home.
My imagining of the spelunking expedition taps into my actual personality
and character. This, I am sure, is why it affects me as it does. It is because of
my (dispositional) claustrophobia that Ifind it distressing to imagine slithering
on my belly through the cramped passages of the cave. The slithering is only
imagined, but imagining it activates psychological mechanisms Ireally possess,
and brings on genuine distress. What Icalled the power of the imagination is
really the power of dynamic forces of ones actual personality released by the
imagination.
To release them is to reveal them. Elevators and small rooms have never
bothered me much, even when I experienced them in real life. But imagina-
tive experiments like the one Idescribed make me realize how susceptible to
claustrophobia Iam. My actual distress exposes psychological mechanisms that
would no doubt come into play should Iactually embark on a spelunking expe-
dition. As a matter of fact, now that Ihave experienced imaginary spelunking,
Iam aware that Iam sometimes uncomfortable in elevators and small rooms.
In performing experiments like this one, people are likely to find themselves
imagining more than what they are specifically asked to. When, in response to
instructions, Iimagine having to squeeze through a long, narrow passageway, it
will probably occur to me, in my imagination, that the passageway is too small to
allow me to pass my companion should Iwant to retrace my steps. Imight then
imagine undertaking one or another course of action:pausing to collect my wits,
or rushing ahead hoping to find a wider place quickly before Icompletely lose
my nerve, or suggesting to my friend that we try slithering backward, or simply
gritting my teeth and going on. Imay find myself, in my imagination, reassuring
my panicked companion that things will be all right while fearing that they wont
be, pretending to a confidence that Idont have. Imay imagine cursing her for
talking me into going on the trip, or berating myself for not resistingor for not
trying out the adventure in an armchair first. Imight find myself, in my imagina-
tion, feeling strangely confident of my ability to cope, or being resigned to my
 Spelunk ing , Simulation , and S lime 277

fate, or hoping to hear the voice of a rescuer, or hoping not to, wanting to rely on
my own resources.
What Igo on to imagine beyond what is called for, like the distress that accom-
panies my imaginings, depends heavily on my character and personalityon
how much self-confidence Ihave, on my propensities to blame myself rather than
others, or vice versa, on whether Iam fundamentally of an optimistic or a pes-
simistic disposition. The additional imaginings reveal features of my personality,
just as the affect produced by the initial imaginings do. It is I, after all, the real I,
who is doing the imagining. And Imay be proud or ashamed of what is revealed.
Imay be proud to find myself imagining acting with cool resolve and unselfish
concern for my friend. If Ishould catch myself abandoning my terrified compan-
ion, in my fantasy, searching for a way out of the cave that bypasses her, this may
be cause for serious concern about my moral character. I can be wrong about
what character traits my imaginings reflect, of course, just as Ican misinterpret
my real actions and feelings. There is danger of self-deception and other errors in
all of these cases. Nevertheless, one neednt be much of a Freudian to accept that
when Iimagine certain things in response to instructions, what else Ithen find
myself imagining, as well as the kinds of affect Iexperience, reflect aspects of my
real character and personality. (The spontaneity of the additional imaginings may
make self-deception less likely, as it sometimes does in the case of dreams and
daydreams.)
None of this is news. But it needs to be emphasized to counter the peculiar
tendency, in some discussions of fiction and the imagination, to think of imagin-
ing as a sterile intellectual exercise.
My imagining experiment is an instance of mental simulation. In imagining
as Idid, Isimulated an experience of a caving expedition. Mental simulation has
most often been invoked, in the recent literature, to explain or help explain how
we acquire knowledge of the mental lives of other people. The intuitive idea is
that we put ourselves imaginatively in the other persons shoesin the shoes
of a real-life spelunker, for instanceand judge, on the basis of our own imagi-
native experience, what she is thinking and feeling, or what she decides to do.
Simulation can also serve to predict what ones own experience would be like
should one really go spelunking in the future, for instance.5 Iam interested now
mainly in simulation itself, apart from its use in ascertaining or predicting the
actual experiences of another person or oneself.

5
Simulation theory is controversial. Isimply record my conviction that it is on the right track.
Many of the points of controversy are not relevant here anyway, for instance, disputes about whether
simulation theory is really distinct from its main rival, the theory theory; worries about the validity of
extrapolations from ones simulated experience to an actual experience of another person; and ques-
tions about how much of our knowledge of other people is based on simulation. There are different
variants of simulation theory as well, which we need not decide among.
278 In Other Shoes

In a study often discussed by simulation theorists, subjects were asked to

suppose that a Mr. Crane and a Mr. Tees are scheduled to take different flights,
departing at the same time from the same airport. Their limousine is caught in
traffic on the way to the airport, and arrives 30 minutes after the departure time.
Mr. Crane is told that his flight left on time. Mr. Tees is told that his was delayed
and left just five minutes ago. The subjects were then asked which of the two is
more upset? Most said that Mr. Tees is.6 How did they arrive at this conclusion?
According to Alvin Goldman, they simulated the experiences of Tees andCrane:

The initial step, of course, is to imagine being in the shoes of the agent,
e.g., in the situation of Tees or Crane. This means pretending to have the
same initial desires, beliefs, or other mental states that the attributors
background information suggests the agent has. The next step is to feed
these pretend states into some inferential mechanism, or other cogni-
tive mechanism, and allow that mechanism to generate further mental
states as outputs by its normal operating procedure. For example, the
initial states might be fed into the practical reasoning mechanism which
generates as output a choice or decision. In the case of simulating Tees
and Crane, the states are fed into a mechanism that generates an affec-
tive state, a state of annoyance or upsetness. More precisely, the output
state should be viewed as a pretend or surrogate state, since presumably
a simulator doesnt feel the very same affect or emotion as a real agent
would. Finally, upon noting this output, one ascribes to the agent an
occurrence of this output state. Predictions of behavior would proceed
similarly. In trying to anticipate your chess opponents next move, you
pretend you are on his side of the board with his strategy preferences.
You then feed these beliefs, goals, and preferences into your practical
reasoning mechanism and allow it to select a move. Finally, you predict
that he will make this move. In short, you let your own psychological
mechanism serve as a model of his.7

What exactly are the outputs of mental simulations? The inputs are imagined
or pretend circumstances and states. Simulation theorists commonly character-
ize the outputs, the results of the operation of ones psychological mechanisms
on this input, as imagined or pretend states also. Goldman, as we saw, prefers
to say that what is experienced by the simulator of Mr. Crane and Mr. Tees is a
pretend or surrogate state of annoyance or upsetness. Others describe outputs

