Imagination, Philosophy and the Arts

IMAGINATION, PHILOSOPHY, AND THE ARTS

IMAGINATION, PHILOSOPHY, AND THE ARTS Edited by

Matthew Kieran and Dominic McIver Lopes

LONDON AND NEW YORK

First published 2003 by Routledge 11 New Fetter Lane, London EC4P 4EE Simultaneously published in the USA and Canada by Routledge 29 West 35th Street, New York, NY 10001 Routledge is an imprint of the Taylor & Francis Group This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” © 2003 Selection and editorial matter: Matthew Kieran and Dominic McIver Lopes; individual chapters: the contributors. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data A catalog record for this book has been requested ISBN 0-203-49864-X Master e-book ISBN

ISBN 0-203-34238-0 (Adobe eReader Format) ISBN 0-415-30516-0 (Print Edition)

CONTENTS

Acknowledgments

vi

List of contributors

vii

Introduction PART I Imagination, narrative, and emotion

1 13

1

Reasons, emotions, and fictions BERYS GAUT

14

2

How I really feel about JFK STACIE FRIEND

35

3

Narrative, emotion, and perspective PETER GOLDIE

55

4

In search of a narrative MATTHEW KIERAN

70

PART II Truth in imagination

89 90

5

Fictional assent and the (so-called) “puzzle of imaginative resistance” DEREK MATRAVERS

6

The tower of Goldbach and other impossible tales KATHLEEN STOCK

106

7

On the relation between pretense and belief TAMAR SZABÓ GENDLER

124

8

Literary fiction and the philosophical value of detail EILEEN JOHN

142

9

Fiction and value ROMAN BONZON

160

Imagining the truth: an account of tragic pleasure JAMES SHELLEY

177

10

v

PART III Sensory imagination

186

11

Seeing twice over CHRISTOPHER WILLIAMS

187

12

Out of sight, out of mind DOMINIC McIVER LOPES

207

13

The imaged, the imagined, and the imaginary DAVID DAVIES

225

14

Film and the transcendental imagination: Kant and Hitchcock’s The Lady Vanishes MELISSA ZINKIN

246

15

The funerary sadness of Mahler’s music SAAM TRIVEDI

260

16

Sculpture and space ROBERT HOPKINS

274

PART IV Afterthoughts 17

293

The capacities that enable us to produce and consume art GREGORY CURRIE

294

Bibliography

306

Name index

318

ACKNOWLEDGMENTS

A volume such as this is impossible without the skill and generosity of many people. First in line for thanks are the authors, who not only contributed their work but also endured the editorial process with patience and good humor. Some of the chapters in this collection were first presented at a conference on Imagination and the Arts held in July 2001 at the University of Leeds. We are grateful to all those who attended the conference for sharing their time and their ideas. We are also grateful for conference funding provided by the Analysis Trust, the British Academy, the British Society of Aesthetics, the Department of Philosophy at the University of British Columbia, and the School of Philosophy at the University of Leeds. R.B.Kitaj kindly granted permission for the use of his work on the dust jacket and Marlborough Gallery in London provided a transparency. We thank Brad Murray, who proved meticulous in preparing the manuscript, compiling the bibliography, and guarding against poor style and poor philosophy. Our final debt is to Tony Bruce and his staff at Routledge for having so ably ushered this book into this print. Matthew Kieran Dominic McIver Lopes

CONTRIBUTORS

Roman Bonzon is Associate Professor of Philosophy at Augustana College, in Rock Island, Illinois. He has written in aesthetics and the philosophy of language. Gregory Currie is Professor of Philosophy at the University of Nottingham. He recently completed a book on imagination, Recreative Minds, and is now working on the psychology and philosophy of imagination. David Davies teaches philosophy at McGill University. He works on art, mind, language, and science, and has just completed a book on the ontology of art. Stacie Friend is Assistant Professor of Philosophy at Washington and Jefferson College. Her research focuses on fiction, reference, and the imagination. Berys Gaut teaches in the Department of Moral Philosophy, University of St Andrews. He publishes in aesthetics, philosophy of film, and ethics, and is the author of Art, Emotion and Ethics, to be published by Oxford University Press. Tamar Szabó Gendler is Associate Professor of Philosophy at Syracuse University. She is author of Thought Experiments and co-editor of Conceivability and Possibility. Peter Goldie is Lecturer in Philosophy at King’s College London. Recent books are The Emotions: A Philosophical Exploration, and, as editor, Understanding Emotions: Mind and Morals. Robert Hopkins is Professor of Philosophy at the University of Sheffield. His publications, which are mostly in aesthetics and the philosophy of mind, include Picture, Image, and Experience. Eileen John teaches at the University of Louisville and writes on topics in aesthetics and the philosophy of literature. Matthew Kieran is Lecturer in Philosophy at the University of Leeds. He is the author of Revealing Art (London: Routledge, forthcoming) and articles in ethics and the Philosophy of Art.

viii

Dominic McIver Lopes teaches philosophy at the University of British Columbia. He is the author of Understanding Pictures and co-editor of the Routledge Companion to Aesthetics. Derek Matravers lectures in philosophy at the Open University. He is the author of Art and Emotion. James Shelley, Assistant Professor of Philosophy at Auburn University, is the author of papers in aesthetics and epistemology. He is currently writing a book on Hume’s theory of taste. Kathleen Stock is a lecturer in philosophy at the University of East Anglia. Her current research focus is the nature of imagination and related questions. Saam Trivedi is Assistant Professor of Philosophy at Simmons College. He has published on interpretation, expressiveness, the ontology of artworks, and the emotions, and is working on aesthetics and ethics. Christopher Williams is Associate Professor of Philosophy at the University of Nevada, Reno. Melissa Zinkin is Assistant Professor of Philosophy and co-director of the philosophy, literature, and criticism program at Binghamton University. She is writing a book on Kant’s concept of force.

INTRODUCTION

Much progress in aesthetics during the second half of the twentieth century may be ascribed to the assumption that an understanding of representation in art provides the key to many problems in aesthetics. A good part of the success of aesthetics during the past twenty years may be credited to attempts to work out how mental representation figures in the appreciation of art—where “mental representation”, when it comes to appreciating art, has often meant “imagination”. (True, the imagination has always formed part of the subjectmatter of aesthetics, but its recent revival can be traced to recent advances in philosophy of language and philosophy of mind.) The seventeen chapters that make up this volume attest to the richness of what might be called imagination theory. But if a rich body of philosophical work is often the child of skeptical impulses, this volume is no exception, for each chapter seeks to extend imagination theory in a new direction. The best introduction to proposals for a change of course makes clear where we are and why we must move forward. Imagination in make-believe Kendall Walton’s Mimesis as Make-Believe revolutionized philosophical thinking about representational art, fiction, imagination, and appreciation. Despite the intense analytic scrutiny these concepts had already endured, the so-called paradoxes of fiction remained seemingly intractable, the concept of imagination had been banished from philosophy of art and philosophy of mind, and scant progress had been made toward understanding appreciation. In place of clarifications or refinements of the existing concepts Walton proposed an entirely new way to carve up the categories. Revolution can be disorienting. Thus, Walton holds that all art is representational, that all representational works are fictions, and that appreciation always involves a kind of fiction-making. Nevertheless, a theorist is right to reconceptualize what he wishes to explain provided that the explanation is illuminating. What precipitates Walton’s recarving is the insight that the categories of representational art, fiction, imagination, and appreciation all connect to makebelieve. This notion appears first as part of the theory of representation that is the

2 INTRODUCTION

principal offering of Mimesis as Make-Believe. Representational works, according to Walton, function as props in games of make-believe. Good models of such games are to be found in children’s play. Suppose that Turner and Desmond are playing Star Wars. In their game, paper tubes are light sabers, trees are droids, skateboards are spacecraft, the garden shed is the Death Star, and the boys are Jedi. Using these props, they make-believe that they destroy the droids with their light sabers and fly their spacecraft to the Death Star: waving a paper tube across a tree is destroying a droid, and riding a skateboard to the garden shed is flying to the Death Star. The role of the props is to determine what is true in the game. When Turner falls off his skateboard, it is true within the game of Star Wars that he has been hit by the enemy. Walton defines a fictional truth as what is true in a game of make-believe. In consequence, fictional truth is not any special kind of truth—there is no special sense in which it is true that Dracula is immortal. A fictional truth is quite simply a fact about what is fictional. Moreover, since props in games of make-believe have the role of generating fictional truths, representations function as props that generate fictional truths. Dracula functions as a prop in a game properly played with works of literature to generate fictional truths. When I read Dracula it is fictional, for example, that I am reading a report of the career of a vampire. Representations function to generate fictional truths and users of representations are meant to imagine these truths. It is fictional that p if and only if there is, in a game of make-believe, a mandate to imagine p. To say that it is a fictional truth that Turner has destroyed the Death Star is to say that the game of Star Wars prescribes that we are to imagine that Turner has destroyed the Death Star. It is fictional that Dracula avoids mirrors because this is what we are asked to imagine when we read Dracula. (Note that we need not actually imagine what a game mandates that we imagine so a fictional truth need not be an imagined truth.) In sum, representations function as props in games of make-believe. Props in such games have the role of generating fictional truths. Fictional truths are what players of games of make-believe are prescribed to imagine. Thus representations function to prescribe imaginings. This network of conceptions has some surprising features. First, since representations are analyzed as props that function to generate fictional truths, all representations are works of fiction. Some of the incredulity this claim may engender is alleviated by a second feature of Walton’s account. Fictional truth is compatible with truth and imagining with believing. It is fictional in the game of Star Wars that Desmond is human but it is also true that he is human; the game prescribes that its players imagine that Desmond is human even as they believe that he is human. Likewise, Dracula’s readers imagine that Transylvania is mountainous even as some of them believe truly that Transylvania is mountainous. In effect, a fictional truth is not, on Walton’s account, anything like what it is in the ordinary sense. It is not, for example, a non-deceptive falsehood. A truth is fictional if and only if there is a prescription in a game of make-believe to imagine it.

INTRODUCTION 3

One might seek empirical evidence in support of this theory of representation or the conceptions of fiction and imagination that comprise it. It might be possible, for example, to locate practices that constitute games of make-believe in people’s interactions with artworks. In the case of literary works these would include “principles of generation” that determine what is fictional given the sentences of the text. However, this strategy is unlikely to succeed unless art-using practices are fairly stable across texts and Walton doubts that this condition obtains. He doubts, for instance, that David Lewis’s reality and mutual belief principles are employed in any systematic fashion to determine what a literary text makes fictional (Lewis 1983). In lieu of a direct argument Walton submits that his proposals best explain certain longstanding puzzles regarding our appreciation of artworks. In particular, the notion of make-believe makes available a novel notion of what Walton calls participation. The key to participation is that the player is a prop in most games of makebelieve, including the games governing our interactions with artworks. Peter Lamarque sets the idea out succinctly (Lamarque 1991:163). John is watching a movie in which a monster attacks New York. There are three ways to characterize his appreciation of the fiction: (1) it is fictional that John believes that the monster is attacking New York; (2) John believes that it is fictional that the monster is attacking New York; (3) John believes that the fictional monster is attacking New York. Walton rejects (3) since it construes John’s belief about the monster as de re, but it is the distinction between (1) and (2) that concerns us here. In (1) John’s belief falls within the scope of the fiction operator whereas in (2) the content of John’s belief (that the monster is attacking New York) falls within the scope of fiction operator but his believing it does not. Walton prefers, wherever possible, giving the fiction operator the wide scope it has in (1). When the fiction operator has wide scope some of what is fictionally true concerns John and some concerns the work. Walton distinguishes two fictional worlds, each accessed by a distinct set of fictional truths. The “work world” is accessed by what is fictional in the work (e.g. that the monster is attacking New York) and these fictions are generated by the representation alone. The “game world” is accessed by fictional truths about the spectator (e.g. that he believes that the monster is attacking New York) and these fictions are generated partly by the spectator himself in so far as he, alongside the work, is a prop in the proceedings. What is true in the game depends on what is true in the work world but the converse does not hold. Participation in many games of make-believe is not, therefore, limited to imagining what is true in the work world; it also includes imagining truths about oneself as a constituent of the game world.

4 INTRODUCTION

Walton submits that this conception of participation delivers solutions to three puzzles: the paradox of emotional responses to fiction, the puzzle of how we can derive pleasure from tragedy, and the puzzle of how well-known stories can sustain suspense. The adequacy of Walton’s solutions to these puzzles has received more attention than any other aspect of his account. But before proceeding to this it is helpful to consider in more detail the theories of imagination inspired by Walton. Imagination as an attitude Walton acknowledges that there are many forms of imagination. However, given the work that it is made to do in his theory, the operative notion of imagination is a prepositional attitude. Two questions to ask are what kind of attitude the imagination is and what principles generate imaginings. The answer to the first question depends on the answer to the second. It is natural to define imagination as that attitude taking fictional propositions as its contents (where a fictional proposition is one that is not true). But on recent theories of imagination, such as Walton’s, imagination is prior to fiction: fictional truths are those that are to be imagined. Thus it is fictional that George W.Bush is the President of the United States wherever it is prescribed that one imagine that Bush is the President of the United States. The contents of imaginings are fictional propositions in the trivial sense that they are to be imagined, not in the ordinary sense that they are a species of falsehood. Since we cannot characterize the attitude of imagination nontrivially by its contents, we must characterize it functionally, by the way imaginings are generated or by the inferential relations obtaining among them. Walton offers one characterization; a second derives from the view that imagination is mental simulation. Simulation theory originated as an explanation of the attribution of attitudes to others. In the version of the theory developed by Gregory Currie, simulation is a mental operation that consists in allowing the belief-desire system to run “offline”, disconnected from its normal sensory inputs and behavioral outputs and from the simulator’s own beliefs and desires, yet taking into account the situation of another person (Currie 1995a). While run off-line, the simulator’s belief-desire system nevertheless continues to operate in an otherwise normal way, generating inferences and causing emotional responses, thereby telling her what mental state she would be in were she in the other’s situation. Suppose that upon seeing a distant hiker suddenly double back in the presence of a bear, I attribute to him the state of fear. According to simulation theory, I do this by substituting for my beliefs (e.g. that my surroundings are safe) beliefs representing the hiker’s environment (that it is bear-infested) and allowing my belief-desire system to process the substituted input. I discover in this way that I would feel fear, so I attribute fear to the hiker—but since my belief-desire system has been run offline, I do not experience fear myself. Currie proposes that imagination is mental

INTRODUCTION 5

simulation. Imaginings are belief-like and desire-like states that are generated by running the belief-desire system off-line. Works of art generate imaginings by providing inputs to the belief-desire system, run off-line. The attitude of imagination is defined by the conjunction of two features of mental simulation. First, imaginings are inferentially insulated from other prepositional attitudes. What is imagined is independent of what is believed or desired. One’s imagining that Holmes inhabits 221B Baker Street is not impugned by any belief one has that is inconsistent with it. Second, imaginings retain belieflike inferential relations to other imaginings (Currie 1995a:149–50). This is the point of describing imagination as the operation of the belief-desire system run offline, for the belief-desire system is what sustains inferential and causal connections among mental contents. But states of imagination rely on the same inference generator and have similar effects on the affective system as do beliefs and desires. Were this not so, simulation would be unreliable as a mechanism for attributing mental states to others. In sum, imaginings are not beliefs because they do not have the right inferential connections to other beliefs and desires, but they may be described as quasi-beliefs (and quasi-desires) because they preserve belief-like (or desire-like) functional and causal connections to other imaginings and to emotions. If I believe that p and I believe that q then I believe that p and q. Hence if I imagine that p and I imagine that q then I imagine that p and q. Likewise, if I believe that p and p is dangerous then I will probably experience fear. I will be put into a state of fear, or something relevantly similar to it, if I imagine that p. According to the make-believe theory, imaginings are prescribed in games of make-believe comprising conditional rules mandating that when a given state of affairs obtains, some proposition is to be imagined. Vermeer’s Woman Holding a Balance is designed to function as a prop in a game of make-believe, which prescribes, in view of the way the picture’s surface is marked, that one imagine that a woman is holding a balance. Imaginings generated in games of make-believe are inferentially insulated from other propositional attitudes. Imagining that Holmes lives at 221B Baker Street does not imply that one does not believe that 221B Baker Street is uninhabited, and believing that 221B Baker Street is a vacant lot and imagining that Holmes lives at 221B Baker Street does not imply either that one imagines or believes that Holmes inhabits a vacant lot. Fictional truths are inferentially elaborated within games of make-believe: some fictional truths entail others. However, what inferential relations obtain between fictional truths is up for grabs, depending only upon the rules of the relevant game of make-believe. Imagining that p and imagining that q may not, in some games of make-believe, entail imagining p and q. According to simulation theory, imaginings are the product of the normal belief-forming mechanism and its inference generator; according to make-believe theory, imaginings are generated in accordance with inference rules that need not match those that generate beliefs.

6 INTRODUCTION

Response and resistance Parts I and II of this volume are taken up with the nature and shaping of our imaginative responses to works and our resistance to them. Tolstoy’s creation of Anna Karenina is not to be understood as asserting that any such person exists, Bram Stoker’s Dracula is not a historical tale or future prediction, and Borges’s Labyrinths is unconcerned with what is strictly possible or not. We are to imagine the characters and states of affairs as represented without any concern for what is or even could be the case. Yet, in doing so we affectively respond to them and their plight in ways that are common to our emotional responses to actual people. We fear Dracula, pity Anna Karenina, and are intrigued by the infinite library. In real life we fear those with great power and evil intent, pity those trapped in lovelorn marriages, and are curious about intellectual puzzles. The only difference is that in the case of fictions our emotional responses are directed toward objects we understand to be nonexistent. Pre-reflectively this all seems straightforward enough. But as Colin Radford originally pointed out, why think that our responses to nonexistent objects ought to be consistent with our responses to actual objects (Radford 1975)? Indeed, what grounds do we have for thinking such responses are rational at all? If we know that Mephistopheles does not exist, it is clearly irrational to be afraid of him. If we believe Christ never existed, it is irrational to pity him. So if it is not part of the content of what we imagine that we believe Dracula, Karenina, or the infinite library to exist, then it is irrational for us to feel for or respond to them. It is important to clarify what emotions are being taken to be on this line of thought. Emotions are held to be constituted by a phenomenological feeling conjoined with a cognitive element such as judgment or belief. I can feel the same way in two cases but my emotions may differ depending on what I believe. For example, in two different cases my palms get sweaty, my heart races, and I feel jittery. Yet, in one case I may be fearful and in the other I may be embarrassed. What explains the difference? In the former case I believe that something is threatening me whereas in the other I believe that I have done something shameful. It is this “cognitive” view of emotion that leads to a paradox. To have the genuine emotion of fear requires the judgment or belief that I am being threatened, but when watching a horror movie I seem to feel fear while believing that I am in no danger at all, for I am well aware that the monster represented on the screen is not going to jump out at me. It is this paradox which pushes Radford to conclude that our emotional responses to fiction are irrational. Walton argues that the paradox can be resolved if we construe our responses to fictions not as genuine emotions but rather as quasi-emotions felt in the context of games of make-believe. Playing such games, we imagine that we are afraid and our imagining this is partly constituted by our imagining that we believe we are in danger. Such responses are not genuine emotions: they have the same feel or affective aspect as genuine emotions but we do not believe Dracula threatens us. Nevertheless our imaginings help explain how and why we make the inferences

INTRODUCTION 7

we do and why we come to care for fictional characters and respond to fictional situations. On Walton’s account, our affective responses to fictions are rational since they do not depend upon judgment or belief. It does seem counterintuitive to claim that our affective responses to fiction are make-believe rather than genuine emotions and Walton’s theory has been criticized on this basis. The very intensity of our “sadness” for the fate of a tragic character suggests that we are genuinely moved (Carroll 1990:73–4). Moreover, we often cannot just switch our affective responses on and off at will, at least if the artwork is any good, which is surely what would be suggested if our responses were make-believe ones. It seems that we genuinely respond to thoughts we entertain and do not merely pretend to do so. The thought theory holds that in entertaining thoughts about what is fictionally true we come to be genuinely sad (Lamarque 1981; Carroll 1990:84–6). But objections to Walton’s account based on appeals to phenomenology are inadequate because it is part of the account that quasi-fear feels the same as genuine fear. What makes our response to Dracula a quasi-emotion rather than a proper one is the difference in their cognitive components (an imagining in one case and a belief in the other). It follows that intensity of feeling cannot be a good criterion for whether an emotion is genuine or not. Consider someone who attends a really good horror movie and is subject to intense and visceral fear, horror, and disgust (which is what makes it a good horror movie). On the way out she is set upon and verbally abused by a stranger. The feelings she is subject to during this incident might be much more attenuated than the feelings she was subject to in the cinema yet in both cases she genuinely feels them. Which emotions are genuine depends upon what she believes rather than the degree of feeling involved. Furthermore, it does not follow from Walton’s account that quasi-emotions can be switched on and off at will within games of make-believe. Walton could argue that as long as one’s involvement in the game is whole-hearted and the prescriptions concerning what we should imagine are of the right sort, we will be prompted to have certain affective responses whether we will it so or not. Walton is committed only to the commonplace recognition that we can stop playing the game and thereby avoid being subject to feelings it prompts. Indeed, he has suggested that the thought theory of emotional responses seems to entail that I genuinely fear a collection of properties or thought contents (Walton 1997:47). This would either be a bizarre result or involve the need to talk of a special kind of fear. So, Walton argues, the thought theory is less faithful than Walton to the phenomenology of emotional responses to art. One might remain skeptical of Walton’s account for a different reason. Affective responses to fiction are being characterized as atypical or peculiar because we are responding to states of affairs, fictional ones, we believe we can do nothing about. But there is a wide range of similar non-fictional cases (Moran 1994). Just think of our emotional responses to the past or to modal facts—how things might or could have been. If Walton were to claim such responses were

8 INTRODUCTION

irrational, which is a hard claim to make out, then there is nothing particular to the paradox in the case of fiction. Alternatively this may provide good grounds for thinking our affective responses to fictions are genuine. What prompts them is perhaps more a function of the expressive aspect of a work than the bare prescription to imagine a certain proposition. There are different ways to present fictional truths and how they are presented to us shapes the nature of our affective responses. Different takes on the same scene may prescribe the same imaginings but solicit highly divergent responses. Consider a movie scene of someone eating dinner. The shots might concentrate on the formality of the diner, the precision with which each mouthful is savored, and the complementary contrast of the different colors and textures of each dish. Or the same scene could be shot in a way that emphasizes the outsized nature of the food consumed, the sheen of juices, the flecked spittle of the diner, the gape of the mouth as the meaty flesh is brought up to be engorged. In the former case the scene is shot in a way that prescribes admiration for the diner’s taste and discrimination; in the latter it solicits repulsion. Thus, the expressive qualities of a work play a large part in shaping our complex cognitive-affective attitudes toward what is represented, in ways that are not strictly relevant to the generation of fictional truths. We should therefore distinguish between prepositional imaginings, where we imagine that something is the case, from the expressively imaginative aspect of a work, which concerns the connections, associations, emotional mood, or tone suggested by the way the propositional imaginings are prescribed. It is the way the scene is presented that prompts the mind to respond in certain ways. The second scene of the diner focuses on the glistening spittle and sheen of the meat juices, prompts us to experience it as seeing something flecked and oozing, not merely to imagine that it is flecked and oozing. Whether they are genuine or quasi-emotions, many of our responses to expressive aspects of a work are prior to rather than part of the content of the imagining. In exploring this possibility, we need a better understanding of non-propositional imagining generally and this is something the chapters in Part III of the volume help provide (more on this below). There are also limits to participation, as the phenomenon of imaginative resistance shows. An example commonly used to illustrate the phenomenon is Leni Riefenstahl’s Triumph of the Will. Riefenstahl’s film is an immensely powerful and beautiful quasi-documentary of the 1934 Nuremburg rallies. The technique, composition, and editing are not only highly innovative but are designed to solicit responses of admiration and awe for Nazism. From the opening shots of the plane descending from the heavens to the sweeping shots of the crowds and the heroic isolation of Hitler, we are presented with an aesthetically appealing intimation of fascism as bold, noble, pure, and righteous. Yet, assuming we hold Nazism to be morally repugnant, it seems we either cannot or will not imagine what the film mandates even while we appreciate its artistry. Our entering into and responding fully to the film is somehow precluded since it

INTRODUCTION 9

would involve entertaining or assenting to values and ideals we take to be deeply wrong. Walton has noted a curious asymmetry (Walton 1994a). We are willing to imagine deviations in fictional worlds from laws of nature but we are much more reluctant to, or cannot, entertain deviations from moral truths. In science fiction, for example, we are prepared to imagine possibilities regarding time travel, multiple identity, or causality which we take to be actually impossible, just because imagining them is mandated. But a story decreeing that genocide is right is insufficient to make us accept that in the fictional world this is so. Walton’s explanation of the asymmetry is based on a supervenience principle. Since we generally assume that moral properties supervene on natural properties, where there is a divergence of moral properties in the fictional world from the actual world we conclude that there is a divergence in the natural properties. Where the fictional world, as it is represented, fails to respect this principle then we are entitled to hold that things could not be as they are represented, no matter what imaginings the work mandates. However, this diagnosis of imaginative resistance looks far too broad if it is supposed to be distinctive to the moral case and it looks as if we can imagine worlds which are at odds with the supervenience principle. After all, even if we assume that the mental supervenes on the physical we do not resist imagining a Cartesian fiction. According to Moran, imaginative resistance is psychological and an upshot of our character; we are reluctant to imagine morally inverted worlds because to do so would be to endorse certain sentiments that should be condemned (Moran 1994). But this explanation looks problematic in a different way. Just to imagine a morally inverted world does not entail the actual endorsement of the relevant sentiments. I might fully imagine certain racist attitudes and values, and allow my responses to proceed on that basis, but it does not follow that I actually endorse them. I may just take up an attitude of fictional assent; if these were the appropriate attitudes and values then I would respond as I do in reading the racist novel. I can, conceptually and psychologically, do this while nonetheless holding that such attitudes and values are actually deeply pernicious and wrong. Heightened resistance to moral falsehoods might reflect a psychological tendency to fear the corruption of our own characters, whereby we tend to react strongly against attempts to get us to imagine some moral falsehoods for fear of coming to internalize them ourselves. But truly great artistry can overcome this reluctance if we are prepared to allow it to do so. This would explain why there remains something important to Plato’s thought that truly great art can be dangerous indeed (Kieran 2003). Making sense of imagination It is an axiom of contemporary theories of imagining that states of imagination are mental states with prepositional contents. Attitude theorists need not deny that there are sensory imaginings, but they do hold that sensory imaginings are

10 INTRODUCTION

propositional imaginings, albeit of a special sort. We have seen that a notion of non-propositional imagining might help account for the rationality of emotional responses to fictions and for the phenomenon of imaginative resistance. Part III of this volume contains several non-propositional accounts of sensory imagination. A simulation theorist may opt to restrict his account to cases of non-sensory imagining and either deny that sensory imagining exists or grant that it exists as a phenomenon independent of prepositional imagining. But Currie takes sensory imagination to be a kind of prepositional imagination and proposes that simulation theory accommodates it, specifically as it is brought to bear in the interpretation of pictures. In his view, grasping what a picture depicts requires that one visually imagine the scene it depicts. It is fictional, for instance, in A Fist Full of Dollars that the Man With No Name keeps on walking despite having been hit by several gunshots. But my imagining this state of affairs while watching the movie differs in obvious and important ways from the imagining that is required to comprehending a description of the event (a literary fiction may evoke a sensory imagining but none is required to grasp the fiction). Currie accepts that the former but not the latter necessarily shares many features in common with seeing a man rise and continue to approach an enemy who is shooting him. It involves “perceptual imaginings”. Perceptual imaginings are a species of propositional imagining unlike that required to comprehend a literary fiction. Since imagination is quasi-belief, we should hew to the principle that “distinctions between kinds of imaginings ought to have their basis in distinctions between types of belief” (Currie 1995a:182). This, after all, is the point of taking states of imagination to be the product of the belief-desire system run off-line. Thus, perceptual imaginings parallel perceptual beliefs. A perceptual belief is a belief with a particular kind of content. First, it is one in which there is a counterfactual dependence between the belief’s content and the perceptible properties of the object of belief (if the object were redder or if it were square instead of round, the believer would have believed that it is redder or that it is square). Second, perceptual beliefs represent objects as having properties that are highly specific and that are “bundled” together in a way that reflects their origin in a sense modality. This distinguishes visual beliefs from auditory beliefs or from beliefs in other sense modalities—visual beliefs, for instance, bundle the specific shape, size, and color of objects; auditory beliefs bundle other types of properties. Two objections may be raised against the hypothesis that sensory imagining is perceptual imagining so defined. First, it does not capture the distinction between perceptual and literary imaginings. A reliable witness in a who-done-it novel may describe the perpetrator of a crime in a way that is counterfactually dependent upon the perpetrator’s features, that bundles them by sense modality, and that is highly specific. The reader’s imagining will be highly specific, bundled by sense modality, and counterfactually dependent on the perpetrator’s features. Yet, this imagining is quite unlike that we have when we see the perpetrator on screen or

INTRODUCTION 11

in the “mind’s eye”. The point may be put another way. We may respond to the who-done-it novel with two kinds of imagining, one required and one optional. The optional imagining is like seeing the perpetrator; the required imagining is like believing a journalistic report. How do these imaginings differ if perceptual imagining is quasi-believing? One might reply by crafting a more nuanced account of the specificity or bundling of properties in perceptual belief and imagining. However, if the second objection is effective, the first goes through, no matter how specificity and bundling are defined. The second objection is that sensory imagining—including the kind of imagining that might be thought to be in play when we watch movies or look at pictures—cannot be characterized as akin to belief, since it shares properties of experience incompatible with belief. Belief is one kind of mental state; experience is another. The differences are myriad, of course, but we should emphasize the following. First, experiences, but not beliefs, are necessarily occurrent. They always have a precise duration and occupy a precise location in the stream of conscious mental events. This is not the case with either beliefs or prepositional imaginings, which are quite often dispositional. (This explains the fact that experiences but not beliefs are the stuff of episodic memory: I can remember that I experienced such and such without recalling the experience itself but I cannot remember having a belief without recalling the belief itself.) Second, experiences necessarily have a phenomenal character (beliefs need not) and a full description of an experience must include what it is like to have the experience. Thus, unlike belief, experience can be more or less vivid. Third, the contents of experience are frequently, if not invariably, belief-independent. This is strikingly brought out in illusions, such as the Müller-Lyer illusion: my seeing the lines as different in length is not corrected by my belief that they are equal in length. Finally, the content of experience is non-conceptual: the properties it represents need not be ones of which the perceiver possesses concepts. A perceiver may have experiences as of shapes, colors, chords, and timbres for which she possesses no concepts. That this is so is consistent with the doctrine that all experience is concept-laden: perhaps some concepts must be applied to experience but representing some properties in experience need not require the possession of concepts of those properties. By contrast, the contents of belief are conceptual: each property represented in belief is necessarily one of which the believer possesses a concept. My seeing the monster in Alien or my running “If She Marries a Taxi Driver” through my head necessarily has a precise duration and is accessible to episodic memory. Sensory imagining also has, by necessity, a phenomenal character— Frank Jackson’s Mary can learn what it is like to experience red by sensorily imagining red or by seeing the blood of the movie monster’s victims (she cannot learn what it is like to experience red by propositionally imagining that she sees a red object.) Its content is non-conceptual—the precise color and consistency of the slime dripping from the alien’s teeth may be properties of which I possess no concepts (DeBellis 1995; Lopes 2000). This content is also belief-independent. I may believe that the monster is a harmless latex dummy and may do my best to

12 INTRODUCTION

resist imagining otherwise, but for all that it looks as if there is an alien monster on the rampage. Interestingly, the contents of pure sensory images are also beliefindependent—the lines in a mental image of the Müller-Lyer illusion appear different in length, even if one believes they are equal (Pressy and Wilson 1974; Berbaum and Chung 1981; Wallace 1984). The hypothesis that sensory imaginings are akin to perceptual beliefs is false if they have certain experience-like features. It should be—and it has been—acknowledged from the start that imagination is not a unitary phenomenon. There are many forms of imagination and each needs its own explanation. But this is no setback. The best first step toward understanding imagination in all its guises is to provide an account of one central form of imagination. Its limitations inform us about other varieties of imagination and guide us in devising accounts of them.

Part I IMAGINATION, NARRATIVE, AND EMOTION

1 REASONS, EMOTIONS, AND FICTIONS Berys Gaut

It is obvious that one can imagine having an emotion. Less obvious is whether one can have an actual emotion directed toward a merely imagined state of affairs. The latter is a quite general issue about the relation of imagination to emotion. A particular application of this issue has received extensive discussion in philosophical aesthetics in the last twenty years, in respect of emotions directed toward fictional objects. Generally the question has been put like this: is it possible to feel real emotions toward events and situations known to be fictional, and if so, how? The so-called “paradox of fiction” is then construed as the problem of how there can be real emotions directed towards known fictions. On the face of it, such emotions are puzzling: when we are afraid of something, for instance, it seems that we must believe that the feared object exists and is dangerous, and also be disposed to remove ourselves from its dangers. Yet, watching a horror film, we do not believe that the depicted monsters actually exist or are really dangerous, and far from rushing in terror from the cinema, we stay and watch with enjoyment. Some aestheticians, including Kendall Walton, Gregory Currie, and Jerrold Levinson, have tried to dissolve the paradox by arguing that (with various nuances and qualifications) it is not, in fact, possible to feel real emotions toward known fictional events: very broadly, what is happening is that the spectator is imagining feeling these emotions (Currie 1990; Walton 1990; Levinson 1997; Walton 1997). These theorists do not deny that spectators may feel real emotions toward real-world analogues of fictional events, nor do they deny that there is something emotion-like going on in spectators when they are moved by fictions: it is simply not real fear, pity, disgust, and so forth, that they are feeling toward the fictional events. Let us dub this view emotional irrealism. Returning a different answer to the question are philosophers such as Peter Lamarque, Noël Carroll, John Morreall, Richard Moran, Derek Matravers, and Richard Joyce, who argue that the phenomenology of spectatorship supports the claim that spectators can feel real emotions toward fictions, and that there are no good philosophical reasons for holding that the phenomenology is non-veridical in such cases (Lamarque 1981; Carroll 1990; Morreall 1993; Moran 1994; Matravers 1998; Joyce 2000). Let us term this view emotional realism.

REASONS, EMOTIONS, AND FICTIONS 15

I argue here for the truth of emotional realism. I also show that there is a distinct puzzle concerning not the possibility but the rationality of fiction-directed emotions. The latter puzzle has often not been as sharply separated from the former issue as it ought to be; for instance, Colin Radford in one of the earliest papers in the debate addressed the question of the rationality of fiction-directed emotions, but his position encouraged a conflation of this with the distinct question of their possibility (Radford 1975). Though distinct, there is nevertheless an important connection between the two issues: it will turn out that the strongest arguments for emotional irrealism stem from arguments based on rationality considerations, so that the puzzle about the rationality of fictiondirected emotions needs to be solved or dissolved in order to resolve the issue of realism. I first examine the possibility puzzle, and show how realism can dissolve it, since it is based on a misconception of what is required for something to be an emotion. I then turn to the rationality puzzle, examine various realist attempts to solve it, and offer my own solution, based on three rationality criteria grounded on the main characteristics of emotional states. Finally, I compare the solution offered here to three recent attempts to solve the puzzles, and argue for its superiority to them. The possibility of fiction-directed emotions Before turning to the paradox of fiction, we need to consider what an emotion is. In one broad, but legitimate, sense of the term, an emotion is a state that is characteristically felt, a state that characteristically has a phenomenology (unlike, say, a belief, which has none). In this sense, we may say that someone is very emotional, and simply mean that he is prone to swings of mood, or tends to feel things deeply. In a narrower sense of the term, an emotion is just one of the various kinds of state that are characteristically felt. It is hard precisely to delineate emotions in this narrower sense from other affective states—indeed, it has plausibly been suggested that emotions do not form a natural class at all (Rorty 1980:1). But it is useful at least to distinguish emotions from moods. A mood, such as happiness or sadness, need not have an intentional object: someone can simply be happy or sad, without being happy or sad about anything. An emotion in contrast has an intentional object: I am afraid of something, I pity someone. According to the dominant (and I would argue correct) cognitive-evaluative theory of the emotions, an emotion not only has an intentional object, but also essentially incorporates an evaluation of that object. So to be afraid of something essentially involves evaluating that thing as dangerous; to pity someone essentially involves construing her as suffering; to be angry with someone essentially involves thinking of that person as having wronged someone about whom one cares. (Cognitive-evaluative theories in various forms are defended by Lyons 1980; de Sousa 1987; Greenspan 1988; Roberts 1988. The particular evaluations mentioned as being essentially connected with each emotion are used for

16 BERYS GAUT

illustrative purposes only; in fact, they may be more complex than is commonly suggested, but that does not matter for our present purposes.) Finally, an emotion is a state that characteristically can motivate actions: I run away because I am afraid, I help someone because I pity her, I strike someone because I am angry with him. There are, then, three main characteristic aspects to an emotion: the affective, the cognitive-evaluative, and the motivational. In most fully fledged occurrent emotions all three are present, but it is not necessary that all three be present for an emotion to exist. Not all of the characteristic features of an emotion are essential to it. In particular, many emotional episodes are non-occurrent: one can be angry with someone for a long time, even when he is absent and one is not thinking about him. One need not have any feeling of anger during these long periods. Nevertheless, if one’s anger does become occurrent, perhaps because the person is present, characteristically one will experience a feeling of anger. According to the cognitive-evaluative theory, an emotion is not however individuated by its particular feeling. The feelings that we experience when we are ashamed and when we feel guilty, for instance, might well be identical, and what individuates them is the content of the respective evaluative thoughts—of having done something shameful, as opposed to having done something morally wrong. Lyons, for instance, holds that an occurrent emotion involves bodily sensations of felt disturbance caused by certain evaluations; but he does not maintain that each emotion has a unique feeling associated with it by which we individuate the emotion (Lyons 1980). A feeling must be present in occurrent emotion, but what feeling it is might vary from occasion to occasion, or from person to person. Turning back to our problem, the paradox of fiction has been discussed chiefly in terms of the emotions of pity and fear, the emotions aroused by tragic representations. But if we cast our net wider, the irrealist case becomes less compelling. Disgust, for instance, is (certainly in the broad sense) an emotion; but it is odd to say that one has to know whether something really occurred if one is to be disgusted at the thought of it. If I describe in vivid and detailed fashion the jam-and-live-mouse sandwich that I breakfasted on the other day, you might well be disgusted, without supposing that I am telling the truth. And if I tell you a joke, you don’t have to wait to find out whether the events I describe really happened before you can find the joke funny. I can also be awed by or feel wonder at a depicted landscape, even though I know that the artist invented it. Nor does there seem to be anything problematic (contra Walton 1990:203) about the possibility of admiring a fictional character, such as Superman. Stronger, though, are the intuitions supporting the irrealist case in respect of fear and pity. How can one really fear something, or fear for someone, if one does not believe that the thing is dangerous or the person endangered? And how can one not be motivated to flee if danger really threatens one? Indeed, how could one enjoy these states if one really were afraid? Similar considerations apply to pitying someone.