6
From an experiment by Daniel Kahneman and Amos Tversky 1982.
7
Goldman 1995:189. Gordon describes the process of simulation rather differently. See Gordon
1995.
 Spelunk ing , Simulation , and S lime 279

as pretended or feigned or imagined or as if emotions and other mental states,

or speak of the simulator as continuing the make-believe.8 Outputs usually
include more than imaginings, however. Ihave already said that when Iimagine
the spelunking expedition, Iexperience intense affect, as well as being induced
to engage in further imaginings. My shuddering, my clammy palms, my cold
sweat, and the sensations that accompany them, are not merely imagined.
There can be no doubt that among the outputs of simulations are imagined
psychological states which the simulator is not actually experiencing. It is only in
my imagination that Ihave beliefs about the topography of the caveknowing
as Ido that there is no cave for my beliefs to be about. Only in my imagination
does it occur to me that the (nonexistent) passageway in which I find myself
when my (nonexistent) companion panics is too narrow for me to pass her. It is
not literally the case that Iintend to turn around at the next wide spot Icome to,
or that Idecide not to, any more than it is literally true that Icontinue crawling
further, or that Istop. Nor do Iactually consider whether to try slithering back-
wards. What Ido in my armchair is imagine forming certain beliefs and inten-
tions concerning my situation in the cave, imagine making certain decisions and
performing certain actions.
This is not to deny that the simulation may result in my forming actual beliefs
and intentions. It might occur to me that if there is a passage in a real cave con-
figured like the one Iam imagining, it would be too narrow to turn around in.
Imight decide to turn back as soon as possible should Ifind myself in a pickle
like the one Iam imagining. Imight swear never to go spelunking in real life.9 But
if Iam simulating, Iwill first form beliefs and intentions in imagination, concern-
ing the particular cave Iam imagining.10 Imight mutter to myself, We had better
turn around next chance we get (not In a real cave like this, it would be best to
turn back), or Gosh, she is blocking my way out of the cave, so Icant go for
help. It is (partly) because of what Ifind myself imagining believing and intend-
ing that I form general or conditional beliefs and intentions (if I do). This is
because what Ifind myself imagining helps me to clarify my interests and desires.
What about feelings and emotions? Aperson simulating Mr. Tees may really
be distressed or upset (in one sense of these terms, anyway). I do not think
it is true, literally, that the simulator is annoyed at having missed her flight. She
didnt miss a flight, and she knows it. Imagine being passed over in favor of your
best friend for a coveted position. Lets suppose that you conclude from your

8
See Gordon 1987, and the essays by Gary Fuller, Jane Heal, Adam M.Leslie and Tim P.German,
Gregory Currie, and Derek Bolton, in Davies and Stone 1995b.
9
William Charlton (1984) holds that to feel for a fictitious person is simply to experience a hypo-
thetical feeling concerning real people.
10
Since Ido not recognize purely imaginary entities, this characterization of my imaginative expe-
rience is misleading. See Walton 1990:Part IV.
280 In Other Shoes

imaginative experience that, were your friend to be offered the job in real life,
you would feel jealous. Ido not think it is literally the case that you are jealous of
him for receiving the offer, when you only imagine that he did.
Looking ahead to the case of Charles and the Slime, we can ask whether, when
Inegotiate the twists and turns of the cavern in my imagination, Iam literally afraid.
Iam claustrophobic, and claustrophobia is a kind of fear. But this claustrophobia is a
standing (or dispositional) condition which Ihad all along and which is merely acti-
vated and revealed by my imaginative experience. It is not part of the output of the
simulation; it doesnt result from my imagining myself crawling around in the cave.
Do I, as Iimagine coming to a particularly tight bend in the cave, fear that Iwont be
able to squeeze through (or worse, that if Ido Iwont be able to squeeze back when
Ireturn)? Do Ifear that my (utterly fictitious) carbide lamp will malfunction, or that
my (equally fictitious) companion will become hysterical? Ithink notalthough
Imay experience these specific object-directed fears in imagination, and my genu-
ine (standing) claustrophobia has a lot to do with the imaginings.
These are not pressing questions, from the simulation theorists perspective.
The simulation works in any case. What simulation requires is that the input and
the output states be analogous to inputs and outputs of the experience being
simulated, and that there be reason to presume a similarity in the processing
whereby the inputs produce the outputs. Many of the inputs of the simulation
are pretend or imagined versions of the corresponding simulated ones, and the
same can be true of the outputsit obviously is true of some of them. You may
simulate an experience of being jealous of your friend, whether your condition,
when you imagine, qualifies as one of actually being jealous of him or only imag-
ining this. In either case you may predict that you would really be jealous should
your friend actually be offered the job you want.
In simulating, ones psychological mechanisms are being run off line. This
means at least that they are disconnected from some of their usual behavioral
manifestations. Sitting in my armchair, Ido not carry out the decisions Iimagine
making, nor do Ibehave, in other ways, as Iwould if Iwere actually in a cave.
What blocks the behavior is my clear awareness that Iam sitting in an armchair
and not actually exploring a cave. Whether ones fear or jealousy or annoyance
are actual or merely imagined, when one engages in simulation, depends on
whether fear or jealousy or annoyance of the relevant kinds require either the
usual links to behavior (actual behavior, or the potential for producing it, or an
awareness of this potential), or the belief that the situation is actual. Again, the
simulation theorist neednt decide.11

11
Anumber of commentators take me to hold that fear requires the belief that one is in danger.
Idecline to endorse this principle in Walton (1990:201202), [although Ihad accepted it in Walton
(1978).]
 Spelunk ing , Simulation , and S lime 281

Back to emotional responses to works of fiction. The movie induces Charles

to simulate an experience of being attacked by a Slime. It induces him to imag-
ine a Slime, and to imagine its going after him. As a result, he enjoysor
undergoesa complex combination of experiences, including both affective
responses (what in Mimesis Icalled quasi-fear sensations) and additional imag-
inings, beginning with imagining himself to be in grave danger. His initial imag-
inings engage certain of his psychological mechanisms, and the results reflect
aspects of his actual personality and character (perhaps a standing propensity
to be repelled by things that are amorphous, shapeless, or slimy, not to mention
concern for his own well-being).12 Different people react differently to horror
movies; the differences in reaction reflect differences in their personalities and
character. Charles imagines fearing the Slime. This is one of the further imagin-
ings that result from the ones initially induced by the movie.
I stand by my contention that it is only in imagination that Charles fears the
Slime, and that appreciators do not literally pity Willy Loman, grieve for Anna
Karenina, and admire Supermannotwithstanding the consternation these
opinions have caused among commentators. Indeed, Ibelieve that these nega-
tive claims depart little from common sense, and should not seem either very
surprising or very momentous. We are now in a position to dispatch some of the
main objections that have been leveled against them.
Concerning Amos Barton, a character in George Eliots Scenes of Clerical Life,
Nol Carroll remarks that given the intensity of our feelings it seems coun-
terintuitive to give up the idea that we are saddened by the plight of Barton.13
This does not seem counterintuitive to me at all. Would a similar but less intense
reaction not qualify as feeling sad (mildly sad) for Barton? If Ihave a mild reac-
tion when Iput myself in Mr. Cranes shoes, and a stronger one when Iimag-
ine being in Mr. Teess situation, is it plausible to say that Iam, literally, more
annoyed at missing a flight, when Isimulate Tees than when Isimulate Crane?
The intensity of ones feelings is no reason to insist that the correct description
of ones experience has to be that of (literally) being saddened by Amos Barton,
or fearing the Slime, or grieving for Anna Karenina. Acar on a hoist simulating
being driven at breakneck speed up a winding mountain road can spin its wheels
like crazy, the engine might overheat or even blow up, without making it at all
counterintuitive to deny that it is actually travelling on a mountain road.
Richard Moran faults my make-believe theory for suggesting a lack of
real-world accountability in our responses to fiction.14 He observes that

12
Cf. Levinson 1990.
13
Carroll 1995: 95. Carroll switches between speaking of being saddened by (the plight of)
Barton and feeling sad for Barton. Itake him to be treating these as equivalent.
14
Moran 1994:93.
282 In Other Shoes

responses of laughter, lust, indignation, relief, delight in retribution, etc. are

normally treated as expressions of genuine attitudes that we actually have, and
are esteemed or repudiated accordingly, as when one chortles at a racist joke. He
evidently thinks that responses consisting in imaginative experiences of the kinds
Iattribute to appreciatorsto the spectator of a movie, for instancewould be
as remote from [the appreciators] real temperament as the events on the screen
are remote from his real beliefs about the world. It is obvious from our discus-
sion of simulation how far off the mark this objection is. There simply can be no
doubt that imaginings often reflect actual attitudes, desires, values, prejudices,
and so forth, and are thus subject to esteem and repudiation. Fantasizing about
torturing kittens may, depending on the circumstances, indicate a cruel nature
as surely as actually doing so would. In order to clarify my interests and desires,
it may help to imagine being faced with a choice between accepting a job as an
accountant and touring with a rodeo. If Ifind myself, in imagination, choosing
the former alternative, this may teach me something about myself. Will Moran
insist that Imust have literally decided to accept the (nonexistent) accounting
job, that if Imerely imagine intending to report to work at the accounting firm,
my experience wont have revealed anything about my actual desires?15
Moran emphasizes the manner in which one imagines, including imagin-
ing with this or that feeling, as opposed to what one imagines. Feelings actu-
ally experienced, as we noted in the spelunking case, are part of the output of
simulations and are revealing of character. They are also, on my account, an
important component of appreciators imaginative responses to fiction. Nothing
follows about whether an imaginer is best understood as actually, that is, literally,
experiencing a given intentional psychological state or merely experiencing it
in imagination. If Iread a story about kittens being tortured, the mere fact that
Iimagine this probably does not, in this case, reflect badly on my moral charac-
ter. If Ishould find myself imagining it with a sense of glee, however, Imay have
reason to worry. The glee is real. But my experience certainly does not have to be
described as actually taking pleasure in the suffering of kittens, in order to signal
a cruel streak in my character.
It is the manner in which Iimagine that is significant, in this case. But in other
cases what Iimagine, even apart from any particular accompanying feelings, may
reflect on my character. Suppose, for instance, that Ifind myself imagining kit-
tens being tortured, on seeing the word kitten on a spelling test, or simply out

15
There is a confusion in Morans observation that although a person typically wont find it a
disturbing discovery about himself that he is capable of imagining this or that fictional truth, he might
well be disturbed by what he finds himself feeling at the movies (93). What betrays ones character
is not what one can imagineone can imagine just about anything, as Moran points outbut what
one does imagine, especially what one finds oneself imagining in given circumstances:what one imag-
ines as a result of mental simulations induced by works of fiction, for instance.
 Spelunk ing , Simulation , and S lime 283

of the blue. Our real selves make themselves felt in what we imagine, as well as in
what we feel and the manner in which we imagine what we do.
Some objectors seem to think that it is just intuitively obvious that Charles
fears the Slime, that readers feel sorry for Anna Karenina, and so forth.16 This is
by no means intuitively obvious to everyone; the simulation theorists I men-
tioned express contrary intuitions about similar cases. It is true that weall
of usreadily describe appreciators as fearing, feeling sorry for, and admiring
characters in fiction, even falling in love with them. Is there a presumption that
what is commonly and naturally said, in ordinary circumstances, is true? At most
there is a presumption that in thus speaking, people express something true. Our
question is whether what is said, taken literally, is true. Saying that Charles is
afraid of the Slime is a way of expressing the truth that it is fictional (true in
the world of make-believe) that he fears the Slime. We also express truths about
what is fictional when we say, There is a horrible green slime on the loose, or
Anna threw herself under the wheels of a train. Whether any of these sentences
is true when taken literally is another matter; clearly the latter two are not. Is
there some presumption that what is commonly and ordinarily said should be
taken literally, rather than in some special way? Not in this context, which already
involves so much make-believe. If there is a presumption at all about how state-
ments like Charles fears the Slime and John grieves for Anna are to be taken,
it is that they, like There is a slime on the loose and Anna threw herself under
the wheels of a train, express what is fictional, what is imagined to be the case.
Moran points out that many common everyday emotional experiences con-
cern what is known to be in some sense nonactual, or things which are not
in the actual here and nowthat our responses to fiction are not special in
this regard. There are feelings directed at modal facts, at things that might have
happened to us but didnt; there are spontaneous empathetic reactions such
as wincing and jerking your hand back when someone else nearby slices into his
hand; there is the person who says that it still makes her shudder just to think
about her driving accident, or her first date; and there are backward-looking
responses such as relief, regret, remorse, nostalgia.17 Moran thinks these are all
paradigms or central instances of emotional experience. His idea seems to
be that responses to fiction of the kinds under discussion are not essentially
different from them, and hence have to be regarded as themselves instances,
maybe paradigms, of emotional response. There is certainly no quarrel so far.
The word paradigm carries unfortunate baggage from the crude paradigm