REASONS, EMOTIONS, AND FICTIONS 17

If the irrealist is right about these matters, he or she owes an account of what is going on in such cases. Perhaps the most developed version of such an account is due to Kendall Walton. Walton holds that it is only fictional that Charles, the spectator of a horror movie, is afraid of the green slime that features in it, and what makes this the case is that Charles “experiences quasi fear [the sensations and physiological reactions normally associated with fear] as a result of realizing that fictionally the slime threatens him. This makes it fictional that his quasi fear is caused by a belief that the slime poses a danger, and hence that he fears the slime” (Walton 1990:245). According to this analysis, spectators actually experience the sensations and physiological reactions normally associated with fear; these reactions do not, however, constitute fear when directed toward fictions. The account thus readily provides an error theory for why spectators might falsely conclude that they really are afraid, even though they are only make-believedly afraid. One evident disadvantage that such an account has over realism is its far greater complexity: whereas the realist holds simply that Charles is afraid of the Green Slime, the irrealist holds that he is in the complex state described above, which would involve his possession of concepts such as quasi-fear. We can indeed ascribe such concepts to spectators if they are necessary to explain the phenomena, even if spectators are not explicitly aware of possessing the concepts (think of how Chomskyans ascribe knowledge of deep grammatical structures to competent speakers, though such speakers have no explicit knowledge of them). But if there is a simpler account available, which adequately explains the phenomena, we ought on general heuristic grounds to adopt that. And this means that, unless the irrealist has good grounds for showing that realism is inadequate, then realism ought to be the preferred account. We have already adverted to the kinds of considerations to which irrealists can appeal; they can be organized around the three aspects of an emotion: the affective, the cognitive-evaluative, and the motivational. First, if a fiction-directed emotion is genuine, how could one enjoy feeling it? After all, real fear is intrinsically unpleasant, and people go out of their way to avoid it. But spectators enjoy horror films, and also a wide variety of other fictions in which they fear for characters. That strongly suggests that such spectators are only make-believedly afraid, not really so. However, this argument fails. For one of the many merits of Walton’s account is that it allows for make-believe fear to have exactly the same phenomenology as real fear—that is what quasi-fear is. And any adequate account of make-believe fear must keep its phenomenology the same as that of fear, because it is a truth of introspection that one can be in a state that feels like fear when watching a horror film. But if one insists that one cannot enjoy fear, because the feel of fear is unpleasant, then one must equally insist that one cannot enjoy make-believe fear, since, sharing the same phenomenology, that too must be unpleasant. Moreover, as I have argued elsewhere (Gaut 1993), it is a mistake to hold that fear is intrinsically unpleasant. According to the cognitive-evaluative theory it is not the feeling that individuates the emotion, but the cognitive content; and this allows for at least some cases where fear can be felt

18 BERYS GAUT

as pleasant (perhaps because of its associated “arousal jag”). Bunjee-jumpers, mountaineers, and many others who are confronted with real danger, can seek out and enjoy fear; the cognitive theory shows how this is possible. The second, cognitive-evaluative aspect of an emotion seems to provide a stronger ground for the irrealist case. Real emotions require their subjects to have appropriate beliefs: real fear requires one to believe that someone about whom one cares is threatened. If a man in a pub tells you a harrowing story about his sister, then you will cease to be harrowed if the man then admits that he made the story up, and does not have a sister (Radford 1975). However, the example does not prove that an emotion requires a belief: if one’s emotion does evaporate on being told that the sister is an invention, that merely shows that if an emotion is grounded on a belief, removal of that belief tends to remove the emotion grounded on it. It does not support the more general claim that having an appropriate belief is a necessary condition for the existence of an emotion. The claim that to have an emotion, a person must have the appropriate evaluative belief, might seem to follow from the cognitive-evaluative theory of the emotions. But in fact cognition is a broader category than belief, including imaginings in a broad sense—unasserted thoughts, construals, seeings-as, and so on. The more restricted version of cognitivism, which holds that emotions require beliefs about their objects has been dubbed judgmentalism, and there are good reasons to reject it. Consider Fido, an old and harmless dog. Due to a traumatic childhood incident, Fred has a neurotic fear of dogs, and Fred fears Fido. Fred does not believe that Fido is dangerous: Fred has, for instance, no tendency to warn others about him. Fred simply cannot help seeing Fido as dangerous, and he responds to Fido with fear. So it is possible to feel fear while not possessing a belief that the object of one’s fear is dangerous (Greenspan 1988: chapter 2). Or consider cases where I vividly imagine my hand being mangled by a machine, and I feel fear, though I firmly believe that my hand is not caught in the machine (Carroll 1990: 78–9); or when I vividly imagine plunging to my death on the rocks below, though I believe that I am safely behind the parapet on the cliff-top. It is also possible to fear merely counterfactual cases (Moran 1994). For instance, it makes sense to be afraid of what would have happened if Hitler had won the Second World War. And a visitor to the CN Tower in Toronto can have the experience of walking on a glass floor several hundred feet up on the observation deck of the tower, seeing the ground far below one’s feet. As one edges gingerly across the floor, one certainly believes that it is safe to do so (otherwise it would be madness to walk across), but the experience of fear is perfectly genuine: the appearance of extreme danger is a very vivid one, and grounds one’s fear of falling. The irrealist will deny the possibility of these cases. First, it will be objected that all of these examples are ones in which one does have a belief, albeit an irrational and probably subconscious one. Several phobic fears may be like this: someone who fears flying may well have an irrational belief that flying is dangerous. However, while this may be true of some phobias, it fails as a plausible

REASONS, EMOTIONS, AND FICTIONS 19

explanation for the full range of cases considered above: I do not, for instance, have a subconscious and irrational belief that Hitler really did win the Second World War, or that I will fall to my death when I walk across the glass floor—if I did have the latter belief, I would not walk across. The better explanation is that the vivid imagining or appearance of danger is sufficient to ground real fear. Second, it has been objected that anti-judgmentalist accounts mistake the intentional object of emotions. Carroll dubs his version of the anti-judgmentalist view the Thought Theory, and holds that Charles fears the thought of the Green Slime, and this thought need not be asserted (Carroll 1990). A related position is defended by Lamarque (Lamarque 1981). Walton has objected to the latter that one does not fear the thought, but the object of the thought—the Green Slime (Walton 1990:203). The anti-judgmentalist should simply agree: the theory should hold that one fears not the act of thinking of the Green Slime, but the thought-content: that is, the Green Slime. Judgmentalism holds that asserted thoughts, that is, beliefs, are necessary for emotions; and that the intentional object of the fear is the belief-content, that is, what the belief is about; antijudgmentalism holds that unasserted thoughts are all that may be required for an emotion, and should relatedly hold that the object of fear is the thought-content, that is, what the thought is about. Finally, an irrealist can object that the Fido case and its ilk show only that it is not necessary to believe that an object is dangerous to fear it; but it is compatible with this that one must believe that the object exists in order to fear it. So fear does not require characterizing beliefs that something is dangerous, but it does require existential beliefs, and hence the paradox of fiction is not solved by the anti-judgmentalist stratagem. Something like this suggestion is made by Levinson (Levinson 1997). However, such a half-way position lacks motivation: fear has as its formal object the dangerous, so the fact that one believes that the object of fear exists, even though it lacks the property of being dangerous, would be no less puzzling than the case where one does not believe that object of fear exists at all. And far from it being the case that fear must be directed to what one believes exists, fear is often felt toward what may happen in the future, not what one believes is presently so. The third, motivational, aspect of emotions grounds what some irrealists have seen as their most powerful argument. Walton remains neutral on Greenspan’s anti-judgmentalism, but holds that there are nevertheless good reasons for thinking that Charles, unlike Fred who fears Fido, is not really afraid, for Charles is not motivated to flee, and “fear emasculated by subtracting its distinctive motivational force is not fear at all” (Walton 1990:102). Similarly, the irrealist can hold that I am not motivated to help Anna Karenina, and that this undermines the claim that I pity her. In reply, the realist needs to get clearer about what the motivational aspect of an emotion involves. Consider a parallel argument to the above in respect of historical victims: it might be claimed that I do not really pity Anne Boleyn, since I am not motivated to help her in any way. But I really can pity Anne Boleyn;

20 BERYS GAUT

and this is so by virtue of the fact that I have a genuine desire or wish that she had not suffered as she did. It is the presence of such a desire or wish that constitutes the motivational aspect of the emotion of pity. If, for instance, I said that I pitied Anne, but had no desire at all that she not suffer as she did, there would be defeasible grounds for questioning whether I really did pity her. But I do not try to help Anne, since I lack the belief that there is something I can do to help her. It is absence of this belief, not of the desire (or therefore of the emotion), that explains why I do not act. So the absence of motivated action does not prove that I do not pity Anne; rather, I lack the appropriate beliefs about being able to help her. In the same way, I may genuinely pity Anna Karenina, since I wish that she had not suffered as she did, but I lack the belief about being able to help her in any way; so I do not try to do so. The point applies not just to historical cases, but also to anyone whom I believe I cannot help. If I watch a live television broadcast of a hurricane sweeping through a small Florida town, I can genuinely pity the inhabitants, though I do not try to get them out of the way of the hurricane, since I know that there is nothing I can do to prevent it from engulfing them. (For a similar set of considerations see Matravers 1991 and 1998: chapter 4.) A parallel argument applies to the case of fear. It is not the case that the desire criterial for fear is a desire to flee, since in the case of certain fears, fleeing is not appropriate—as in the case of my fear that I will lose my job. Rather, the desire is not to be threatened or endangered. And here again, spectators do not flee, since they lack the appropriate beliefs to motivate flight. Charles does not believe that he is endangered, but only imagines it; and given this, his desire not to be threatened does not motivate flight. However, this desire, imagined to be frustrated, may express itself in his reactions: covering his eyes, flinching, moaning, and so forth. Charles may in addition desire to flee the Green Slime, but even then he is not motivated to flee it, since he knows that no action available to him could count as fleeing it, since he cannot physically remove himself from the Slime: it occupies a fictional space, distinct from the physical space he occupies, and no point of this physical space is closer or more distant from the Slime than any other. (He may of course flee the cinema, but this action counts as fleeing the image of the Slime, not the Slime itself.) So again, the motivational aspect of fear is constituted by the presence of a desire or wish; lack of the motivated action of flight may be explained by the absence of appropriate beliefs, not by the absence of the emotion. (Note that I have not maintained that motivated action invariably requires an appropriate belief as well as a desire—for Fred only imagines Fido to be dangerous, and this imagining, combined with his desire not to be endangered, is sufficient to motivate him to flee Fido. An imagining here takes the role of a belief in motivation; but—to anticipate the later discussion—this looks like it is an aspect of Fred’s irrationality, and that in rational persons, an appropriate belief is required together with a desire to motivate action. Since the present section discusses the possibility of fiction-directed emotions, I am assuming here that spectators of fictions are rational, so that if they lack the appropriate belief, they will

REASONS, EMOTIONS, AND FICTIONS 21

not be motivated to act. The rationality claim will be defended in the next section.) The irrealist may object that we do not genuinely pity fictional characters, for we have desires that require them to suffer. I do not genuinely pity Anna, since I want to read Anna Karenina because it is a great tragic novel, and Anna’s suffering is necessary to its being a tragedy. If someone produced a Bowdlerized version of the novel, in which Anna goes off to live happily ever after with Vronsky, I would not want to read it. So my feeling of pity is illusory: I really do want Anna to suffer. One can in general desire that some state of affairs occur, insofar as it has some property, and also desire that the same state of affairs not occur, insofar as it has a different property. I want to eat this chocolate cake insofar as it will taste delectable; I also want not to eat the chocolate cake insofar as it will make me fat. Having such conflicting desires is possible, indeed common, and not in the least irrational. Moreover, if I decide overall that the latter desire is the stronger, and I do not eat the cake, that does not show that the desire to eat the cake is not real or strong—it is, as witnessed by my pangs of regret as I reluctantly refuse a slice. In the same way, I may genuinely desire that Anna not suffer, insofar as she is loveable and the suffering is overall undeserved, while also desiring that Anna suffer, insofar as her suffering is a necessary constituent of a great novel. The greater strength of the latter desire does not prove that the former does not exist. So my desire that she not suffer, and my pity for her, are genuine (Neill 1993 gives a somewhat similar response). We have seen that irrealist arguments drawing on all three aspects of an emotion fail to show that we cannot have fiction-directed emotions. The realist replies in each case drew not just on discussion of fictional cases, but also crucially on non-fictional cases, such as emotions felt toward counterfactual cases, or to real life cases, and on emotions felt toward historical personages or toward contemporaries whom one cannot help. It emerged that it is not a requirement that one has an appropriate belief for one to have a genuine emotion, a mere imagining may suffice; and we also saw that the motivational aspect of an emotion can involve merely the presence of a wish or desire. And as we noted, the greater complexity of Walton’s make-believe account gives reason to prefer the realist account, if realism is at least as adequate an account of the phenomena. Realism is, in fact, a better account of the phenomena, being grounded on an antijudgmentalist theory of the cognitive aspect of emotions, and a better understanding of the motivational aspect of emotions. There are other irrealist accounts besides Walton’s. Most prominently, Currie has offered an account of make-believe emotions directed toward fictions, which holds that such states are simulated emotions; these are emotions which are running “off-line”, that is, without their normal connections to the input system of beliefs and the output system of actions (Currie 1990: chapter 5). Currie’s account has the advantage that it lacks some of the complexity of Walton’s, eschewing talk of quasi-emotions, for instance. But, in light of the account of

22 BERYS GAUT

emotions offered above, there is reason to deny that if an emotion is detached from its normal inputs and outputs, it is only a simulated emotion. Emotions can be detached from beliefs (Fido-type cases) or not motivate actions (Anne Boleyntype cases) while being perfectly genuine. More recently, Currie has suggested that the issue between realism and irrealism rests on little more than a verbal point about how one decides to employ emotion-terms (Currie 1997). Putting his point in the terminology adopted here, whether one uses emotion-terms so that an emotion requires a belief and a motivated action to exist, or a mere imagining and a desire, is a matter of an essentially verbal choice. However, even this reconciliationist approach is ruled out by the account defended above: for Currie’s proposal would make it merely an essentially verbal choice whether I say that I really pity Anne Boleyn, or Fred really fears Fido, yet as we have seen, there are good reasons to hold that both are genuine emotions. The rationality of fiction-directed emotions There is a related problem for realism, which is at least as pressing as that considered above. Suppose that the anti-judgmentalist is right about the nature of emotions, and that beliefs are not required for emotions to exist. Nevertheless, it looks plausible to hold that appropriate beliefs are required for emotions to be rational. Even if I can fear Fido without believing that he really is dangerous, it seems that for my fear of him to be rational, I must believe that he is dangerous. (I may have an irrational belief that he is dangerous, but if so, this is a case of theoretical irrationality, and the emotion itself is derivatively irrational by virtue of this theoretical irrationality.) If I say “I’m afraid of Fido, though he’s perfectly harmless” this may describe a possible state, but it also seems to be a selfconviction of irrationality. Yet, in the case of fiction-directed emotions, we do not believe that fictional entities exist, or are suffering, or are dangerous; and that seems to render these emotions irrational. For anyone who believes that our emotional engagement with fictions is valuable, this is not a welcome result. Moreover, though we argued against irrealist accounts, if it transpired that the cost of rejecting them were that spectators were systematically irrational, then a principle of charity in interpretation would come to bear. If we had to conclude that emotional responses to fictions were irrational, there would be a strong reason, given the pervasive nature and importance of such responses, to reinterpret them so that they were only imagined emotions, and hence to preserve the rationality of spectators’ responses to fictions. This would put pressure on the arguments of the last section, and it might be better to conclude that fictiondirected emotions are make-believe. We need first to say a little about what rationality is. In its most general characterization, rationality is a matter of sensitivity to reasons; specifically, it is a matter of being able to recognize reasons and respond appropriately to them. It covers reasons to believe (theoretical rationality), reasons for action (practical rationality), and reasons to feel (affective rationality), the latter being our

REASONS, EMOTIONS, AND FICTIONS 23

particular concern. There is a distinction between two standards of demandingness of rationality. Sometimes when we wonder what rational persons would believe, do, or feel, we wonder what they have most reason to believe, do, or feel. This requires the deliberator to bring to bear all relevant reasons in the appropriate domain, and this can be a very demanding task, far beyond the capacity of an ordinary person. If one wonders what one has most reason to believe about the truth of a particular mathematical proposition, one may have to engage in mathematical reasoning of a complexity that few can follow. Call this notion maximal rationality. When we call someone irrational, on the other hand, the accusation cannot be of failing to be maximally rational, for this would be an extremely demanding standard, given all of the reasons that are potentially relevant in any particular domain, and most people on most occasions might well be judged to be irrational by this standard. Rather, we must be adopting a lower standard in making this accusation: the irrational person is a person who fails to achieve some minimal standard of rationality. The reasons that he or she fails to recognize or to respond to appropriately would be those, which are readily graspable, being clear and evident. How to specify this notion more exactly is a moot issue, but for our present purposes, we need not attempt to do so. The main point is to distinguish these two standards of irrationality, and relatedly to note that not all failures to grasp reasons are instances of irrationality. Irrationalism How is the realist to respond to the claim that fiction-directed emotions are irrational? The most straightforward realist position is simply to embrace the objection, and to be an irrationalist about fiction-directed emotions. Emotional responses towards fictions are irrational, and charity should not lead us to deny that they occur. There are, after all, natural responses that are irrational: people tend to be systematically more optimistic about how well their lives will go than is epistemically warranted. And love, as Proust and Stendhal have made us well aware, can have many irrational aspects to it, involving projections of one’s hopes and fantasies onto the loved one, which may have little foundation in reality. But that is simply how love is, and it would be wrong to try to interpret away these features of our lives in the interest of ascribing greater rationality to ourselves. Similarly, perhaps we have a natural tendency to respond to fictions with real emotions, but this is nevertheless irrational. Something like this seems to have been Plato’s view in the Ion. Socrates says of the recitor of verses: There he is, at a sacrifice or festival, got up in holiday attire, adorned with golden chaplets, and he weeps, though he has lost nothing of his finery. Or he recoils with fear, standing in the presence of more than twenty thousand friendly people, though nobody is stripping him or doing him damage. Shall we say that the man is in his senses? Ion: Never, Socrates, upon my

24 BERYS GAUT

word. That is strictly true. Socrates: Now, then, are you aware that you produce the same effects in most of the spectators too? Ion: Yes, indeed, I know it very well. As I look down at them from the stage above, I see them, every time, weeping, casting terrible glances, stricken with amazement at the deeds recounted. (Plato 1961:535 d-e) Socrates, then, does not doubt the reality of fear and sorrow in these cases: he simply thinks that they are irrational. Radford’s preferred response to the problem also seems to be a version of irrationalism (Radford 1975). The emotional realist can hold that fiction-directed emotions are real, but irrational, and fall into a class of other real but frequently irrational emotions or attitudes. However, this response is not very plausible. First, the claim is extremely counterintuitive, since all instances of emotions directed toward fiction would have to be held to be irrational, since we always lack the requisite evaluative and existential beliefs about fictions. It is not that the responses sometimes err in being irrational (as is the case with optimism and love); they always so err. And these responses are a core part of our general response to fictions, so that it is likely that almost every response to fiction would be tainted by irrationality. This is hard to accept. Second, the objection from the principle of charity cannot be easily evaded. For emotional irrealism, as we have seen, has a story to tell about how imagined emotions have affective features that explain why we confuse them with real emotions; the account provides an error theory of why a spectator might falsely conclude that he really is afraid. And since that error theory is not wildly implausible, a principle of charity points toward the conclusion that emotional irrealism is correct. In contrast, it is much more implausible to interpret away the irrational aspects of optimism and love—wild optimists are not experiencing quasi-optimism, and people irrationally in love are not experiencing merely quasilove. The pressure exerted by charity in favor of reinterpretation is much greater in the fiction case, so irrationalism should not seek shelter in the occurrence of irrational responses in real life. A final problem with the irrationalist line is that claiming that fiction-directed emotions are irrational leaves us with the difficulty of making some of the discriminations required in assessing viewers’s responses to works. Suppose you fear some dreadful monster in a horror film, but Bertie fears an innocent victim in that film; Bertie’s fear is irrational, in a way that your fear is not. But if all emotions directed toward fictions were irrational, we could not make this sort of discrimination. Yet, we clearly can, do and ought to make such discriminations, so we are not and ought not to be irrationalists.

REASONS, EMOTIONS, AND FICTIONS 25

Control The realist might hold that the reason why fear of Fido is irrational has nothing to do with the absence of an evaluative belief, so the absence of this belief does not tend to show that fear of Dracula is also irrational. Rather, the crucial difference between the two cases lies in relation to the control that subjects have over the respective emotions. Fred’s neurotic fear of Fido is something he cannot control: he sees Fido and is consumed with fear. But Charles, watching the Green Slime, can control his fear: he can turn his attention away from it, thus abolishing his fear; nor does the fear overwhelm him. It is lack of emotional control that makes Fred’s fear irrational, not lack of an evaluative belief. However, it cannot be lack of control that makes an emotion irrational. For consider my terror of the pit-bull terrier that is howling ravenously and unrestrainedly in front of me. My fear is out of my control, but it is also completely rational: I may be about to be viciously attacked. Moreover, the control criterion will not invariably separate fictional from nonfictional cases, since emotions felt toward real situations can sometimes be as much under our control as are emotions felt toward fictions. Fred may well be able to remove his fear of Fido by avoiding Fido and not thinking about him. In just the same way, Charles may decide that the horror film has got too grisly, close his eyes, and think of something pleasantly soothing. Strategies for coping with fear, by diverting one’s attention to something else and engaging with pleasant thoughts, are equally available in both cases. Lack of control is in fact more likely to be explained by experiencing an extremely powerful emotion, whether produced by a fictional or nonfictional situation—the power of the response temporarily robs one of one’s ordinary coping abilities for dealing with emotions. Aspectual criteria The correct solution to the problem, I propose, begins by acknowledging the complexity of emotional states, with their three aspects of the motivational, cognitive, and affective. Given this complexity, it would be surprising if the criteria for rationality (in the minimal sense of not being irrational) were not similarly complex, and related to these three aspects of an emotion. The cognitive criterion There is a salient difference between Fred’s fear of Fido and Charles’s fear of the Green Slime—Fred believes that Fido exists (even though he does not believe that he is dangerous), but Charles does not believe in the existence of the Green Slime. This difference looks like a plausible reason for why Charles’s fear is rational, and Fred’s is not. So consider this criterion:

26 BERYS GAUT

(C1) the rationality of fear of objects believed to exist requires one to believe that they are dangerous; and the rationality of fear of objects merely imagined to exist requires one (correctly) to imagine that they are dangerous.1 Fred believes that Fido exists, but he does not believe that he is dangerous, hence his fear is irrational. Charles only imagines that the Green Slime exists, but he correctly imagines it as dangerous, hence his fear of it is rational in this respect. And the criterion is not ad hoc: the distinction in the rationality criterion derives from the difference in the cognitive states of imagination and belief. The criterion also gives the correct answer about Bertie, who fears not the evidently vicious monster, but the evidently harmless victim in a horror film (she bears a strong resemblance to his aged aunt, of whom he is terrified). In the case of fictions, a large part of what makes it correct to imagine that someone is the victim, and not the monster, is laid down by the internal norms of the fiction—by what the artists made fictional by their actions. Given these norms, Bertie has imagined incorrectly. The criterion is general: it draws the distinction not between fictional and reallife cases, but between cognitive states, which are imaginings, whether or not they are associated with fictions, and cognitive states, which are beliefs. Incorrectness in imagining is possible not just for fictions that are generated by an author independently of the audience, but also for self-generated imaginings, because incorrectness can be fixed by the implicit prescriptions governing such imaginings. Imagining a calm and peaceful scene on a beach, I am suddenly struck by fear of the passing shadows of children, even though these children are in no way dangerous in my fantasy. That counts as an irrational fear. And contrast the rational fear of what might have happened had Hitler won the Second World War with the fear of what would have happened had Italy won the 1974 World Cup. The latter fear isn’t rational, unless I have ancillary imaginings that would make the team dangerous—such as that they were really a terrorist organization planning to blow up the stadium after they had won. The criterion thus yields intuitively correct results; but it may seem vulnerable to objections. First, consider Frank, who has no neurotic fears of Fido, but instead decides to play a game of make-believe featuring the trusty old dog, in which he imagines Fido to have the temperament of a maddened Rottweiler on speed, and to possess not his actual toothless old gums, but glinting dental machinery with which to effect his murderous dispositions on the hapless Frank. Frank, imagining Fido thus, is terrified. On the cognitive criterion, since Frank believes that Fido exists, for his fear of Fido to be rational, he should believe that Fido is dangerous. But he doesn’t. So his fear of Fido counts as irrational. But surely his fear may be perfectly rational: perhaps Frank is a psychotherapist, and is trying to figure out what it is like to be gripped by fear. In this scenario, what is the intentional object of Frank’s fear? Not Fido as he actually is: for, knowing that he is playing a game of make-believe, Frank’s feelings about the real Fido may (and should) not have changed in the least: Frank

REASONS, EMOTIONS, AND FICTIONS 27

may pat and stroke Fido without the slightest tremor when he steps out of his game. Rather, it is the make-believe Fido that he fears, he of the terrible temperament and the deadly fangs. Since this Fido is a merely imagined being, he can rationally fear him according to the cognitive criterion. Contrast this with Fred’s neurotic fear of Fido. What makes it possible for Fred to fear Fido is that he imagines the dog to be dangerous, even though he does not believe that he is. In seeing him as dangerous, he may imagine him as having the vicious temperament and deadly capacities that Frank imagines him to possess in the game. But the difference in the case of the phobia is that Fred fears the actual Fido: he shrinks from his touch, he goes clammy at the sight of him. The intentional object of Fred’s neurotic fear is actual-Fido, whom he believes to exist: but the intentional object of Frank’s rational fear in the game is imaginedFido. (It is possible, of course, that as a causal consequence of his imaginings, Frank might come to fear the actual Fido, for instance, fearing him outside the game; he is then irrational on the cognitive criterion; and that is correct, for in this case he has induced Fred’s irrational phobia in himself) There is a close connection to another case here. Suppose I meet Tom Cruise, who plays Dracula in a film, at a party, and on seeing his warm smile and outstretched hand, I cringe with fear and shrink from his touch. What I have reason to fear is Dracula, not Cruise—Dracula is imaginary, and given the properties ascribed to him in the film, it is correct to imagine him as dangerous. If the intentional object of my fear is however Cruise, my fear is irrational, for I believe that Cruise exists and is not dangerous. I am irrational because I have confused fictional persons with real ones, and the responses appropriate toward fiction are in this case inappropriate when directed at reality. A similar scenario can be constructed for pity: I may irrationally pity an actress who played the part of Tess of the D’Urbervilles. Second, it may be objected that this solution actually shows that we are no longer dealing with the same emotion directed toward objects imagined to exist and believed to exist, and therefore in the former case perhaps the emotion itself is only imagined. For if the rationality-conditions differ in the two cases, then the emotions must differ too: since mental states are normative, any mental state is constituted by a specification of those conditions under which it is rational. However, consider the case of belief: belief is rational (justified) with respect to mathematical statements only when these are proved a priori; but for statements about the number of chairs in a room, attempting to ground them a priori would be irrational. So the rationality-conditions for belief are distinct in the two kinds of cases. Nevertheless, “belief” means just the same when belief is about a priori propositions as when it is about a posteriori ones. So the distinction in the rationality-conditions between emotions directed toward entities only imagined to exist and towards entities believed to exist does not show that there are two different kinds of emotions involved. Moreover, the difference in the rationalityconditions need not be traced to a difference in the emotion: the

28 BERYS GAUT

difference can be traced to a difference in the mode of cognition of their objects, that is, to whether they are imagined or believed to exist. The motivational criterion Suppose that you imagine ghosts to exist, you imagine them as dangerous (correctly, given their well-known propensities), and you fear them. You don’t believe that the ghosts exist or are dangerous; so in respect of the cognitive criterion, your fear is rational (i.e. not irrational). But are you really rational? You might be—like Frank, you might be engaged in a project of imaginative understanding, and want to grasp what is going on in the minds of the ghostfearing tribe that you are currently living with and studying, without irrationally holding their beliefs. But suppose that you are so caught up in this imaginative project, that you start looking in the dark corners of your room for signs of ghostly inhabitancy, huddle under your bed when you hear rumbling noises from outside, and so on. At that point, you have slipped into irrationality, because you are now engaged in trying to avoid those non-existent ghosts; but you have not violated the cognitive criterion. (In respect of the two ways you were fearing ghosts, before you were like Charles, watching the horror film, and you were rational; now you are like a spectator who flees screaming from the cinema, trying to get away from the Slime.) So there must be a further criterion for rationality, concerning motivations. If you actually believed that ghosts exist and are dangerous, your glances under the bed would not count as irrational (though your beliefs probably would). So your fear counts as irrational by virtue of its motivating actions to avoid ghosts, even though you have no motivation-relevant beliefs—in this case, that ghosts exist and are dangerous. And this suggests a motivational criterion: (C2) an emotion is irrational if it motivates action, though its subject lacks motivation-relevant beliefs. Motivation-relevant beliefs are those that ascribe the relevant evaluative property to the object of the emotion (e.g. believing that something is dangerous, in the case of fear) or those that concern the possibility of the subject acting toward the evaluated object (e.g. believing that one can do something to help, in the case of pity). This criterion also picks out a further aspect of Fred’s irrationality. Fred is motivated to flee Fido, though he lacks the belief that Fido endangers him. So he is motivated to act by his desire not to be endangered and his vivid imagining that he is endangered. As I noted earlier, motivated action requires not necessarily desire and belief, but desire and some cognitive state, which may be an imagining, that ascribes the evaluative property to an object. And I also suggested that if the motivation to flee in the case of fear were present, though the belief about being endangered were not, then that would show that the motivation is irrational. So it is with Fred: his motivation is irrational, since he lacks the belief

REASONS, EMOTIONS, AND FICTIONS 29

(about being endangered) that would rationalize it. His fear violates the motivational criterion. Frank’s emotion in contrast is not irrational, since he also lacks the belief, and is not motivated to flee Fido. Is the motivational criterion tenable? First, it may seem that it sits badly with the cognitive criterion, since it renders it otiose. What makes Fred’s response to Fido irrational is not his emotions, it might be objected, but what he is motivated to do: run away from him. If like Charles in the fiction case, Fred were not motivated to flee, then whatever his emotional state, he would be rational. Appeal to the cognitive criterion is unnecessary. But even if Fred were not motivated to flee by his fear of Fido, his fear would still be irrational, since he has no reason to fear the dog. Emotions involve evaluative thoughts, and the rationality of these thoughts can be assessed independently of the motivations, which they ground in certain circumstances. This can also be true of responses to nonfictions: if I pity the Nazis, and not their victims, while knowing that it is the victims, not the Nazis, who suffered, then my pity is irrational, even though I am not motivated to help either group, since I know they are beyond the reach of my actions. The motivational criterion cannot do all the work. Second, people watching a horror film may flinch, moan, go pale, shiver, and so on; these are actions motivated by fear. Since spectators do not believe that they are threatened, these motivated actions will count as irrational on the motivational criterion. But they are the natural accompaniments of fear, and hard to suppress. So almost all fear would count as irrational on this criterion—and the same will apply mutatis mutandis to other emotions by virtue of their natural accompaniments. We should distinguish between expressive and goal-directed actions. The activity just mentioned, insofar as it involves actions and not merely involuntary behavior, is expressive activity, the natural bodily and behavioral tendencies of the emotion. Goal-directed actions in contrast aim at some end cognized by the agent, and the actions seek to achieve that end. So contrast the case of blanching and shivering at the Green Slime, with the case where I carefully check under my cinema seat for signs of ectoplasmic goo adhering to it. The latter action is irrational, the former not. The motivational criterion should be understood to govern actions insofar as they are goal-directed, and not insofar as they are simply expressive. The affective criterion Suppose that I believe (correctly) that I am in considerable danger of losing my job; this motivates me to do what I can to avoid losing it. Cognitively and motivationally, my fear is rational. But I am tormented by the prospect of losing my job, I brood on it and suffer immensely, though none of this affects my job performance. I am suffering to no point, and the emotion is irrational, yet it passes the cognitive and motivational tests. So we need a further criterion, which governs the affective (or more broadly evaluative) dimension of an emotion:

30 BERYS GAUT

(C3) an occurrence of an emotion is irrational if experiencing it involves suffering to no point. This criterion is supported by a wide range of cases. It governs not just actual, but also imagined situations, or what we might call “negative fantasies”. Suppose that I am very prone to torment myself with fear and dread by imagining myself being roundly humiliated whenever I meet people for the first time. I imagine suitably luridly embarrassing and humiliating scenarios, though I am not motivated to avoid meeting people because of this. On the cognitive and motivational criteria, my fear is rational. Yet, it is in fact irrational: there is no point to this self-torment, and I ought to stop it. So the affective criterion correctly governs these kinds of cases. Note in contrast that if I am prone instead to engage in positive, pleasant fantasies, the warm feelings that I have toward them do not seem irrational at all. The affective criterion also explains an aspect of the irrationality of phobias: one reason we hold them to be irrational is because they are disabling in various ways—the person in the grip of the phobia suffers, and he or she is prevented from performing otherwise worthwhile activities (such as going outside, or meeting people, or enjoying the pleasures of canine company). Some instances of suffering are worthwhile because they have a point (one may believe them to be character-improving, for instance). But these cases are not like that: this is suffering without a point. And the affective criterion provides yet another reason why we deem Frank’s encounters with Fido to be rational and Fred’s not. Frank, if he feels intense fear, may suffer; but his suffering has a point, for it allows him to understand his patients better; Fred’s suffering in contrast has no point—indeed, it prevents him from enjoying time with that loveable old hound. One might suppose that some kind of proportionality criterion would do the job better here—a criterion that holds, for instance, that one ought to feel greater fear the more in danger one is in. But that is too crude. If I am climbing an extremely dangerous mountain, I have reason not to feel a very high degree of fear, since that might be so distracting that my climb is rendered even more dangerous. Fear presumably has as its point the preservation of the subject who fears; yet in this case, feeling it in proportion to the danger might well make my preservation less likely. This fear not only has no point, it undermines the point of having fear itself. We should understand the “has no point” criterion as also including the case where the suffering undermines the point of the emotion itself. Finally, note that the pointlessness must, according to the general constraint on minimal rationality stated earlier, be clear and evident. If it transpired, for instance, that feeling certain emotions, such as jealousy, were always prone to produce psychic damage with no compensating advantage, then it would not necessarily follow that feeling jealous is always irrational—the invariably damaging nature of jealousy might not be clear and evident, but might be discoverable only via complex empirical work in psychology. In such cases, the ability to see this would

REASONS, EMOTIONS, AND FICTIONS 31

be more like the ability to grasp complex mathematical reasons; and failure to recognize them would not count as an instance of irrationality. I have argued that there are at least three criteria for emotional rationality, each associated with a particular aspect of what it is to be an emotion. Violation of any of them grounds a charge of irrationality, in the sense of a failure to be even minimally rational. Perhaps this is not a complete list of criteria; for even if each is associated with an aspect of an emotion, it does not follow that each aspect has only one criterion associated with it. And there may be ways to finesse and refine the particular criteria that I have offered. Recall too that these criteria are only meant to capture minimal rationality: if one asks what a maximally rational emotional response requires, then that will require a different and more demanding answer (quite plausibly appealing to something like Aristotle’s phronimos). Nevertheless, these three criteria do capture, at least in outline, a large part of our reasons for holding occurrences of emotions to be irrational. And it is clear that fictiondirected emotions can be rational according to each of these criteria. Charles can really feel fear: he correctly imagines the Slime as dangerous; he is not motivated to flee the Slime, since he does not believe that it is dangerous; and he is not suffering to no point (indeed he is likely to be enjoying the experience). Fictiondirected emotions can be rational according to a general set of criteria that govern the rationality conditions of emotions both toward situations believed to exist and toward situations that are merely imagined. That being so, we have shown that realism is correct. Comparisons I conclude by comparing the solution offered here with three other recent solutions to the paradox of fiction, due to Alex Neill, Derek Matravers, and Richard Joyce, briefly displaying the advantages of the present account over theirs. Neill adopts a judgmentalist version of the cognitive-evaluative theory of the emotions (however, he says at one point that emotions require one to adopt a perspective on something, which may be a concession to the anti-judgmentalist view—see Neill 1993:4). However, he maintains that the belief necessary for the existence of the emotion need not be belief that its object has the appropriate evaluative property, but may be instead belief that it is fictional that its object has that property. The latter condition allows it to be possible to fear for or pity a fictional character. But fear for oneself is not possible, since one neither believes, for instance, that one is threatened by Dracula, nor does one believe that it is fictional that one is so threatened, since one knows that nothing fictional can threaten a real person (there is an “ontological gap” between fiction and reality, preventing cross-world action). Moreover, the reason that one does not flee Dracula is that the desire to flee from something requires either that one believe that one is threatened by it, or that one believe that it is fictional that one is

32 BERYS GAUT

threatened by it; but neither of these is the case. Hence Charles can stay firmly in his seat, despite the marauding Green Slime. The main problems with this account derive from its judgmentalist account of the emotions. As Fido and the other cases demonstrate, it is not necessary for an emotion to exist that one have the relevant belief—Fred does not believe that Fido is dangerous; nor does he believe that it’s a fiction that Fido is dangerous (indeed, that thought would be a source of comfort to him). Also, Fred has a desire to flee Fido, and indeed may act on that desire, even though he lacks both of the beliefs just mentioned. It is possible for action to be motivated by a desire and an imagining, not a desire and a belief—the latter, as noted earlier, is a condition for the rationality of the motivation, not for its possibility. Moreover, there is no good reason to deny the possibility of fear for oneself. When the Slime stares out of the screen at Charles, it is fictional that the Slime threatens him (this is akin to a theatrical aside). The case is unusual, since fictional characters do not generally perform such asides, but when they do so, imagining that one is threatened is prescribed, and hence it is fictional that one is threatened. So it is correct to believe that it is fictional that one is threatened, and hence even on Neill’s account it should be appropriate for one to feel fear. But there is a subtler worry: in such cases, it seems that it cannot be myself for whom I fear, since it is not true that the Slime threatens me, but at most it threatens my fictional or possible counterpart (I, as I exist in the fictional world) (Neill 1992: 96). However, this is, in fact, sufficient for fear for oneself. For consider a case where I narrowly avoid skidding in my car, and afterwards fear what might have happened to me, had I hit that tree. Here, one might insist that strictly speaking, it is a possible counterpart of myself for whom I fear (myself when I hit the tree), rather than myself as I actually am (unharmed). But nevertheless we speak quite naturally and appropriately of such a case as one in which I fear for myself. My concern for myself includes not just myself as I am, but also myself under various possible circumstances. I do not cease to care for myself merely because certain of my properties alter, for example, I get older. Hence when we talk of selfconcern, this includes concern for possible counterparts of ourselves, and so there is no reason to deny that I am afraid for myself when I contemplate myself in various possible situations. Matravers in contrast to Neill endorses anti-judgmentalism; and he also notes that the inability to act toward historical figures shows that lack of motivated action does not undermine the existence of fiction-directed emotions (Matravers 1998: chapters 3 and 4). These are, as we have seen, central parts of the solution. But Matravers in addition embraces as a core part of his solution the report model for fiction—the claim that the appropriate game we play with a work of fiction is to imagine it as a narrated documentary representation, which is non-transparent (i.e. whose function is not solely to produce belief in its audience). But the report model is unnecessary, since, as we have seen, anti-judgmentalism and the motivational point are all that are required to defend realism. The report model is also deeply problematic, and false of many fictions, particularly cinematic ones. Its

REASONS, EMOTIONS, AND FICTIONS 33

truth would require that every fiction film have a narrator, and be fictionally a documentary. Yet, most films do not have narrators (even implicit, visual ones); and to suppose that they are fictionally documentaries would require bizarre imaginings about why, say, the couple making love privately in the room do not notice that there is a film crew in the room with them, or how a modern camera crew can be filming in, say, an ancient Roman amphitheatre (is this a time-travel movie?). Responses open to supporters of the report model are available, but they are flawed, and the model should be rejected (Gaut forthcoming). Joyce (2000) distinguishes clearly between the problems of the possibility of fiction-directed emotions and their rationality. He solves the former by adopting an anti-judgmentalist view (though he does not confront the problems due to lack of motivated actions). The rationality puzzle he addresses by distinguishing between emotions that we do not control, which if counterproductive are phobias, and emotions, which are under our control, and these he holds are common in the case of fiction-directed emotions. Emotions under our control are subject to the norms of instrumental rationality, so they are rational if subjects are justified in believing that having them serves their interests. Since fiction provides valuable experiences in respect both of pleasure and of providing us with opportunities to learn, having fiction-directed emotions can be rational. Joyce’s criterion is related in part to the affective criterion developed earlier, since it is sensitive to the question of whether an emotion is in a certain way counterproductive. And his argument that it is not necessarily the absence of belief that makes something a phobia (since phobias can be founded on irrational beliefs as well) is well taken. But his instrumentalist solution fails. Suppose that you offer me a large sum of money, in order that (via the processes of active imagination) I induce in myself a serious fear of doorknobs. I succeed in doing this, and will not touch doorknobs, shrink away when I see one, and so on. My production of the emotion in myself was instrumentally rational (I gained something to my benefit), but the emotion itself is still irrational. It violates, for instance, the motivational criterion—I am motivated to avoid doorknobs, even though I do not believe they are dangerous (or if I do believe they are dangerous, my fear is derivatively irrational, as being founded on an irrational belief). So it can be instrumentally rational to induce in oneself an emotion that is irrational. Joyce’s account suffers from the same general defect as the other two just examined—it seeks to provide one simple criterion for a complex phenomenon, and so ends up in oversimplifying the rationality constraints to which emotions are subject. Realization of the complexity of these constraints is the key to seeing how fictiondirected emotions can be rational. Acknowledgments Ancestors of this chapter were delivered at the Universities of Aarhus, Oslo, Toronto, and Wisconsin, Madison; a more recent version was presented at the

34 BERYS GAUT

Imagination and the Arts conference. I am grateful to participants at all of these occasions for their questions and comments, as well as to Patricia Greenspan and Jerrold Levinson. Note 1 The proposal, of course, generalizes: for each emotion, one should substitute in the formula the evaluative property, which individuates that emotion—for pity, suffering; for anger, having been wronged, and so on. Note that in some instances of phobias the subject may have irrational beliefs about the feared object, which he believes to exist; in these cases, the phobic fear would pass the cognitive criterion, but the emotion would be derivatively irrational; that is, the fear would be irrational by virtue of being grounded on an irrational belief. A related criterion has also been suggested by Eva Dadlez (Dadlez 1996). However, she holds that it is a necessary condition for an emotion to be irrational that it be directed at a real object; but this is not correct, as the cases of Bertie and of the fear of ghosts show.