16
This seems the only way to understand Berys Gauts (1992:298)declaration that my theory
fails to respect the phenomenology of our responses to art:we sometimes are genuinely afraid of
fictional monsters.
17
Moran 1994:78.
284 In Other Shoes

case arguments of the 1950s and 1960s, but all of the examples mentioned,
including the reactions to fiction, are indeed clear and obvious instances of emo-
tional response. Are they clear instances of emotions whose objects are things
not present? Regret, remorse, and nostalgia are, but these are plainly irrelevant
to the issue at hand. The fact that regret and nostalgia are feelings about events
long past has no tendency to suggest that appreciators literally pity or admire
people they know do not exist and never did, or fear things that they are utterly
certain pose no present danger.
Some of Morans other examples are more interesting, and more like the fic-
tion cases, although it is unclear how they are supposed to contribute to his argu-
ment. Is the person who shudders on recalling her automobile accident terrified
of the truck that she remembers careening into her car years previously? Moran
thinks that, as a paradigm, this example is not or should not be regarded as
paradoxical. Whatever he means by this, it certainly does not follow that she
is, literally, afraid of the truck when she recalls the accident. (Does a persons
wincing in empathy with a friend who cuts his finger, indicate that she actually
feels pain herself?)
Moran admits that we might find these cases puzzling. But if we do, he
says, we would thereby lose what was supposed to be distinctive about the
fictional case.18 That fiction must be understood to be distinctive is a require-
ment of Morans own manufacture. To reiterate a prominent theme of Mimesis,
make-believe and imagining are pervasive in human experience, by no means
confined to our interactions with works of fiction.19 This theme is reinforced by
our discussion of mental simulation, as it occurs in everyday life as well as in our
experiences of fictionand by Morans own examples. Asignificant advantage
of the make-believe theory is that it allows us to see fiction as continuous with
the rest of life.
Vivid memories like the ones Moran discusses involve imagining; one relives
the remembered experience. The shudders result from vividly imagining the
truck careening into ones car, despite being fully aware that that is not now hap-
pening. One is terrified of the truck in imagination; there is no need to insist that
one is (also) actually terrified of it. This is another instance of mental simulation;
one simulates ones own past experience. To empathize with anothers pain may
also be to imagine being cut and feeling pain oneself; one may (automatically,
without reflection) imagine oneself in the others situation.
After all the ink that has been splattered on the question of whether apprecia-
tors experiences include emotions of various kinds vis--vis fictional characters

18
Moran 1994:78. See also p.80.
19
See, for instance, Walton 1990:7.For other applications of notions of pretense or make-believe,
see Clark 1996, Clark and Gerrig 1990, Crimmins 1995, Pavel 1986, and Rosen 1994.
 Spelunk ing , Simulation , and S lime 285

and situations, it may be disappointing to learn that it doesnt much matter. It

doesnt; not for our purposes. The positive side of my account of appreciators
responses to fictiontheir imaginative participation in games of make-believe
is much more important.19a
This imaginative participation, we now see, consists (in part) in mental
simulation. Appreciators simulate experiences of being attacked by monsters,
of observing characters in danger and fearing for them, of learning about and
grieving for good people who come to tragic ends, of marvelling at and admir-
ing the exploits of heroes. We simulate these experiences, including the fear
and grief and admiration, whether these emotions are construed in such a way
that appreciators, literally, experience them, or in such a way that they merely
imagine doing so. In either case, appreciators bring much of themselves to the
make-believe; their actual psychological makeup, attitudes, interests, values,
prejudices, hangups, and so forth, come powerfully into play. And this some-
times makes their experience of the fiction a deeply moving one. The connec-
tion with simulation is especially helpful in cashing out the suggestion which
Ireiterated throughout Mimesis (and which many others have made as well) that
appreciating works of fiction and engaging in other make-believe activities are
important in helping us to understand ourselves. So far we have no reason to
suppose that Charles literally fears the Slime, or that any of us, in normal circum-
stances, fears for or grieves for or admires purely fictional characters. There is no
reason to cook up a theory specially designed to make it come out true that we
do experience these emotions.
Even if we were to hold that appreciators do, literally, experience them, we
would need to recognize that they experience them in imagination as well. If an
old shed counts as a stagecoach in a game of make-believe, a child sitting on the
edge of its roof might imagine sitting on a stagecoach and driving the horses.
Suppose that, in place of the shed, the child uses an actual abandoned stage-
coach for her game, and imagines driving horses as they pull the coach from one
town to another. In this case also she is sitting on a stagecoach, in her imagina-
tion, even though she really is doing so as well. Charles imagines a Slime ooz-
ing toward him, and, in his imagination, it threatens him. It would be strange
to deny that he fears it in his imagination also, even were we to decide that his

19a
[R. M.Sainsbury (2010) thinks the view that make-believe, understood as Walton under-
stands it ... constitutes an essential ingredient in an account of the nature of fiction collapses if we
can accept that ... emotional states [such as Charless fear of the slime] are real. (19). Idisagree most
heartily. Sainsbury ignores the many features of fiction other than appreciators emotional responses
that the make-believe theory helps to explain. Concerning Sainsburys rejection of the make-believe
account of fiction, see also Walton (1978), p 262 and footnote 11a, in this volume. Paul Harris (2000,
esp. ch.3) cites numerous psychological studies demonstrating important continuities between chil-
drens make-believe and adults reading of literary fictions.]
286 In Other Shoes

experience counts as one of actually fearing it.20 In that case his actual fear would
be incorporated in his make-believe in the way that the real stagecoach is incor-
porated in the childs game. Surely spectators of Romeo and Juliet not only realize,
in imagination, the tragedy that befalls the young lovers, but also grieve for them
in imaginationwhether or not we suppose that their experiences amount to
grieving for them in reality.
Some who insist that appreciators experience fear or pity or grief toward fic-
tional characters and situations admit that the emotions they experience are of a
different kind from the fear or pity or grief that one feels toward real people and
situations. We might think of fear-of-fictions, for instance, as a variety of fear dif-
ferent from fear-of-perceived-dangers. Some hold that the intentional object of fear
or pity (what the Slime in Charles fears the Slime refers to) is something of an
entirely different kind in the fiction cases:a Fregean sense, a collection of prop-
erties (Carroll), or a kind of imagining (Lamarque). Some say appreciators of
fiction are frightened by thought contents (Lamarque), or that their pity involves
beliefs and desires about what is fictional rather than what is actual (Neill).21
The stated rationale for describing appreciators as experiencing fear, pity, and
so forth, of different kinds, rather than as imagining fearing and pitying, is to
underscore the similarity to the real-life cases. But in one way this exaggerates
the differences. Appreciators imagine having pity or fear of ordinary, everyday
kinds, with ordinary kinds of objects. Iimagine pitying a person who really suf-
fers a tragedy, or admiring someone who actually performs heroic deeds, or fear-
ing a monster that really poses a threatnot thought contents or Fregean senses
or collections of properties. The words that may come to me as Iimagine are
not, Oh, that poor thought content! or Yikes! Ahorrible fictitious slime, or
What a dangerous collection of properties, but simply, Oh, that poor waif!,
Yikes, a horrible slime!, or What a dangerous situation! The view that we
really do fear, pity, and admire fictitious entities, but with fear, pity, or admira-
tion of a special kind, fails to account for the phenomenology of our experiences,
in particular for the close analogies they manifestly bear to possible real-life
experiences. To account for this, we need to recognize that we imagine feeling
fear, pity, and admirationfear, pity, and admiration of the kinds we might actu-
ally feel in real life. Once we recognize this, there is little reason to insist that we
also, really, fear, pity, and admire fictional characters.

20
Those who hold, on ordinary language grounds, that only what is not true or not believed to be
true can be imagined, misunderstand ordinary language. To say that a person imagines such and such,
sometimes carries a conversational implication that it is not true or not believed. But it is obvious
that much of the content of our dreams and daydreams and games of make-believe is known by us to
be true. Ordinary language should not be decisive in constructing a theory, in any case. Cf.Walton
1990:13.
21
Carroll 1990:8486, Lamarque 1991:164, and Neil 1993.
 Spelunk ing , Simulation , and S lime 287

References
Carroll, Nol. 1990. The Philosophy of Horror, or, Paradoxes of the Heart. London:Routledge.
Carroll, Nol. 1995. Critical Study:Kendall L.Walton, Mimesis as Make-Believe. The Philosophical
Quarterly 45:9399.
Charlton, William. 1984. Feeling for the Fictitious. British Journal of Aesthetics 24:20616.
Clark, Herbert H. 1996. Using Language. Cambridge:Cambridge University Press.
Clark, Herbert H., and Gerrig, Richard J. 1990. Quotations as Demonstrations. Language
66:764805.
Crimmins, Mark. 1995. Quasi-singular Propositions: The Semantics of Belief Reports. The
Aristotelian Society, supp. vol. 69:195209.
Davies, Martin, and Tony Stone, eds. 1995a. Folk Psychology: The Theory of Mind Debate.
Oxford:Blackwell.
Davies, Martin, and Tony Stone, eds. 1995b. Mental Simulation: Evaluations and Applications.
Oxford:Blackwell.
Feagin, Susan L. 1996. Reading with Feeling:The Aesthetics of Appreciation. Ithaca, N.Y.:Cornell
University Press.
Gaut, Berys. 1992. Book Review of Kendall L. Walton, Mimesis as Make-Believe: On the
Foundations of the Representational Arts. Journal of Value Inquiry 26:297300.
Goldman, Alvin I. 1995. Empathy, Mind, and Morals. In Davies and Stone 1995b, 185208.
Oxford:Blackwell.
Gordan, Robert M. 1987. The Structure of Emotions: Investigations in Cognitive Philosophy.
Cambridge:Cambridge University Press.
Gordan, Robert M. 1995. Simulation without Introspection or Inference from Me to You. In
Davies and Stone 1995b, 5367. Oxford:Blackwell.
Harris, Paul L. 2000. The Work of the Imagination: Understanding Childrens Worlds.
Oxford:Blackwell.
Kahneman, Daniel, and Amos Tversky. 1982. The Simulation Heuristic. In Judgment under
Uncertainty:Heuristics and Biases, 2017, edited by Daniel Kahneman, P. Slovic, and Amos
Tversky. Cambridge:Cambridge University Press.
Lamarque, Peter. 1991. Essay Review of Mimesis as Make-Believe. Journal of Aesthetics and Art
Criticism 49:16166.
Levinson, Jerrold. 1990. The Place of Real Emotion in Response to Fictions. Journal of Aesthetics
and Art Criticism 48:7980.
Matravers, Derek. 1991. Art and the Feelings and Emotions. British Journal of Aesthetics
31:32231.]
Moran, Richard. 1994. The Expression of Feeling in Imagination. Philosophical Review
103:75106.
Neil, Alex. 1993. Fiction and the Emotions. American Philosophical Quarterly 30:113.
Pavel, Thomas G. 1986. Fictional Worlds. Cambridge, Mass.:Harvard University Press.
Rosen, Gideon. 1994. What Is Constructive Empiricism? Philosophical Studies 74:14378.
Sainsbury, Richard M. 2010. Fiction and Fictionalism. NewYork:Routledge.
Walton, Kendall L. 1978. Fearing Fictions. Journal of Philosophy 75: 527. Reprinted in this
volume.
Walton, Kendall L. 1990. Mimesis as Make-Believe:On the Foundations of the Representational Arts.
Cambridge, Mass.:Harvard University Press.
 ACKNOWLEDGMENTS

The mentions in texts and footnotes in these pages do not begin to do justice to
the debts Iowe colleagues, students, friends, and participants in discussions on
many occasions, for ideas and advice about the essays in this volume. Among
those who havent received enough credit are Carola Barbero, Stuart Brock,
Malcolm Budd, Sarah Buss, Herb Clark, Gregory Currie, Joe Dubiel, Susan
Feagin, Stacie Friend, Manuel Garcia-Carpintero, Mitchell Green, Marion
Guck, David Hills, Robert Howell, Fred Kroon, Peter Lamarque, David Lewis,
Shen-yi Liao, Dom Lopes, Derek Matravers, Fred Maus, Patrick Maynard,
Bryan Parkhurst, Jerome Pelletier, Ian Proops, Peter Railton, Nils-Hennes
Stear, Kathleen Stock, Alberto Voltolini, Alicyn Warren, George Wilson, James
Woodbridge, Richard Woodward, Stephen Yablo, Eddy Zemach. I gratefuly
acknowledge permissions, granted by editors and publishers, to reprint some
of my previously published essays. The original sources of the essays in this
volume are as follows:
Empathy, Imagination and Phenomenal Concepts. New in this volume.
Fictionality and Imagination:Mind the Gap. This is a revised and expanded
version of Fictionality and Imagination Reconsidered, in Fictionalism to
Realism: Fictional and Other Social Entities, ed. Carola Barbero, Maurizio
Ferraris, and Alberto Voltolini (Newcastle upon Tyne: Cambridge Scholars
Publishing, 2013), 926. Published with the permission of Cambridge Scholars
Publishing.
Two Kinds of Physicality in Electronic and Traditional Music. Copyright
2012. From Bodily Expression in Electronic Music: Perspectives on a Reclaimed
Performativity, ed. by Deniz Peters, Gerhard Eckel, and Andreas Dorschel.
Reproduced by permission of Taylor and Francis Group, LLC, a division of
Informa plc.