2 HOW I REALLY FEEL ABOUT JFK Stacie Friend

Here are some unsurprising responses to works of fiction: pitying Anna Karenina and feeling sad at her death; hoping that Strether remains in Paris and experiencing disappointment when he leaves; admiring Mr Smith and feeling suspense while he filibusters on the Senate floor. Common as they are, however, these reactions are puzzling: how is it possible for us to be moved emotionally by characters we know do not exist and events we know never transpired? It would appear impossible, or at least highly irrational, for you to pity my sister if you believe I do not have one, or to be angry with me for selling your trousers if you do not believe that I did. So it would seem that we ordinarily take beliefs to be requisite to our experiencing certain emotions—which makes our responses to fiction that much more difficult to comprehend. The most well-known and controversial solution to this paradox of fiction is Kendall Walton’s. According to Walton, because our pity of Anna—to take one example—does not involve belief, it is not genuine pity, but rather make-believe or imagined or, in Walton’s preferred terminology, quasi-pity (Walton 1978, 1990). On the face of it, this claim is rather unintuitive: my pity does not feel any less genuine than my pity of actual people, and I am not pretending to have an emotion the way an actor might do. But in spite of what some critics assume, Walton is not making the patently implausible claim that readers of Anna Karenina are faking their pity or pretending to feel pity, or that their experience must be less intense than genuine pity, or that they have conscious control over their emotions, or that they are not experiencing anything at all. To the contrary, Walton agrees that readers experience something essentially related to pity; what he denies is that this something is full-fledged pity. Even so, Walton’s opponents argue that we can resolve the paradox of fiction while preserving the intuition that our response to Anna is ordinary, run-of-the-mill pity; and they claim that retaining this intuition explains more than Walton’s approach. In my view, the arguments of Walton’s opponents depend on idiosyncratic features of examples involving purely fictional characters like Anna Karenina. Although the debate is usually couched in terms of such characters, the same basic problem arises when the individuals represented in fiction are real. We might detest Richard III as Shakespeare portrays him, even though we do not believe that this portrayal is accurate; and we might hope for the success of the Apollo 13

36 STACIE FRIEND

mission in Ron Howard’s film, even though we know that it failed in reality. It would be a mistake, then, to assume that the paradox of fiction arises only when the objects of our emotions do not exist: what is really at issue is the fact that we respond emotionally to fiction in ways that are not explained by our beliefs, but instead by what we imagine. That this is the crux of the paradox of fiction becomes clear only when we consider fictions about real persons and events. And I contend that once we turn our attention to these cases, Walton’s theory proves significantly more explanatory than the opposition. I defend this claim through an examination of our emotional responses to the real-life characters of Oliver Stone’s movie JFK, arguing that unless we draw a distinction between emotions along Walton’s lines, we cannot explain central features of our engagement with the fiction. In particular, I will argue that the distinction proposed by Walton accounts for important facts about cognitive organization that are easy to overlook when the debate over the paradox of fiction focuses on invented characters. In the opening section I outline the debate between Walton and his critics, whom after Noël Carroll I designate as thought theorists (Carroll 1990). Both Walton and the thought theorist accept the general thesis of cognitivism, according to which emotions essentially involve propositionally contentful states. In what follows I assume the truth of cognitivism; what is at issue in my chapter is the debate between narrow cognitivists such as Walton who maintain that the contentful state must be a belief, and broad cognitivists such as Carroll who deny this. I then take up JFK, arguing that the thought theory cannot resolve the apparent conflict between my emotions in response to the film and my emotions based on what I believe about the real people and events it depicts. I argue that Walton’s approach better explains the fact that my belief-involving emotions take priority in my cognitive life—a priority necessary to maintain rationality. I next consider and reject the reply that conflicts between my emotions can be avoided if we accept that my responses to the film are directed, not at the real people portrayed, but instead at fictional characters. Finally, I suggest that the priority of belief-involving emotions, and with it their claim to be the only genuine emotions, derives from their place at the foundations of practical reasoning. Puzzle and pretense Walton’s solution to the paradox of fiction cannot be understood in isolation from his more general theory of make-believe. Starting from examples of novels, plays, films, paintings, and sculptures, Walton aims to construct a theory of what he calls representational art or fiction. Walton defines a work of fiction as a work whose function it is to act as props in certain games of make-believe. The connection with children’s games of make-believe, where the props may be dolls and toy trucks, is intentional; for Walton there is continuity between the two types of games. The idea is that works of fiction are designed to prescribe imaginings about their content, and imagining what is prescribed is participating in the game of

HOW I REALLY FEEL ABOUT JFK 37

make-believe authorized by the work. What we are supposed to imagine is what is “true in the fiction”. We must agree that works of fiction are not created so that we can register that such-and-such is the case in some distant fictional world; rather, we are supposed to engage imaginatively with them, making believe that the events narrated really have taken place, that the people described really do exist, and so on. On Walton’s theory our imaginings about the content of the work are generated by the more basic games of make-believe we play with fictional representations, which involve imagining about ourselves. Specifically, we imagine that in engaging with the fiction we are learning about actual fact, and we respond to the fictional events from within this pretense. In reading Crime and Punishment, for example, we imagine ourselves to be reading a nonfiction report of actual events, thereby generating fictional truths about our own experiences. In particular, we imagine, of our reading the novel, that this very experience counts as learning about actual people and situations. Similarly, according to Walton, in looking at a bust of Napoleon, we imagine, of our seeing the bust, that this very experience counts as looking at Napoleon. Our psychological, emotional, physical, and verbal responses within the scope of these imaginings all constitute aspects of our participation in the relevant games of make-believe. Suppose I find myself pitying Hardy’s Tess because Angel rejects her. I am perfectly aware that there are no such people as Tess Durbeyfield and Angel Clare, and no such event as Angel’s rejection of Tess. Nonetheless, it is clear that Tess of the d’Urbervilles prescribes that I imagine that there are such people and that Angel rejected Tess. I imagine, of my reading the novel, that I am learning about these real people and their situations. Within the context of so imagining, I learn that Tess has been rejected; in reality, of course, this is not the case. Still, my imagining is sufficiently vivid that I have a sinking sensation, tears well up, I find it hard to turn the page. Facts about my real feelings—specifically, the psychological and physical state I am in due to my imagining that Angel rejected Tess, which Walton terms quasi-pity—make it fictionally the case that I pity Tess; from a perspective outside the game of make-believe, however, this experience is not genuine pity. Just as it is the fact that I (really) read a novel that makes it fictionally the case that I read a nonfiction report, it is the fact that I (really) experience quasi-pity that makes it fictionally the case that I experience pity. And because it is only fictionally the case that I am reading a true report or that Angel mistreats Tess—the imaginings that explain my quasi-pity—it is only fictionally the case that I pity her. As even this brief overview indicates, Walton’s introduction of quasi-emotions is a natural consequence of his theory of make-believe. While the theory is by no means uncontroversial, it is commonly accepted that the concept of make-believe is central to the institution of fiction, and that we can understand truth-in-fiction as prescriptions to make-believe or imagine the content of the work. That is, Walton’s critics agree that I imagine, rather than believe, that in reading Tess of the d’Urbervilles I am learning about actual events such as Angel’s rejection of Tess.

38 STACIE FRIEND

At the same time, however, they think that my emotional response to reading the fiction is genuine compassion. By contrast, Walton argues that because this response is based on what I imagine rather than what I believe, it cannot be genuine compassion. Walton’s defense of this claim appeals to a particular version of cognitivism about the emotions (Walton 1990:200–4). According to cognitivism, a propositionally contentful state constitutes an essential component of an emotion and partially individuates it. This cognitive component is what distinguishes emotions from mere “feelings”, which are just affective states. In paradigmatic cases, the cognitive component of an emotion is a belief: in pity the belief that someone has suffered, in fear the belief that one is in danger, and so on. At least some of these emotional states also have a particular affective component, involving phenomenological and physiological responses, as well as associations with action: for example, fear usually involves a feeling of anxiety and an increase in heart rate and perspiration, along with a desire to flee whatever is feared. The complex of these aspects is what makes the emotion the emotion it is. Although there is general agreement that in the central cases the cognitive component of an emotion is a belief, consensus diminishes when we come to the question of whether the belief-requirement is conceptually necessary. Narrow cognitivists such as Walton argue that for a particular experience to count as, say, genuine pity, the propositional content that someone has undeservedly suffered must be believed. One reason for adopting this view is the connection between belief and the sort of motivational force that would lead to action: one will not desire to help without believing that there has been undeserved suffering. Another reason is the observation that we typically assess emotions as irrational when they are not explained by a belief: no matter how vividly I imagine that my friend Eric stole my watch, it would be irrational to yell at him. Even if one rejects the constitutive claim of cognitivism—that a propositional attitude is literally a component of an emotion—it remains the case that my imagining Eric to be a thief provides no explanation for genuine anger. By contrast, neither feature seems to apply to emotions in response to fiction: our “pity” of Tess has no motivational force, nor do we count ourselves irrational in spite of the lack of belief; in other words, the response departs substantially from what we ordinarily expect of the emotion. In spite of these considerations, there remains a conviction that our emotional responses to fiction are no different in kind from the emotions we experience in at least some other contexts. Thought theorists therefore reject narrow cognitivism, holding that if an emotion involves a cognitive component—in the case of pity a thought or proposition with a content like “that person is suffering undeservedly”—this thought could be entertained, supposed, imagined, or whatever, and the emotion would still count as genuine pity. Noël Carroll contrasts a belief, which is “a proposition held in the mind as asserted” with a (mere) thought, which is “a matter of entertaining a proposition in the mind

HOW I REALLY FEEL ABOUT JFK 39

unasserted,” as when we suppose or imagine that something is the case (Carroll 1997:209). Moreover, it seems to be indisputable that emotions can be engendered in the process of holding propositions before the mind unasserted. While cutting vegetables, imagine putting the very sharp knife in your hand into your eye. One suddenly feels a shudder. You need not believe that you are going to put the knife into your eye. Indeed, you know that you are not going to do this. Yet merely entertaining the thought, or the propositional content of the thought (that I am putting this knife into my eye), can be sufficient for playing the role in causing a tremor of terror. For emotions may rest on thoughts and not merely upon beliefs. (Carroll 1997:209) Thus, on the thought theory, there is nothing out of the way in our terror at the merely imagined thought of plunging a knife into our eye; and similarly, there is nothing problematic in taking our pity of Tess to be genuine even though we know there is no such person who suffered. But Walton can simply accept Carroll’s contention that in both cases, our response is caused by something we imagine; establishing a causal relationship is still insufficient to show that the resultant state is an emotion. Because Carroll agrees that “a central component of the emotions is a cognitive state” (Carroll 1997:209), if one’s “shudder” or “tremor of terror” is to count as a full-fledged emotion Carroll must demonstrate that the propositional content imagined constitutes a component of the experience. A causal connection does not fulfill this task. There are, however, clearer cases of genuine emotions where beliefs or motivational force are lacking, which can be adduced by thought theorists as evidence against Walton. Thought theorists frequently invoke phobias: while Sally’s phobic fear of Fido is irrational because she does not believe that Fido poses any danger to her, it still counts as genuine fear. In response to this kind of example, Walton stresses the motivational component of the emotion, the fact that Sally always avoids Fido (Walton 1990:201–2). Thought theorists then point to our emotional responses to nonfiction representations: in these cases there is belief, but due to spatial or temporal distance no motivation to act. If something occurred a hundred years ago or a thousand miles away, we are as little capable of doing anything about it as we are in the case of fiction (Matravers 1998:69–73; Gaut, Chapter 1, this volume). Thought theorists conclude that if these responses, where one or another central component of an emotion is missing, nonetheless count as genuine examples of particular emotions, so should our responses to fiction. At this point the debate between Walton and the thought theorist seems merely to be a matter of terminology: one side wants to classify our response to Tess’s fictional plight as one among several species of the genus genuine pity, while acknowledging that it departs from the paradigm case involving belief; the

40 STACIE FRIEND

other side takes this departure to be sufficient to classify the response as quasi-pity. Although much of the opposition to Walton’s theory apparently results from discomfort about the terminology, what matters is not the choice of label, but whether or not classifying the emotions one way or the other explains the phenomena we want to explain. After all, both Walton and the thought theorist agree in their basic description of our emotional responses to fiction: that these responses can have the same physiological and psychological profile as emotions that involve beliefs; that they do not involve beliefs, but only imaginings; and that they lack motivational force. Furthermore, Walton and the thought theorists agree that we need to draw a distinction between those cases where lack of belief entails irrationality, as with phobias, and those cases where lack of belief does not entail irrationality, as with responses to fiction. Walton has a simple explanation of the distinction: only full-fledged emotions require beliefs to be rational, while quasi-emotions do not. The thought theorist cannot have recourse to this explanation, for she maintains that both phobias and responses to fiction are species of genuine emotions. It is for this reason that once Berys Gaut (Chapter 1, this volume) has finished arguing that the emotions are genuine, he must then address the problem of how to explain why emotional responses to what we imagine are not irrational in the way phobias are. To understand the problem this poses for the thought theorist, let’s reconsider Carroll’s example. One reason that a causal connection between imagining putting a knife in one’s eye and a tremor of terror is insufficient to show that the result is an emotion is that causation is the paradigm of a nonrational relation. My reaction to this episode of imagining looks instead to be an involuntary physiological response, a response that provides no reason to stop cutting vegetables. Indeed, if I told you I had decided not to cut the vegetables on this basis, you would think I was either irrational or that I was deceptively trying to get out of cooking. By contrast, if I believed that I might suffer an epileptic seizure and really put the knife into my eye, I would be irrational, given the risk, to cut the vegetables. Carroll, like Gaut, must provide some additional explanation of the fact that although I am genuinely terrified in both cases, I am rational to continue cutting when my terror results from what I imagine, while I would be irrational to continue cutting when my terror results from what I believe. This whole approach misses the virtue of Walton’s account. Walton’s invocation of quasi-emotions is designed, not primarily to stress the differences between these experiences and genuine emotions, but rather to stress the essential connection between them. If we take ordinary, belief-based pity and what Walton calls quasi-pity both to be full-fledged emotions in their own right, we have to start explaining why they differ in central respects, for example, why only one kind is irrational when belief is lacking. If, on the other hand, we take quasi-pity to have the relationship to real pity that Walton suggests, we can see just how they are connected. On Walton’s theory, one genuinely experiences a certain emotional state, involving an imagined content and just those physical and

HOW I REALLY FEEL ABOUT JFK 41

psychological features that characterize genuine pity, and one imagines of that experience that it is an experience of genuine pity. Again, this construal follows from Walton’s overall account of our experience of fictions. One genuinely reads a novel, and one imagines of one’s reading the novel that it is reading a true report. One genuinely sees a picture of an elephant, and one imagines of one’s seeing the picture that it is seeing an actual elephant. Put more concisely, one “imaginatively sees a real elephant”. In Chapter 1 Gaut suggests a reason for rejecting this approach to emotions. Addressing a puzzle about how we can learn through imagination, he proposes that having actual emotions toward imagined situations can help us learn what we ought to do. If we accept that our imaginative engagement with fiction gives rise to genuine emotions, Gaut says, we can explain our learning as a case of education through actual experience. Presumably, Walton’s insistence that we do not experience genuine emotions in response to fiction leaves him faced with the puzzle. It is not clear to me why describing my experience in reading Tess of the d’Urbervilles as “genuine pity toward an imagined person” provides a better explanation of how we learn than describing it as “imagined pity toward a real person”, given that the phenomenology of the latter is exactly the same as the former. What is clear, however, is that Gaut’s construal is unavailable when we consider our responses to works of fiction about real people. As a result, I will argue, such works pose a serious problem for the thought theory. To see why, let us turn to JFK. Keeping cognitive order JFK is Oliver Stone’s 1991 controversial movie about the assassination of John F.Kennedy. The film takes as its subject Jim Garrison, the real New Orleans District Attorney who prosecuted the only trial related to the assassination, and is based largely on Garrison’s 1988 book On the Trail of the Assassins. In the movie Garrison (Kevin Costner) is a noble person willing to make sacrifices in the pursuit of the truth: specifically, the truth that Lee Harvey Oswald (Gary Oldman) was merely a patsy in the military-industrial-governmental conspiracy to assassinate Kennedy. The investigation by Garrison and his team uncovers a plot arranged in New Orleans by the civic leader Clay Shaw (Tommy Lee Jones). In the movie as well as in reality, Garrison brought Clay Shaw to trial in 1969 for conspiring to murder President Kennedy. It is more than clear that in response to the film we are supposed to imagine that Jim Garrison is a person of inestimable integrity, whose dedication to the truth is genuinely heroic, and that Clay Shaw is a shady, arrogant individual whose acquittal depends on the continued operation of the conspiracy. It is important to recognize that imagining in these ways differs from our imagining, in response to Hardy’s novel, that Angel rejects Tess. Because Jim Garrison and Clay Shaw, along with most of the other characters in JFK, are real people, the movie prescribes imaginings about them—imaginings that refer to the

42 STACIE FRIEND

real individual. That is, we do not merely imagine that there is someone the movie is about, who is named “Jim Garrison” and who prosecuted someone named “Clay Shaw”; rather, we imagine, of the real Garrison and the real Shaw, that the one prosecuted the other. Compare little Peggy’s game with her dolls. Let’s say that in Peggy’s game, it is fictional—that is, participants in the game are supposed to imagine—that a doll in the sink is a baby having a bath, that a doll in the sock drawer is a baby in her crib, and so on. Imagining, of the particular dolls involved in the game, that they are babies bathing, sleeping, etc., is to be contrasted with imagining that there is some baby bathing, sleeping, and so on. Walton calls the real people, places, things, and events depicted in works of fiction objects of representation; fictions generate de re fictional truths about the objects they represent. So just as the rules of Peggy’s game make it fictionally the case, of a particular doll, that it is a baby asleep in her crib, JFK makes it fictionally the case, of the real Jim Garrison, that he was a noble, self-sacrificing individual pursuing the truth. That Stone would also like us to believe that this is so is obvious; the propagandist element of the film is hard to miss. However, people with any independent information about Garrison are unlikely to change their beliefs as a result of watching the movie. Having read a number of articles undermining Garrison’s version of events, I have formed a very different opinion of the man, perhaps “the most thoroughly discredited” of any conspiracy theorist even among conspiracy theorists themselves (Wicker 1991). Garrison and his office bribed witnesses, took statements made under hypnotic suggestion as testimony, and otherwise presented no evidence for the guilt of Clay Shaw, who was acquitted in under one hour. After years of prosecution, Shaw was left bankrupt and died an early death (Epstein 1992). Let us say that I believe that Jim Garrison’s obsession with proving a government conspiracy and his cavalier attitude toward the truth ruined a perfectly respectable man’s life. Call the set of a person’s beliefs about the world her belief stock. It is part of my belief stock that Garrison was a contemptible person. Right now, as I am writing this, such beliefs about Garrison are occurrent: that is, I am consciously aware of them. Most of the time, however, the beliefs are dispositional: that is, most of the time I am not explicitly thinking anything whatever about Jim Garrison or the Kennedy assassination. Even when I am thinking about something else entirely, though, it would be an accurate description of me to say that I believe that Garrison’s theory was false. Most of the beliefs in our belief stocks are dispositional in this way. Now, these beliefs about Garrison and Shaw clearly contradict what the movie prescribes that I imagine about them. This contradiction need not pose a problem: that we are able to imagine that real things are different from the way we believe them to be is basic to the nature of imagination and to the practice of fiction. I imagine that Garrison is admirable, while I believe he was contemptible. If we were unable, in response to science fiction, to imagine the world to operate with different physical laws; or in response to horror films, to imagine that there are monsters; or in response to

HOW I REALLY FEEL ABOUT JFK 43

historical novels, to imagine that we are privy to the secret thoughts of historical personages, the institution of fiction would be difficult to explain. Let us agree, then, that even while watching JFK and imagining as prescribed (occurrently), my dispositional beliefs stay constant. Contrast my case with the response of Naïve Nellie. Nellie is completely persuaded by Stone’s propaganda, coming to believe that Garrison was noble while Shaw was part of a conspiracy that reached to the highest levels of government. Both of us engage in imagining that Tommy Lee Jones is Clay Shaw, that Kevin Costner is Jim Garrison, and so on. And, depending on one’s theory of film experience, both of us imagine either that we are seeing the real events unfold (Walton 1990), or that we are watching a nonfiction report of those real events (Currie 1990:92–8; Matravers 1998). (In several parts of the film, of course, we are watching documentary footage, edited by Stone.) But we go on to do different things with the contents of our imaginings. The cognitive difference between us can be described by employing the contrast between incorporating and compartmentalizing new information (Gerrig 1993:207–24). Nellie incorporates the content of the film into her stock of beliefs about the world. She includes in her belief stock such propositions as “Garrison was noble” and “Shaw was a conspirator”. Of course I form beliefs in response to the movie as well, beliefs like “in JFK, Garrison is noble” and “in JFK, Shaw is a conspirator”. But in my case the contents remain attached to the representation, and are thereby compartmentalized. This is in sharp contrast to my belief that Garrison was contemptible: once I decide that the nonfiction articles I’ve read about Garrison are accurate, I can forget these sources of my belief entirely and assimilate their contents into my belief stock. If I did the same with the content of JFK, my rationality would be impugned, because I would believe both that Garrison was contemptible and that he was noble—and it is a basic constraint on rationality that we avoid contradictory beliefs. Instead I believe that Garrison was contemptible and that the movie portrays him as noble. And though in response to the movie I imagine that Garrison is noble, I continue to believe that he was contemptible. Keeping the beliefs in our belief stocks organized in this way reflects their role in practical reasoning: our actions will not be successful if our beliefs fail to represent the world correctly. The same kinds of issues arise with emotions as with beliefs. It should come as no surprise that based on my beliefs, I feel contempt for Garrison and disgust at his cavalier attitude for the truth, while I pity Shaw, a good man wrongly accused by an obsessed prosecutor. Most of the time these emotions are not occurrent, because I am not thinking about Garrison; still, even while I am thinking about something else, it would be correct to say that I feel contempt for him. These emotions thus form part of my emotional stock, and they change in predictable ways correlated with changes in my beliefs. If I came to believe that Garrison had, in fact, been nobly pursuing the truth, I would be irrational to continue to feel contempt for him. In other words, there is a rational connection between my emotions toward, and my beliefs about, particular individuals and events.

44 STACIE FRIEND

In response to the film, however, I experience emotional responses contrary to my ordinary attitudes. I sympathize with Garrison and his cause, and I am convinced that Shaw is an evil conspirator. I want Garrison to win and it comes as a shock when the jury acquits. All this in spite of the fact that I already knew, going into the movie, that no one has ever been convicted for conspiring to murder Kennedy. To keep cognitive order, it seems that I must keep these emotions separate from my ordinary attitudes toward Garrison, just as I keep my imaginings about Garrison separate from my beliefs about him. If I do not, I will find myself admiring Garrison for his noble pursuit of truth and condemning him for his disregard of truth, surely an irrational state of mind. Avoiding contradictions among our emotions can be just as important to rationality as avoiding contradictions among our beliefs, because they too play a role in practical reasoning: inconsistent desires will prevent us from attaining our goals. By contrast, Nellie can incorporate her emotional responses to the film into her stock of emotions without inconsistency. Based on the beliefs she forms in response to the movie, she genuinely admires Garrison. Walton has a simple explanation of the fact that I remain as rational in my emotions as Nellie, since there need be no conflict between my genuine contempt and my quasiadmiration: it is not literally true that I admire Garrison, but only fictionally true; really I detest him. The thought theorist must describe the case differently. On her view, both my admiration and my contempt are genuine emotions, one of which is explained and partly constituted by what I believe, and the other of which is explained and partly constituted by what I imagine. So I feel admiration for Garrison because I imagine that he is admirable, and I feel contempt for him because I believe that he is contemptible. But this way of putting it implies that imagining, of Garrison, that he was admirable is sufficient to explain genuine admiration of him. That just seems false. Compare the following case. Suppose you believe Nelson Mandela to be a great man and you admire him. Now, if I made up a story in which Mandela tortured kittens, and you knew this story to be pure fantasy, it would be inexplicable for you to change your feelings toward him. Similarly, my imagining that Garrison was a noble pursuer of truth would not explain a change in my feelings toward him. That is, my emotional stock should remain exactly the same, even while I am imagining that Garrison is different from how I believe him to be. Clearly the thought theorist will have to find some way of keeping apart the contempt and admiration, such that they do not conflict. On the thought theory, the fact that my admiration is explained by what I imagine makes no difference to the analysis; it is just the same sort of emotion I would experience based on a belief. The problem is just that, based on my beliefs, I experience an apparently incompatible emotion, namely contempt. The thought theorist responds that there is no difficulty in principle in admiring and condemning the same person, so long as one is responding to different features of the person (Gaut, Chapter 1, this volume). Thus I admire Bill Clinton for his record in foreign diplomacy, but I feel

HOW I REALLY FEEL ABOUT JFK 45

contempt for him due to his inability to keep his pants up while in office. However, the same sort of approach does not apply to my feelings about Garrison, because these respond to exactly the same feature of the man: his attitude toward the truth. The thought theorist might reply that in one case the feature is real, and in the other case imagined. But then we are back to the problem that imagined features of a person provide no rational explanation for genuine emotions. To remove the sense of conflict, the advocate of the thought theory could say that different episodes of thinking about Garrison—one imagining, one believing— result in different emotions toward him. This way of representing the situation accords with Carroll’s description of the case where I imagine putting a knife in my eye. We could assume that insofar as I experience genuine terror in response to my imagining, the terror is temporary; once I recall that I am not going to stab myself, I can continue cutting vegetables. In other words, it is only so long as I am occurrently imagining stabbing myself in the eye that I experience terror. Similarly, the thought theorist might suggest, I genuinely admire Garrison only so long as I am occurrently imagining that he is nobly pursuing the truth. The rest of the time I feel contempt for him. Because I do not experience these emotions at the same time, there is no conflict. However, if I believe that Garrison was contemptible and I feel contempt toward him, there is no reason to think that I stop believing he is contemptible and feeling contempt for him as soon as I start imagining him to be as the movie portrays him. Suppose I am watching the film at exactly the moment two of my friends discuss my feelings about the Kennedy assassination. If Lauren tells Anthony that I reject Garrison’s version of the conspiracy theory and that I detest the man, we do not want to say that she has given a false description of my attitude just because I am emotionally caught up in the film. It is better to say that all the while I continue to detest Garrison, even while I am imaginatively construing him as heroic. In other words, my belief-involving contempt has priority over my imagination-involving admiration. This is because it forms part of my stock of emotions, the ones that guide my actions in accord with my stock of beliefs. And because it forms part of my emotional stock, it remains constant so long as I continue to believe that Garrison was contemptible. As a result, it is fair to say that regardless of what thoughts about Garrison I happen to be imagining, all the while I really feel contempt for him. Walton’s theory explains this feature of my interaction with JFK. Because my admiration is experienced in the context of my game of make-believe with JFK, it is only fictionally the case, rather than actually the case, that I admire Garrison. This is perfectly consistent with its being true that I really detest Garrison. It is difficult to see how the thought theory can provide an equally satisfactory account.

46 STACIE FRIEND

Real and unreal individuals The thought theorist does have another strategy open to her, however. She could argue that my emotional responses to JFK are not directed at the real Garrison, but rather at Garrison-as-he-is-in-JFK. If this means that I experience genuine admiration only insofar as I imagine Garrison to be as portrayed in the movie, we have not progressed beyond the last option. So it must mean that my emotions are directed at different things: my contempt is for the real Garrison, while my admiration is for the fictional Garrison—a fictional character based on the real person. Similarly, Gaut (Chapter 1, this volume) suggests that in imagining that one’s friendly dog Fido is dangerous, the resultant fear is not of the real Fido, but of “the make-believe Fido,…a merely imagined being.” Leaving aside the question of what it means to be “a merely imagined being,” let us agree that if the object of my admiration were not identical to the object of my contempt, there would be no conflict. I have argued elsewhere that it is a mistake to construe our emotional experiences in response to fictions about real individuals as directed instead toward fictional characters (Friend 2000a), so I will outline only briefly some of the problems that arise on such a construal. First, if we agree that my imaginings in response to the film are imaginings about the real Garrison, it is ad hoc to say that my emotions—which are supposed to be explained by, and partly constituted by, my imaginings—are directed at a fictional character. On this construal, I respond to the film by imagining, of the real Jim Garrison, that he is nobly pursuing the truth; and as a result of so imagining, I admire a fictional character for nobly pursuing the truth. But if it is (the real) Garrison who is the noble pursuer of truth in my imaginings, then it should be the same Garrison, imagined as noble, whom I admire. The obvious reply to this consideration is to deny that my imaginings themselves are about the real Garrison. According to this reply, we take the Garrison-of-JFK to be a fictional character modeled on, but not identical to, the real Garrison, thereby placing him in the same category as Tess, Angel, Raskolnikov, and their ilk. Yet there are good reasons to reject this approach. To begin with, it would mean committing oneself to the very strong—and to my mind indefensible— claim that works of fiction never prescribe imaginings about actual people, places, things, or events. Consider how one feels about London as it is portrayed in Bleak House, or about Napoleon as he is portrayed in War and Peace. Should we say that Dickens and Tolstoy were inventing fictional characters modeled on the real city and person, simply to account for conflicts in our emotional responses to the fictions? Clearly not. Second, this solution fails to account for the difficulty people might have in imagining something contrary to what they believe, as with historians of the English monarchy who face a psychological obstacle in imagining as prescribed by Richard the Third. It is precisely because we recognize that we are supposed to be imagining about a real person that such obstacles arise. Finally, leaving aside these more general considerations, to deny that JFK is about the real Jim Garrison—that is, to deny that JFK refers directly to the historical individual—

HOW I REALLY FEEL ABOUT JFK 47

is simply to misinterpret the movie. This is not only because Stone wants to persuade us that Garrison was, in reality, a noble pursuer of truth. It is also because the use of documentary footage in the film requires our making the identification: More than halfway into “J.F.K.,”…New Orleans District Attorney Jim Garrison and his wife, Liz, are seen watching a television documentary about Mr. Garrison’s investigation of the events of Nov. 22, 1963, in Dallas. The documentary’s anchorman is heard charging that the District Attorney used improper methods to get witnesses to support his case against the New Orleans businessman Clay Shaw for his part in a supposed conspiracy surrounding the murder of President Kennedy. Kevin Costner, portraying Mr. Garrison, suggests by facial expression and dialogue that the charge is unfair and rigged to destroy his credibility—thus attacking the credibility of the documentary. (Wicker 1991) In describing this scene, Wicker assumes that Costner portrays the real Jim Garrison, and for good reason. There can be no doubt that the documentary— a real NBC broadcast, which was aired on 19 June 1967—refers to the real Jim Garrison. But there also can be no doubt that we are supposed to imagine the subject of the documentary to be identical to the man watching it in the living room with his wife. Such aspects of the film undermine attempts to distinguish the Garrison-of-JFK from the historical person. None of what I have said provides a knockdown argument that JFK refers directly to the real Jim Garrison and prescribes imaginings about him. And there are, of course, any number of philosophers and literary theorists who deny either that works of fiction refer at all, or that they prescribe our imagining directly about real individuals. I happen to think that if we can resolve the apparent conflict between my imagination-involving admiration and my belief-involving contempt, while also maintaining that JFK prescribes imaginings about the real Jim Garrison, we ought to do so; and Walton’s claim that my contempt is a genuine emotion while my admiration is merely a quasi-emotion does just that. But because there is so much controversy over reference in fiction, I would like to indicate why taking Garrison-as-he-is-in-JFK to be a fictional character still would not suffice to resolve the conflict. Consider this question: why do the problems I have outlined for the thought theory seem to arise only when we are dealing with real individuals? That is, why does it seem more plausible to say that I genuinely pity Tess because Angel rejects her than it is to say that I genuinely admire Garrison because of his noble pursuit of truth? Because in the former case there seem to be no conflicts between what I imagine and what I believe. My imagining that Tess suffered and my consequent pity for her appear to function cognitively the same way as my believing that

48 STACIE FRIEND

Garrison was contemptible and my consequent contempt for him. If I detached my imaginings about Tess from their source in Hardy’s novel and incorporated them into my stock of beliefs, they would not conflict with any other beliefs about Tess —and the same applies to my pity for Tess, which conflicts with no other feelings about her. If this is right, we can say that my imaginings and feelings about Tess are dispositional: even if I am thinking about something else entirely, it would be accurate to say that I pity Tess. The situation is not so simple, however. First, it is not true that I could assimilate my imaginings about Tess into my stock of beliefs without causing conflicts. After all, one of the things I imagine in response to Hardy’s novel is that Tess exists, and I certainly do not believe that. In my belief stock I might have the propositions “Tess does not exist” and “in Hardy’s novel, Tess exists”, but I could not have both “Tess does not exist” and “Tess exists” without being irrational. Recall that my belief stock is my set of beliefs about the world; taken together, these beliefs constitute my representation of how the world really is. Nowhere in this representation of the world will we find reference to such people as Oliver Twist or Raskolnikov, to such places as Eldorado or Lilliput, or to such events as Angel’s rejection of Tess. We will find reference to works of fiction that prescribe imaginings about these fictional people, places, and events; but the contents of those imaginings remain attached to their sources and thus compartmentalized. Now, some theorists argue that we should include Oliver et al. in our representations of the world, as abstract or nonexistent objects. From this perspective, works of fiction inform us about real entities about which we can form beliefs, thereby allowing for smooth assimilation into our belief stocks. There are numerous problems with these approaches (see Friend 2000b), but in the present context the difficulty is that they still fail to remove all conflicts among our emotions. When the same character is portrayed differently in different fictions, emotional responses can change: if I admire Odysseus’s devotion to Penelope in the Odyssey, I will feel quite differently about him in response to Dante’s Inferno (Canto 26) according to which he preferred further adventures to returning home to his wife. We could claim that these are two different characters, but such a move not only appears ad hoc, it runs counter to our ordinary experiences of fiction. In the present case, identification of the character across works is necessary to the correct interpretation of Dante’s epic: if we did not recognize the Ulysses identified by Dante as the very same one imported by Virgil from Homer, we would not possess the background necessary to explain (and thereby, for Dante, to justify) the punishment he receives in the Inferno. And in spite of the changes undergone by James Bond over the course of his career, the popularity of the Bond films depends on our assuming that they are about the same character. The same can be said for any other serial fiction. Indeed, the problem of conflict would not be resolved even if we insisted that for each film or each epic poem there are distinct fictional characters, because many individual works of fiction are inconsistent, whether by error—in Turgenev’s Fathers and Sons, Katya Odintsov is said to be both eighteen and