289
290 A c k n o w l e d gments

Thoughtwritingin Poetry and Music. Copyright 2011 New Literary History,

The University of Virginia. This article was first published in New Literary History
42.3 (2011), 455476. Reprinted with permission by Johns Hopkins University
Press.
Its Only a Game!:Sports as Fiction. This is a revised and expanded version
of Sports as Fiction, Michigan Philosophy News (Fall 2009):58, copyright
2009 Kendall Walton.
Restricted Quantification, Negative Existentials, and Fiction, Dialectica, 57/2
(2003):241244. Copyright 2005, John Wiley and Sons.
Existence as Metaphor? in Empty Names, Fiction and the Puzzles of
Non-Existence, ed. Anthony Everett and Thomas Hofweber (Stanford:Stanford
University Press, 2000), 6994. Postscript new in this volume.
Projectivism, Empathy, and Musical Tension. Credit: Walton, Kendall.
Philosophical Topics 26: 1 & 2Projectivism, Empathy, and Musical Tension.
Copyright 1999 by The University of Arkansas Press. Reproduced with the
permission of the University of Arkansas Press, http://www.uapress.com.
Listening with Imagination: Is Music Representational? The Journal of
Aesthetics and Art Criticism 52/1 (1994):4761.
Metaphor and Prop Oriented Make-Believe. The European Journal of Philosophy
1/1 (1993):3957.
Understanding Humor and Understanding Music, in The Interpretation of
Music: Philosophical Essays, ed. Michael Krausz (Oxford: Oxford University
Press, 1993), 259270. By permission of Oxford University Press. Also pub-
lished in The Journal of Musicology, 1/1 (1993):3244.
What Is Abstract about the Art of Music? The Journal of Aesthetics and Art
Criticism, 46/3 (1998):351364.
The Presentation and Portrayal of Sound Patterns. From Human
Agency:Language, Duty, and Value, ed. by Jonathan Dancy, J.M. E.Moravcsik,
and C.C. W.Taylor. Copyright 1988 by the Board of Trustees of the Leland
Stanford Jr. University. All rights reserved. Used with the permission of Stanford
University Press, http://www.sup.org.
Fearing Fictions. Credit: Walton, Kendall. The Journal of Philosophy, 75/1
(1978):527.
Spelunking, Simulation, and Slime:On Being Moved by Fiction, in Emotion
and the Arts, ed. Mette Hjort and Sue Laver (NewYork:Oxford University Press,
1997), 3749. By permission of Oxford University Press, USA.
 INDEX
the acceptance rule, 190n27
acousmatic listening. See reduced listening
acting, 70, 187, 220, 259265, 267268, 269
Allen, Woody, 212n10
apparent (fictional, implied) artists, 3642,
4950, 52, 155156. See also music:musical
personae; narrators (in literature)
appropriation (of anothers words), 6174
architecture, 44, 50, 142n48, 208, 212
Arnheim, Rudolf, 223
Bach, J.S., 56, 154, 188, 242243, 244
Barber, Samuel, Adagio for Strings,121
Bargh, John A., 83n10
Baron-Cohen, Simon, 4
Barth, John, 219n33
Bavalas, Janet Beavin, 133n36
Beethoven, Ludwig van, 121, 154, 155156,
213n13, 222. See also Marthoven,
Ludwig van
Bell, Clive, 44, 216
Berry, Wallace, 120121n10
Blackburn, Simon, 114116
Boghossian, Paul A. and J.David Velleman, 128
Boretz, Benjamin, 203
Brock, Stuart, 83n10
Budd, Malcolm, 166, 221n37, 224n45
bumper stickers, 6768
Calvino, Italo, 4243, 169170
Camp, Elisabeth, 194n32a
Carroll, Lewis, Jabberwocky, 105
Carroll, Noel, 275, 281, 286
Casals, Pablo, 192n30
Cavell, Stanley, 43, 169
Charlton, William, 279n9
Chion, Michel, 39, 43
Clark, Herbert H., 7n16, 13n28
and Richard Gerrig, 101, 284n19
claustrophobia, 11, 136, 276, 280
291
color, 7, 49n20, 114116
Cone, Edward, 5557, 154
Cook, Nicholas, 204n7, 205n8
Cooke, Deryck, 212n11
Coleridge, Samuel Taylor, 126n21
concepts, possession of, 140, 192
Coplan, Amy, 3n6, 4, 11n24
Cortzer, Julio, 196
Crimmins, Mark, 90n3, 284n19
culinary dishes, 212n10, 230, 237, 239
Currie, Gregory, 1, 3, 26, 34, 50, 274n3
and Ian Ravenscroft, 18, 29, 30
dance, dancing, 44, 49, 50, 155, 156,
168169
silent raves, 48n18
Darwall, Stephen, 148n51
Davies, Stephen, 126n22, 148n52, 156n12
Davis, Lydia, Wife One in Country, 57n10
daydreams, 18, 28, 257, 269, 277
de Zayas, Marius, 193
Dickinson, Emily, The Belle of Amherst, 268
dolls, 2122, 91, 103, 161, 175, 264
Donnellan, Keith, 8586n4
dreams, dreaming, 5, 22, 78n7, 153, 190n27, 257,
268n15, 269270, 277
Dubuffet, Jean, 37
Eisenberg, Nancy and Janet Strayer, 23
Ekman, Paul, 134
Elliott, R.K., 5960, 71n38, 150n54, 166,
172n38, 227n51
emotion, affect. See also fiction:emotional
responsesto
ascertaining others by ones own, 130131.
(See also mental simulation)
emotional contagion, 9, 11, 131, 143,
147148, 149
empathy, 116, 40, 57, 60, 65, 71, 118150,
171172, 198199, 283284
292 I n d e
empathy (Cont.)
with inanimate objects, 50, 142, 143
situational,15
sort-of,14
Evans, Gareth, 103, 108, 109
Everett, Anthony, 87n6a
existence, 8488, 89117, 253, 269, 273
expression and expressiveness in music, 36,
4041, 118150, 152, 156, 165174, 189,
211229. See also Elliott, R.K.; emotion,
affect; Levinson, Jerrold; Robinson, Jenefer
and in poetry, 5474
arousal (evocation) theories of, 152, 166,
225226
in the service of representation, 144147
fiction
emotional responses to, 2, 7576, 167,
252272, 273288
vs. nonfiction, 17, 56, 6365
perceiving in accordance with a, 9899