HOW I REALLY FEEL ABOUT JFK 49

twenty-one—or by design—in Robbe-Grillet’s Jealousy, the main character is both a customs official and not a customs official, both honest and dishonest. Just as one might respond differently to the same fictional character in different fictions, one could conceivably experience apparently contradictory emotional states depending on variations in the way a fictional character is represented from one part of a work to the next. What these considerations indicate is that so long as we wish to maintain the view that our affective responses to fiction are genuine emotions, we are faced with the problem of conflicts among these emotions—a problem that does not go away even if we construe all of these emotions as directed toward fictional characters, and even if we take fictional characters to be real. More importantly, these cases highlight the distinctive feature of cases involving beliefs about real individuals. If one is describing my feelings about Garrison, there is a clear priority to be given to my contempt over my admiration; this is why Lauren is right to say I detest Garrison even while I am watching JFK. By contrast, there is no reason to prioritize either of my feelings about Odysseus, no answer to the question, “do I really admire him or do I really detest him?” It is not that I am ambivalent, that I simply cannot decide how to feel about the character. Rather, it is because my different responses to Odysseus occur within the scope of different imaginings; neither can be detached from their sources and incorporated into my emotional stock. Just as the thought theorist must offer some additional explanation of why emotional responses to fiction are not irrational despite the lack of belief, she must offer some additional explanation of why certain emotional responses take priority over others in our cognitive organization. Once again Walton has a straightforward explanation of the data: only my belief-involving contempt for Garrison is a genuine emotion, and this explains why it is the sole feeling that takes priority. By contrast, my feelings about Odysseus and my responses to JFK count as genuine only within the confines of certain games of make-believe. Moreover, the kinds of object-directed emotional states on which I have focused so far—contempt for Garrison, pity of Tess—are not the only ones where conflict arises and prioritization is relevant. It will be recalled that in watching JFK, I hope that Garrison will win his case, and I am shocked and disappointed when the jury acquits Shaw. These responses are puzzling in their own right. First, it cannot literally be true that I hope the real Garrison will win the case, because the real case is long over. Second, it is not literally true that I hope the fictional Garrison will win; to the contrary, I think the film would be much worse, and my experience of it much less enjoyable, if it had a happy ending. On the thought theory, my desire that Garrison win because I am sympathetic to him, and my desire that he lose because that would make a better film, are in direct conflict, though one might outweigh the other (Gaut, Chapter 1, this volume). But it is quite intuitive to say that really I want the film to be structured a certain way, and that it is only within the context of my imagining that I want Garrison to win. It is a common feature of our experience of fiction that we can

50 STACIE FRIEND

experience the same or similar emotional ups and downs each time we see a movie or read a book. On Walton’s theory, whenever I see the film I imagine that I am learning about these events for the first time (Walton 1978, 1990:258– 62). This is consistent with its being the case that all the while, I know how the movie ends and do not genuinely wish it to end any differently. Similarly, it would be incomprehensible for me literally to be shocked and disappointed when Shaw is acquitted. Because I could not literally be shocked by an event I know already to have transpired, it must be the case that I am shocked that Garrison loses only within the context of imagining that I am learning about these events for the first time. To make sense of this, we must assume that I am imagining the events to unfold before my eyes, or at least to unfold concurrently with my watching the movie. But it is not merely that I go from one episode of imagining to another—from imagining that Garrison might win the case to imagining that he has lost—responding to each separately. In addition to imagining about Garrison and Shaw, I must also imagine about myself: in particular, I must imagine that in watching the movie I am finding out about the events. Only this can explain how I am shocked “when” Shaw is acquitted. One might think, though, that my shock on the first viewing of the film still counts as genuine, because I do not know how the plot will unfold. But this only goes to show that I am shocked that in the film Garrison loses his case; in other words, that I am genuinely surprised the film turned out a certain way. Compare Neill’s proposal that Charles, who appears to be “terrified” of the Green Slime in a scary movie, is really just startled by the images and music of the film (Neill 1991). Because this analysis provides a different object for Charles’s emotional state, it lends no support to those who wish to claim that Charles is genuinely afraid of the Slime. Similarly, my genuine shock upon finding out that the film develops a certain way— which is just what I am ignorant of before seeing it—lends no support to those who would claim that I am genuinely shocked that Garrison loses his case. To the contrary, the fundamental features of my response lend support to Walton’s claim that I experience only quasi-shock, rather than the genuine article. Emotions in practical reasoning I have emphasized the priority that certain emotional states have in our cognitive organization, maintaining that because they can be detached from their sources and incorporated into our emotional stock they have a good claim to being the only genuine emotions. But why are this prioritization and incorporation so important? How do they justify denying the status of “genuine emotion” to other experiences? Above, I suggested, briefly, that the answer has to do with the contribution of certain emotions to practical reasoning. We must avoid contradictions among the beliefs in our belief stocks because a failure to represent the world correctly can prevent successful action. Similarly, we must avoid

HOW I REALLY FEEL ABOUT JFK 51

contradictions among our emotions because inconsistent desires can prevent us from attaining our goals. This explanation, however, seems to leave me vulnerable to one of the charges leveled by thought theorists against Walton. It will be recalled that, in reply to the thought theorist’s appeal to phobias as genuine emotions without belief, Walton stresses the significance of action: “fear is motivating in distinctive ways, whether or not its motivational force is attributed to cognitive elements in it…. Fear emasculated by subtracting its distinctive motivational force is not fear at all” (Walton 1990:201–2). If this point could be generalized to other emotions, it looks as though on Walton’s account, the importance of belief to emotion is just that it plays a role in explaining motivation. And if that is right, the fact that works of nonfiction often fail to motivate actions should imply that the emotions we experience in response to these works, though they involve beliefs, are not genuine—surely an unintuitive result. In writing Robinson Crusoe, Defoe drew heavily on the journal of the reallife castaway, Alexander Selkirk. A reader of Selkirk’s journal could surely be moved to emotion at his suffering. However, the fact that Selkirk was marooned in 1704 (and died in 1721) makes it impossible— really impossible—for the reader to help him…. Hence, if the absence of such a connection [to action] is no reason to deny that we feel emotions towards Alexander Selkirk, it is no reason to deny that we feel emotions towards Robinson Crusoe either. (Matravers 1998:69) Matravers goes on to argue that motivational force is a characteristic only of our emotional responses to events or situations with which we are confronted; when we encounter them instead through representations, whether fictional or nonfictional, there is no motivation to act (Matravers 1998:69–73). That we may be motivated by Selkirk’s journal to avoid sea voyages does not alter this fact; the same can be said of reading Defoe’s novel. The issue with which Matravers is concerned is instead whether or not we are motivated by our pity to help Selkirk, which clearly we cannot be. He concludes that unless we are willing to deny that our pity of Selkirk is genuine pity, we should not deny that our pity of Crusoe is equally genuine. I have serious doubts that Walton wants to generalize his claim about the importance of motivational force from defining fear to analyzing other emotions. Some emotions, like admiration, typically do not motivate us to act even when we are confronted with their objects, and most are not connected to action as tightly as fear. If I had met Jim Garrison before his death in 1992, an observer would not question my contempt on the grounds that I failed to spit in Garrison’s face; by contrast, if Sally claims to fear Fido but always behaves as if he were the gentlest creature on earth, we would be right to question the genuineness of her fear. There are no grounds for assuming that Walton would move from the case of fear to the

52 STACIE FRIEND

implausible conclusion that we only ever experience genuine emotions when we are motivated to action. But does not this mean that Matravers is right when he says that connections to action have nothing to do with whether an emotion should count as genuine? No, because the connection to action need not be direct. Remember our friend Nellie, who in watching JFK comes to believe everything that I merely imagine. We can safely say that Nellie believes that Garrison was a noble pursuer of truth and that she genuinely admires him, that she genuinely despises Shaw, and that she is genuinely shocked and disappointed to learn that Shaw was acquitted But she is not motivated to act on any of these emotions: there is nothing she can do about the trial, which was over in 1969; there is nothing she can do about Shaw, who died in 1973; and even in 1991, when Garrison was still alive, she was (let us assume) separated from him geographically. Matravers is correct to say that the lack of motivation does not undermine the claim that Nellie’s admiration, despisement, disappointment, and shock are all genuine emotions. But this does not mean that these emotions have no connection to action whatsoever. To the contrary, because Nellie incorporates the beliefs she forms in response to JFK into her belief stock, and because she incorporates her emotional responses to JFK into her emotional stock, they form part of the foundation for her practical reasoning. Even if Nellie is prevented from acting directly on these emotions because the events depicted in the movie are long over, they still play a rationalizing role with respect to other actions she might take, a role that distinguishes them from the quasi-emotions I experience. For example, if Nellie were passionate enough, she might join the ranks of conspiracy theorists and try to defend Garrison’s reputation. Her decision to do this would be explained by her genuine admiration for Garrison. By contrast, the fact that I admire Garrison in the context of imaginatively responding to JFK would provide no rational explanation of my taking the same action. My contempt, on the other hand, would explain why I write letters to the editor denouncing Nellie’s campaign. It is true that in response to works of fiction we are often motivated to act in ways indirectly connected to the people or situations represented in the fiction. Thought theorists sometimes point out that fictional portrayals of poverty might be just as likely to motivate me to donate to a charity as nonfiction portrayals, and in both cases my response may be very indirect: “my pity for a particular starving person on the news may cause me to send a cheque to a famine relief agency which may not even operate in the relevant part of the world” (Matravers 1998: 70). But I could decide to send a check to a famine agency that would help the particular starving person I pity; I could even decide to fly to the region, find that person, and give her food. The point is not that I probably will not do these things; the point is that if I did, only pity founded on the belief that that person is suffering would provide an explanation of my behavior. Contrast this with my response to a television advertisement in which an actress playing a starving person appeals to the viewer for donations to a famine relief agency. Suppose that

HOW I REALLY FEEL ABOUT JFK 53

I am moved to tears by the portrayal. Even so, it would be irrational, if not incomprehensible, for me to fly out to Los Angeles, find this actress, and give her food, all on the basis of my feeling of pity in response to the advertisement. And a good explanation of this fact is that I do not really pity her: my experience is just quasi-pity, and counts as genuine pity only within the context of my imagining that the actress is a starving person. It makes sense to give priority to those affective states that play the role in our practical reasoning that only belief-involving emotions do—to reserve the classification genuine emotion for them. While these emotions may not motivate us to act in direct response to the situations that cause them, the fact that they form part of an emotional stock rationally correlated with our set of beliefs about the world means that they explain certain actions in a way that imagination-involving affective states cannot. Only those emotions that can be detached from their original contexts and incorporated into our emotional stocks can contribute in this way to practical reasoning. And the importance of this function is reflected in our descriptions of how people really feel. As we have seen, in any case where there is a conflict between an emotional state involving belief and an emotional state involving imagination, the former takes priority: it is how one really feels so long as one continues to have the relevant beliefs, and regardless of what one might be imagining at the moment. Thus, I genuinely feel contempt for Garrison even while watching JFK. By contrast, there is no priority to be given to either of my quasi-emotions toward Odysseus, because neither can be detached from its game of make-believe and assimilated into my emotional stock. I conclude that Walton’s distinction between genuine and quasi-emotions accounts for defining features of our interaction with works of fiction that are not explained by the thought theory. His claim that my response to JFK constitutes genuine admiration for Garrison only within my game of make-believe captures the sense in which this response shares features of real admiration, while also recognizing that it does not play the same role in my cognitive life as the genuine emotion. For each puzzle posed by the assumption that imaginings are sufficient for genuine emotion—how to distinguish the rational from the irrational emotions, how to account for the priority of certain emotions—Walton has a solution that follows naturally from his larger theory of make-believe. At the center of his account of our engagement with fiction is the claim that we do not stand outside the fictional world looking in, learning about distant fictional events and experiencing emotions as a result; rather, we engage in imaginings about ourselves, and only within the context of these imaginings do we respond emotionally. This explains why our quasi-emotions do not conflict with the genuine emotions that rationally motivate many of our actions. Acknowledgments David Hills first drew my attention to such “epistemological attitudes” as shock and surprise in our experiences of fiction. In developing the ideas in this chapter,

54 STACIE FRIEND

I have benefited from numerous discussions with Kendall Walton, as well as informal conversations with several participants at the conference Imagination and the Arts. I have also profited from the comments of Dominic Lopes and Matthew Kieran.

3 NARRATIVE, EMOTION, AND PERSPECTIVE Peter Goldie

Sometimes, in thought and talk, we are engaged with the actual—with events that are actually going on around us. But, reflective creatures as we are, our concerns are often directed toward events that are now past, or events that might come to be in the future, or events that might have been, or events that are fictional. Much of our daily lives is taken up by imaginative engagement with non-actual events in this way: we read fiction, we see films, we daydream, we remember things happening, we hypothesize about how things might have happened, and we make plans for the future. And, in all these cases, we can respond emotionally to what we imagine. My thesis is that there is a centrally important feature that unites and explains such engagement with non-actual events: the engagement is with a narrative, and the engagement is achieved through imaginatively adopting two kinds of perspective that the narrative involves (see Goldie 2000, 2002a, b, forthcoming). Roughly (the details will become clear as I progress), a narrative is something that can be narrated; it need not be narrated though, for a narrative can be just “thought through”, as, for example, when one remembers or imagines a sequence of events. It is more than just a bare chronicle, but a representation of the events, and of the people involved, which is organized, shaped, and colored in a way that gives coherence, meaningfulness, and emotional import to the represented events (see Genette 1980; Ricoeur 1984–; Bal 1997). Narratives can involve internal and external perspectives or points of view. The internal perspective is involved where the narrative represents or otherwise indicates the perspective, including the thoughts, feelings, and emotions, of one or more of the people who are internal to the narrative. Then, external to the narrative, there is the narrator’s own external perspective. (I leave to one side here the difficult issues concerning external narrator, implied author, and actual author; see Booth 1961 and Iseminger 1992. For the purposes of this chapter, the external narrator of fictional works can be taken to be the authorial figure implied by the narrative, although I would not want to be committed to the idea that the interests of a reader of fiction should be restricted to such a notional figure.) The external perspective is distinct from the internal perspective even where, as in autobiographies, the two perspectives are those of one and the same person. And the external perspective is always there, in spite of sometimes seeming evanescent, always shaping and coloring the narrative,

56 NARRATIVE, EMOTION, AND PERSPECTIVE

and indicating the narrator’s own evaluation and emotional response to what happened. A narrative, when it is narrated, can thus reveal internal and external perspectives, and the narrative will thereby invite the reader or audience to share in these perspectives and to respond emotionally to the events as they are represented in the narrative. (From now on I will not mention audiences unless the context requires it. Remarks about the reader should be taken to apply equally to audiences.) If all this is right, then there should be much in common amongst our emotional responses to non-actual events, whether fictional or real life, and to the extent that there are differences, these should be explainable. My main aim in this chapter is to demonstrate this by showing that the idea of imaginative engagement with a narrative, and specifically with the narrator’s perspective, external to the narrative, plays a central explanatory role. I will focus on just three aspects of our emotional responses to non-actual events, each of which needs to be explained. They are as follows. First, the relation between imagination, emotional response, belief, and action needs to be explained. Second, there is the phenomenon, familiar to readers of fiction, of being emotionally “carried along” by what we imaginatively engage with. Third, also familiar to readers of fiction, is the phenomenon of emotional resistance: sometimes we are willing to be emotionally “carried along” only so far, and we reach a point where resistance sets in. All three of these features have been discussed recently in relation to our imaginative engagement with fiction; I will show how they relate more generally to our imaginative engagement with narrative, whether fictional or otherwise, and, by doing this, show the special explanatory importance of the narrator’s external perspective. I An emotional response to actual events in real life—to the here and now— paradigmatically bears a very direct connection to belief and to action. For example, my fear of the snake that is slithering toward me involves (at least typically) recognition of the snake as frightening, and the emotion immediately motivates me to avoid it; or, if I see someone who is near or dear to me being threatened in some way, my feelings of compassion involve recognition of the threat to the other person, and will immediately motivate me to help him or her (see Goldie 2002a). Emotions are thus both world-guided and action-guiding. Things seem different when we respond emotionally to fiction. My fear of the looming monster in the horror movie does not motivate me to run out of the cinema (cf. Walton 1990); or, if I listen to the weekly radio soap opera, The Archers, and hear of Mark Hebden being killed in a car crash, I may get a lump in my throat, but my feelings of compassion do not motivate me to rush to Ambridge, or send flowers, or ring up the BBC to see if I can be of any help. The reason why there is this difference in action-motivation is, it is often said, to be explained by a difference in belief. In real life, if I see a monster approaching me and feel fear, I believe a monster to be actually in front of me, or if I see a friend

PETER GOLDIE 57

in danger and feel compassion, I believe someone to be actually threatened in some way. Whereas, in fiction, I have no such beliefs. No wonder I do not act. This contrast, between our paradigmatic emotional responses to actual events in real life and to fictional events, has led many to argue that our emotional responses to fiction are not real emotions but quasi-emotions, involving make-beliefs or imaginings rather than beliefs, and make-beliefs do not have the same connection to action as do beliefs (cf. Currie 1990, 1997; Walton 1990, 1997). But the trouble with this stark division between two types of feeling response, emotional and quasi-emotional, is that our responses to real life non-actual events would also be consigned to quasi-emotions: we would only have quasi-emotional responses to accounts of things that have happened in the past, or to things that we imagine, remember, daydream, hypothesize, and plan. Here are some examples, in none of which are the relevant beliefs involved. (1) I imagine someone being held down by a kidnapper and having his little finger cut off, and I feel fear. (2) I imagine my future self, old, decrepit, and senile, with saliva trickling down my chin, and I feel disgust and horror at my future state. (3) I think back on the evening’s dinner party, and I feel what Nietzsche nicely calls bonheur de l’escalier—a happiness that I did not experience at the time, but now feel toward what I remember as I make my way home. (4) I read a vivid historical account of how slaves were treated on slave ships in the eighteenth century and I feel deep compassion toward their plight and anger at how they were so mistreated. (5) I remember something very silly that I did some time ago (by thinking through what happened), and I feel embarrassed. If we want to say, as I do, that in all these five cases I am responding with genuine emotions, not quasi-emotions, then we should allow that genuine emotions need not involve beliefs in the relevant sense, but can also involve imaginings, rememberings, and so on. (The line of argument I am running here is closely related to Richard Moran’s constraint that any solution to the so-called paradox of fiction should also deal with other types of emotional response to the non-actual (Moran 1994). My approach, as Greg Currie notes, may be substantially the same as the appeal to quasi-emotions (Currie 1990:211–13; 1997:70). However, words seem to me to be important here.) So we can now say that we can have genuine emotions directed toward the non-actual, whether fictional or real life. Then one might go on to try, as Greg Currie has done, to unify our emotional responses to non-actual events, both fictional and real life, by appeal to just one kind of imagining, namely simulation (Currie 1990, 1997). Simulation, according to Currie, is (and these terms are equivalent according to his terminology) empathizing with someone, or putting oneself into someone’s shoes, or taking the role of someone (where the someone can be me at some non-actual time or place). Simulation in this sense is what I will also call, following Richard Wollheim, central imagining, or imagining “from the inside” (Wollheim 1984). (I introduce this further term for the sake of a contrast with acentral imagining, as will become clear in a moment.) If I simulate, or empathize with (or centrally imagine) Albert, Currie says, “while Albert

58 NARRATIVE, EMOTION, AND PERSPECTIVE

believes p and desires q, I, in empathizing with him, come to believeI p and desireI q, where ‘believingI’ and ‘desiringI’ denote states that bear systematic resemblances to believing and desiring…. [T]hey are like real beliefs and desires in respect of content…and they are like real beliefs and desires in terms of internal causal role, but unlike them in terms of external causal role…. I-states are blocked off from behavior. They are, as people sometimes say, ‘off-line’” (Currie 1997: 67–8). Thus, according to Currie, if Albert related to me an account of how his finger was cut off by a kidnapper, I would simulate Albert, believingI that my finger is about to be cut off and so forth, and as a result I might feel fear. Here, the emotion that I experience matches that of Albert during his experience: fear. But what if I did not feel fear, but rather compassion or pity toward Albert as he tells me his story? If I did, this would involve—and this is the disagreement with Currie—a different kind of imaginative engagement with non-actual events: not simulation or central imagining from Albert’s perspective, internal to the imagined scene, but acentral imagining—that is, perceptually imagining from no point of view within the imagined scene (see Wollheim 1984; Goldie 2000, 2002a). Thus, in this example, acentrally imagining Albert having his finger cut off, I feel compassion or pity, and not fear; I do not centrally imagine the narrative unfold from Albert’s perspective internal to the narrative, but rather I imagine it from an external perspective, responding to the narrative in the way that the narrator, Albert, encourages me to do. Thus we now seem to have two distinct kinds of imagining of the non-actual. In many other cases, as can be seen by looking back at the five examples that I have just given, we can acentrally imagine things from a perspective external to what is narrated. In (1), I could have acentrally imagined the victim, and felt compassion or pity, as I did for Albert. In (2), when I imagine being old, decrepit, and senile, I do not simulate or centrally imagine my future self, for if I did I would not feel disgust and horror (for my mental state then is beyond such self-reflection at how I have degenerated); rather, I (that is, me now) imagine my future self acentrally, from a perspective external to the imagined scene, and the disgust and horror that I now feel at imagining myself in that state bear no “systematic resemblance” to my mental state at the imagined time, as would be the case if I were simulating myself. One can put the point like this: the disgust and horror that I now feel at myself, old, decrepit, and senile, is a response to what I imagine; it is not part of the content of what I imagine. In (3), casting my mind back to the dinner party earlier that evening, I acentrally remember myself and the other guests from a point of view, say, behind the chair in which I am seated, so that I (that is, me then) am as much part of the content of what is imagined as are the other guests, and, surveying the imagined scene, I (that is, me now) experience happiness in a way that I did not at the time. Thus I have a perceptual memory of what happened, and an emotional response to it, without centrally imagining myself or anyone else in the imagined scene. Perceptually remembering past events is often like this. Similarly, in (4), the compassion and pity that I feel toward the slaves does not involve centrally imagining being one of them, for

PETER GOLDIE 59

compassion and pity are emotions directed toward the suffering of another. The acentral imagining in this case does not exclude the possibility that I might also, as part of the same imaginative project, simulate, or centrally imagine being a slave, and this might, so to speak, enliven my thoughts about them, and thus enliven my compassion. (If we are fairly similar in personality to the protagonist, and if the situation is one that we can readily imagine being in, then we may well be more likely to centrally imagine the protagonist.) However, it is not, I think, coherent to centrally imagine and acentrally imagine at one and the same time, although we do have an ability to flit from one to the other with great facility. In (5), when I remember the silly thing I did, again I could both centrally and acentrally imagine the scene—the embarrassment could be both personal embarrassment and what is called witness embarrassment, the latter arising through my (that is, me now) acentrally imagining my earlier self. So I do not think we can unify our emotional responses to non-actual events by appeal to just one kind of imaginative engagement, namely simulation or central imagining. Rather, when we respond emotionally to a narrative, sometimes we imaginatively engage with the narrative by simulation or centrally imagining one or more of the people internal to the narrative, and sometimes by acentrally imagining, sharing the perspective of the external narrator. Consequently, imaginative engagement with the non-actual does not require make-beliefs or I-beliefs of the sort introduced by Currie: for example, my compassionate response to the slaves involves beliefs about their plight, not Ibeliefs, and these beliefs are part of what is involved in my feeling of compassion at their plight. Indeed, the whole idea of our engagement with the non-actual being in some sense “off-line” begins to seem unmotivated. For our minds are directed toward the non-actual in all sorts of different ways that are not properly unified by the notion of I-beliefs and of thinking off-line. This is demonstrated by the fact that, when we think about, imagine, remember, daydream, hypothesize, or plan, and respond emotionally to it, our emotional responses do often have a very clear connection with action (although obviously the connection is not as direct as it is in respect of real life that is actual, such as my fearfully avoiding the snake). This is in part why it is important to remember that there is a range of cases within the non-actual. In some cases, our emotions will be more directly action-guiding: for example, at a meeting I might avoid asking a question having felt embarrassed at the thought of how I might appear stupid and be laughed at. In other cases, our emotions will be more obliquely action-guiding: for example, my compassionate emotional response to the vivid account of life and death on the slave ships might motivate me to give money to Christian Aid, and perhaps to feel and behave with more tolerance toward black African Americans. Then, with this range of cases in mind, when we turn to non-actual fictional examples, we can see that in many respects there is really not so much of a contrast between what is non-actual and real life and what is non-actual and fictional. Sometimes engagement with fiction will involve centrally imagining or simulating one or more of the characters; but, as in real life non-actual, it need

60 NARRATIVE, EMOTION, AND PERSPECTIVE

not. For example, the compassion that I feel on hearing of the fate of Mark Hebden in The Archers arises through acentrally imagining what is being narrated and responding emotionally to it in the way that the narrative encourages. And sometimes the emotional response to fiction can be connected to action in very much the same sort of way as with real life non-actual—for example, on hearing of Mark Hebden’s death, I might come to feel and behave with more tolerance toward farmers than I would otherwise have done, just as I might give money to Christian Aid on hearing of the treatment of the slaves. (This is not to succumb to what Currie 1990 calls the transfer strategy as a way of dealing with the paradox of fiction; this is the idea that we respond to the fate of fictional characters as surrogate objects standing in for real people.) One might object here that I have lost the distinction, within non-actual narratives, between what is fictional and “the rest”. But this is wrong. The essential distinction remains that what makes a narrative fictional, and thus sets it off against all other sorts of narrative, is its being narrated, and read, as part of a special sort of practice, which invites the reader to imagine or make believe that what is being narrated actually happened, even when it is known that it did not. This distinction remains entirely unaffected by what I have said (see Currie 1990; Lamarque and Olsen 1994; and Goldie forthcoming). (There are, however, particular aspects to the creation and understanding of fictional narratives that give imagination a more central role than in narratives of factual matters; it is not merely contingent that fictional narratives are mostly imaginary (Savile 1998).) So far then I have tried to show that there are two kinds of imagining involved in our engagement with narratives, and that acentrally imagining from the narrator’s external perspective has a special role in explaining the place of belief in our emotional responses to narrative, and in explaining the connection of emotion to action. II The second feature of our emotional responses to narrative that I want to discuss and explain is as follows: when we respond emotionally to the fate of characters in fictional narratives, we allow ourselves to be “carried along” emotionally by the story, and we seem to do this more than we would if the narrative were not a work of fiction. There are various manifestations of the phenomenon in our responses to fictional narratives, as some examples will show. Let me begin by returning to the person who gets extremely involved emotionally with The Archers. Let us now call him Freddie. Freddie feels deep compassion toward the friends of Mark Hebden when they hear of his death. And yet, when Freddie is told that his cousin’s friend has been killed in a car crash, he feels little emotion toward his cousin in his tragic loss, perhaps thinking that a bit too much fuss is being made about the whole thing; after all, we are all going to die sooner or later. This example illustrates a difference merely of degree of compassion. But then there is the person who feels compassion toward the fate of a fictional

PETER GOLDIE 61

character, whereas, if he heard about a similar fate in real life, he would not just feel less compassion, but none at all—indifference—or even another sort of emotion entirely—contempt, perhaps. Matthew Kieran has an apt example of this, of “someone who once was a Roman Catholic but is now a confirmed atheist reading of what happened to Charles Ryder in Evelyn Waugh’s Brideshead Revisited. Given her secular conversion she firmly believes that morality categorically cannot depend upon the commands issued by God. Nonetheless, in reading Brideshead Revisited she responds with sympathy, admiration, awe and ultimately affirmation to the culmination of the novel” (Kieran, Chapter 4, this volume), whereas in real life she would do no such thing, perhaps considering Ryder to be a fool. Greg Currie also gives an example of this phenomenon, which he calls the problem of personality: his own response as reader to the fate of Jago, a character in C.P.Snow’s The Masters. In the fiction, Jago wants the Mastership of a Cambridge college. Currie says that Jago is the sort of person that he would have little time for in real life, and someone that he would never lend his support to in trying to attain a Mastership: an intellectually mediocre character, sometimes rather pathetic. Yet, Currie says, “nonetheless, I wanted Jago to win, and it was with growing dismay that I watched the decline of his fortunes” (Currie 1997:64). My explanation of what is going on in these examples appeals to the narrator’s external perspective. This shapes and colors the narrative, indicating the narrator’s evaluation of the narrated events, and at the same time, the narrator’s emotional response to them. (The idea of a narrator “indicating” an evaluation and an emotional response is intended to be vague: it should include, e.g. the possibility of the narrator showing his evaluation and feelings, without specifically stating them; and it should include the possibility of the narrator unintentionally showing his evaluation and feelings.) The narrator’s external perspective thus encourages the reader, through acentral imagining, to share this perspective on the events that are narrated, and to have the same sort of emotional response. Now, there are two ways in which a reader’s emotional response to a narrative can be appropriate. First, it can be appropriate in the sense of being concordant with the narrator’s external response, as evidenced principally in the text of the narrative. This is the response that the narrative invites the reader to share in—the response that the work, so to speak, calls for. I will call this narrative appropriateness. And second, the reader’s response can be appropriate in the sense of being taken by the reader to be, in fact, appropriate to the events that are narrated. I will call this ethical appropriateness (at this point I do not want too much to hang on the word “ethical”). And these two ways of being appropriate can diverge in particular cases, as they did in the examples given by Kieran and Currie. In both these examples, the narrative is inviting the reader to share in the narrator’s external response to Ryder and to Jago, endorsing the values that Ryder and Jago endorse. Thus, the narratively appropriate reader response to Brideshead and to The Masters is one that is concordant with that of the narrator, sharing the narrator’s perspective. Kieran’s hypothesized reader of Brideshead, and Currie, as reader of

62 NARRATIVE, EMOTION, AND PERSPECTIVE

The Masters, both have these narratively appropriate responses. But clearly Kieran and Currie consider their responses to be ethically inappropriate, being grounded in values that they do not share: Kieran’s reader does not now share Waugh’s Catholicism; and Currie does not share Snow’s “outlook” (see Currie 1997:72– 3). Thus, in each case, the emotional response that was narratively appropriate was also, according to the values of Kieran and Currie, ethically inappropriate. Their responses were thus discordant. Currie puts forward his own explanation of what he calls the problem of personality. It is different from mine. His is based exclusively on the notion of simulating, empathizing with, or taking on the role of someone. It goes as follows. When Currie reads The Masters, he simulates a hypothetical reader of fact whose outlook is like that of the implied author and who is reading about the fate of a Cambridge academic called Jago, and this hypothetical reader of fact is, in turn, simulating Jago; and because of what Currie claims to be the transitivity of simulation, this is equivalent to Currie simulating Jago (Currie 1997:69). The upshot is that Currie finds himself sharing the outlook of the implied author. This seems to me to have a number of difficulties. First, it is too complex an account of what is actually, and phenomenologically, much more simple and direct. Second (although this would obviously not be considered an objection by Currie), it does not give a place for the reader to take a direct external perspective through acentral imagining. And third, his account depends on simulation being transitive (and transparent), and this is doubtful as it depends in part on people being alike psychologically. The examples I have given so far are all from our responses to fictional narratives. But the phenomenon also arises, albeit to a lesser degree, in our experiences of real-life narratives. To show that it does arise, let me give an example from my own experience of reading a work of history, where I was drawn into and carried along by the narrator’s own values and emotional responses, in spite of not sharing these values. Antonia Fraser’s recent history of Marie Antoinette is unashamedly sympathetic to her and to her plight (Fraser 2001). We are, for example, encouraged to understand Marie Antoinette’s rather clumsy approach to the political situation in France, and her delights in the pleasures of the court, as being the products of her upbringing, rather than being typical of what Nancy Mitford called a woman of “monumental stupidity,” who was “frivolous without being funny” (cited in Fraser 2001:422). And I found myself to a considerable extent, just as in Kieran’s and Currie’s examples, having the narratively appropriate response to what was related, in effect taking Fraser’s (and thus Marie Antoinette’s) side much more than I would have done if I had been reading some other less favorable account of her life, such as one by Nancy Mitford. So we have a story of real life taking just the same form as those earlier fictional ones (cf. Gaut 1998; Matravers 1998). The phenomenon is also quite familiar in reading diaries and autobiographies, as well as in hearing the shorter stories that people tell about themselves in the pub or over the dinner table: one finds oneself at the time taking the narrator’s side more than one would on

PETER GOLDIE 63

reflection, and more than one would if what happened had been related from some other less sympathetic perspective. So this discordance between the two sorts of external response of a reader (the narratively appropriate response and the ethically appropriate response) can be found in factual narratives just as it can be found in fiction. And now it is possible to see at least two reasons why we find this phenomenon to a lesser degree with factual narratives. The first reason is simply that the sustained coloring and shaping of a story takes some considerable time and skill if it is to achieve the effect that concerns us here—if we are to allow ourselves to be carried along emotionally by the narrator’s values as revealed in his external perspective in spite of our discordant ethical responses. It is precisely the expressive qualities of the successful fictional work and the artifice involved that achieve these effects; it is not as though the artifice gets in the way (cf. Moran 1994:85–6). And these expressive qualities are found more often in fictional works than in factual narratives (although I would maintain that they are far from uncommon in wellwritten histories, biographies, and autobiographies, such as, to take some recent examples, Orlando Figes’s The People s Tragedy, Antony Beevor’s Stalingrad, and Ray Monk’s two-volume biography of Bertrand Russell). The second reason why we more readily allow ourselves to be emotionally carried along by the fate of characters in fiction is that we know them not to be real. When we are engaged with a work of fiction, as Freddie was with The Archers, we know that we can “switch off” more or less at will. When we make believe that what is being narrated actually happened, and engage emotionally with the narrative, sharing in the narrator’s external perspective, we do not lose sight of the thought that this is, after all, fiction. Like Freddie, we can, at least typically, allow ourselves to a considerable extent to indulge our profound feelings for humankind, and let our sentiments run away with us without concerns about their connection to action. Again, though, the phenomenon is not restricted to our engagement with fictional narratives, and it will vary from case to case across all sorts of narrative—we might equally feel we can safely indulge our sentiments when we read an account of the plight of slaves in the eighteenth century. This may not be a particularly attractive feature of the human personality, but it is one we should face up to. I suspect that this phenomenon, of allowing ourselves to indulge our feelings in our responses to fiction, when they are typically less directly connected to action, also sometimes applies to our less savory attitudes. For example, we may feel “safe”, up to a point, in allowing ourselves to indulge our hidden sadistic side when we read de Sade’s Juliette: we allow ourselves to have the narratively appropriate response—the one that goes along with de Sade—whilst holding on to the thought that this response is itself ethically inappropriate. But, in our experience of fictional works there is sometimes a resistance to this sort of emotional engagement, and it is to this third aspect of our emotional response to narratives that I now turn. It will emerge that emotional resistance is not unique to fictional works; the phenomenon is found too in our

64 NARRATIVE, EMOTION, AND PERSPECTIVE

responses to narrative of real life non-actual events. So hopefully this discussion will further bolster my central thesis about narrative, and about the role of the narrator’s external perspective. III Hume, with a specific sort of example in mind, describes the phenomenon of emotional resistance as follows: Where any innocent peculiarities of manners are represented…they ought certainly to be admitted; and a man, who is shocked with them, gives an evident proof of false delicacy and refinement…. But where the ideas of morality and decency alter from one age to another, and where vicious manners are described, without being marked with the proper characters of blame and disapprobation; this must be allowed to disfigure the poem, and to be a real deformity. I cannot, nor is it proper I should, enter into such sentiments; and however I may excuse the poet, on account of the manners of his age, I never can relish the composition…. We are displeased to find the limits of vice and virtue so much confounded: And whatever indulgence we may give to the writer on account of his prejudices, we cannot prevail on ourselves to enter into his sentiments, or bear an affection to characters, which we plainly discover to be blameable. The case is not the same with moral principles, as with speculative opinions of any kind…. Whatever speculative errors may be found in the polite writings of any age or country, they detract but little from the value of these compositions. There needs but a certain turn of thought or imagination to make us enter into all the opinions, which then prevailed, and relish the sentiments or conclusions derived from them. But a very violent effort is required to change our judgements of manners, and excite sentiments of approbation or blame, love or hatred, different from those to which the mind from long custom has been familiarized. And where a man is confident of the rectitude of that moral standard, by which he judges, he is justly jealous if it, and will not pervert the sentiments of his heart for a moment, in complaisance to any writer whatsoever. (Hume 1985) As Hume points out in this passage, our imagination happily goes along with “speculative errors”; moreover, we are happy to imagine magical and miraculous goings-on: for example, the time machine that works in Back to the Future, and the miraculously cured disfigurement on the face of the priest in Graham Greene’s The End of the Affair. And we may also go along up to a point with ethically inappropriate responses, so long as they do not go deep. (“Innocent peculiarities of manners” are all right.) But where the narrator’s perspective invites us to share in values and responses that we consider to be ethically inappropriate in a significant

PETER GOLDIE 65

way, we seem to reach a point where resistance sets in. There seems to be something special about morality here (see Moran 1994; Walton 1994a). In Hume’s view, we ought to consider the fictional work to be aesthetically less praiseworthy in virtue of this ethical deficiency. So, taking the hint from Hume, I will discuss this phenomenon of emotional resistance in the context of ethicism; this is, in the words of Berys Gaut, the view that “ethical assessment of attitudes manifested by works of art is a legitimate aspect of the aesthetic evaluation of those works” (Gaut 1998:182; cf. Carroll 1998b). Consider a fictional work—a novel—about a camp guard at Auschwitz who is assiduous at his tasks, and who takes great pride in his achievements, and in the achievements of those of his fellow guards who share in his values; he considers any signs of compassion for the inmates to be weak and contemptible. Yet let us assume that it is also made clear through the narrator’s external perspective that this man’s thoughts, feelings, and emotions were terribly, horribly wrong. What should ethicism have to say about this novel? In Gaut’s words, continuing the sentence quoted earlier, “if a work manifests ethically reprehensible attitudes, it is to that extent aesthetically defective, and if a work manifests ethically commendable attitudes, it is to that extent aesthetically meritorious” (Gaut 1998: 182). If this were all there were to ethicisrn, then this work would be aesthetically defective as it manifests the camp guard’s ethically reprehensible attitudes at the internal level. However, as Gaut makes perfectly clear, what matters here are the ethical attitudes that are revealed at the external level, and not those at the internal level. He says, “it is important to distinguish between the evil or insensitive characters represented by a work and the attitude the work displays toward those characters. Only the latter is relevant to the ethicist thesis” (Gaut 1998:188). So the answer to the question, according to Gaut, is that this novel about the camp guard will be aesthetically meritorious, and this is so only because of the attitudes revealed at the external level. If we are engaged fully with a fictional narrative, “carried along” by the narrator’s perspective, we do not just make-believe that the ethical values that the work invites us to share are correct; we really enter into the narrator’s values, as we are invited to do, and become really committed to the thought that what is narrated about their characters and their actions is, say, blameworthy, or praiseworthy. Having such narratively appropriate responses is part of what is involved in being a “good” reader. But when we are antecedently committed to different and opposing ethical values, we can begin to resist having the narratively appropriate responses, as they are discordant with what we know to be ethically appropriate. As Hume says in the passage cited above, “we cannot prevail on ourselves to enter into his [the narrator’s] sentiments, or bear an affection to characters, which we plainly discover to be blameable.” It is now easy to see how emotional resistance can also arise with real-life narratives. If I told you a story of my boorish behavior last week in a way that invited you to find it amusing and harmless, you would resist entering into my sentiments, as you would if, at a party, I suggested that you should plan to