fictional characters, fictional entities, 2, 84117.
See also existence; fictionalism
fictional worlds, 18, 21, 31, 32, 273. See also
gameworlds
secondary worlds, 3233
fictionalism, 89113, 114117, 175195
Blackburn, Simon and Lewis, David
on,114116
fictionality (truth in fiction), 1735, 257259. See
also principles of generation (principles of
implication, principles of make-believe)
clustering of fictional truths, 1819, 3034, 257
of contradictions, 1819, 3033
of illusions, 2224
illusions of, 2325
film, 22, 71, 167, 168, 172, 208, 223, 253. See also
point-of-view shots (infilm)
film music, 24, 48, 73, 144, 146, 152, 209
Flaubert, Gustav, 216
food. See culinary dishes
free indirect discourse, 101102
Friend, Stacie, 17n3
Freud, Sigmund, 203, 277
Fuller, Gary, 132n32
game worlds, 32, 3334, 100, 144, 164,
264, 269
without work worlds, 65, 72n41, 173174
games, 7583, 91, 214, 270. See also make-believe;
game worlds
Gaut, Berys, 283n16
gestalt therapy, 270
Glennie, Evelyn, 47
Goehr, Lydia, 249250
Goldie, Peter, 3n6, 4, 5, 8n19
Goldman, Alvin, 1, 8n19, 9n22, 11n24, 132,
135n40, 140n46a, 278
Gombrich, Ernst, 104
x
Goodman, Nelson, 6n12, 7n16, 118, 184,
212n11, 226, 235n6
Gordon, Robert, 132n32, 133, 135n40, 278n7
green slime, 167, 252268, 274, 280, 281, 283,
285286
Greenberg, Clement, 216n20
Guck, Marion, 157n16
Handel, George Frideric, 210, 222
Hanslick, Eduard, 151, 210, 214215, 216n23,
217218, 219, 220n36
Harris, Julie, 268
Harris, Paul, 274n3a, 285n19a
Haydn, Franz Joseph, 209
Heal, Jane, 6, 7, 132, 139
Hegel, G.W. F., 171n36
Henriques, Julian, 53
Hills, David, 139n46, 183n14a, 194n32a
Hoffman, Martin, 11n24
Hume, David, 128, 138139,
humor, 37, 38, 196207, 282
Huxley, Aldous, 62n24, 212
idioms,106108
illustration
by pictures, 21, 263264
by music, 124, 188, 209, 219, 222223
imagination. See also reliving an experience
and belief, 1719
and empathy, 116. (See also empathy)
and fictionality, 1735. (See also fictionality
(truth in fiction))
imagining centrally (Wollheim), 45
implicit imagining, 8283, 153154, 172
nondeliberate imagining, 190, 153, 190, 257,
258, 262
other-shoe imagining, imaginative
identification, etc., 1, 35, 9, 11, 13, 60.
See also mental simulation
parallel (vs. fancy), 35, 10, 1112
imaginative resistance, 6566, 71, 73
ineffability, 67, 8, 216, 217, 221222,
228229
interpretation,68
of works by performances, 238244, 249
Isenberg, Arnold, 202
Jackson, Frank, 201n4
Javanese gamelan, 120, 121n11
Johnston, Mark, 122, 182
Jumonji, Bishin, Untitled, 2425, 26, 27
Kafka, Franz, Metamorphosis, 1819
Kahneman, Daniel and Amos Tversky, 278
Kandinsky, Wassily. 41, 44n10
Kaplan, David, 7n15,
Keats, John, 58, 68
Kerman, Joseph, 155156, 200, 215
Kivy, Peter, 6970, 151, 152n2, 222n40
 Ind e x 293
Klee, Paul
Camel in Rhythmic Landscape, 4445
Rhythmical, More Rigorous and Freer, 4546
Kraft, William, Momentum, 37, 156, 234
Kroon, Fred, 90, 105106
Krumhansl, Carol 120n7, 123n16
Lakoff, George and Mark Johnson, 183184
Lamarque, Peter, 286
and Stein Olsen, 83n10
Lehrer, Keith, 6n12
Lerdahl, Fred and Ray Jackendoff, 120, 124
Levinson, Jerrold, 5859, 71, 145n49, 150n54,
245n14, 246, 281
Levitin, Daniel, 44n11, 49, 51
Lewin, David, 155n8, 157n13
Lewis, David, 19n7, 30n19, 114116
Ligeti, Gyrgy, Requiem, 234
Lipps, Theodor 50n21, 142n47
literature, 25, 55, 5657, 69, 215216, 237238. See
also narrators (in literature); poetry; reporting
the unreported (untold tales); theater
contrasted with music, 42, 49, 52, 59, 151, 152,
153, 155, 159, 163, 164165, 174, 210, 214,
217, 221, 222, 241
Loar, Brian, 6n11
Luce, William, The Belle of Amherst, 267268
make-believe. See also prop oriented make-believe
childrens, 83, 91, 92, 103, 105106
participation in, 92, 99102, 119n3, 161,
163164, 174, 227, 264, 265266, 270, 272,
273274, 285
unofficial games of, 100101, 103104,
111113, 116, 178
Mansfield, Katherine, A Dill Pickle, 2526
Marthoven, Ludwig van, 231232, 235, 239240
Matravers, Derek, 2n4, 274n3a
Maus, Fred, 160
Meinong, Alexius, 8687
Meltzoff, A.N., 133n36
mechanics of generation. See principles of
generation (principles of implication,
principles of make-believe)
memory, 14, 60, 7071, 177178. See also
re-living an experience
and mental simulation, 135139
traces, 12, 135139, 149
Mendelssohn, Felix, 221n37
mental simulation, 12, 118119, 125n20,
130150, 270, 273287
metaphor(s), 2526, 2728 89113, 226,
175195. See alsoidioms
applied to music, 123124, 129, 157, 187,
215, 217
dead, 109, 185186
framing effect of, 184185, 192
meta-representation, 2027, 3134
Meyer, Leonard, 51, 154, 158
Miller, Arthur, Death of a Salesman, 211
Mishima, Yukio 193
Monet, Claude, 208
La Dbcle, 127128, 142
Moran, Richard, 184186, 192, 281284
motor mimicry, 4950, 133134, 142143,
148149, 171
Mozart, Wolfgang Amadeus, 124, 148, 247
Piano Concerto K. 488, 157159, 161
music. See also expression and expressiveness in
music; musical performances
abstract (absolute, nonrepresentational), 37,
144, 150, 151174, 208229
analysis, 199207, 215, 248,
electronic, 3651, 53, 236, 246
musical personae, 5574, 119, 147148,
155n10, 172
musical portrayals of space, 157, 160, 162163,
174, 186, 187, 188190, 215
musical tension (and release or