66 NARRATIVE, EMOTION, AND PERSPECTIVE

drive home, blind drunk as you are, and that you should feel proud of taking on such a challenge whilst drunk. As in the fictional examples, the emotional response that I, the narrator, invite you to share in—the narratively appropriate response— is discordant with what you consider to be ethically appropriate, and resistance sets in: you will not be carried along by my view on these matters. There are, however, some difficulties and complications with ethicism. First, it ought to be relevant to the ethicist thesis just how well the internal perspective is presented. Consider two different stories about the camp guard, both of which display, at the external level, ethically appropriate attitudes toward the narrative. One reveals the guard’s thoughts, feelings, and emotions in such a sensitive way that the reader is able to enter into his attitudes, and imagine feeling pride in such work, and contempt for colleagues who feel compassion for the inmates. The other story is told in a way that not only leaves the reader emotionally cold, but also leaves him mystified as to how anyone could behave in such a way. Would we not want to say that the first of these stories is an aesthetically better story, and that it is aesthetically better in part because of the tension that is set up between the attitudes revealed at the internal and the external level? In reading the first story, the reader enters into the mind of the camp guard, and sees just how someone like that (someone unnervingly like him, the reader) could have such feelings, and thereby becomes acquainted with badness in a way that the first story fails to achieve. And it is a better story for that (cf. Devereaux 1998). The same sort of point can be made about real-life narratives. When we read Christopher Browning’s Ordinary Men: Reserve Police Battalion 101 and the Final Solution in Poland, we are told what happened with great sensitivity to the motivations and emotional responses of the perpetrators of the terrible slaughter of innocent civilians in Poland, and yet it is entirely clear that the narrator’s external attitudes are appropriate. Second, we need to consider narratives that reveal ethically inappropriate attitudes at the internal level but that are dispassionate at the external level. An immediate response to this is that dispassionateness is an inappropriate attitude here. And this is right: it is a profound mistake to think that dispassionateness is a merit in all cases. If, as Aristotle says, one should have the right feelings in the right way and at the right time, then there are occasions where one should have strong feelings; it is wrong to be dispassionate. I presume that Gaut would agree with this. But here is the complication. What of those occasions where dispassionateness (or at least trying to be dispassionate) is taken to be a merit: for example, in scientific enquiry, or in journalistic reporting? Perhaps we need to ask a further question: Why is this story being told at all? Stanley Milgram’s account of his obedience experiments (Milgram 1963, 1974), much of which is narrative in form, is praiseworthy for its dispassionateness in the interests of science (although Milgram’s own perspective is never entirely obscured from view). But surely we should take a different view of a dispassionate court report of a rape trial in The Times where the rape is related in full detail, or of a detailed but dispassionate report in The National Enquirer of a judicial execution by lethal

PETER GOLDIE 67

injection. In such cases, surely, the very publication of the narrative is ethically inappropriate: regardless of its dispassionateness, it is published with the unexpressed intention of engendering in its readers a lurid interest, and an ethically inappropriate emotional response; indeed, the narrative could surely still be ethically inappropriate even if the external response had the appearance from the text of being ethically appropriate; reports in The Sun seem to me to be often like this: “see pages 2, 3, 4, 5, and 6 for more on this unspeakably [sic] outrageous assault on an innocent young girl”. Some stories are better not told; others are better not told in such detail. The third point goes deeper, and I can only touch on it here. According to ethicism, as formulated by Gaut, if a narrative reveals an ethically inappropriate attitude at the external level, then this necessarily reduces its aesthetic value. Perhaps this is too strong. Perhaps, as Matthew Kieran puts it, “there can be and indeed are works whose value as art is enhanced in virtue of rather than despite their morally defective character” (Kieran, forthcoming). No doubt sometimes the ethical values revealed in a work can be such that emotional resistance sets in, so that we resist having the narratively appropriate emotional response, and, because this resistance can adversely affect our ability to appreciate the work qua work of art, its aesthetic value is reduced. But perhaps there are other works of art—call them dangerous— that impress on us a profound aesthetic appreciation of the appeal of what is ethically bad in a way that could not be fully achieved if the narrator’s external perspective were ethically appropriate—if it were, in effect, moralizing. Such dangerous works—dangerous because of their potential effect in the wrong hands—may have an aesthetic value that is, as Kieran has argued, enhanced in virtue of its ethically inappropriate values at the external level: only because of this can we come thereby to be truly acquainted with the beauty of badness. As Tolstoy said in the “Kreutzer Sonata,” it is amazing how complete is the delusion that beauty is goodness, and, one can now add, that badness is ugly. Examples are difficult here. Let me suggest two, where the main characters are morally bad, but the narrator’s perspective does not encourage the condemnation that it ethically should. In A Bout de Souffle, Jean-Luc Godard’s film, we are persuaded to see Michel, the character played with such style by Jean-Paul Belmondo, as really very likeable and charming; but here is someone who steals from his girlfriend, who kills policemen, and is generally bad. And in Mozart’s and da Ponte’s Don Giovanni, we are persuaded to see Don Giovanni in at least an ambiguous light: again, we come to see much that is likeable—even admirable— about someone who has on his record 1,003 women in Spain alone. In both, I suggest, the work would be aesthetically of less worth if it moralized about its main characters. Leaving such possibilities to one side, I need now to address the following question: at what point, in a particular case, does one cease to be carried along by the narrative, coming to a point at which emotional resistance sets in? Obviously, such a question does not have a simple answer, and, as I have already emphasized, there is no systematic divide between fictional and nonfictional narratives. All

68 NARRATIVE, EMOTION, AND PERSPECTIVE

sorts of factors will be relevant in a particular case—for example, how imaginative one is; how emotionally “malleable” one is; what the subject matter is; the artifice of the narrative. But I would like to focus on just two factors: first, what ethical values one has; and second—an issue I have already discussed in part— how close and direct the connection is between emotional response and action. It reveals something about my values that I have, perhaps, less resistance toward a narrative that invites me to think of boorish behavior as meriting amusement than I have toward being invited to find wanton physical cruelty amusing; and perhaps being proud of driving whilst drunk is somewhere in between these. But emotional resistance can reveal more than this about the reader, for sometimes we might reach a point at which we ought to resist, given our considered ethical values, but we do not resist: the narrative has a terrible fascination and resonance for us. As I suggested earlier, we might find out to our surprise that we are carried along a lot further with de Sade’s Juliette than we thought we would be, and further than we think we ought to be, given our considered views about sadism. We begin to wonder whether we are after all capable of doing things like that, and whether it is not just a matter of luck that we have not been put in circumstances where the opportunity arises (cf. Devereaux 1998). In some such cases, the discordance and tension between the narratively appropriate response and the ethically appropriate response can become almost unbearable: one tries to stop reading or to turn away one’s eyes—not out of disgust or horror at the content of the narrative, but rather because one knows one is being drawn in, like a moth to the flame. The second factor that can affect emotional resistance that I want to discuss is how close and direct the connection is between emotional response and action, which, I have said, will vary from case to case, but typically, across comparable cases, will be less direct where the narrative is fictional. For example, Currie, in reading C.P.Snow’s novel, was willing to be carried along by the narrator’s values, and to respond accordingly; he knew that sympathy with the fate of Jago was ethically inappropriate, as Jago was intellectually mediocre in Currie’s view, but this did not seem to constrain his actually caring about Jago, and wanting him to succeed. But, as I suggested earlier, presumably Currie would have shown more emotional resistance if he was told that a real-life Jago-character was trying to get a chair in his own university department, or in some other university in which he has valued friends. And this would be so even if he were told this by someone who was an avid and eloquent supporter of “Jago”. What is relevant is only that Currie thought him mediocre, and that real-life Jago’s appointment would impinge on Currie’s future intellectual life. Freddie’s case is rather different. Freddie, like Currie, is more willing to be carried along emotionally by what is fictional (feeling deeply for the friends of Mark Hebden in The Archers) than he is by the real-life story (feeling little or nothing for his cousin whose friend has died). And Freddie, like Currie, is responding differently in the two cases because of the different connections between emotional response and action (as well as, perhaps, the extraordinary artifice of

PETER GOLDIE 69

the British Broadcasting Corporation). But Freddie, unlike Currie, has revealed something ethically unattractive about himself: he fails to be carried along by the real-life narrative not because he is invited to respond emotionally in a way that he considers to be ethically inappropriate—in a way that goes against his ethical values; rather, he fails to be carried along because, selfishly, he does not want to be drawn into action. One might put it like this: merely thinking that his cousin needs help puts less psychological pressure on Freddie to help than if he were genuinely emotionally engaged. The thing about Freddie is not only that he does not really care about his cousin; he does not really care about Mark Hebden’s friends either—he only pretends to care, indulging himself in an overly sentimental way, because he knows that they are not real-life characters, so there is no risk of being drawn into action. But remember the point that the connection to action is not necessarily less direct where fiction is concerned than it is in comparable real-life narratives: fictional and real-life narratives could have the same sort of content and the same sort of connection to action. For example, we might resist entering very far into the emotional responses encouraged by Leni Riefenstahl in The Triumph of the Will, her documentary of the 1934 Nuremberg rally; but I dare say we would be just about equally resistant to a fictional movie, set roughly in the present day, involving the same sort of people as Hitler and his followers, and inviting the same sort of response as Riefenstahl did to Hitler. In both these cases, there is clearly a direct connection between emotional response and action, for the characters portrayed could so easily have their counterparts in present day real life. In contrast, we might not resist being carried along by a narrative that sought to get us to see that there was something glorious about Aztec human sacrifice, in spite of our considering human sacrifice to be terrible, and this would be equally the case whether the narrative was a work of fiction set in those times or a historical account of what actually happened. In both these non-actual narratives, fictional and real life, the connection to action is, one hopes, remote: a real-life world, with me in it, containing racist National Socialists is a much closer one than a world containing organized rituals of human sacrifice. Thus, by considering three aspects of our emotional responses to narratives of non-actual events—the connection to belief and to action, our willingness to be emotionally carried along by the narrative, and our emotional resistance—I hope to have bolstered my thesis, with which I began this chapter, that the notion of narrative, and the narrator’s external perspective that is involved in narrative, have a centrally important explanatory role. Acknowledgments A number of people have helped me with earlier versions of this chapter. I would especially like to thank Matthew Kieran and David Papineau for their comments, and, as editors, Matthew (again) and Dom Lopes for all their help. My debt in this chapter to the writing of Greg Currie and Richard Moran will be evident.

4 IN SEARCH OF A NARRATIVE Matthew Kieran

What’s involved in the appreciation of narrative artworks? How do we come to understand characters, what they are thinking, feeling, what their underlying attitudes are, what the nature of their character is? In virtue of what do we come to care about them? In trying to address these questions I will go some way toward articulating a conception of narrative understanding. Part of the thrust of the argument is negative since it aims to show that imagination, construed in terms of simulation, cannot be as central as some recent authors have suggested. Through showing how and why this is so, we will arrive at a more complex grasp of the nature of narrative appreciation. But I will also go on to give a more positive characterization of the ways in which we arrive at a deep and full understanding of fictional characters in artistic narratives. In seeking to distinguish and clarify some of the ways we do so I will discuss notions of narrative understanding, character, empathy, sympathy, imaginativeness and expressive qualities. Expressivity In relation to narrative appreciation it has been claimed that simulation, which involves taking on pretended or simulated beliefs, is fundamental (Currie 1995a; Feagin 1996; Oatley and Gholamain 1997). The claims made for simulation theory are most clearly articulated by Gregory Currie who holds that simulation is central to both working out what is fictionally the case, primarily with respect to a character’s experience, and with regard to how and why we care about and affectively respond to fictional characters (Currie 1995a:153–4). Of course there are many cases where we do not use simulation, since we are told what state a character is in or the type of character, and we can understand our responses to those passages without taking into account simulation (Carroll 1998a). Furthermore it seems unproblematic to grant that some degree of narrative understanding is required for simulating a character or the sort of beliefs and feelings a character is represented as having. So I take the simulation thesis really to be the more nuanced claim that:

IN SEARCH OF A NARRATIVE 71

(1) When I want to really understand the nature of a character’s experience and their attitude toward their own experience (what their character is really like), then I need to simulate. A deep understanding of fictional characters requires simulation, though a shallow understanding of them need not. (2) In order to capture the full nature of our affective responses to a narrative, we must understand the simulation process that we go through as readers— because that simulation process is central to our acquiring an understanding of characters. Now we are often told much about a character or they are represented as being certain types of people. With respect to situations, characters and allusions we have a stock of schemas, often based on certain prototypes, which enables us to recognize certain states, characteristics and make assumptions about various interrelations therein (Carroll 1998a). There is a fair amount of psychological evidence to support the claim that we standardly employ prototypes, schemas and general categorization in making sense of the states people are in and attributing certain characteristics to them (Farr and Moscovici 1984; Fiske and Taylor 1984). Thus rhetorical devices, drawing the readers attention to particular features of the character, their behavior and the use of both direct and indirect discourse enables us to identify characters as being certain types of people. The representation of actions, facial expressions, gestures, body language and tone of voice characteristic of certain states are all used in ways that, given what we discern about the context, give rise to a recognition of the state the fictional character is in. Importantly such characterization is often used to help the reader identify more fundamental features of a fictional person’s character. For example, when the reader first encounters the Steele sisters in Austen’s Sense and Sensibility they are represented as paying an overly felicitous attention to Lady Middleton’s spoilt children. Their endurance and excessive affection are clearly marks of fundamental features of their character: prudential self-interest allied with a craven willingness to be sycophantic. Furthermore, the imagery and allusions deployed by an author may intimate to the reader what type of person they should recognize the fictional character to be. In Anna Karenina Vronsky is characterized as a splendid animal, much given to affected physicality and the self-admiration of his physical features. Thus we are already given a fair idea of the kind of person Vronsky is, the kind of charm he possesses, the ways in which he loves athletic games, his regiment and St Petersburg society. For, amongst other things, he is the kind of narcissistic person who loves things, activities and people by virtue of the way they reflect his own self-adoration. Even if we assume claims (1) and (2) are perfectly true there is a question to be asked about how important they are. Most fiction bestsellers can be read without having to simulate much because the authors often do not set up situations or characters in such a way that simulation would be profitable or interesting enough. Simulation hardly seems to offer much in relation to works such as Harry Potter or the works of Jeffrey Archer. So even if the forementioned claims are right

72 MATTHEW KIERAN

and simulation is required for a deep understanding of characters in say the novels of Henry James, for most narrative reading simulation is probably quite worthless. Furthermore, to see why these claims cannot be right consider the opening to Dickens’ Hard Times: “Now, what I want is, Facts. Teach these boys and girls nothing but Facts. Facts alone are wanted in life. Plant nothing else, and root out everything else. You can only form the minds of reasoning animals upon Facts: nothing else will ever be of any service to them. This is the principle on which I bring up my own children, and this is the principle on which I bring up these children. Stick to Facts, Sir!” The scene was a plain, bare monotonous vault of a schoolroom, and the speaker’s square forefinger emphasized his observations by underscoring every sentence with a line on the schoolmaster’s sleeve. The emphasis was helped by the speaker’s square wall of a forehead, which had his eyebrows for its base, while his eyes found commodious cellarage in two dark caves, overshadowed by the wall. The emphasis was helped by the speaker’s mouth; which was wide, thin, and hard set. The emphasis was helped by the speaker’s voice, which was inflexible, dry and dictatorial. The emphasis was helped by the speaker’s hair, which bristled on the skirts of his bald head, a plantation of firs to keep the wind from its shining surface, all covered with knobs, like the crust of a plum pie, as if the head had scarcely warehouse-room for the hard facts stored inside. The speaker’s obstinate carriage, square coat, square legs, square shoulders—nay, his very neckcloth, trained to take him by the throat with an unaccommodating grasp, like a stubborn fact, as it was— all helped the emphasis. “In this life, we want nothing but Facts, Sir; nothing but Facts!” The speaker, and the schoolmaster, and the third grown person present, all backed a little, and swept with their eyes the inclined plane of little vessels then and there arranged in order, ready to have imperial gallons of facts poured into them until they were full to the brim. Our understanding of Gradgrind, arrived at via the prescribed imaginings, is not a function of imagining myself to be in the schoolmaster’s position, imagining myself to be him in what I imagine or indeed imagining myself to be anyone at all. So even if what is going on here is a case of imagining myself to be in the fictional world it does not seem that it is or must be a case of simulating any of the characters, whether it be the schoolmaster or one of the onlookers, since I may imagine that I am peripherally observing the scene. Furthermore, it does not seem as if any form of imagining concerning myself is prescribed or required—I may just imagine that the scene unfolds as it is characterized. In either case no fictional character is being simulated in any way. It might be objected that even where we are just imagining the scene nonetheless we are simulating—we are not imagining

IN SEARCH OF A NARRATIVE 73

the scene from the perspective of one of the fictional characters but from that of the implied narrator. So centrally imagining a character is not notably required to grasp their experience and we need not advert to the simulation of such a narrator as opposed to recognizing the state Gradgrind is represented as being in and what his putative attitudes are. Thus central imagining, whether of a fictional character or narrator, is just not required to both grasp and affectively respond to the character of Gradgrind as portrayed. Furthermore, and just as crucially, I need not imagine “being someone” in the weak sense of just “taking on” the sorts of thoughts and responses Gradgrind is portrayed as having in order to understand him. What information is unavailable to us without simulation here? Here’s a case where we can arrive at a deep understanding of a character. When I read the opening passage of Hard Times I pick up on Gradgrind’s harsh, brutal and emphatic tone, his austere, rigidified, overly rationalistic conception of education and his underlying attitude of commercialism. In part apprehending the way Gradgrind is portrayed as being expressive of these attitudes is a function of the other possible ways in which Gradgrind could have been portrayed. His tone of voice could have been represented as being tentative, wavering and mournful. He could have been represented as saying that it was unfortunate that the school system required the drumming in of facts. He could even have been represented as being somewhat sanguine about this form of education. Whereas what Gradgrind does is identify himself wholeheartedly with the application of the conception of education represented. Furthermore had the manner of Gradgrind’s behavior been different, though the same beliefs and desires might have been expressed, we would have a different understanding of him. So what Gradgrind’s actions are expressive of is partly a function of what other actions and manner of so doing he could have been represented as enacting but did not. If he thought this form of education inhuman or harmful to the children then his tone and choice of words would not have been eulogistic but sorrowful. If he had thought that “possession of facts” were by no means what was required for life he would have been derogatory or perhaps even abusive of the school before him. By contrast if he were merely sanguine about the state of affairs he might merely have shrugged the scene off and not bothered to comment on it. Given the conventional artistic devices used, and the expressive tone and tenor of the representation, I come to see that Gradgrind is proud of the school system, its application, underlying ideals and identifies wholeheartedly with a broadly utilitarian attitude toward life and others. Already we have a deep and sophisticated grasp of Gradgrind. But in no way was simulation required to arrive at such an understanding. I’m not simulating Gradgrind or an observer or a narrator or even just simulating holding the kind of thoughts Gradgrind is represented as having. In the very first few paragraphs we feel as if we have got into Gradgrind’s mind and character very deeply indeed without any simulation. No doubt someone might be tempted to claim that we still rely on simulation here in the following way. Even in cases where the author writes “Gradgrind is proud” or the expressive tone and tenor of the characterization suggests this, if we

74 MATTHEW KIERAN

had not done some appropriate and quite specific simulating in the past we would not understand the author. So though there may be no present simulating, past simulation is required. To understand “is proud” we have to simulate it. So later on when we encounter the term or its associated expressive aspects we do not have to simulate just because we already have some simulation-based understanding to hand. But not only does the simulation account now rest on speculative child psychology but, more importantly, it is uninteresting. For the interesting and robust claim was that the process of simulating the beliefs, desires and attitudes of a character or, more minimally, simulating the same sort of beliefs, desires and attitudes that such a character has, was required in engaging with the work in order to understand and appreciate the narrative. Blocking simulation and the role of memory At least in certain kinds of cases it is important not to simulate. For simulating a character’s states may not only be unnecessary but may actively distort our understanding of them. One might be tempted here, as an obvious example, to cite instances where we are prescribed to imagine that a character is subconsciously motivated or self-deceived. It is common in narratives for fictional characters to be mistaken about their own mental states, dispositions and character whereas it may be clear to the reader exactly what is motivating the character and why. But, one might think, were one simulating the character this would not be the case. Why? Because a reader’s successful simulation of a character would mimic the structure of the unconscious motivation or self-deception and the reader would thus arrive at the same mistaken self-understanding that the fictional character has. But this is standardly not the case precisely because the reader appreciates the character in ways they themselves do not—and this cannot be a function of simulating them. Now the example here will not quite do the work. For, it may be replied, this is to confuse what is being simulated with the object of the simulation. Consider, for example, Olivier’s famous portrayal of Hamlet. Hamlet, who for the sake of argument is the object of our simulation, is portrayed as being unconsciously motivated for Freudian reasons. Of course, if we were seeking to simulate that which Hamlet is conscious of, in terms of his thoughts and desires, we would not simulate the darker Freudian drives Hamlet is portrayed as having. But precisely because Hamlet is portrayed in a manner that suggests an overweening dependence on his mother, an overly felicitous attention to her every state and whim, apparently jealous overreactions to her sexual affections placed elsewhere and a sensual attention to the look and feel of her, the performance prescribes us to imagine that he is so unconsciously motivated. And, the thought goes, just as we can make sense of someone acting out the part of someone so unconsciously motivated so we too can simulate such unconscious motivation. And a veridical simulation would indeed do just that. The difference lies not in what is being simulated and the state the fictional character is portrayed as being in but, rather,

IN SEARCH OF A NARRATIVE 75

the asymmetry between what we are aware of informationally regarding what gets factored into the simulation and what, fictionally, Hamlet is aware of as shaping his beliefs and desires. Notice, first, that the reply relies on us already having grasped much about the character, including that they are self-deceived in certain ways, so it is far from clear that simulation is required to understand the character rather than attention to the expressive aspects of the work and performance. Second, despite the simulationist reply, if I am too successful in simulating the states of characters represented then I may not be in an appropriate position to discern their important features. The parallel with acting here may be instructive. I once heard recounted the following tale: Dustin Hoffman and Laurence Olivier were acting together on Marathon Man and Olivier was beginning to grow impatient with Hoffman continually delaying shooting because he felt he could not quite get “inside” the skin of his character. After one interminable delay Olivier finally asked Hoffman what his problem was and Hoffman duly replied that he could not “feel” what the character was supposed to be feeling. To which Olivier rather acidly replied “Well, my dear boy, you could always try acting.” The sharp point of the barb is that what is crucial in acting concerns whether or not the manner of acting is appropriately expressive of certain states and cognitive—affective attitudes. And not only need this not but often had better not involve simulation. Simulating Hamlet’s states may preclude the actor from acting—Daniel Day Lewis’s famous walk out halfway through a performance of Hamlet followed by his absconding from the production as a whole might be a salutary lesson here. For Day Lewis, renowned for his emphasis on method acting, apparently could not cope with playing Hamlet precisely because it paralyzed his capacity to perform the play. Less dramatically, simulations of the states of dramatic characters would often preclude actors from performing as they are required to in order to convey or express the characters’ thoughts and feelings. Just consider, for example, how common it is for characters represented as being in a state of uncontrollable grief to articulate in complex verbal form the nature of their grief, enunciated clearly, with rhetorical repetition of key words for emphasis. Yet uncontrollable grief (and presumably any veridical simulation thereof) is usually manifested in sobbing and wailing that muffles speech and paralyzes complex verbal thought in the face of sheer horror at the loss of whatever or whomever is being mourned—it is uncontrollable. In these kinds of cases simulation is not merely unnecessary but serves to preclude the appropriate dramatic expression of the character’s states and attitudes required by the play. Now just as an actor’s simulation of the state of a dramatic character may preclude them from enacting the requisite expressive aspects of their performance, so too the simulation of a character by a reader may preclude them from being able to discern features that are crucial to their understanding of the character and their situation. If one were to simulate certain states of characters, such as blind terror, utter subsumption of self in sexual ecstasy, sheer panic, uncontrollable grief, blind anger or all consuming fury, then

76 MATTHEW KIERAN

presumably it would be very difficult indeed if not nigh impossible to switch attention to other features of the narrative. For the point about such states is precisely that they are supposed to be all consuming and uncontrollable. Hence in such cases simulating the particular character’s experience is neither required, since one can grasp the characteristics of such states without simulation, nor sufficient for understanding the character since the reader would then be blocked from attending to other crucial expressive aspects of the text. Perhaps this is to overstate the case somewhat. After all the simulationist could acknowledge that if one were to spend the whole time simulating such states then one would get nowhere. But surely the point is that to really understand the rage or grief of a character some simulation has to be undergone but this is consistent with my being able to read on with a vivid, tangible sense of the relevant state. So I start to feel something akin to rage or grief building up in me, identifying the states ostensively, and am thereby reminded as to what states in that locale really are like without thereby delving too deeply into them. Now I know the types of feelings involved and what the character is feeling since I’ve just been subject to something like it myself. If I had not simulated then I would not understand the depth or true nature of the state the character is in. So simulation here may be required only to “tease” oneself with such emotions or ones in their locale and then read on with the vivid memory of them at hand. But without the vivid memory all I’d have is a very shallow understanding indeed. But if this is the response I do not see what the appeal of simulation is here. First, the claim has been weakened from simulation of the sort of states the character is subject to being required, to the simulation of states that are in the same sort of locale or region. Thus simulation of the same sort of state is not now required to ostensively identify it. Second, and more significant, simulation is not required to do this. Indeed more often surely we advert to memory. In the case of familiar emotions we just remember what they are like—I might recall the feeling of betrayal say in which case I am not simulating an emotion but recalling it. In the case of unfamiliar emotions we grasp the contours of the state from the description and the way in which it is thus relatable to those emotions we are familiar with and have been subject to. Character traits We often apprehend the character or personality traits of fictional characters in what we imagine but it would be odd to think these are simulated. The logic of character traits is not susceptible to simulation. For character traits as such are not experienced since, minimally, they are dispositions to act in certain ways. Now it could be objected that character traits can be simulated since though they are dispositions, they are dispositions to have certain emotions or respond emotionally in certain ways. But this cannot be right. At best one can simulate emotional states consistent with certain character traits, but then there are many emotions consistent with a particular trait, and any given emotional state is consistent with

IN SEARCH OF A NARRATIVE 77

many character traits. So we need to say more about the relations between traits and emotional states. First, not all character traits are even in the first instance reducible to the disposition to feel certain emotions. A considerate person is one who is disposed to take care to register and treat as important the feelings, thoughts, needs and desires of others. Such a disposition does not connect to or have any internal connection to any particular emotional state. Someone can be considerate without feeling any particular way about the state of affairs or person their attention is directed toward. And clearly one can be considerate whether one feels sad, angry, sanguine, resentful or a whole host of other emotional states. Second, even those character traits that look as if they are amenable to being treated as temperaments cannot be so characterized. Pride, for example, looks as if it is a disposition to feel self-satisfied. But being disposed to feel self-satisfied is not the same kind of thing as pride. One can, after all, be self-satisfied without being proud in any way. Pride, rather, is constituted by a tendency to overestimate or over value the nature of one’s achievements or character. And one can be proud without feeling self-satisfied, be proud whilst yet feeling unhappy at what one has failed to achieve or proud without any concomitant affective feeling whatsoever. Character traits are, minimally, internalized dispositions to act, respond or apprehend in cognitive-affective terms in accordance with a regulative principle. From the third-person perspective the character trait of pride may be captured in terms of someone who always overrates their thoughts, actions and achievements. The content of the injunction from the character’s viewpoint, which is internalized and regulates a proud man’s thoughts, actions and attitudes, might then be captured by something like the principle “always value highly your character and achievements”. And the internalized disposition to regulate oneself thus cannot be simulated. One can only come to the appropriate judgments concerning what traits a character possesses, and their particular nature, by coming to grasp certain narrative relations that obtain as their character is shown to develop. Consider, for example, Austen’s Pride and Prejudice. Thematically the book develops from a scenario where both Elizabeth and Darcy are proud in different respects and are prejudiced against each other as a result. Their characters develop from the various injuries to their pride and the subsequent insight afforded them, ultimately leading to the restoration of a less partial pride to each other. From what we are told about Darcy, his behavior and manner of expression we learn that Darcy possesses highly idealized standards of propriety and has received unqualified approval and admiration from his parents. Hence we come to see his pride. But as the story develops we discern that his pride is both narcissistic and insecure. His defensive manner, his habit of scorning others and solicitations of praise suggest the particular nature of his pride. Hence it is intelligible just why he should both assume Elizabeth would admire him as all others apparently do and why it is a severe blow to him that he should be rebuffed. Only by coming to see his behavior as Elizabeth has viewed it does he come to the self-realization that he

78 MATTHEW KIERAN

has failed to live up to his high ideals and only when he reshapes himself accordingly and wins her approval does his pride come to be better founded. Simulation is neither necessary nor sufficient for grasping the fundamental traits of Darcy’s character, since traits cannot themselves be simulated, and yet understanding his traits is crucial to our grasp of particular episodes, his character development and the thematic development of the novel. Narrative appreciation How do we arrive at narrative understanding then? In a molecular and symbiotic way that makes use of many of the features I’ve adverted to. First of all we are guided by what states the work explicitly tells us a character is in, the type he is identified as being and what we are prescribed to imagine about them. In imagining a character’s behavior, their facial expressions, gestures, intonation we come to an initial judgment. For example, that Gradgrind is feeling bullish or that Othello is horrified. Any deeper understanding requires a grasp of what the objects of their cognitive-affective thoughts, feelings and attitudes are and the ways in which they think of and respond to them. There are many ways in which a work can do this but we should at least consider a few. A work may prescribe us to imagine a character in terms of certain expressive qualities we relate to concepts we already possess. For example, we know that people are animals, and as such are partly driven by animalistic drives and appetites, but this is not always at the forefront of our apprehension of people or characters. An author may seek to make this vivid in the apprehension of a character’s behavior, as animalistic, in order to give us a deeper understanding of the base drives or vulgar aspects that dominate him. In portraying the behavior of a character as bestial, say, we may be prescribed to imagine him as being akin to a leader of a pack of dogs on the prowl, to imagine his glances as wolverine, to imagine the deference of others in terms of the recognition of a pack leader and so on. Second, a work may seek to extend or modify our initial apprehension of the states of a character. For example, as the narrative structure of Othello progresses I come to see Othello’s reactions in certain situations as being crucially interrelated. I see his slowness to anger, his romantic and trusting nature, his horror at Iago’s initial speculations, I see his increasing agitation as being related to a personal sense of insecurity that Iago cultivates and exploits and as I come to see how he descends into jealousy proper I come to understand the interrelations therein that explain the highly noble nature of both his jealousy and his actions. Hence it is a great tragedy. Our narrative understanding can also become increasingly refined by coming to apprehend expressive qualities in terms of newly acquired concepts or ever finer distinctions amongst concepts we possess. For example, we imagine Darcy’s reaction to Elizabeth’s rebuff as one of anger but are only later struck by the particular kind of anger involved and what it stems from—the way in which his

IN SEARCH OF A NARRATIVE 79

sullen pout, glowering facial expression, arms crossed and body deliberately turned away are expressive of a petulant anger, which is the mark of injured pride. Our narrative understanding of a character can be further deepened by the sudden apprehension of new connections between and interrelations amongst expressive features. Austen, for example, often represents a character glancing elsewhere whilst conversing as being expressive of a desire to seek socially significant people out. She also tends to represent over solicitous flattery as being expressive of a yearning desire to please or be well thought of and overly felicitous shows of formality and manners as expressive of a particular kind of vulgarity of mind. But in reading Sense and Sensibility one is suddenly struck by the way in which Lucy Steele is portrayed as exhibiting all these qualities and the ways in which they are all interconnected; how her seeking out of others whilst in conversation stands in stark incongruity with the showy drawing of attention to formal courtesies, how the highlighted show of manners serves to block rather than ease social interaction and how the insensitivity to ease social interaction expresses a vulgar lack of interest in and respect for others. This is not to mention the figurative, symbolic and metaphoric allusions that are commonplace or the many uses of different perspectives on a character’s actions which serve, indirectly, to afford the reader a much deeper grasp of character. Whether episodic or over a more extended range, our grasp of a character gives rise to and symbiotically interacts with our narrative understanding. For understanding is not merely a matter of knowing what mental state a character is in at t1, followed by knowing what mental state they are in at t2 and so on. It requires, furthermore, some degree of overarching understanding as to how and why those states are connected with each other and thus make sense in relation to one another. I do not deny that simulation may play a role here—imagining that I am in a character’s place or that I am them may aid me in following the kinds of responses that help to make sense of their consequent mental states. But again this already depends upon our imaginative apprehension and understanding of the particular character and the kind of person they are or are likely to become. Someone who lacks imaginativeness in this sense lacks neither standard powers of observation nor the capacity to imagine they think or feel something—for example, in reading a novel they can adequately imagine a character seeing certain things, where they move their hands when they talk and can imagine holding the beliefs or having the feelings being reported to them. Rather what they lack, or at least possess only to a very low degree, is the capacity to appreciate the manner of the portrayal, to see them as an instantiation of finely variegated types of people, to discriminate between the appropriate associations or connotations of different uses of language, to grasp an action in terms of relevant contrasts and metaphorical associations and make the right kinds of judgments concerning the nature and interrelations of character traits to the character’s states. Any kind of understanding of particular characters primarily depends upon the expressive qualities or manner of their behavior and actions as represented. One

80 MATTHEW KIERAN

then attempts to make sense of what one apprehends in their imagined action, what one recognizes it as expressive of and the various associations, analogies, connections, types and what one may metaphorically perceive the action in terms of in relation to some narrative. The narrative understanding arrived at is then assumed as a defeasible explanation that best makes sense of how and why the person’s manner is as it is and how their mental states relate both to their situation and their basic character traits. Only then are we in any position to attempt any kind of imaginative simulation or reenactment regarding their mental states with any hope or degree of verisimilitude. Thus it is crucial that we distinguish imaginative apprehension, where we recognize the expressive qualities of a character’s action, from the fictional reenactment of a character’s mental states involved in imaginative simulation. The former concerns the different ways in which the prepositional attitudes and cognitive-affective states generally may be expressed. Our imaginative apprehension of them is required in order to even be in a position to attempt an imaginative reenactment of the mental states of a particular character in the first place—for such imaginative recognition is prior to, constrains and informs imaginative simulation. Hence someone whose powers of imagination in the sense of entertaining mental states is great may, nonetheless, not be very good at all at understanding particular characters in literary narratives since their powers of imaginative apprehension may be coarse and uncultivated. In such a case the simulation will be insufficiently constrained and informed by a discriminating imaginative apprehension of the manner of the represented actions, behavior and character. Of course, we are sometimes at a loss as to what a character is thinking and feeling precisely because the characterization proffered, deliberately or otherwise, is relatively thin. In such cases we no doubt naturally do ask ourselves “what on earth would I have to think and feel in order to be doing that”. And here the role of simulation may come to the fore. But these kinds of cases do not constitute the norm. What is needed is a rich, complex theory of how the above features of narrative understanding are interlinked and this is something simulation as such cannot provide. Sympathy for fictional characters The grounds for skepticism regarding the centrality of simulation to grasping what is fictionally the case regarding characters do not, however, affect the claims that simulation is central to explaining how and why we care about fictional characters. In what follows I shall concentrate on the paradigmatic cases of sympathy and empathy. Stephen Darwall has argued that sympathy generally depends upon empathic simulation and the degree to which one cannot empathize is the degree to which one is unable to be sympathetic toward another (Darwall 1998). If true then empathic simulation must be what explains how and why we are sympathetic with respect to fictional characters. Susan Feagin, in contrast to Currie’s more generalist aspirations, claims simulation is merely one of the many processes involved in narrative appreciation (Feagin 1996). Feagin

IN SEARCH OF A NARRATIVE 81

seems undecided as to whether sympathetic or anti-pathetic responses involve empathic simulation but she has argued that simulation gives us an informative account of empathy. Let us consider sympathy. Initially it may seem puzzling as to why one might think that sympathy for fictional characters should be primarily explained in terms of our empathic reenactment of their states and situation. After all, sympathy seems to involve taking up a cognitive-affective attitude toward a character’s situation that involves recognizing and commiserating with their difficulties. Sympathy is thus tightly tied to a concern for the good and well-being of a character. But why think this must be tied to the simulation of what a character is thinking and feeling? For one can recognize the nature of a character’s plight as represented, care for them and thus respond sympathetically, without having simulated the same sort of suffering or feelings the character is represented as having. I sympathize with the children in Gradgrind’s school, for example, but I do not imagine myself experiencing the same sort of pain, deprivation and suffering. For what is crucial is not that I simulate their experience as represented but that I care about it. However, Darwall argues that sympathy should properly be construed as an emotion of a particular form that depends upon empathic simulation. Darwall’s characterization of sympathy is as follows: It is a feeling or emotion that (a) responds to some apparent threat or obstacle to an individual’s good or well-being, (b) has that individual himself as object, and (c) involves concern for him, and thus for his well-being, for his sake. Seeing the child on the verge of falling, one is concerned for his safety, not just for its (his safety’s) sake, but for his sake. One is concerned for him. Sympathy for the child is a way of caring for (and about) him. (Darwall 1998:261) Darwall claims, with reference to C.D.Batson’s experimental psychological work on the “Empathy-Altruism Hypothesis”, that sympathy naturally arises from the capacity to empathize (Batson 1991; Batson and Shaw 1991). In one of Batson’s experiments students are played an audio tape of “Carol” who they believe to be one of their class members. They hear that Carol has missed a month of term because she’s been in hospital as the result of an accident. The subjects of the experiment are then asked if they will help. There’s a “difficult escape” scenario where Carol will return to class next week and an “easy escape” scenario where she will be studying at home. According to Batson, where the subjects are drawn into empathizing by imagining what Carol is feeling the tendency to help shown even in the “easy escape scenario” is markedly high. The likely explanation, according to Darwall, is that the high empathy subjects feel vicarious distress as a function of their empathic reenactment of what Carol is feeling. They thereby come to be distressed not merely at Carol’s plight but on her behalf, that is, come to desire relief for her sake. Thus empathy gives rise to sympathy; the latter is a higher-order emotional response dependent on the former. An important feature

82 MATTHEW KIERAN

of Darwall’s characterization of sympathy, with which I agree, is that the form of sympathy is individual-regarding. The desires and feelings bound up with sympathy do not just have prepositional objects, for example, that Carol gets better and makes up her classes. Rather the form of sympathetic concern is partly constituted by the non-propositional “indirect objects” it takes, for example, I desire for Carol’s sake that she get better and make up class. We can see how easily this kind of explanation in the ordinary case smoothly transfers across to the case of fictional characters. Where a work manages to get me to empathically reenact the plight of a fictional character I am more likely to be distressed for them as opposed to merely recognizing that a fictional character is in a dire state of affairs. Darwall is not alone in characterizing sympathy as an emotion. For example, Peter Goldie has argued that sympathy is a basic emotion, though he does not claim that sympathy depends upon empathy (Goldie 2000). However, it strikes me as odd, to say the least, to conceive of it thus. Sympathy seems more like a standing state or condition that is constituted by an attitude toward someone, albeit one where we are concerned for another person for their own sake. The assumption that sympathy is an emotion seems to rest on a conflation of sympathy with compassion or the occurrent emotional response of pity. But that it is a conflation can be shown by the following considerations. First, there is no distinctive phenomenology that constitutes the affective aspect of sympathy. My sympathy for someone can be concomitant with the affective phenomenology of fear, since I am afraid for them, sadness, since I am sorry for them, anger, since I am enraged on their behalf. It’s not even obvious that being sympathetic for someone precludes certain more positive emotional responses such as feeling delighted for someone. After all, I might be sympathetic for someone because they find it hard to choose between living in London or Rome. I would truly sympathize if a friend of mine had to make such a decision but the sympathy would be concomitant with delight for them that they are faced with such an enviable (though I believe a rather difficult) choice. Second, having sympathy for someone does not entail feeling pity, compassion or fellow feeling for them. A friend tells me of their romantic troubles and I may be moved to intercede on their behalf. I am disposed to take pity on them because I recognize their plight and they are my friend. But this does not depend upon my having any occurrent emotional response whatsoever. I may neither feel the desperation they themselves are subject to nor sorrow at their desperation and yet may still be strongly moved to act because I desire what is good for them and recognize their plight. Indeed, I may even have sympathy for their plight and not act at all if I judge that continuance of their romantic entanglements would clearly be bad for them. The crucial point here though is that having sympathy for someone is conceptually independent of occurrent emotional responses. Sympathy is thus better characterized as a standing state or condition constituted by an attitude of concern for someone for their own sake (one desires what is best for them), which disposes one to act in their

IN SEARCH OF A NARRATIVE 83

interests. As such it is conceptually independent of rather than constituted by any occurrent emotional response. Still, it might be asked, why does it matter that sympathy proper is a standing state rather than an occurrent emotional response? The form of Darwall’s argument seems unaffected and the claim, that sympathy arises as a function of the empathic reenactment of a person’s or fictional character’s pitiful plight, still stands. But this is not the case. If sympathy were an emotion then Darwall’s claim would have more plausibility than if we construe it, as I have argued we should, as a standing state. Consider the following contrast between sympathy proper and an emotion such as fear. If I am normally afraid of being threatened with physical violence then I both should and will, ceteris parabis, feel fear where I empathically reenact the position of another who is so threatened. This is because the emotional response, in both cases, arises as a function of the intrinsic properties of this kind of situation. There is a relation of fit between situations that present themselves as being highly threatening and feeling the emotion of fear. This is not the case with sympathy. If I am normally sympathetic with respect to my friend’s romantic entanglements it neither does not nor should it necessarily follow, ceteris parabis, that I am sympathetic with respect to someone else who is similarly romantically caught up. This is so even were I to pay attention to all the intrinsic properties of the comparative situations and were to find them alike in all respects. It’s important to realize that this is not merely a matter of psychology for the difference here is blameless. Why is this? What matters in the case of sympathy, in a way that does not hold true in the case of fear, is what the external relations are that obtain between the person in the romantic entanglements and myself. In the case of my friend I have an attitude of sympathy because I care for him and I care for him because I take certain relations to hold between us—of trust, affection, self-revelation and an overarching attitude of mutual regard. In the case of just anyone such relations do not hold and thus I am neither constrained as a matter of normative or psychological necessity to be sympathetic. Of course, I can come to be sympathetic with respect to someone who is not a friend in such a situation. But what this depends upon is a matter of the relations that are made salient as holding between myself and the other person. So the standing state of sympathy with respect to someone, given they are in a pitiful plight, is a function of the relations perceived to hold between myself and themselves as opposed to being a straightforward function of empathic reenactment. It is the crucial importance of perceived external relations and overarching attitudes toward those relations that explain why empathy, in and of itself, is neither necessary nor sufficient for sympathy. I can be sympathetic for someone without having empathized with them since my overarching attitude toward them and my perception of their relations to myself may be such that where their plight is made salient to me, for example, I may observe it or it may be reported to me in more or less vivid terms, I am sympathetic toward them. Conversely I can empathize with someone, in the sense of imaginatively reenact what I take them to be thinking and feeling, and yet fail to be sympathetic because of my

84 MATTHEW KIERAN

overarching attitude and how I perceive the relevant external relations to be. Empathizing with an enemy or someone I consider to be feckless may, for example, only serve to enhance my contempt for them rather than inculcating sympathy. Darwall himself does concede, with reference to Batson’s study, “it might be that the subjects had some standing desire or principle to aid others in need and that empathy simply makes Carol’s need more evident than it would otherwise have been.” It is just that he thinks the “more likely possibility is that the vicarious distress that high empathy subjects feel comes to have a new object, namely Carol and her predicament” (Darwall 1998:274). A more adequate way of characterizing the alternative possibility being dismissed is that the subjects may already have had a standing disposition to be sympathetic with respect to those they perceive to be in bad circumstances and closely related to themselves. Nonetheless, the simulationist may reply, where the subjects empathically reenacted Carol’s predicament there was a higher percentage of students who perceived their relations to Carol as being more relevantly close to themselves than in the scenario where they were merely informed of her predicament. Thus, as a psychological rather than a conceptual matter, empathic reenactment of another’s plight is more likely to render salient the perception of external relations upon which sympathy depends. There is no conceptual dependence of sympathy upon empathy, nor does sympathy arise as a straightforward function of empathy but empathy is, putatively as a matter of psychological fact, more likely to highlight the kind of external relations required to give rise to (but of themselves insufficient for) sympathy. It is important to note that the experiments themselves do not justify even this conclusion. Reports of Carol’s situation were contrasted with the audio-tape of her. But it does not follow that what does all of the work in the audio-tape case is empathy, since much of the relevant work may just be psychological presentism induced by hearing Carol’s voice, the expressiveness in her voice and the expressive characterization of her experiences. Moreover, it’s not obvious that a brief second hand report of her condition is the right kind of contrast. If subjects had been presented with, for example, a second hand characterization of her plight in the style of Dickens, say, they might well have showed the same responses or perhaps an even higher response rating of sympathy as those subjects exposed to the audiotape. For the vividness and expressiveness of the characterization of her experience and the ways in which this helped to make salient her relations to the subjects of the experience could be just as powerful without calling upon them to empathize with Carol’s experiences. Nonetheless, I think there is some reason to hold that empathy is, as a psychological matter, one way that the perception of relevant external relations required for sympathy can be made salient. To that extent there is some truth in the thought that empathy can help to explain our sympathy for others (whether actual or fictional). Nonetheless, empathizing is neither necessary nor sufficient for sympathy and is merely one way amongst

IN SEARCH OF A NARRATIVE 85

many of making salient the relevant external relations required (but insufficient) for sympathy. Empathizing with fictional characters We have seen that empathy at least can play a role in the cultivation of sympathy (though not nearly as strong or direct a role as has been suggested), so now the question arises as to whether the simulationist analysis of empathy is the right one. Feagin’s simulationist analysis of empathy is as follows; we empathize with someone (actual or fictional) where we (a) come to “share” the feelings, typically at least, of a reflectively identifiable protagonist in terms of structurally analogous activities or processes and (b) the phenomenological quality of the experience is the same for the protagonist and the empathizer (Feagin 1996:94–109). In other words empathy is experiencing the same thoughts and feelings as someone else, where “same” means identical both in terms of kind and felt quality, as a result of the same kind of emotional interactions or thought processes that the other person went through. For example, if my friend tells me of his romantic woes I empathize with him to the extent that I empathically reenact what gave rise to his feeling so sad and the reasoning that gave rise to his despair. It is important to note, as Feagin herself emphasizes, that in the fictional case the simulation of a process can be carried out even though there is not the same “input” for the simulation as there typically is when one participates in “real” life. This is just to recognize the point that ordinarily our empathic reenactment of someone else is based on how we perceive them and their situation to be whereas in the case of literature, for example, it is a matter of what we are told by the words on the page. As articulated the position requires some slight fine tuning. It is common enough to assume that empathizing with someone requires that there be a reflectively identifiable protagonist (Wollheim 1984). Yet this is too strong. This is not merely because we can and sometimes do find ourselves spontaneously empathizing with someone without already having identified them as someone to empathize with. I may suddenly find myself empathizing with the frustrated person in the queue next to me or may unexpectedly find myself empathizing with a fictional character like Hannibal Lecter. But this is merely to point out that those we empathize with are not always identifiable prior to the act of empathizing. Rather, as Peter Goldie has argued, it is because we sometimes can and do empathize with a type of person even though we cannot refer to nor have an identifiable particular individual in mind (Goldie 1999). Goldie’s example concerns his experience of walking in the Pyrenees and suddenly imagining what it must have been like for a Roman soldier struggling up the mountains in the heat. As he says, one can think of him empathizing with “‘that soldier’ or ‘him’, whilst acknowledging that one knows nothing particular about him that enables one to individuate him from others of the type; all one needs for the imaginative project is a characterization of the type. Rejection of the reference condition also

86 MATTHEW KIERAN

makes coherent the project of the detective, seeking constantly to build up a substantial characterization of ‘the murderer’, in order to better predict what ‘he’ will do, and thus to find out his identity and be able to refer to him” (Goldie 1999:418). Still, this is to quibble about the formulation of the claim rather than a challenge to the simulationist analysis of empathy. So let us take the more appropriate characterization to be cashed out in terms of, minimally, a retrospectively identifiable individual or type. Does the simulationist analysis hold up? On the simulationist account of empathy the phenomenological quality of the experience shared by the person (or type) being empathized with and the empathizer must be the same and the upshot of structurally similar reasoning and emotional processes. But this cannot be right. For example, shame and horror loom large in Anna Karenina’s emotional narrative. The two-page chapter following the scene of Anna’s seduction uses the term horror ten times or more. Yet one need not “feel” horror nor have reenacted the same emotional narrative and reasoning processes as she is represented as having gone through in order to empathize with her. We need only recognize that horror is what she is represented as feeling, grasp how and why this is so and identify with her situation. And, in light of the above considerations, it does not seem as if simulation is required to do that. A weak way of making the objection is to advert to phenomenological differences within the same emotion type. I may “feel” the same emotion type as Anna Karenina in empathizing with her and yet there may be phenomenological differences since the intensity of Anna’s felt emotion and its overall shape, the depth and permeation of the feeling in relation to her mental economy, may be markedly different from that which I empathically reenact. But such differences are ones of degree rather than kind and as long as simulation of the same emotion type is involved that is all that the simulationist needs. A more profound way of making the point is to show that empathy as such does not require the empathic reenactment of the same type of emotion, hence one can empathize with someone and yet the phenomenological differences can be ones of type rather than just matters of degree, nor need it depend upon having reenacted the same mental narrative. If true, then the simulationist’s identity claim must be false. Many cases of empathy seem to arise from the perception and characterization of someone and their situation in terms that are related to our own thoughts, emotional responses, attitudes and past experiences. Consider the following varied kinds of cases: someone describes their psychological unease at rejecting their religious beliefs; someone describes how they were humiliated by a rude customer in a shop; someone expresses how unutterably bored and frustrated they are with the pettiness and fecklessness of certain people they work with; someone describes how nice it was to be on holiday in Italy; how let down they feel after being betrayed by a friend; how annoyed they feel at being taken for granted by someone. Ordinarily it is easy to empathize with someone describing such states of affairs precisely because we can relate them to similar, though not necessarily the

IN SEARCH OF A NARRATIVE 87

same, types of experiences we ourselves have had. But though we might empathize, in the sense of sharing fellow feelings with the person, none of these cases need depend upon or be a function of having simulated what the person “must have” thought and felt. Furthermore, the better and more informative characterization of the experience afforded the less likely it is we would need to do so given the expressiveness and salient relations highlighted. Indeed some of my reasoning may well be highly disanalogous. For example, reflections about my own experiences or those of my friends, parents and relations may be brought to bear in a way that leads me to have similar kinds of responses to the situation as represented and identify them with hers. Hence, to return to the case of Anna Karenina, I may well empathize with her because I relate her experiences to those of my own or perhaps even those of my friends and family. I have some sense of what it is like to be treated coldly, to give in to temptation and to be betrayed and may come to share the feelings she is represented as having to the extent that I relate her experiences as represented to my own or those of others I care for. Thus I may empathize without having structurally mirrored the mental narrative and processes she is represented as having undergone. Furthermore, for similar reasons, we often can and do empathize with someone though we are not simulating the same type of feelings they are experiencing. After all, we can empathize with someone in the sense of extending our fellow feelings toward them without it being the case that we are simulating the pain, horror, shame or guilt we take them to be subject to. Empathy, I would suggest, is more appropriately characterized as involving the recognition of the states and situation a character is represented as being in, a grasp of what this may be like and an identification with the desires, aspirations and attitudes of the character. All that is required here is adversion to features of our own past experience, familiarity with the experience of others and an ability to extrapolate and interrelate from such experiences as they are thought to hold with respect to the character one identifies with. I am not denying that an empathic reenactment of the mental narrative of someone is sufficient for empathy. The problem is that the simulationist analysis conflates a typical means of empathizing with an analysis of empathy itself. So simulation is a typical and important means of empathizing with people (rather than being constitutive of empathy). The further question remains. How does the simulative means of empathizing help, if at all, to explain how and why we care for fictional characters? It follows from the argument thus far that it cannot be the only mechanism that explains the phenomenon. Characters can be made vivid, we can come to identify with and care for them just in virtue of the imaginative way they are described to us, how their situation is relatable to our own past experiences and our cognitive-affective attitudes. Nor is empathy achieved via simulation a sufficient condition for caring for or indeed hating them. After all, I may simulate the mental narrative of a character and remain bored and indifferent to them precisely because I find there to be nothing of interest or value therein. Nonetheless, at least where the mental narrative of a character is of

88 MATTHEW KIERAN

interest to us and where the nature and plight of a character is either beyond our own experience or difficult to see how it is easily relatable to our own experience, simulation can have a role to play in bringing home to us their plight, the nature of their experience and thus engaging our cognitive-affective responses with respect to them. We should not underestimate the extent to which we can easily be made to love, hate, admire, loathe or simply be beguiled by fictional characters based on the same kinds of concerns we are moved by in real life. We are interested, for example, in those who are exceptional, witty or who resemble ourselves in some respect. Furthermore, the kinds of situations, conflicts and choices facing the characters serve to heighten our interest and care for them. To return to Othello, we care deeply for him because he is represented as being of such an exceptionally noble and admirable character and is drawn into a situation where any ordinary person would have felt jealous long before he succumbs. On the basis of the above considerations it seems we have little reason to grant either that simulation is standardly required for the deep apprehension of character or that it is the canonical ground for grasping the experiences, states and traits of dramatic characters. This is not to deny that simulation has a role to play. Given a rich narrative understanding of a character simulation may help deepen our understanding in terms of what our expectations about that character’s dispositions are likely to be. It is just that, as I have argued, simulation cannot be the central process by which we come to understand particular characters. For deep understanding can be achieved without simulation, necessarily depends upon the possession of a rich narrative understanding and makes use of many narrative features with respect to which simulation is otiose or inapplicable. Simulation cannot be the central mechanism that affords narrative understanding. Furthermore, empathizing is neither necessary nor sufficient for sympathy and is merely one way amongst many of making salient the relevant external relations required (but insufficient) for sympathy. As a typical means of realizing empathy simulation does have a role to play in engaging our cognitiveaffective concerns for fictional characters, but it is neither the only nor the central psychological mechanism required to do so. Narrative appreciation and understanding is far more rich and complex than adversion to simulation theory seems to allow. Acknowledgments I would particularly like to thank Dominic Lopes, Bryan Frances and R.R.Rockingham Gill for their helpful comments and exchanges on drafts of this chapter.

Part II TRUTH IN IMAGINATION

5 FICTIONAL ASSENT AND THE (SOCALLED) “PUZZLE OF IMAGINATIVE RESISTANCE” Derek Matravers

In his 1994 paper “Morals in Fiction and Fictional Morality” Kendall Walton poses the following question: Can an author simply stipulate in the text of a story what moral principles apply in the fictional world, just as she specifies what actions characters perform? If the text includes the sentence, ‘in killing her baby, Giselda did the right thing; after all, it was a girl’ or ‘the village elders did their duty before God by forcing the widow onto her husband’s funeral pyre,’ are readers obliged to accept it as fictional that, in doing what they did, Giselda or the elders behaved in morally proper ways? Why shouldn’t storytellers be allowed to experiment explicitly with worlds of morally different kinds, including ones even they regard as morally obnoxious? There is science fiction; why not morality fiction? (Walton 1994a:37) There does seem to be a curious asymmetry here. Before I can spell it out more precisely, I shall need to fill in some background. In doing this, I broadly follow Walton in his earlier work, Mimesis as Make-Believe (Walton 1990). For the moment, I shall take as my example written fiction (the novel), although I will make some remarks about other sorts of fiction later. In reading a fiction, a reader engages in an imaginative project. Within this imaginative project there are some propositions to which he assents (“Emma Woodhouse married Mr. Knightley”) and some propositions to which he does not assent (“Emma Woodhouse wore an eyepatch”). Let us call this pro-attitude to such propositions, a pro-attitude that is part of the imaginative project of reading a fiction, “fictional assent”. (Those troubled by the apparent split infinitive are invited to consider “fictionally assent” as a compound verb.) The asymmetry I want to discuss is as follows: (1) Within the imaginative project of engaging with fiction F, a reader can fictionally assent to p, where p is some nonmoral proposition that he does not believe holds.

FICTIONAL ASSENT AND IMAGINATIVE RESISTANCE 91

(2) Within the imaginative project of engaging with fiction F, a reader cannot fictionally assent to q, where q is some moral proposition that he does not believe holds. To take an example, let p be the proposition that people can be instantaneously transported from a spaceship to the surface of the earth and q be the proposition that there is nothing wrong with female infanticide. A reader can fictionally assent to the former, but cannot fictionally assent to the latter. That is, readers who come across “in killing her baby, Giselda did the right thing; after all, it was a girl” cannot fictionally assent to it. Even within the fiction, female infanticide is wrong. This bald statement of the asymmetry needs to be hedged with explanations and qualifications. By a “moral proposition”, I mean a proposition asserting something to be a moral truth. Speaking in the realist idiom aids clarity, but is not an essential part of the problem. Antirealists too will need to have some way of characterizing the asymmetry (see Walton 1994a:38). By a nonmoral proposition I mean every other proposition. Nonmoral propositions include assertions about fictional characters’ moral beliefs. For example, it is unproblematic for a reader to fictionally assent to claims such as “these people believe female infanticide is wrong”. For the purposes of the contrast, I have classed all nonmoral propositions together. This is provisional: we shall see later that some sorts of nonmoral propositions (e.g. those that assert that something is or is not funny) should perhaps be placed on the other side of the divide. Second, I shall not be providing arguments that this asymmetry exists. There seems to be enough agreement that it does exist for it to be worth considering. If any readers do not recognize it at all, they should simply think themselves lucky and stop reading this chapter. I shall, however, be returning to consider the circumstances in which the asymmetry becomes less stark. Third, I have stated the asymmetry in terms of what a reader “cannot” do. In order not to beg any questions, I mean this—initially at least—to be taken as neutrally as possible. It should cover all possibilities from simply psychological difficulty (perhaps hardly “cannot” at all) to logical impossibility. The problem itself goes back to Hume’s “On the Standard of Taste.” Walton (as we have seen), Richard Moran and Tamar Gendler have recently essayed solutions (Moran 1994; Gendler 2000). I have also benefitted from hearing a paper by Greg Currie on the subject, which has not at time of writing, been published (Currie 2002). Walton’s solution, albeit only tentatively endorsed, starts with a supervenience thesis about moral facts. My account of Walton is indebted to Gendler. Moral properties depend or supervene on ‘natural’ ones and, I believe, in the relevant manner (whatever that is); being evil rests on, for instance, the actions constituting the practices of slavery and genocide. This, I suggest, is what accounts (somehow) for the resistance to allowing it to be fictional that slavery and genocide are not evil.

92 DEREK MATRAVERS

(Walton 1994a:45) This is not, however, enough. Even if I believe that A depends upon B, it does not seem to follow (in general) that within the imaginative project of engaging with fiction F, a reader cannot fictionally assent to some proposition that A exists in the absence of B. I believe that mental states supervene on brain states, but can engage with the science fiction that has explicit Cartesian presuppositions. Walton (of course) realizes this, and so provides us with a second step. We still need an explanation of why we should resist allowing fictional worlds to differ from the real world with respect to the relevant kind of supervenience relations. My best suspicion, at the moment, is that it has something to do with an inability to imagine these relations being different from how we think they are, perhaps an inability to understand fully what it would be like for them to be different. (Walton 1994a:46) Gendler glosses the solution as follows: “if there are things that are unimaginable, then we certainly cannot imagine them, so we should not be surprised that our capacity to make-believe at will runs out at precisely those points” (Gendler 2000: 66). She construes Walton as claiming that the reason we cannot imagine moral supervenience relations being different to how they are is that we think it conceptually impossible that they could be. She refers to this as the impossibility hypothesis, and goes on to reject it by showing—successfully, I think—that conceptual impossibility does not preclude imaginability. I take Walton’s point to be slightly different. It is not so much that moral supervenience relations being different to how they are is conceptually impossible, but rather that we could not fully understand what it would be like for them to be different. Gendler herself stresses this point in the course of her discussion, before going on to concentrate on impossibility. (We could not fully understand what it would be for slavery to be morally sound; we would simply lose our grip on what the terms meant.) Focusing on the failure of understanding, however, fares no better than focusing on conceptual impossibility. Consider Walton’s earlier example: that female infanticide is morally sound. To believe this one would need to believe that being female is a reason to kill a child. Could we fully understand what it would be to endorse this? We need to consider two possibilities. The first is that there are certain things true in the fiction that makes this belief easy to understand. For example, the fiction might be set in a world in which female children are innately evil and inevitably cause death and destruction. In such a case, we would fully understand female infanticide but we would also be able to fictionally assent to it. This case (although it has some interesting ramifications that we will investigate later) is not relevant to our current concerns. Walton had argued—in short— that the propositions that we could not fully understand are those to which we cannot fictionally assent. This is a proposition that we can fully

FICTIONAL ASSENT AND IMAGINATIVE RESISTANCE 93

understand but, once the complexities are spelled out, it is also a proposition to which we can fictionally assent. The proposition to which we are unable to fictionally assent is that female infanticide is morally sound in a world like ours in all relevant respects. In such a world, the reason for infanticide is that the baby is female. What, in this scenario, is beyond our understanding? The inhabitants of that world have a reason for their belief. We are familiar with prejudice; it is not beyond our understanding that being black, or being a woman might, for some, be sufficient reason for a difference in treatment. As Bernard Williams and others have stressed, views such as racism and sexism are not necessarily cognitively flawed (Williams 1985: 115– 16). Although I accept that moral facts supervene on “natural” facts, I cannot see any reason to believe a priori that I will always fail to understand the position of someone who sets up the dependence relations differently. Failure of understanding cannot be the explanation of failure of fictional assent. One moral of the above discussion is that cases of imaginative resistance arise only with those fictions that share a background with the actual world. Gendler refers to such fictions as “nondistorting”. By contrast, “distorting” fictions are those “where the mirroring between the fictional and the actual world is more complex” (Gendler 2000:77). Imaginative resistance arises only in those fictions that are nondistorting in the relevant respects. Gendler uses this contrast to provide her own solution to the problem. She begins with the Aristotelian point that fictions can give us knowledge: readers of realistic narrative fiction “export from the fictional world fictional truths that [they] take to be not merely truths in the story” (Gendler 2000:76). Propositions that are taught, or perhaps simply made clear, in the story are added to the reader’s stock of knowledge in the same way as the reader might add knowledge gained by testimony. Gendler goes on: “so my hypothesis is that cases that evoke genuine imaginative resistance will be cases where the reader feels that she is being asked to export a way of looking at the actual world which she does not wish to add to her conceptual repertoire” (Gendler 2000:77). Why should this raise particular problems for morality? If this were so, would not imaginative resistance be evoked by nonmoral propositions that we knew to be false? Gendler’s reply to this is twofold. First, moral claims are categorical; if they are true, they are true in all possible worlds. Hence, “fictional moral truths clamor for exportation, in a way that other sorts of fictional truths do not” (Gendler 2000:78). Morality, however, is not the only sphere in which truths are shared between the fictional and the actual world. There are also “the truths of logic, mathematics and—in most genres—the laws of physics and psychology and even etiquette” (Gendler 2000:78). Why, then, would we not encounter the resistance problem when asked to fictionally assent to a proposition in any one of these spheres that we knew to be false? Gendler’s reply is that the very fact that the proposition, although true in the fiction, is false is a signal that the author does not intend that we export it. As we are not being asked to export it, imaginative resistance will not arise. The reason for the asymmetry in the moral case is that

94 DEREK MATRAVERS

“because we recognize that there are instances of actual moral disagreement, when we encounter fictional truths that concern deviant morality, we cannot assume their deviance is an indication that the author does not wish them to be exported” (Gendler 2000:78). I have a couple of problems with Gendler’s account. The first is whether the distinction between distorting and nondistorting fictions can be made clear. Much science fiction has the curious property that the more outlandish it gets, the more moralistic it gets. Early Star Trek, for example, seemed to have no exportable propositions save the moral rectitude of the American way. Evelyn Waugh’s Decline and Fall is nondistorting in its facts, distorting in its minor moral themes and nondistorting in its major moral themes. It might emerge that the distinction between distorting and nondistorting is so closely bound up with the distinction between what can and cannot be exported that it could not be used as an explanation for it. That, however is speculation. The real problem with Gendler’s account lies in the relation between fictionally assenting to q and exporting our assent to q. Let q be the proposition that female infanticide is morally sound, and F be some nondistorting fiction. Putting Gendler’s account into my terminology, we have the following: (i) In reading fiction F, we are invited to fictionally assent to q. (ii) In fictionally assenting to q, we are being invited to export q into our beliefs about the actual world, (iii) We resist exporting q into our beliefs about the actual world and hence resist fictionally assenting to q. The obvious objection is to (ii). What argument is there for this being the connection between fictionally assenting to q and exporting q? The alternative is simply to regard q as true-in-F (which, from “the inside” during our engagement with F, will be the pro-attitude to q of fictional assent). Gendler’s answer, as we have seen, is that moral truths are true in all possible worlds. Hence, assenting to q in the world of the fiction entails assenting to q in the actual world. Furthermore, because we are familiar with actual moral disagreement, disagreement with fictional morality does not block the invitation to export. However, Gendler herself earlier maintained that whether some proposition p is or is not true-in-F is independent of whether or not p is a necessary truth. Hence, being a necessary truth would not, by her own lights, be a sufficient mandate to export. A more puzzling aspect of the theory is (iii) does not follow from (ii). What does follow is: (iii•) We resist being invited to export q into our beliefs about the actual world and hence resist fictionally assenting to q.

FICTIONAL ASSENT AND IMAGINATIVE RESISTANCE 95

Why should being invited to export q into our beliefs about the actual world be something to resist? By hypothesis, q is something that we find morally objectionable in the actual world. Hence, there is no chance at all—we have to suppose otherwise my resistance would disappear and there would be nothing to explain—that I am going to export q. Hence, there is no chance at all of q making it into my conceptual repertoire. Why, then, should I resist fictionally assenting to it? It poses no danger to me. Gendler actually says that “the reader feels that she is being asked to export a way of looking at the actual world” (my italics), but that hardly seems a reason for withholding fictional assent. Even if I am being asked (which, as indicated, I doubt) that does not commit me to anything. Unless I object to being asked, my resistance to fictionally assenting to q is unexplained. My suggestion, which I will explore for the rest of this chapter, is that the explanation of our imaginative resistance lies in the nature of the authority of the fictional narrator. In order to do that, we need to look at the nature of our engagement with fiction. There is no standard and accepted view on how we engage with fiction, which is probably because we do so in a number of ways. I have argued elsewhere that the usual way in which we engage is best described by the “report model” (Walton 1990; Matravers 1997). Greg Currie describes this as follows: Our make-believe is not merely that the events described in the text occurred, but that we are being told about those events by someone with knowledge of them. Thus, it is part of the make-believe that the reader is in contact, through channels of reliable information, with the characters and their actions, that the reader learns about their activities from a reliable source. To make-believe a fictional story is not merely to makebelieve that the story is true, but that it is told as known fact. (Currie 1990:73) There are a number of issues that need to be sorted out in defending such a view, such as unreliable narrators, fictions that have as a content events that could not have been observed, the prospect of multiple interpretation of narratives and epistemological problems surrounding the (so-called) “omniscient narrator”. Apart from the last, I am not going to discuss these problems as they have been exhaustively covered elsewhere. I will discuss omniscience as it is particularly germane to our problem. It is rarely (if ever) the case that it is part of the content of a fiction that there is an omniscient narrator. That is, it is rarely (if ever) the case that one of the characters in a fiction is possessed of supernatural epistemological capacities that enables them to discover the intimate thoughts and feelings of the characters that are reported on the page. Rather—as Kendall Walton has pointed out—the omniscient narrator is a device that belongs to the structure of the project of our imaginative interaction with fiction:

96 DEREK MATRAVERS

…it may be understood that whatever he is portrayed as saying is true, that if fictionally he says that p, the fictionality of p itself is automatically implied, regardless of what is or is not fictional about his knowledgeability and honesty. It will be fictional that he knows that p, I suppose, and (in our games) that we take his word for it, if it is fictional that he says that p… The advantages of this arrangement are obvious. It enables the author cleanly and crisply to establish fictional truths about the characters and events described—including fictional truths about their innermost thoughts and feelings—and to provide readers with sure access to them. (Walton 1990:360–1) The device of the omniscient narrator is a natural concomitant of the report model. As Currie says, the reader needs to be “in contact, through channels of reliable information, with the characters and their actions…the reader learns about their activities from a reliable source.” The omniscient narrator—or, as I call him, the “fictional narrator”—operates according to the following principle: (3) Within the imaginative project of engaging with fiction F, a reader will fictionally assent to p if p is asserted by the fictional narrator. This, however, seems to compound the problem of imaginative resistance. We know that (3) does not hold for all propositions. What we need is an explanation for the difference between: (4) Within the imaginative project of engaging with fiction F, a reader will fictionally assent to p, where p is some nonmoral proposition that he does not believe holds, if p is asserted by the fictional narrator. (5) Within the imaginative project of engaging with fiction F, a reader will not fictionally assent to q, where q is some moral proposition that he does not believe holds, even if q is asserted by the fictional narrator. The explanation for this lies in finding reasons for the limits on the authority of the fictional narrator. According to the report model, in the actual world, I, Derek Matravers, am sitting in my chair reading a Penguin paperback. In my engagement with the book, I, Derek Matravers, am sitting in my chair reading a report of what is going on elsewhere. In the actual world, I do not believe the content of the fiction (putting aside difficulties concerning sentences that are true both in the fiction and the actual world). In the engagement with the fiction, I do believe the content of the report. The explanation works by applying the explanation of our actual-world engagements with reports that we believe, to our imaginative engagement with fictions that we do not believe. What, then, are the limits on the authority of the narrator of a report in the actual world? Recall the last time you were reading the foreign correspondent of a decent newspaper, reporting on a series of characters and events of which you

FICTIONAL ASSENT AND IMAGINATIVE RESISTANCE 97

knew nothing. Assume you trust the correspondent to be reporting the truth. Then, it might seem that the beliefs you formed would have been guided by the following principle: (6) Within the project of engaging with the newspaper F, a reader will assent to p if p is asserted by the foreign correspondent. This, however, is far too liberal. You will only assent to those propositions to which the correspondent stands in a privileged epistemological position: namely, those propositions they know in virtue of the special position they are in. You will assent to their assertions about the characters and events about the place from which they are reporting, but will have no reason to assent to their assertions otherwise. In particular, you will have no reason to assent to their assertions about the truth or otherwise of the moral beliefs held in the place from which they are reporting—if they are inconsistent with your current moral beliefs—as your epistemological relation to those propositions is exactly the same as theirs. The principle by which your belief formation would have been guided would rather have been as follows: (7) Within the project of engaging with a newspaper report, a reader should assent to some proposition p on the grounds of reading the newspaper report if: (a) p is asserted by the foreign correspondent; (b) the reader trusts the correspondent; and (c) that the foreign correspondent is in a sufficiently privileged epistemological position with respect to p. As I have said, the report model is just the claim that the details of our engagement with fictions is explained by what happens with our engaging with reports in the actual world. Hence, the limits on the authority of the fictional narrator should be analogous to (7). (8) Within the project of engaging with a fiction F, a reader should assent to some proposition p on the grounds of reading the fiction if: (a) p is asserted by the fictional narrator; (b) the reader trusts the fictional narrator; and (c) that the fictional narrator is in a sufficiently privileged epistemological position with respect to p. We can take (a) as given. It is worth pausing over (b). As Walton argues, the device of the omniscient narrator is part of the structure of our engagement with fiction. It is necessary as a clean and crisp way of providing the reader with sure access to the fictional world. Putting unreliable narrators to one side, within the game this structural feature is manifested as the reader trusting the narrator to report truly. In short, the report model can accommodate—even explain—the “omniscient narrator” without strain. It is, however, (c) that is important for us here. For, as with the actual-world case, this divides assertions concerning

98 DEREK MATRAVERS

characters and events in the fictional world (for which the narrator’s position is privileged) from assertions concerning the truths of morality (for which the narrator’s position is not privileged). As the fictional narrator has no epistemic authority in this area, their word alone carries no weight. That is the explanation of the difference between (4) and (5). In short, the asymmetry between (4) and (5) is exactly what one would expect if the report model were true. Hence, the fact that there is such an asymmetry lends further support for the model. In what remains of this chapter, I am going to give three further considerations in support of my conclusion and then consider two objections. The first consideration is that, if one considers those engagements with fiction that are not modeled by the report model, the asymmetry disappears. Take, for example, an actor preparing to play the Younger Son in The Revenger’s Tragedy. (Hugh Mellor noticed that the asymmetry does not apply in this case. The example was originally Iago. The reason for the change is that it is not clear to me that Iago does have moral beliefs that are inconsistent with ours. Rather, he is motivated to pursue ends that—even by his own standards—are evil. The Younger Son, however, does have moral beliefs that are inconsistent with ours.) As part of this preparation they might try to imaginatively identify with the character. The imaginative project might be described as follows (I is the fictional character): (9) Within the imaginative project of identifying with I, an actor will fictionally assent to p if p is believed by I. This is an imaginative project entirely distinct from those considered earlier. Here there is no asymmetry. The Younger Son believes that rape is morally acceptable (“My fault was sweet sport which the world approves”). This is something to which the actor will need to fictionally assent if he is to imaginatively identify with the character. I shall assume what I think is true, namely, that an actor can do this and hence that the problematic asymmetry does not arise in this case. It is difficult to see how this could be explained by Walton. Walton’s tentatively expressed thought was that it is difficult to imagine the dependence relations in moral supervenience being different to how they in fact are. This, however, does not seem to be a problem for the actor. Gendler’s claim, as I argued above, is more complex. If fictionally assenting to p during the imaginative project of identifying with the character involves the actor being asked to export p, then we should expect the asymmetry. Hence, whether or not Gendler can account for this example depends on whether the actor is being asked to export p. She argues that in works of realistic fiction, “regulations concerning exports will be extremely lenient: in general (though there will be numerous exceptions), if something is true in the fictional world it will be true in the actual world” (Gendler 2000:76). Whilst perhaps not quite realistic, The Revenger’s Tragedy would certainly seem to qualify as nondistorting. If this is right, and there is an injunction to export truths in The Revenger’s Tragedy; then on Gendler’s account

FICTIONAL ASSENT AND IMAGINATIVE RESISTANCE 99

the actor is being asked to export p. As such, we would expect—contrary to fact —there to be imaginative resistance. On my solution, however, as the actor’s imaginative project does not involve a fictional narrator, there would be no reason to expect the asymmetry. The second consideration is that my solution is able to account for other examples—such as the following by Kendall Walton—of resistance to fictional assent. Consider a really dumb joke, like this one: “Knock, Knock. Who’s there? Robin. Robin who? Robbin’ you! Stick ‘em up!” It is not easy to see how it could be fictional that this joke is hilariously funny (in circumstances just like ones in which, in the real world, it would be dumb), how one could reasonably allow it to be hilarious in a fictional world, while thinking that is it actually dumb. (Walton 1994a:44) We need to separate two propositions: (i) The joke is held to be funny by the inhabitants of the fictional world, (ii) The joke is funny in the fictional world. The fictional narrator is in a privileged epistemological position with respect to (i), but not with respect to (ii). Consider the parallel in the actual world. There would be no difficulty, in reading a trusted foreign correspondent’s report, in assenting to the claim that the French (bizarrely) find Walton’s joke funny. There would, however, be resistance to assenting to the claim that in France the joke really is funny. The correspondent is epistemologically privileged with respect to the former, but in no better position than their reader with respect to the latter. Hence, the report model suggests that a reader can fictionally assent to (i) and not to (ii), which is exactly what intuition suggests. David Davies has pointed out to me that there are other propositions, apart from morality and humor, for which the asymmetry in question applies. For example, that something is a good reason for something else. My account suggests that the asymmetry will apply to all those propositions for which testimony would not be sufficient for belief, in the kind of circumstances illustrated by the example of reading the foreign correspondent. The third and final consideration is that the solution is able to accommodate the fact that the problematic asymmetry changes when we cease to see the world of the fiction is relevantly similar to the actual world. Here is Gendler’s example: Suppose instead of: (5) ‘In killing her baby, Giselda did the right thing; after all, it was a girl.’ we had: (6) ‘In killing her baby, Giselda did the right thing; after all, it was born on January 19.’