relaxation), 118150, 157, 163164 165,
167, 213
musical works, 199, 230251
physical responses to, 4250, 53, 70, 156,
167169, 172173
program music, 151, 153, 157158,
209, 215
musical performances, 37, 39, 52, 53, 7172,
155156, 230251
and copies of novels, 237239, 241242
correct and incorrect, better and worse, 230,
234236, 239, 241242, 245246
hearing patterns in, 239240
narrators (in literature), 2, 2930, 3133, 5474,
162, 163, 165n24, 227. See also apparent
(fictional, implied) artists; music:musical
personae
negative existentials, 8488, 8990, 10211
Newcomb, Anthony, 123, 160n20,
Nichols, Shaun and Stephen Stich, 18, 34, 138.
See also Stich, Stephen and Shaun Nichols
Nisbett, Richard E., 83n10
Nunberg, Geoffrey, 107n25
One Thousand and One Nights, 22, 27, 33
ontology. See existence; fictional characters,
fictional entities;
Palestrina,209
Parkhurst, Bryan, 171n36
Parsons, Terence, 85n3, 8687
participation in make-believe. See
make-believe:participation in
Pavel, Thomas, 284n19
performing literary works, 6970
phenomenal concepts, 1, 5, 8, 9
picture(s), depiction. See visual arts
Plato, 37, 151, 153
294 I n d e
poetry, 5474, 209, 238
point-of-view shots (in film), 13, 22, 33
Pratt, Carroll, 225
pretense, 13n28, 9092, 95102, 104105,
115116, 179180, 259, 265266, 270.
See also imagination; make-believe; mental
simulation
and existence, 84, 102116
and poetry, 6466, 70n36,
and sports and competitive games, 7576,
77n4, 78, 8283
principles of generation (principles of implication,
principles of make-believe), 19n7, 116,
185, 190, 191, 192193, 258. See also the
acceptance rule; the reality principle
projection, projectivism, 118, 128130. 133, 141,
143, 146147, 212
prop oriented make-believe, 90, 9295, 9699,
100102, 103, 104110, 114116, 157, 158,
175195
propositions, pretended expression of, 9798,
104105, 115116
proprioception,50
props (for make-believe), 33, 78, 92, 160, 161,
174, 259, 264, 274. See also prop oriented
make-believe
Proust, Marcel, 54
Psalm 23, 63, 6667, 68
Puig, Manuel, Heartbreak Tango, 57n10,
165nn2425
purring and tail wagging, 60, 148
quasi-fear, 256, 260262, 263, 267, 281
defined,253
quotation, 61n21, 101
Rahn, John, 199
Railton, Peter, 14n29
Rashoman (Kurasawa), 22
Ravenscroft, Ian, 137138. See also Currie,
Gregory:and Ian Ravenscroft
the reality principle, 18
reduced listening, 39, 43
reification of sounds, 4243, 53, 169170
reliving an experience, 12, 14, 284
reporting the unreported (untold tales),
2833
rereading (literary) fiction, 271272
restricted quantification, 8487
Rhamphorhynchus, 2830, 31, 32
Richards, I.A., 184, 186n21
Ridley, Aaron, 126n22, 148n52
Robbe-Grillet, Alain, 216
Robinson, Edwin Arlington, Richard Cory, 64
Robinson, Jenefer, 5859, 6061n20, 7172,
123n16, 126n22
Rosen, Gideon, 284n19
Russell, Bertrand, 86n5
x
Sainsbury, Mark, 87n6a, 101n16, 262n11a,
285n19a
samples, 57, 815,
Santayana, George, 118, 128
Schachter, Carl, 148
Schenker, Heinrich, 154, 215, 244, 248249
Schoenberg, Arnold, 121, 124, 128
Schopenhauer, Arthur, 164, 167n31, 218, 219n34
Scruton, Roger, 129n26, 188,
Schubert, Franz, Erlknig, 56
seeing the unseen, 2833
Sessions, Roger, 119
Shakespeare, William, 62, 96
Shelley, Percy Bysshe, 58, 68
Shoemaker, Sydney, 122, 129
Shusterman, Richard, 50
simulation. See mental simulation
Skolnick, Deena and Paul Bloom, 18n5
somatic experiences of sound, 4751, 170n34a
sound patterns, 161, 211, 230251
sonic dominance, 53
speechwriting, 54, 6063, 6769, 7274
spelunking, 34, 56, 10, 1314, 136, 275277,
279280
sports and competitive games, 7583, 181, 183
Stear, Nils-Hennes, 77n4
Stella, Frank, 209
Sternberg, The Docks of NewYork, 223
Stevens, Wallace, 168
Stich, Stephen and Shaun Nichols, 136n41. See
also Nichols, Shaun and Stephen Stich
Stravinsky, Igor, Rite of Spring, 44, 120121
Sullivan, J.W. N., 222
Sumarsam,120
suspense,271272
suspension of disbelief, 268
Sutton, A.Anderson, 120
Swift, Jonathan, A Modest Proposal, 38
tension. See music:musical tension (and release
or relaxation)
test of time, the, 219220
theater, 70, 182, 208, 220
and childrens make-believe, 259262
and sporting events, 77, 7980, 83
Thomasson, Amie, 85, 87nn66a
thoughtwriting,5474,
titles, 124, 150, 152, 214, 218
Tolstoy, Leo, 2
Tovey, Donald Francis, 209nn34
tragedy, 252, 270271
paradox of tragedy, 79
and sporting events, 7980
truth in fiction. See fictionality (truth in fiction)
Twain, Mark, The Adventures of Huckleberry Finn,
31, 33
Twas the Night Before Christmas, 64, 67, 68
Tye, Michael, 8n20
 Ind e x 295
Urmson, J.O., 236n7
Van Gogh, Vincent, 42, 144145, 150, 177
Van Inwagen, Peter, 85, 87n6, 111113
Vermeer, Jan, Woman at a Virginal, 2021, 26,
27, 181
Verstehen, 2, 4, 9, 198
visual arts, 3, 2034, 37, 4142, 4546, 4950,
52, 123, 127128, 144145, 154155,
162164, 174, 194. See alsofilm
music contrasted with painting, 208227
Vivaldi, The Four Seasons, 121
Warren, Alicyn, 45, 48n18, 71n39
Weinberg, Jonathan, 15
We Shall Overcome, 67, 68
West, Rebecca, 212
Wilson, Forrest, 143n48
Wittgenstein, Ludwig, 201n4
Wollheim, Richard, 45, 59, 125n19, 129130,
208n1
Wolterstorff, Nicholas, 234n4
Woodward, Richard, 27n13
Yablo, Stephen, 90n3.98
Zuckerkandl, Victor, 119120