100 DEREK MATRAVERS

or: (7) ‘In killing her baby, Giselda did the right thing; after all, it was a changeling.’ For me at least, neither (6) nor (7) evokes the same sort of response that (5) does. I said above that in reading (5), my and Walton’s first instinct was to reject the invitation to make-believe that ‘in killing her baby, Giselda did the right thing; after all, it was a girl’… But this is not my first instinct with regard to (6) or (7). There my first instinct is to say: “How interesting! I wonder what this world is going to turn out to be like, this world in which killing one’s baby is the right thing to do, so long as the baby is born on January 19 or is a changeling.” In fact, in light of (6) and (7), I can almost feel my imaginative resistance to (5) evaporating. So long as I am not inclined to take (5) as making a claim about the way the world is, I am perfectly willing to grant it the autonomy I grant to other sorts of makebelieve. (Gendler 2000:75) I almost share Gendler’s reaction to the three examples. Almost, because (6) and (7) still leave with me a feeling of moral disquiet. A fact is being adduced as a reason to kill the baby. Clearly, this fact does not at all serve as a reason for infanticide in the actual world. However, rather than fictionally assenting, I suspect that, in reading the novel, I would suspend judgment and move on. I do not, however, think this is a case in which the asymmetry evaporates. Let us revert to our earlier example in which it is true in a fiction that female children are innately evil and inevitably cause death and destruction. In this world it might be the case that female infanticide is morally justified. Similarly, it might turn out (as Gendler suggests) that being born on January 19, or being a changeling might serve, similarly, as a justification. The proposition to which we are being asked to fictionally assent is: “female infanticide is morally justified in situations in which female children are innately evil and inevitably cause death and destruction”. That proposition seems to me to be true, and, as such, is one to which I am able to fictionally assent. Let us consider one of Gendler’s propositions: “female infanticide is morally justified in situations in which female children are changelings”. In the absence of further background (filled in by the epistemologically privileged fictional narrator) I am not sure whether this is morally acceptable or not. Hence, in the absence of further background, I am not sure whether I can fictionally assent to it or not. If it transpires that what is fictional in that world justifies the moral claim (if, e.g. it is fictional in that world that changelings are innately evil and inevitably cause death and destruction) then I could fictionally assent to it. However, if it transpires that what is fictional in that world does not justify the moral claim (if, e.g. the only difference between a changeling infant and a human infant is skin color) then I could not fictionally assent to it. In these cases the asymmetry is still in place.

FICTIONAL ASSENT AND IMAGINATIVE RESISTANCE 101

Michael Tanner discusses a more complicated case; fictions that express a “worldview”. A worldview is a comprehensive view of the actual world. This will include factual claims (whether true or false) and a moral dimension. A Nazi worldview, for example, will include a raft of claims about genetics and history, together with racist views, and views about the sublimation of individual goals to the greater destiny of the state. A Christian worldview will include claims about historical events, together with views concerning the primacy of faith, hope and charity. Tanner makes the point that readers are rarely asked to fictionally assent to single moral falsehoods; rather, falsehoods come in the context of the expression of a worldview. A worldview with which we agree will not cause a problem, as fictionally assenting to it will not cause a problem. What about worldviews with which we disagree? There are three ways in which a reader can react. First, by not fictionally assenting to the view. Indeed, this reaction is suggested by the fact that the fictional narrator is not in any epistemically privileged position with respect to worldviews. This would simply be a manifestation of the asymmetry discussed above. Second, the reader could fictionally conditionally assent. Schematically, consider factual claims a, b, c and d to be true-in-F. It is also true-in-F that the moral claims q, r, s and t follow from those factual claims. Hence, q, r, s and t are also true-in-F. The reader thinks that, actually, q, r, s and t are false. This, according to the above, will provoke resistance to fictionally assenting to q, r, s and t. However, the reader might want to fictionally assent to: If a, b, c and d, then q, r, s and t. This, of course, is not a moral proposition and so resistance to fictionally assent does not arise. The problem with this reply is that fictional conditional assent is unlikely in all but a small number of cases. The reason is that there would not be a lot left of the project of imaginatively engaging with fiction if, within the scope of the project, we did not assent or dissent. We do not usually treat fiction as a thought experiment whereby we reason within the scope of an assumption. That would be to regard fiction as a branch of logical reasoning, concerned with establishing connections between propositions. In reading the novel, we are not being invited to fictionally assent to a conditional; we are being invited to fictionally affirm the antecedent. Refusing to do the latter is refusing to engage with the fiction. Tanner’s view is that one of the reasons art is troubling is that we do find ourselves fictionally assenting to moral propositions we think are actually false. Crucially, worldviews are an amalgam of factual and moral claims in complex interrelation. As they cannot be separated from each other, the reader is moved seamlessly from fictional assent to the former to fictional assent to the latter. Furthermore, as morally troubling fiction usually expresses morally troubling worldviews rather than isolated morally troubling propositions, this is the usual case. What we find with the great novelists (and some who aren’t so great) is not so much assertions about genetic differences, though they can occur, as in

102 DEREK MATRAVERS

some of the works of D.H.Lawrence, but rather depictions of the world which, if they compel us, do so by making us share their perspective, so that we find ourselves taking up moral positions which may surprise or even shock us, but which seem inevitable once we have agreed to imagine life on their comprehensive terms. (Tanner 1994:58) To ascertain what the report model would think about this, we need to consider the actual-world scenario. What of a reader who is confronted with a documentary report of a morally alien comprehensive worldview? Such a situation is exactly that which provokes the problems involved with ethical relativism. Do moral judgments extend to morally alien comprehensive worldviews, or do they stop on the boundary? This is not the place to discuss the problem of relativism in order to formulate an injunction about what the reader of a documentary ought to do. Obviously, one simply needs to slot in one’s solution to that particular problem here. Our interest is in how plausible this is transferred to the fictional case. The report model, which argues that the relation to fiction is modeled on the relation to documentary, suggests that Tanner’s reader also faces the problem of relativism. How would this be manifest in the reader’s experience of fiction? I am going to assume that the approach to relativism taken by Bernard Williams, specifically his distinction between real and notional confrontations, is true (Williams 1974–75). Broadly, a confrontation is real for some person S, if there is something that would count as S fully assenting to or even living according to that view. A confrontation is notional if the view is so alien that nothing would count as S fully assenting to or living according to it (this is not exactly Williams’s distinction, but it will do for our purposes). To use Williams’s examples, we are in notional confrontations with “the life of a Greek Bronze Age chief, or a mediaeval Samurai” (Williams 1974–75:140). Notional confrontations do not bother us; we do not feel we need to judge actions to be right or wrong. Real confrontations, by showing us an alternative to how we live, do bother us. The analogous difference is clearly manifested in the fictional case. Tanner describes two sorts of engagement with fiction. First, “we are not…in any serious ways challenged or offended in those cases where we can’t make reasonably strong connections between a fictional world we encounter and our own” (Tanner 1994:63). Tanner implies that we are “challenged or offended” in those cases where we can make a “strong connection.” Second, “the more distant the culture represented, the less likely we are to read about it in any other than in an anthropological spirit” (Tanner 1994:63). These are precisely the reactions we would expect a reader to have first in a real confrontation and second in a notional confrontation. We stand to Geoffrey of Monmouth’s Arthurian tales in a notional confrontation, and hence we approach it in “an anthropological spirit.” Such a life is not an option, not least because women are simply not like that any more. We stand to Saint-Exupery’s Night Mail, or Kerouac’s On the Road in a real

FICTIONAL ASSENT AND IMAGINATIVE RESISTANCE 103

confrontation. Both articulate comprehensive moral views that are alien to, but not out of reach of, most of us. This is what makes them such challenging and unsettling books. Once again, the parallel between the actual and the fictional case would explain this systematic difference in the kind of engagement we have with fiction. The first of the two objections I will consider questions whether the distinction between what is and is not within the epistemic competence of the narrator coincides with the asymmetry we are trying to explain. Consider again the realworld case on which the fictional case is modeled. The argument has been that a reporter does not have any epistemic authority concerning moral truth (or truths about what is funny and so on) in faraway places. However, there are other areas where we would not cede authority to a narrator: namely, necessary truths. (There are, arguably, different kinds of necessary truth. I shall take the strongest— the truths of mathematics—as they will be the most difficult for my account.) No newspaper report would be sufficient evidence for me to believe that in some obscure part of the world two plus two equals five. My proposal does not rest on moral propositions being necessary truths (if they are) but on the fact that the position of the narrator does not give him a privileged epistemic position in the moral realm (as it does in the nonmoral realm). Hence, the issue is the ground for fictional assent. So the objection here needs two stages. First, there is a class of propositions that are believed by the reader to be necessary truths. Second, the reader will believe, with respect to those propositions, that because they are necessary he is in the same epistemic position as the narrator. The role of necessity in this objection is only to neutralize any apparently epistemic privilege the narrator might enjoy. If this is right, then we should find that readers resist fictionally assenting to the negation of any proposition they believe to be a necessary truth. This would be an objection, of course, only if readers did not resist fictionally assenting to the negation of any proposition we believe to be a necessary truth. However, there seem to me all kinds of cases here. Dickens was not able to solve his characters’ financial problems by allowing them to borrow two pounds here and two pounds there, combining them to make five pounds, paying back the four they had borrowed and keeping the pound. We would resist fictionally assenting to this, as Dickens has given us no grounds for assenting to the proposition that two plus two equals five. However, there are (or could be) fictions in which time travel occurs, Gödel is refuted and two plus two does equal five. Sometimes we do fictionally assent to the negation of propositions we believe to be necessary truths. What would explain this would be if the fictional narrator used the privileged epistemic position they do have to describe a scenario in which we can assent to the impossible happening. Recall, all that is needed for my proposal is that the person being reported to grants the narrator privileged epistemic status with respect to a proposition that, in a cool hour, he would believe to be a necessary truth. Hence, if the narrator evokes magic, God or other such doers of the

104 DEREK MATRAVERS

impossible, then the epistemic imbalance is restored and my account faces no problem. Other devices diverting the focus of the person being reported to might be obfuscation (it is not clear what the Gödel refuter actually could say), avoidance of the issue (the person traveling back in time never does kill his or her own mother), or distracting the reader’s attention. In these cases the narrator spells out a world in which we are happy to assent to taking his word about the occurrence of the impossible. In this I agree with Gendler’s discussion of this issue. When we imagine the things that, on reflection, we realize to be conceptually impossible, we imagine them in ways that disguise their conceptual impossibility. So when God gets angry and causes twelve no longer to be the sum of two primes, we are considering ‘twelve is the sum of two primes’ primarily with regard to one of its features, namely, that it is a proposition of which human beings are categorically certain only as a consequence of their hubristic arrogance… It is as a result of lot of local bits of conceptual coherence that the global incoherence is able to get a foothold. (Gendler 2000:69) The second objection I will consider is that I have begged the question. I have assumed a model of engaging with fiction (the report model). Readers have their actual persona (the “A-reader”) and their fictional persona (the “F-reader”). The A-reader fictionally assents to all and only those propositions the F-reader believes. As the A-reader is not identical to the F-reader (they have different mental states) the question arises as to what principle is appropriate in the construction of the F-reader’s mental states. By assuming the report model, I have assumed that the appropriate principle is the same as that which governs believing propositions on the grounds of testimony in the real world (in the circumstances characterized by the foreign correspondent)—let us call this “the testimony principle”. Why is a different principle not appropriate? That is, why is it not appropriate (e.g.) that an assertion by the fictional narrator is sufficient warrant for F-reader to believe any proposition, including moral propositions? The question of why the testimony principle is appropriate just is the question of why there is the problematic asymmetry. Why should our interaction with fiction be governed by the testimony principle (that entails the asymmetry) rather than some other principle (that does not)? To an extent, I think this objection is unfair. The asymmetry is, initially at least, presented as something surprising. What I have shown is that it is not surprising; that it is entailed by a standard account of our interaction with fiction that was proposed independently of consideration of this particular problem. Undoubtedly, however, the larger question is raised: namely, why our interaction with fiction takes this form at all. To answer this I would need to show that there are good independent reasons for preferring the testimony principle. That is, I

FICTIONAL ASSENT AND IMAGINATIVE RESISTANCE 105

need to show that the testimony principle has benefits that other principles would not have. (Recall that I do not claim that the testimony principle governs all our interactions with fiction; it does not cover, e.g. an actor identifying with a character.) Defending the testimony principle takes us quickly into the value of fiction, in particular, the relation between fiction and the moral life; questions too complicated for me to resolve here. There is, however, a simple line of argument that builds on the thought—that I will not defend here—that the value of fiction is bound up with our vicarious experience of unfamiliar situations. It is necessary, if we are to learn anything from fiction, that we can transfer the insights garnered in reading a fiction into our lives. It is only through the experience (as F-reader) of how I would react within a fiction to a situation, that I learn (as A-reader) how I would react within the actual world to the situation. The F-reader must have those principles by which I negotiate my way around the world, because only then will the insights garnered be relevant to me (as opposed to someone with a different set of principles). Hence, the F-reader must share the A-reader’s moral reactions (broadly construed). Contrast this with engaging in a fantasy. I could construct an F-reader that did not share my moral views. I fantasize about my passage through the world, unconstrained by the moral straightjacket under which I usually labor. The representation provides no resistance to any scenario I imagine for the sake of the immediate gratification of our desires. While this no doubt has its place, it is difficult to see the activity as sharing the value we typically associate with fiction. Unless representations resist our fantasies they will, to quote Bernard Williams, “be pathetically or repulsively impoverished” (Williams 1985: 220). The problem for the objector is to describe an alternative to the testimony principle that will allow the value of fiction to be something more than the value of fantasy. I do not think it can be done; it has to be us to whom the events are being reported. Acknowledgments I am grateful to comments given to me on presentation of this chapter at the Art and Imagination conference in Leeds; in particular to David Davies, Robert Hopkins (especially for showing me an early draft of a paper on morality and testimony) and Stacie Friend. I am also grateful to the editors and to Alex Barber, Peter Cave and Gregory Currie who read and commented on the manuscript.

6 THE TOWER OF GOLDBACH AND OTHER IMPOSSIBLE TALES Kathleen Stock

The claim that there are no or few constraints operative upon what one is able to imagine is often casually assumed to be true. For instance, in a recent article Jonathan Cohen writes: there appear to be very few restrictions on the class of thoughts I can entertain…. For example, I can entertain the thought that the moon is made of green cheese (although I do not believe this thought). I can also entertain the thought that the moon is not made of green cheese, and I can entertain the conjunction, i.e., the thought that the moon is made of green cheese and the thought that the moon is not made of green cheese. (Cohen 1999:67; see also de Sousa 1987:26; Walton 1990:64) More recently, Tamar Szabó Gendler has argued explicitly against the claim that “a judgement of conceptual impossibility renders a scenario unimaginable” (Gendler 2000:66). She ingeniously argues for the conclusion that, even where a proposition p expresses a conceptual impossibility, the agent can imagine that p. Gendler’s arguments have a sympathetic audience in Derek Matravers, who endorses them in his paper “The So-called ‘Problem of Imaginative Resistance.’” In this chapter, I shall offer some reasons to be cautious about Gendler’s conclusion. Seeming to imagine the impossible Gendler introduces her argument by telling a story entitled “The Tower of Goldbach,” in which at one point it is fictionally true that twelve is not the sum of seven and five, and, at another point, that twelve both is and is not the sum of seven and five (Gendler 2000:67). First, God, in a fit of anger, makes it the case that twelve is not the sum of seven and five (i.e. the sum of two primes). This means that certain mathematicians, desperately in search of twelve righteous souls to satisfy God’s wrath, cannot find that number, even though they find five in one place and seven in another. Eventually, after consultation with Solomon, God makes it the case that twelve both is and is not the sum of seven and five. Gendler argues that, in the context of the story, the conceptual impossibilities that twelve

TOWER OF GOLDBACH 107

is not the sum of seven and five, and that twelve is and is not the sum of seven and five, are true, and hence imaginable (Gendler 2000:68). Whether such conceptual impossibilities are true in the story is not a question that concerns me here. Even if it is accepted that a conceptually impossible proposition is true in the story, it need not follow from this that a conceptually impossible proposition is imaginable. And, I shall argue, it does not so follow. The first implicit reason Gendler apparently offers for thinking that one can imagine a conceptually impossible proposition is that, when a person introspectively scrutinizes the contents of her imaginings, it appears to her that she can do so (Gendler 2000:68). She purports to show that one can have an imagining whose content is expressed by a conceptually impossible proposition by relating “The Tower of Goldbach” and inviting the reader to imaginatively engage with the propositions it contains. The underlying strategy here appears to be to lead the reader of the story into realizing that it seems to her that she has imagined a conceptual impossibility; from this realization, ostensibly, it is a short step to concluding that she has imagined a conceptual impossibility, and hence that imagining the conceptually impossible must be possible. Hence Gendler’s presentation of a story that makes certain conceptually impossible propositions fictionally true, and with which the reader can (at least partially) imaginatively engage, is apparently intended to have rhetorical force. Yet at this stage the only grounds offered for thinking that one can imagine that p, where p expresses a conceptual impossibility, is that sometimes, as when reading the story, it appears to one as if this is the case. The problem here is that it is presupposed that one can judge the nature and content of one’s mental states by how they appear to one. Whereas in fact, there are good reasons to deny that the content of one’s mental state necessarily coincides with how it initially appears. There are many familiar cases where, in the light of a fresh perspective, or new evidence, one revises one’s initial judgment as to the nature of a past or present belief, emotion, desire or memory. Ascriptions of mental content to oneself can be made as part of a process of selfinterpretation: for instance, when the eponymous heroine of Austen’s Emma, upon reflection, becomes “acquainted with her own heart” and realizes that she loves Knightley (the point and its exemplification are from Wright 1998:15). One can attribute a belief to oneself that, one subsequently comes to realize, one did not actually have: such as a belief in another person’s trustworthiness. Most relevantly, in the course of thought experiments, philosophers often ask themselves “what is it that I’m imagining here?” Even when self-ascriptions of content do not involve such overt interpretation, they do not guarantee their own truth, since we tend to discount an ascription, even in one’s own case “if accepting it would get in the way of making best sense of the subject’s behavior” (Wright 1998:17). For instance, we are unlikely to accept an apparently sincere self-ascription of the belief that the floor is dangerously unstable, where the ascriber continues to walk on it untroubled. We will not accept self-ascriptions of a desire for x

108 KATHLEEN STOCK

where an agent consistently fails to act so as to procure x, even in the absence of other overriding reasons. Generally, initial self-ascriptions of mental content are revisable under certain circumstances, and Gendler does not provide us with any good reasons to think imagining is different in this respect. In the absence of such reasons, it remains possible (though not yet established) that in the overall context of a subject’s behavior, it is appropriate to reject as false a subject’s claim that she is imagining a conceptually impossible scenario. Hence on its own, the undoubted truth that it seems to some people that they can imagine conceptual impossibilities does nothing to establish Gendler’s conclusion. Disguising conceptual impossibility Gendler claims that her story is able to make the reader imagine that a conceptually impossible proposition is true because it is possible for the relevant impossibility to be disguised from the reader, in virtue of her attention being alternately focused upon properties of the fictional objects present in the situation, whose conjunction in thought is not conceptually impossible. “We focus now on this aspect of what it is to be twelve, now on that aspect, in a way typical of fictional understanding in general” (Gendler 2000:67, footnote 19). This cannot mean that in imaginative engagement with fictions, one’s attention alternates from one imagined aspect of a situation to the next imagined aspect, without considering the consequences of the conjunction of these aspects, for clearly this is false. In engaging with fictions as directed, we are both expected and likely to imaginatively consider in conjunction many or even most of the different aspects of the situations described or shown in the fiction. It is precisely this considering of conjunctions of propositions and what follows from them in which imaginative engagement with fiction principally consists. We treat fictions as narratives, rather than lists of isolated propositions. Hence where in a fiction two propositions jointly entail what counts for me as a conceptual impossibility, I am unlikely to remain ignorant of the fact. Presumably then, Gendler means that one deliberately can ignore the fact that one knows that a proposition expresses a conceptual impossibility, by selectively concentrating in thought on aspects of the situation that “disguise” this fact to us: that is, only those aspects, which when thought about conjointly, do not generate any contradictions. Whether automatic or deliberate, the important point is that directing such selective mental attention towards a limited set of properties of a situation constitutes imagining about that situation for Gendler. She claims of her story: So when God gets angry and causes twelve no longer to be the sum of two primes, we are considering ‘twelve is the sum of two primes’ primarily with regard to one of its features, namely, that it is a proposition of which human beings are categorically certain only as a consequence of their hubristic arrogance. When the mathematicians’s search concludes with their having

TOWER OF GOLDBACH 109

found five righteous souls in one town and seven in another, we are willing to accept that this does not give us twelve righteous souls because we are thinking of it as: ‘number of righteous souls required for God to lift the decree.’ (Gendler 2000:69) This passage underlines the fact that one can imaginatively engage with a proposition—say, the proposition “there are five righteous souls in one place and seven righteous souls in another”—even whilst being prepared to make only certain of the inferences one would normally make about that proposition, were it mentally engaged within the context of a belief. I will return to this point in the next section. In order to imagine the impossible as Gendler describes, we minimally require some grip on some aspect of the situation described by the conceptually impossible proposition, reflection upon which distracts us, or at least, does not remind us of the impossibility of the proposition. In many cases, there is no such aspect, as Gendler herself admits. She notes that one cannot imagine what is conceptually impossible of an object that falls under a “one criterion concept” (Gendler 2000:73, footnote 25), where we are asked to “imagine away” the only condition that is relevant to determining which objects fall within the extension of the concept. For instance, I cannot imagine that a square is a circle: for if a square does not have four sides of equal length it ceases to be a square, and there is no other aspect of a square to focus upon whilst imagining it as a circle that might distract me from this fact. Gendler suggests that, in contrast, objects with a plurality of constitutive conditions can be the subject of imaginings which deny some of those conditions. This, it is suggested, is possible because in such cases there are aspects of the imagined situation one can focus upon to distract oneself from its overall conceptual impossibility. An initial problem with this suggestion is that thinking does not often work like this. In fact, it seems evident from examination of the first-person case that a deliberate mental injunction to avoid thinking about some aspect of a situation or object usually results in the opposite of the desired effect. However, this does not establish that it is impossible to ignore the presence of contradictory elements in one’s thinking, but only that it is very difficult. Imagining and theories of concepts At this point I would like to situate Gendler’s claims in the context of a discussion about the nature of concepts. In doing so I do not intend to endorse any particular theory about concepts myself, but only to draw out the implications of the view Gendler proposes against recent developments in cognitive science. Imagining, I take it, is a form of thinking. So Gendler’s model entails that one can think something about a certain object O (which, it should be stressed,

110 KATHLEEN STOCK

does not imply that there is an O that one is thinking about), even though one is not disposed to make the inferences one normally would be disposed to make, were one to think about O in belief contexts. This follows from Gendler’s claim that one can selectively attend to only certain features of O, whilst “ignoring” other features of it, where this act of selective attention constitutes imagining something about O; in conjunction with the assumption that mentally “ignoring” certain features of O is equivalent to not being disposed to think of them. The latter is a reasonable assumption. Where x is a property of O, it is true that simply [not thinking consciously about x] is not equivalent to [not being disposed to think about x], but as I have argued above, Gendler has to be claiming that in ignoring x, we do more than this; we actively and deliberately avoid thinking about it. If we can do this successfully, then presumably it follows that for the period that one deliberately avoids thinking about x, one is not disposed to think about x. So Gendler thinks it possible to apply a concept O in imaginative thought whilst not being disposed to attribute to the subject of one’s thought certain of the conditions constitutive of O-hood. This view is in apparent tension with several prominent theories of concepts. Most familiarly, it is inconsistent with the Classical Theory of concepts, which states that concepts encode conditions necessary and sufficient for their own application (Laurence and Margolis 1999:9). On the Classical Theory, what it is to think of an object O, whether in imagination or otherwise, is to be disposed to attribute to the subject of one’s thought all the conditions necessary and sufficient for application of the concept O. The Classical Theory of concepts has been largely discredited in its original form (Laurence and Margolis 1999:14–27), and this might be thought to negate the objection. However, the idea that concepts are partially constituted by necessary conditions lives on in various amended versions of the Classical Theory still much in currency among cognitive scientists and semanticists. Modern Prototype theories of concepts often model concepts as having two components: an identification procedure with a prototype structure, and a “classical core” of necessary conditions (Laurence and Margolis 1999:33). Meanwhile, what has been called the Neoclassical Theory of Concepts is gaining respectability amongst semanticists and philosophers (Laurence and Margolis 1999:52–4), and again expresses a commitment to necessary conditions partially governing the application of concepts. In the context of any theory holding that certain necessary conditions at least partially govern concept application, the following model of conceptual impossibility seems appropriate: a conceptual impossibility arises if a proposition states or entails that a concept O applies to a situation, yet that situation (either explicitly or by implication) fails to satisfy one or more conditions individually necessary for O’s application. This being the case, theories that cite necessary conditions as partially or wholly constitutive of the conditions of application of a concept are incompatible with

TOWER OF GOLDBACH 111

Gendler’s account. It cannot be possible to entertain a proposition expressing a conceptual impossibility at all, whether in imagination or otherwise, so long as conceptual impossibility is construed as I have just outlined it, whilst what it is to think about entity O, falling under concept O, is construed (at least partially) as being disposed to attribute to the subject of one’s thought all of the conditions necessary for O’s application. Classical and Neoclassical theories of concepts are both varieties of Inferential Role Semantic theories. Inferential Role Semantic theories state that one’s concepts are individuated, by some or all of the inferences one is prepared to draw (Fodor 1998:13). For Classical and Neoclassical theories of concepts, the inferences that (at least partially) constitute a concept are inferences to those states of affairs cited by the necessary conditions governing the concept. To suggest, as Gendler does, that one can imagine a conceptually impossible situation whilst ignoring its impossibility is to liken imagining to looking at a picture whilst ignoring certain parts of it. I can look at a picture whilst ignoring certain parts of it, an action which will not affect a correct description of what it is that the picture actually represents in its totality. This is because the representational content of the picture is fixed independently of what I see in it or the inferences I am disposed to make. However, Inferential Role Semantic theories entail that the content of what I imagine is not fixed independently of the inferences I am disposed to make. If these theories are true, it is not so much accurate to say that I “ignore” certain features of my imagining as I do not imagine them to be there in the first place. Where it seems as if I “focus” only on certain “parts” of my imagining, in fact those parts constitute the content of my imagining (see Martin 1998:103). To think the contrary is, on this view, to falsely construe an imagining as a picture to which we can limit our cognitive access at will, while the content of the imagining remains the same. Another way of analyzing the problem, then, is that Classical and Neoclassical theories of concepts are inconsistent with Gendler’s claims, because she is arguing that one can imagine something about, say, the number seven, whilst refusing to draw the inferences that on such theories would be considered constitutive of employment of the concept seven. Classical and Neoclassical theories of concepts are not the only options, however. Another currently prominent theory of concepts is the Theory-theory of concepts. This states that the identity of a concept is determined by its inferential role in a mental theory possessed by the thinking subject. On this view concepts are “individuated in virtue of the inferences they licence based on their role in the theories that embed them” (Laurence and Margolis 1999:45). Unsurprisingly, this too looks incompatible with Gendler’s position. Like the previous theories, the Theory-theory is a version of an Inferential Role Semantic Theory (Fodor 1999:112). A concept is constituted by the set (or some sub-set of) the inferences licensed from it and to it, which in turn is determined by the larger mental theory in which it is embedded. In order to class an object as falling under a particular concept, a person accesses a mentally represented theory that

112 KATHLEEN STOCK

determines which inferences to particular properties are licensed from and to the concept, and then judges whether those properties are satisfied by the object. Meanwhile, empirical evidence establishes that both developed and developing thinkers (Gelman and Wellman 1999) have a tendency to psychological essentialism (Medin 1989:1476–7). That is, they believe that category membership is established in virtue of possession of certain internal or intrinsic features. This data, when connected to the Theory-theory, implies that categorization under a particular concept is a matter of ascertaining whether an object has some internal or intrinsic feature, which according to one’s mental theory, is necessary to its falling under the concept in question (Laurence and Margolis 1999:46). This again appears inconsistent with Gendler’s implied claim that one can have a thought—an imagining—a constituent of which is a concept O, and yet at the same time remain indisposed to draw in thought the inferences to those (internal/ intrinsic) properties that, according to one’s mental theory, are constitutive of something’s counting as a O-token. If it is part of my mental theory that the essence of water is H2O, so that, according to my theory, I take it to be a necessary inference from “that is water” to “that is H2O”, then presumably on the Theory-theory, I cannot employ the concept water in imaginative thought whilst refusing/not being disposed to make the inference to H2O. Another prominent theory of concepts is the Prototype Theory, which characterizes concepts as representations wherein are encoded the properties that objects falling under them are statistically likely to possess. Categorization of objects under a given concept occurs in virtue of a similarity judgment between an object proposed for categorization and a prototype. An instance is understood to fall under a concept if “the representation of the instance and the representation of the category are judged to be sufficiently similar” (Laurence and Margolis 1999:29). Whereas, I have argued, a Classical or Neoclassical theory entails that where a concept O is governed by necessary condition n, one cannot imagine a situation where O applies but n does not, the Prototype Theory says that concepts have no individually necessary conditions. Hence clearly on this view it is possible for a certain concept to be constituent of one’s thought, whether in imagining or other contexts, without having to draw an inference to any one particular property of the object that falls under the concept. This looks promising for Gendler’s theory, although the promise is slightly mitigated by the fact that problems associated with the Prototype Theory have forced many of its supporters to attribute to concepts a “classical core” of necessary conditions that determines the conditions constitutive of category membership (Laurence and Margolis 1999:32–43). On the other hand, concepts still have disjunctively necessary conditions on the Prototype Theory. In other words, the Prototype Theory gives rise to the following model of conceptual impossibility: a conceptual impossibility arises if a proposition states or entails that a concept O applies to a situation, yet that situation (either explicitly or by implication) fails to satisfy a sufficient number of the conditions disjunctively necessary for O’s application. The Prototype Theory

TOWER OF GOLDBACH 113

says that if there are not a sufficient number of a set of certain specific features shared between a proposed instance and a prototypical member of a conceptual category, the instance will not fall under the concept. Presumably this entails that I cannot coherently think of or imagine, say, the knave of hearts from Alice in Wonderland as a playing card (Gendler’s example), since there are not sufficient properties shared between a knave of hearts and a playing card for the former to count as an instance of the latter. This conclusion can be generalized to rule out many propositions as conceptual impossibilities, and hence as off-limits for imagining. So what the Prototype Theory has to offer in terms of the possibility of imagining conceptually impossible situations is not sufficient to legitimize all of the sorts of imaginings Gendler thinks are possible. The final theory of concepts I will mention here is the one that seems most conducive to Gendler’s view of how imagining works. This is the Conceptual Atomist view, according to which concepts are primitive representations whose content is determined not by inferential role but by the concept standing in a nomological causal relation to an external object in the world (Fodor 1999). Since on this view, concepts are unstructured primitives with no necessary or statistically reliable inferential relations to any other properties, it seems that there are no limits to the ways in which concepts may combine in thought, including in imagination. There is a problem with this view, however, which threatens to undermine the support the theory appears to give to Gendler’s position. This is the fact that in most thought contexts (Laurence and Margolis 1999:68, footnote), and especially in the context of imagining, concepts which are the constituents of thought do not stand in any nomological relation to objects in the world. It may be true that the belief “that is a cow” stands in a lawful relation to the property of being a cow; but clearly it is not the case that the imagining “the cow jumped over the moon” stands in a lawful relation to the property of being a cow (let alone the property of a cow jumping over the moon). Imagining by its very nature is resistant to any attempt to incorporate it into an externalist theory of content, since for any proposition p, imagining that p is not reliably linked to any particular sensory inputs, nor any particular behavior, verbal or non-verbal. Furthermore, if Conceptual Atomism is assumed to be right about how conceptual content is individuated in belief contexts, then this seems to lead to the problematic consequence that nothing appears to guarantee stability of conceptual content between belief- and imagining-contexts. If concepts invoked in the belief that p are individuated by existing in a nomological causal relation to event p, then the same concepts cannot be involved in imagining that p, since no such relation exists in that case. For these reasons, it is doubtful that a Conceptual Atomist account can say anything useful about the employment of concepts in imaginative contexts. The theory apparently most conducive to Gendler’s conclusions does not have the resources to explain what occurs when one imagines, let alone to legitimize

114 KATHLEEN STOCK

Gendler’s conclusion by outlining the mechanism and processes implicated in the imagining of conceptual impossibilities. In this section I have examined Gendler’s claims against a background of the current leading players in theories of concepts. I have suggested that several problems face any attempt to assimilate her conclusion into currently prevalent theories of concepts. At the very least, I have established that Gendler has some work to do in further elucidating her claims before her conclusion is shown to be consistent with current thinking in cognitive science. How do I know what I am imagining? I now wish to leave aside the question of the theory of concepts appropriate to Gendler’s view, and move on to an epistemological question. Assuming that it is true that to have a thought is (roughly) to exist in relation to a particular mental representation, presumably what it is I am imagining in a particular instance is ultimately to be settled with reference to what set of mental representations I am related to in thought. However, acknowledging this is not particularly relevant to the epistemic question of how it is I know what I am imagining. It is clear that we do not have immediate and unproblematic access to the content of our beliefs. As my earlier remarks suggest, to some extent analysis of one’s beliefs is an exercise in interpretation. One often works out what beliefs one has by working through the inferences one is prepared to accept. Of course, in the case of identifying what one believes, dispositions to certain inferences are not the only relevant source of information: sensory inputs and behavioral outputs are relevant as well. In the case of imagining, just as in belief, it seems reasonable to suppose that we do not have immediate and unproblematic access to the content of our imaginings, and that to some extent analysis of one’s imaginings is an exercise in interpretation. There is no reason to think imagining is different in this respect to beliefs. There is one important difference between imaginings and beliefs in this area, however: in contrast to the case of beliefs, sensory inputs can be of no help in allowing one access to the content of one’s imaginings, since as earlier remarks have suggested, for any proposition p, imagining that p is not reliably correlated with any particular sensory input. Nor, for any proposition p, is imagining that p reliably correlated to any form of nonverbal or verbal behavior (Ishiguro 1966: 157). I can imagine that I am flying to the moon whilst lying motionless in bed, walking, driving or even talking about something else completely. Hence it seems that behavioral information will be of little help in helping to ascertain what it is that one imagines. I suppose it might be suggested that one knows what it is that one imagines by “observing” the content of a mental image experience with some “inner eye”. However, it cannot be the case that the content of an imagining can be known by reference to the content of a mental image experience. This is for two reasons.

TOWER OF GOLDBACH 115

First, having a mental image is not necessary for imagining that some state of affairs is the case. This can be shown from imaginative interactions with literary fictions. Contrary to what is often asserted, only relatively infrequently does one have mental images as one imaginatively engage with literary works. This is partly because literary fictions, even in descriptive passages, do not provide one with enough fictional detail to generate a mental image of what they describe. It is also because often fictions ask us to imagine states of affairs for which there is no appropriately associated image. For instance, where I imagine that a character is in a mental state, where in the actual world the state is not typically attached to any particular physiognomic behavior, I need not and do not associate any particular mental image with this imagining, since in such a case it is not clear what mental image I should invoke. This point can be generalized to cover any fictional proposition which describes a state of affairs to which no characteristic perceptual experience is typically associated. So, though the having of a mental image may accompany imagining in relation to fictions, it is not necessary to it. The second reason why it would be false to think that one could know what one imagines by reference to the content of a mental image experience is that, notoriously, the content of an image on its own is always undetermined. An image on its own is consistent with multiple interpretations: is that a snow storm or a nuclear fall out? Is that a woman, or an automaton disguised so that it is indiscernible from a woman? And so on. Of course, in actual practice I do not need to ask such questions in order to establish what it is I imagine. Moreover, if I did need to ask them, I could never arrive at a satisfactory answer, since I would have access to no independent criteria of what counted as the “right” interpretation of my mental image. There is no question of a mental image experience grounding the content of what I imagine, since even where I have a mental image in the course of imagining, the content of what I imagine cannot be ascertained by reference to that image alone. So it seems that in the case of ascertaining what one is imagining, both in first and third personal cases, the inferences one is disposed to make play a particularly informative role. Indeed, it seems that there is no other potential source of information, since causal relations to external objects, sensory inputs and behavioral outputs have all been ruled out, as well as information available in the form of mental image experiences. This being the case, a problem arises for Gendler’s claim that one can imagine a conceptually impossible situation by focusing only on certain aspects of that situation. Recall that Gendler argues that one can imagine, say, that five righteous souls plus seven righteous souls does not make twelve righteous souls by thinking of “twelve righteous souls” as “number of righteous souls required for God to lift the decree” (Gendler 2000:69). Elsewhere she says that one can imagine that a banana is a gun by focusing “on certain similarities, such as shape, while ignoring others, such as internal complexity” (Gendler 2000:69).

116 KATHLEEN STOCK

(Actually, here Gendler explicitly is talking about pretending that a banana is a gun, which, presumably, involves imagining that a banana is a gun. I do not wish to deny that one can pretend that a banana is a gun, nor that one can imagine that I am holding a gun, whilst using a banana as a prop. I want to deny that one can imagine that a banana is a gun, in the sense that one imagines that there is an object such that it is both a gun and a banana.) It is a consequence of Gendler’s general position that, where p and q are not synonymous, one can justifiably claim to know that one is imagining that p, even though the inferences one is prepared to draw from what one imagines are the same as those one would be prepared to draw from imagining that q. I can know that I am having an imagining about seven righteous souls, even though I am prepared to draw inferences in imagination only to claims about the number of righteous souls required for God to lift the decree, and not to inferences involving the number seven in relation to the number five. I can know that I am imagining that I am holding a gun, even though the only inference I am prepared to draw from this imagining is to the imagining I am holding something with a gun-like shape. This being the case, the following epistemological question arises: where p and q are non-synonymous propositions, and where I claim to imagine that p, and yet am disposed only to make those inferences consistent with imagining that q, what epistemically justifies my claim that I am imagining the proposition p? When I claim to be imagining something about seven righteous souls, and yet am “focusing” in thought only on inferences connected to the thought of the number of righteous souls required for God to lift the decree, what justifies the claim that I am imagining something about seven righteous souls, and not rather about the number of righteous souls required for God to lift the decree? When I am disposed to make only those inferences that equally I would make from the imagining that I am holding something with a gun-like shape, what justifies my claim that I am imagining that I am holding a gun? It seems to me that the answer is: nothing. Having ruled out sensory inputs, behavioral outputs and access to mental images all as useless in giving one epistemic access to the content of one’s imagining, it seems that the imaginative inferences one is disposed to draw to and from an initial imagining that p are the only legitimate route to epistemic verification of what exactly it is that p stands for. This being the case, where one claims to imagine that p, and yet those inferences are compatible with imagining that q, nothing in particular justifies the claim that one is imagining that p over the claim that one is imagining that q. This leaves it indeterminate whether I am imagining that p or that q. So when it is claimed by Gendler that a person imagines that seven righteous souls and five righteous souls do not make twelve righteous souls, by “focusing” in thought on the proposition the number of righteous souls required for God to lift the decree does not make twelve righteous souls, I claim that in such a case nothing justifies the claim that one is imaginatively engaging with the former proposition rather than with the latter. Hence it is indeterminate which is being imagined.

TOWER OF GOLDBACH 117

This does not yet establish that we should say that it is the latter proposition that is being imagined, rather than the former. More generally, it is not yet established that we should say that a conceptually possible proposition is being imaginatively engaged with, rather than an impossible one. In the next section, I will offer some considerations suggesting that we should analyze apparent cases of imagining the conceptually impossible as cases of imagining the conceptually possible. Against imagining the conceptually impossible The first of these considerations is a straightforward argument from economy. In constructing our model of the mental, it is surely theoretically simpler, and hence desirable, if possible, to hold that imaginings are governed by the same, or at least, some subset of the same conceptual constraints that govern beliefs and other thoughts. A second source of considerations is other theoretical commitments we already have. For one thing, if we abandon the idea of imagining as constrained by our concepts, then this undermines the legitimacy of certain other theoretical commitments we have in philosophy. Traditionally in philosophy, it has been presumed that there is a close connection between what one is able to imagine, and the concepts that one has. Imagining has been thought of as an instructive means of ascertaining the limits of one’s concepts via the use of thought experiments. If Gendler’s arguments were sound, and imagining was not governed by conceptual constraints, then this strategy would be removed as a potential source of insight. Hence one reason to assume that imagining is confined to the conceptually possible is that it leaves a major and apparently extremely fruitful argumentative strategy in philosophy intact. The legitimacy of continuing to use thought experiments is not all that is at stake. In Off-line Simulation Theory, currently the most prominent attempt to offer an account of imagining in the terms of cognitive science, much of the weight of the conclusion that imaginings are “off-line beliefs”, processed through the same cognitive mechanisms as real beliefs, rests on the observation that the roles of beliefs and imaginings are similarly inferentially constrained. Off-line Simulation Theory argues that the activity picked out by our application of the expression “imagining that p”, where “p” describes some state of affairs, is to be understood as identical to the activity of feeding “pretend” or “quasi” or “offline” beliefs that p into some cognitive mechanism, issuing in some other pretend or quasi- or offline mental state with the content that q, where, had I fed the real belief that p into the same mechanism, or some other relevantly similar mechanism, a corresponding real belief with the content that q would have been issued, ceteris paribus. (Arguments for and against Off-line Simulation Theory are found in Davies and Stone 1995a, b.) Gendler’s claims are inconsistent with this conclusion, for they suggest that there is an asymmetry between the inferential roles of, say, the belief that there

118 KATHLEEN STOCK

are twelve bottles on the wall, and the imagining that there are twelve bottles on the wall. When I believe that there are twelve bottles on the wall, assuming that my concept corresponds to the standard concept of twelve, it will follow that I will also believe that the number of bottles on the wall is the sum of seven and five bottles; but if Gendler’s arguments were true, it would follow that I could imagine that there are twelve bottles on the wall without thereby having to imagine that the number of bottles on the wall is the sum of seven and five bottles. Hence a potential asymmetry opens up between the inferential roles of a belief that p and an imagining that p: the inferential role of the imaging that p does not replicate that of the belief that p, since that of the imagining that p need not be constrained by one’s current concepts, while that of the belief that p must be. Offline Simulation Theory depends on a symmetry between the inferential roles of beliefs and imaginings, which Gendler’s arguments undermine. So the plausibility of Off-line Simulation Theory as an account of imagining is fundamentally threatened by Gendler’s claims. Again, in tandem with the objections already stated, this provides further motivation to resist her conclusion. In this section I have offered two apparently compelling kinds of reason to analyze alleged cases of imagining the conceptually impossible as cases of imagining the conceptually possible: an argument from economy, and two arguments citing other theoretical commitments. The kind of strategy I advocate will not work, however, unless I also am able to give a detailed analysis of what it is one is doing when it seems to one that one is imagining the conceptually impossible. I turn to this task in the next section. Possible defeaters What, then, is it that one is imagining, when it seems to one that one is imagining a conceptual impossibility; for instance, as one reads Gendler’s story? Michael Tye, following a suggestion of Stephen Yablo (Yablo 1993), has suggested that one’s interpretation of the content of a particular imagining is subject to possible “defeaters”, in the light of which one should reassess one’s view (Tye 1995: chapter 10). Whether or not there are possible defeaters to a particular interpretation of the content of an imagining is partially determined by what else the imagining subject is prepared to say about what she imagines. If a subject claims that she imagines that water is not H2O, and: in response to questions, she replies that she is imagining a world in which there is a colorless, tasteless liquid that comes out of taps and fills lakes but that is not H2O. Now we have a possible defeater…it is not unreasonable to suppose that she is really just imagining that something superficially resembling water is not H2O rather than water itself is not H2O. (Tye 1995:186)

TOWER OF GOLDBACH 119

Tye thinks that possible defeaters should be postulated only in cases where one claims to be able to imagine something that is metaphysically impossible (for an apparently similar view see Kripke 1980:113). I deny this claim. The existence of metaphysical facts of which I am unaware on their own do not provide any good reason for one to deny that the content of my imagining is as I take it to be. If what I can imagine was limited by what is metaphysically necessary, then it would follow that I could not imagine states of affairs that unbeknownst to me were metaphysically impossible. If this were true, then we would have to say that prior to the discovery of the chemical composition of water, one could not imagine that water had any other physical composition than H2O. This seems very implausible. The main reason for the postulation of possible defeaters is primarily philosophical: Yablo and Tye wish to forge a link between imaginability and metaphysical possibility, in order to salvage claims about metaphysical possibility made on the basis of imaginability. I, on the other hand, say that if salvaging such claims leads us so far from our intuitive grasp of the concept of imagining, so much the worse for those claims. Despite this, I wish to appropriate an amended notion of “possible defeaters” for my purposes. Though I deny that we should search for a possible defeater to a claim that one is imagining a metaphysically impossible situation, a better ground for postulating possible defeaters to an account of imagining is where the subject claims to imaginatively engage with a proposition that is conceptually impossible. When someone sincerely claims to be able to imagine a conceptually impossible situation, I claim that usually there will be a conceptually possible defeater to their interpretation of the content of what she imagines. Upon questioning, her description of what she imagines will be consistent with a description of a conceptually possible situation; in which case, for the reasons I have given in the previous section, we should attribute the latter to her as the content of her imagining. In her story, Gendler purports to show that it is possible to imagine that twelve is and is not the sum of seven and five by inviting her readers to imagine that Solomon, with the support of God, decrees that this was the case. This being the case, several possible defeaters to the reader’s claim “I imagine that twelve is and is not the sum of seven and five” immediately occur. Possible defeaters to this analysis might be that one is imagining as follows. (1) That Solomon said that twelve is and is not the sum of seven and five. (2) That most people believe that twelve is and is not the sum of seven and five. (3) That one asserts that twelve is and is not the sum of seven and five. (4) That one believes that one has epistemic warrant for accepting that twelve is not the sum of twelve and five. (Walton finds himself “strangely tempted” by the thought that one can imagine accepting that genocide is morally permissible, but that one cannot imagine that it is true that genocide is morally permissible. I take it that in imagining accepting that genocide is morally

120 KATHLEEN STOCK

permissible, one imagines that one believes that one has epistemic warrant for accepting that genocide is morally permissible (Walton 1994a:49).) (5) That the utterance “twelve is and is not the sum of seven and five” refers to a true state of affairs, though it does not have the meaning we attribute to it in the actual world. (Yablo discusses a sense of “conceivable” according to which p is conceivable “if one can imagine believing something true with one’s actual p-thought” (Yablo 1993:24).) I suggest that, if a reader thinks that she is imagining that twelve is and is not the sum of seven and five as she reads Gendler’s story, the content of her imagining rather should be explicated in terms of one of these or some other conceptually possible “defeater”, where the precise choice will depend on what else the person is disposed to think or say about what she imagines. As a general rule, if a person claims to imagine that p, and on further questioning she can describe the content of her imagining only in terms of proposition q, where q is conceptually possible and consistent with the falsity of p, then she has not yet established that she imagines that p. In this case a possible defeater to the supposed content of her imagining will be q. If I claim to imagine that twelve is not the sum of seven and five, and upon questioning describe the content of what I imagine in terms of the description “Solomon decrees that twelve is not the sum of seven and five”, then since the latter proposition is consistent with the falsity of the former, the content of my imagining is most coherently interpreted in terms of the latter proposition. No decrease in the value of imagining It might be feared that in denying that one can imagine the conceptually impossible, I am somehow curtailing the freedom of the imagining subject, by placing overly rigid constraints on what we can imagine, and thereby decreasing the potential value that imagining as an activity may have. However, this is not the case. Even if it is true, as certain theories of concepts claim, that concepts are partially governed by individually necessary conditions, it remains true that most of the properties we attribute to objects in the actual world are contingent, so that we are free to imagine their absence at leisure. We can imagine them away or attribute them to unfamiliar objects with which they are (contingently) never associated in the actual world. Hence there continues to be great scope for imagining situations unlike the ones we are used to experiencing. Furthermore, if the Prototype Theory of concepts is true, then concepts have no individually necessary conditions (though they do have disjunctively necessary ones); this brings still greater imaginative freedom. Another residual worry might be that my arguments entail that one cannot imaginatively engage with certain fictions, interaction with which has up until now been thought of as both possible and extremely rewarding. Gendler cites propositions such as the owl and the pussycat went to sea in a beautiful pea-green boat,

TOWER OF GOLDBACH 121

and Frosty the snowman sings as expressing conceptual impossibilities that nonetheless one obviously can imagine to be the case (Gendler 2000:70). Even granting that the concepts snowman, owl and pussycat have individually necessary properties as constitutive conditions of their application, it is not clear to me that these propositions describe conceptual impossibilities. The most plausible candidate for the necessary condition governing snowman is made of snow; and this does not seem conceptually incompatible with singing, even if a singing snowman is a physical impossibility. Similarly, as far as I can see, for any properties plausibly proposed as necessary to the concepts owl and pussycat, going to sea is not inconsistent with the fulfilment of those properties. Meanwhile, if it is true, as the Prototype Theory states, that such concepts have no individually necessary but only disjunctively necessary conditions, then it seems even less plausible that these propositions express conceptual impossibilities. Of course, certain fictional propositions do unambiguously express conceptual impossibilities; for instance, in Through the Looking Glass, Lewis Carroll asks the reader to imagine that there is jam tomorrow and jam yesterday—but never jam today (Carroll 2000:206). It seems to me that usually in this sort of case the author is playfully drawing the reader’s attention to the unimaginability of such propositions, rather than seriously asking her to entertain them. Whatever the intention, in such cases, where a fiction requires one to imagine that p and p expresses a conceptual impossibility, then one is nonetheless able to imagine a closely related, conceptually possible scenario without any loss of value or enjoyment in one’s imaginative activity. I can imagine that the Queen had jam yesterday and will have jam tomorrow, but not today; and then I can imagine that tomorrow the Queen will have had jam the day before, and will have jam the day after that. But this is not to be confused with imagining what Carroll asks us to imagine: that both of these states of affairs might be true at the same time. I take it that some of the fundamentally important features of our imaginative engagement with fictions are their ability to extend our notion of possible states of affairs; to enrich our conceptual apparatus; to allow us to explore our emotional and cognitive reactions to possible combinations of events we have not encountered or thought of before; and to learn about the world, or other worlds. On the other hand, I see no possible cognitive benefit from imagining, or at least, trying to imagine, conceptual impossibilities. Hence I conclude that no impoverishment of the role of fiction in our lives is implied by our denying that we can imagine the conceptual impossibilities that fictions occasionally (and usually playfully) “ask” us to imagine. “Imaging” the conceptually impossible An objection to the account I have offered might be that I place too much emphasis on only one sense of imagining. There is another sense of imagining, it might be argued, where it is possible to imagine the conceptually impossible. Just as it is possible to have a visual experience as of S, where S is a conceptually

122 KATHLEEN STOCK

impossible scenario, so too is it possible to imagine that S, in the sense of have a mental image of S, where S is a conceptually impossible scenario. Just as, it is argued, concepts and perception come apart, so too may having a mental image and one’s conceptual scheme come apart. One response to this is that mental content that is nonconceptual does not therefore count as conceptually impossible content. This seems plausible: conceptually impossible content is content that violates the conditions upon the concepts it purports to utilize; and this is not the case with nonconceptual content, since nonconceptual content does not purport to utilize any concepts at all. In any case, the analogy between visual perception and mental imagery upon which this objection relies is insufficiently strong to warrant the objection’s conclusion. The content of one’s visual experience as of p (typically) is determined at least partially by the objectively detectable visual characteristics of some object in the world. Furthermore, the objectively detectable visual characteristics of objects in the world may exist in combinations not allowable in terms of one’s conceptual scheme. It is for this reason that it is conceivable that one might have a visual experience as of p, externally caused by the kind of situation described by p, where prior to that experience one had thought it conceptually impossible that p. For instance, it might follow from one’s conceptual scheme that it is conceptually impossible that what looks moving also looks still; yet when one looks at the “Waterfall illusion”, where an after-image of a waterfall is projected onto a stationery object, one might have a visual experience as of a moving yet still object (Crane 1988:142). In contrast, in the course of imagining, the content of a mental image is not determined by objectively detectable visual characteristics of objects in the world in the same way as is the content of visual experience. Though, of course, my mental image as of p may “have” certain characteristics F and G in virtue of the fact that at some prior time I have perceived an actual p or ps exhibiting F and G, usually I am not forced to picture p with F and G merely on the grounds that p’s exhibit F and G in real life. Mental images in the course of imagining (as opposed to remembering) are not directly dependent upon features of the world in the way that visual experience is. Hence there is no reason analogous to the reason in the visual case for claiming that the content of a visual image can outrun one’s conceptual capacities. This is reflected in the fact that when I have a visual experience, it may make sense for me to ask “what is it that I am seeing?”—what are the objective facts about the visual characteristics of the object which is the source of my visual experience? In contrast, it does not make similar sense to ask of a visual image experience, “what is it that my image is an image of?” This is because mental imagery is always had under an interpretation; one cannot have a mental image where one has no idea what it is a mental image of (Wollheim 1984:63). The content of a mental image of a p, at least in the course of imagining, is determined by one’s conceptualization of it under the description p, and not by any external

TOWER OF GOLDBACH 123

facts about some actual p (though things may be different with memory images, and in particular after-images). For this reason, though, in virtue of the relation of visual experience to external objects, it is plausible that the content of visual experience may outrun one’s conceptual scheme, it is not by the same token plausible that the content of a mental image may outrun one’s conceptual scheme. One may rejoinder that, having seen the Waterfall Illusion, one can have a mental image as of a still- yet moving-looking object, where one takes such a scenario to be conceptually impossible. Yet it is not clear that, where one’s conceptual scheme deemed such a scenario impossible, prior to the encounter with the Waterfall (or some other similar illusion) one could have had a mental image as of an object looking still and moving at the same time. And, where one is able to have such an image after an encounter with the illusion, rather than analyze such an image as an image as of a conceptually impossible scenario, it seems more plausible to argue that an object’s looking still yet moving is now conceptually possible for that person, and hence imageable: that the encounter with the illusion has extended her conceptual scheme. The illusion has (literally) shown her that such a thing is possible. (The illusion shows how it is possible that something can look still and look moving at the same time; not that it is possible that something is still and moving at the same time. Similarly, the drawing Ascending and Descending by M.C.Escher does not show that it is conceptually possible that there can be an endless staircase that both ascends and descends, but that something can look as if it is such a staircase.) I conclude that the objection under consideration does not present us with good reason to suppose that one can imagine that p, in the sense of have a mental image of p, where p expresses a conceptually impossible proposition. This is both because nonconceptual content is not equivalent to conceptually impossible content and because in any case it is implausible that mental images have nonconceptual content. In this chapter I have examined the claim that one can imagine the conceptually impossible against a background of recent developments in theories of concepts, and have pointed out that the claim sits uneasily against many currently prevalent views. I then have attacked Gendler’s model of what happens when one supposedly imagines the impossible from a different angle, by arguing that the claim that one “focuses” one’s attention on only certain aspects of an imagined situation undermines the epistemic justification for the claim that one is imagining that situation at all. I have offered reasons to analyze apparent cases of imagining the conceptually impossible as cases of imagining the conceptually possible, and outlined a model of how this might be done. Since, as I have argued, this strategy is simpler than Gendler’s, results in no reduction in the value or enjoyment potentially experienced through imagining, whilst safeguarding existing theoretical commitments in a number of areas of philosophy, I suggest that it is the correct one.

7 ON THE RELATION BETWEEN PRETENSE AND BELIEF Tamar Szabó Gendler

Introduction Overview By the age of two, children are able to engage in highly elaborate games of symbolic pretense, in which objects and actions in the actual world are taken to stand for objects and actions in a realm of make-believe. These games of pretense are marked by the presence of two central features, which I will call quarantining and mirroring (see also Leslie 1987; Perner 1991). Quarantining is manifest to the extent that events within the pretense-episode are taken to have effects only within that pretense-episode (e.g. the child does not expect that “spilling” (pretend) “tea”1 will result in the table really being wet), or more generally, to the extent that proto-beliefs and proto-attitudes concerning the pretended state of affairs are not treated as beliefs and attitudes relevant to guiding action in the actual world. Mirroring is manifest to the extent that features of the imaginary situation that have not been explicitly stipulated are derivable via features of their real-world analogues (e.g. the child does expect that if she up-ends the teapot above the table, then the table will become wet in the pretense), or, more generally to the extent that imaginative content is taken to be governed by the same sorts of restrictions that govern believed content. At the same time, from the same early age, both quarantining and mirroring are subject to systematic exceptions. Quarantining gives way to its opposite—I call this contagion—most strikingly in cases of what I call affective transmission (e.g. a child who imagines a bear on the staircase may be reluctant to go upstairs alone), but also in cases of what I call cognitive transmission (e.g. a child who has been playing at “birdwatching” may perceive a partially hidden squirrel in a nearby tree as having birdlike features). And mirroring gives way to its opposite— I call this disparity—as a result of the ways in which imaginary content may differ from believed content: in being incomplete (e.g. there may be no fact of the matter (in the pretense) just how much tea has spilled on the table), and in being incoherent

PRETENSE AND BELIEF 125

(e.g. it might be that the refrigerator serves (in the pretense) as a mathematicaltruth inverter). There is a tendency among philosophers and psychologists to subscribe, tacitly or explicitly, to a conception of pretense that gives pride of place to quarantining and mirroring.2 Ordinary successful pretense, on such a picture, involves segregated off-line processing and is hence unlike belief in its motivational force and cognitive contributions (thereby exhibiting quarantining); but it involves the processing of belief-eligible content and is hence like belief in its subject matter (thereby exhibiting mirroring). Much important headway has been made in explaining these striking features of pretense (in the work of Currie, Goldman, Gordon, Harris, Leslie, Lillard, Nichols and Stich, Perner, Walton, and others). But such accounts leave other equally conspicuous features of pretense largely unexplained. While ordinary pretense is, in many ways, unlike belief in its motivational force and cognitive contributions, the differences are not absolute. In particular, there are at least two sorts of cases where the contributions to subsequent cognitive processing made by imagining p and believing p differ at most in degree (thereby exhibiting contagion): in their role in evoking affective responses (affective transmission), and in their role in activating schemata, generating attentional filters, and altering evidential standards (cognitive transmission). And while the subject matter of ordinary pretense is, in many ways, like the subject matter of belief, the similarities are not absolute. In particular, there are at least two sorts of ways in which the content of what is potentially imagined may differ systematically from the content of what is potentially believed (thereby exhibiting disparity): in its potential to remain deeply underspecified (incompleteness), and in its potential to involve contradictory content (incoherence). Standard conceptions of pretense emphasize the ways in which ordinary cases of pretense exhibit quarantining and mirroring (and hence resemble belief-like states in terms of subject matter but differ from them in terms of motivational force and cognitive contributions), but they tend to neglect the ways in which ordinary cases of pretense also predictably manifest both contagion and disparity (and hence differ from belief-like states in terms of subject matter but resemble them in terms of motivational force and cognitive contributions). But, I will suggest, it is fully consonant with our general knowledge of how human beings learn from and act in the world that normal cases of pretense should exhibit not only high degrees of quarantining and mirroring, but also that they should exhibit elements of contagion and disparity. Indeed, as I will argue later, the sources of all four tendencies can be traced to fundamental aspects of our cognitive architecture. Preliminary remarks Quarantining and contagion seem to be principally a matter of near-universal largely automatic processes: our general incapacity to believe at will is a manifestation of the degree to which quarantining occurs beyond the reach of our

126 TAMAR SZABÓ GENDLER

cognitive control, and our somatic responses to affect-laden stimuli—whether real or imaginary—are a manifestation of the degree to which the same is true of contagion. At the same time, higher-order exploitation of our susceptibility to contagion is one of the key elements of “self-help”—as evidenced by examples ranging from Pascal’s trick of coming to believe in God by acting as if one already did, to the advice that fills books like Success is a Choice: Ten Steps to Overachieving in Business and Life (Pitino with Reynolds 1997). And a second-order capacity for quarantining (keeping track of what is real and what is merely imaginary) plays a central role in the guidance of action. Deviations from these general patterns seems correlated with wide-ranging disturbances in a variety of cognitive and emotional skills: for example, an extreme tendency towards contagion tends to be associated with systemic forms of psychopathology such as schizophrenia; and complete immunity from (affective) contagion seems to occur only among those with damage to the ventromedial prefrontal cortex. Mirroring and disparity, by contrast, seem to be the result of capacities exercised primarily at the level of cognitive control: our general sense of the “free play” of the imagination is a manifestation of the degree to which mirroring exploits the flexible nature of symbolic representation, and our corresponding ability to countenance incompleteness and inconsistency within the context of imaginative exercises is a manifestation of the degree to which the same is true of disparity. At the same time, certain involuntary constraints seem to govern both: these are manifest in the case of mirroring by the sense of “naturalness” connected with certain associations, and in the case of disparity by the corresponding sense of incoherence. Again, deviation from these general patterns tends to be correlated with systemic disturbances: the inability to balance mirroring with disparity seems uniquely associated with the cluster of incapacities that characterize autism. In this chapter, I will do little more than to sketch, in preliminary form, what the four features amount to, tentatively proposing several hypotheses concerning the relations among them, and suggesting a number of avenues for further investigation. In the first section, I will provide some developmental background. In the main sections, I will spell out in somewhat more detail the ways in which quarantining/contagion and mirroring/disparity are manifest in young children’s games of pretense, and how analogues to them govern more complicated exercises of imagining. Two final preliminary notes: first, although there are contexts in which it is important to distinguish imagining and imagination on the one hand from pretending or pretense on the other, the distinction is unimportant for the purposes at hand; I will thus use the terms largely interchangeably. Second, it is a complicated and important philosophical question what precisely belief that p amounts to. For the purposes of this chapter, I am interested only in one small aspect of this issue, namely the contrast between belief-like attitudes on the one hand, and pretense-like (or make-belief-like) attitudes on the other. Of the key differences between these attitudes, two are, for the purposes of this

PRETENSE AND BELIEF 127

chapter, particularly salient. The first is that whereas belief is what might be called a receptive attitude (the agent takes herself to be responding to something about the world itself), pretense is—at least at its core—a productive attitude (the agent takes herself to be projecting something onto the world). The second is that belief is— at least at its core—an attitude intimately connected (via desire) to action (one who desires B, and believes that doing A will bring about B will, ceteris paribus, do A), whereas pretense is—for the most part—disconnected from action (one who merely makes-believe that doing A will bring about B will not generally, for that reason, do A). So, generally speaking, when S believes that p, she holds p to be true, and she takes the truth of p into consideration when deciding how to act; by contrast, when S pretends or makes-believe that p, she does not (for that reason) hold p to be true, and so does not (for that reason) make decisions about how to act that take p into consideration. One indication of what one believes is what one is willing to assent to or assert in normal or high-stakes circumstances (i.e. what one reports oneself to believe). But nonverbal “reports”, in the form of actual or intended actions in normal or high-stakes circumstances, autonomic responses to certain stimuli, and other sorts of physiological or behavioral indicators may also be gauges of belief. (I suspect that it is the tendency to focus on the former to the exclusion of the latter— motivated, perhaps, by an instinctive avoidance of anything even tenuously associated with behaviorism—that has led many philosophers to an oversimplified picture of both belief and pretense.) In the discussion that follows, my use of the term “belief-like attitude” or, occasionally the shorthand “belief” will be both loose and stipulative. I will credit someone with the belief(-like attitude) that p if she acts—outside the context of an explicit episode of pretense—as p-believers generally do. That this is inadequate as a general account of belief should be so obvious as to make it clear that my ambitions in this regard lie elsewhere. Early childhood pretense: developmental background By the time normal children reach the age of fifteen months, they are capable of engaging in primitive games of make-believe—acting, for instance, as if a piece of cloth or coat collar were their special bedtime pillow (instances of unconscious symbolic representation may occur much earlier—see Piaget 1945/1962:96 and chapters 6 and 7). By eighteen months, many show signs of tracking rather elaborate games of pretense initiated by others—for instance, being able to identify which of two dolls that have been “washed” by an adult experimenter is “still wet” and engaging in the requisite “drying” activity (Walker-Andrews and Kahana-Kelman 1997). By twenty-two months, these skills become quite widespread (Harris and Kavanaugh 1993; Harris 2000: chapter 2), and by 24–28 months, most children are able to participate fully in such games—for example, pouring “tea” for a stuffed cow from an empty plastic teapot, feeding a toy pig some “cereal” from an empty bowl, giving a toy monkey a “banana” when there

128 TAMAR SZABÓ GENDLER

are no (real) bananas in sight, and so on (Walker-Andrews and Kahana-Kelman 1997; Harris 1994, 2000: chapter 2). Around this same age (24–28 months), children show themselves readily able to generalize on the basis of others’s pretend stipulations—if they are told, for instance, that a particular yellow block represents a banana and a particular red block represents a cookie, they require no further prompting to engage in a pretense where yellow blocks in general represent bananas, and red blocks in general represent cookies (Harris 2000: chapter 2 reporting Harris and Kavanaugh 1993; Walton 1990). They show themselves readily able to suspend such stipulations as soon as a new episode of pretense begins—the bricks that represent bananas or sandwiches in one game can without difficulty come to represent bars of soap or pillows in the next game. They show themselves ready to credit imaginary objects with causal powers much like those of their real-world analogues—if Teddy eats one of the (wooden brick) bananas, he will no longer be hungry; if he is bathed in a (cardboard box) bathtub, he will emerge wet (Harris 2000). And they are ready to describe situations from the perspective of the imaginary world—when asked to express what happened after (literally) an experimenter holds a stuffed animal in such a way that the animal’s paws grip an empty plastic teapot and hold the teapot above the head of some other stuffed animal, children are happy to report the event as: “Teddy poured tea on Monkey’s head” or “Monkey’s all wet— he’s got tea on his head” (cf. Harris 2000). During the year that follows, most children develop the capacity to engage in complex coordinated games of joint pretense with others (Perner et al. 1994:264). And well before the age of four, they have figured out how to keep track of different individuals simultaneously engaging in different games of pretense—recognizing, for instance, that if you pretend the pebbles are apples and I pretend the pebbles are plums, you will be baking an apple cake while I bake a plum cake (Perner et al. 1994:264). These capacities are accompanied by a parallel capacity to keep track of what is pretend, and what is not. The fifteen-month old does not give any indication that she comes to think that pieces of cloth are pillows. To the contrary, awareness of the merely pretend status seems explicit: even as she indicates the pillow-like status of the cloth by rehearsing her “going to sleep” routine—lying down on her side and repeatedly closing her eyes—she emits a giggling “no-no” (Piaget 1945/1962: 96; Harris 1994:257). Moreover, when bedtime comes around, she gives no indication that she expects the piece of cloth to be the surface on which she rests her head. That is, even with children as young as fifteen months, there is no indication of what Alan Leslie has termed “representational abuse”, that is, no indication that the child comes to believe that actual-world objects have or will come to have features of the pretend objects that they serve to represent (Leslie 1987). And by the age of three, children are able to articulate this difference— noting, for instance, that a child with a real dog will be able to see and pet the dog, whereas a child with a pretend dog will not (Wellman and Estes 1986; Estes, Wellman, and Wooley 1989; Harris 2000: chapters 2 and 4; see also Bouldin and Pratt 2001 and references therein). While there is some evidence that there are

PRETENSE AND BELIEF 129

instances when the real-pretend boundary is difficult for children (and even adults) to keep track of—in ways that I will discuss below—it is a crucial feature of games of pretense that, for the most part, the boundary poses no difficulties whatsoever. Children as young as fifteen months old are completely capable of recognizing that the world is one way (e.g. that there is a piece of cloth in front of them) and pretending that it is another way (e.g. that there is a pillow in front of them).3 Quarantining and mirroring It seems, then, that even very young children are able engage in games of actingas-if symbolic pretense that are carried out (a) without resulting in representational abuse, that is, without producing the expectation that real-world objects have the characteristics they are supposed to have in the context of games of make-belief, and without between-game permeability, that is, without stipulations from one game of pretense being automatically assumed to carry over to other games of pretense, and (b) without resulting in realm-mixing, that is, without producing the expectation that the actual world will be transformed in ways that accord with actions in the pretense. And it seems additionally that even very young children are able to engage in rule-governed games of pretense that require: (c) the capacity to understand and make use of generative rule-governed pretend stipulations and (d) the capacity to apply certain real-world causal relations to actions within the context of the game. These features can be captured by a pair of principles concerning quarantining on the one hand, and mirroring on the other. The (prop-based) principle concerning quarantining specifies how what is pretended affects what is believed: it says that when one pretends that X is Y, things that are believed to be true of Y do not come to be believed to be true of X merely because they are pretended to be true of X. The (prop-based) principle concerning mirroring specifies how what is believed affects what is pretended: it says that when we pretend X is Y, X is—in the pretense—taken to have the effects and features that Y is—in reality—believed to have. So, for example, the prop-based quarantining principle tells us that when a child pretends that a block is an apple, she does not thereby come to expect that the features attributed to the block within the pretense episode (e.g. being sweet and edible) will hold of the block in reality (or, more generally, in any scenario outside the pretense episode). And the prop-based mirroring principle tells us that when the child pretends of an empty teacup that it is a full teacup, the attributes believed to hold of the full but not the empty teacup prior to engaging in the pretense (e.g. being filled with a drinkable liquid) will be pretended to hold of the empty teacup while engaging in the pretense. What holds for prop-based pretense holds more generally. One might generalize the quarantining principle to say that things do not come to be believed merely because they are pretended; and one might generalize the mirroring principle

130 TAMAR SZABÓ GENDLER

to say that things that are pretended are the sorts of things that could be, in principle, believed. If the quarantining and mirroring principles held universally, the real-pretend boundary would be completely permeable in the real-pretend direction, and completely non-permeable in the pretend-real direction. Pretending and believing would be exactly alike, except that one would take place “on-line” while the other took place “off-line”. For the mirroring principle would guarantee that the same content p could be the object of a belief or of a makebelief, and that no content would be such that it could be the object of only one or the other attitude. And the quarantining principle would guarantee that makebelieving p could never, in itself, bring one to the belief that p. Contagion and disparity But the quarantining and mirroring principles do not hold universally. In regular and predictable ways, merely pretending p does seem to cause (what can be characterized as something sufficiently like) the belief that p; and in regular and predictable ways, pretending that p does not seem to be the same thing as makingbelieve what one would, as a matter of fact, believe if p were (actually) the case. That is, successful pretense is also characterized by the contagion principle: some things do come to be believed—or treated as if they were believed—merely because they are pretended, as well as by the disparity principle: when one engages in the pretense (imagines) that p, then what one makes-believe (imagines) to be true in the make-believe world differs in some way from what one believes would be true in the actual world if p. Contagion and disparity occur in cases of pretense that are manifestly nondefective: cases where the pretender is explicitly aware that she is engaged with a realm of make-believe, and where the make-belief is elaborate and rich. That is, contagion and disparity are features of successful pretense, albeit features with an importantly circumscribed role. The systematic project of articulating exactly what that circumscription amounts to will need to wait for another venue. In the remainder of the chapter, I merely offer some data points around which a successful theory would need to be built. Examples of contagion Affective contagion Contagion is most strikingly manifest in cases involving what I call affective transmission: cases where mere contemplation of an emotionally charged situation causes the thinker to behave in a way consistent with the belief that the situation is sufficiently probable so as to influence prudent behavior. So, for example, a child who has been pretending that there is a monster in the shower may well be

PRETENSE AND BELIEF 131

reluctant to enter the bathroom (as may an adult who has just seen Psycho). This common experience has been noted by a number of philosophers (see, e.g. Hume 1739/1978:1, iii, 13:248; Ryle 1949:259; Price 1960/1969:308–9), and is borne out by laboratory research. In one typical experiment, children were shown and permitted to inspect two opaque empty boxes, and then asked to imagine either that one of the boxes is occupied by a nice and friendly rabbit, or that one of the boxes is occupied by a mean and horrible monster. As expected, children’s verbal reporting exhibited typical features of quarantining: when asked whether there was really a monster or rabbit in the box, the children were readily able to confirm that they were “just pretending” that there was. But subsequent nonverbal responses to the situation were more complicated. The experiment continued with the researcher asking whether she might leave the room to get the child a little gift. In four cases —all cases where the child had been asked to imagine a monster—the child was unwilling to let the researcher depart, despite repeated verbal and visual reassurance concerning the box’s emptiness; in the remaining cases, the researcher stepped out, and (videotapes reveal) nearly half of the children opened one or both boxes, showing a marked tendency to focus on the box containing the imaginary creature. When, subsequent to her return, the experimenter asked the children about their action, a considerable proportion of the children who had opened the boxes maintained that they had done so because they wondered whether, after all, there was something in the box (Harris et al. 1991; Harris 2000: 173–80). Although the interpretation of these data is not uncontroversial (see Bourchier and Davis 2000a, b), it seems likely that affect plays at least some role. If, for instance, children are asked to imagine that there is a pencil in a box when there are no other pencils in the room, and a person comes into the room looking for a pencil, children show no inclination to hand the visitor the box with the imagined pencil in it. (Whether this outcome is the consequence of some sort of cognitive override is, to my knowledge, an open question: I am not aware of research concerning the question of whether children in this situation exhibit some sort of momentary hesitation during which, e.g. they look toward the box but decide not to reach for it—and, if so, whether this hesitation can itself be traced to some sort of affect-based response. If so, this would show the phenomenon of contagion to be more rather than less widespread, supporting further the main theses of this chapter.) In any case, examples where there is a clear failure of override tend to be cases that are emotionally charged (in the sense that they involve either issues of personal safety, emotional significance, or empathy), at least according to studies done so far. In a widely reported study performed by Rozin and Nemeroff, adults were presented with two bottles, and invited to pour sugar into each one. Subjects were then asked to affix a “sugar” label to one bottle, and a “sodium cyanide” label to the other. Although subjects were happy to report that both bottles contained the same thing, namely sugar, and happy to concede that the choice of

132 TAMAR SZABÓ GENDLER

labels was purely arbitrary, many nonetheless showed a marked reluctance to eat from the bottle labeled “cyanide” (Rozin and Nemeroff 1991; Lillard 1994:221). Although this tendency might be overridden in a high-stakes situation—it would be surprising to hear that a subject was unwilling to eat from the “cyanide” bottle even in exchange for a significant sum of money, or unwilling to make use of its contents when sugar was apparently required for some important purpose— its existence in the low-stakes situation is nonetheless striking. For even if the instinct to take the cyanide label as reporting some actual fact about the world is superseded by the realization that the label is misleading, the instinct is nonetheless there to be overcome. (For an alternative analysis of the case, see Velleman 2000:276, n. 65—since our uses of “belief” differ in precisely the ways relevant to our apparent disagreement, our views may be less distant than they initially seem.) It is possible, of course, that the presence of the label plays some role in the subject’s reasoning, providing apparent evidence that is processed as realityindicative until it is overridden by the recollection that it is not. This hypothesis accords well with much of the heuristics and biases literature, which seems to suggest that we initially process all information as if it were a sign of the background circumstances normally associated with such a phenomenon (think of how easily we are jolted into action by a watch we know to be five minutes fast). But even if this is the explanation, it remains the case that some overridden cue— the “danger” indicator invoked by the presence of the “cyanide” label—feeds into the agent’s decision-making system in a way that allows it to play an actionguiding role. So while we may not have isolated the precise source of motivation in the sugar/cyanide case,4 it seems clear that some feature of the situation that the agent recognizes (perhaps on reflection) to be merely imaginary nonetheless plays some role in guiding her behavior. And, indeed, other research suggests that similar contagion occurs in cases where the imaginary cue is purely internal; moreover, it suggests that (nonoverridden) contagion occurs in circumstances where the subject is directly emotionally affected by the imaginary scenario, and where the emotional involvement concerns the avoidance of risk. In a study of British voters, Nigel Harvey instructed subjects to pretend that they were supporters of a certain political party (party A) and then asked them whether they would undertake a slightly devious action (lying to a pollster) that would benefit the party they imagined themselves to be supporting (party A) and harm another party (party B) (Harvey 1992). Some of those asked to engage in the pretense actually were supporters of party A (actual A-supporters); others were actually supporters of the rival party B (actual B-supporters); still others actually supported neither party A nor party B (neutrals). When the study was conducted during a nonelection period, there were no differences among the three groups: 80 percent of actual A-supporters, 80 percent of actual B-supporters, and 80 percent of neutrals reported that they would (in the pretense) undertake the A-benefiting action. By contrast, when a similar study

PRETENSE AND BELIEF 133

was conducted during the election period, the results differed strikingly: 70 percent of actual A-supporters and 70 percent of neutrals reported that they would (in the pretense) undertake the A-benefiting action—but only 40 percent of actual B-supporters reported that they would do so. (Note that in the first study, “party A” was the conservative party whereas “party B” was the labor party; in the second study, the roles were reversed. Harvey argues convincingly that this difference is immaterial to the studies” outcomes, as a pair of studies in which both parties played both roles would presumably show.) Harvey suggests that this may be a case of “decoupling failure” where “wishful thinking impairs belief-desire reasoning.” The reluctance of actual B-supporters even to pretend to engage in A-benefiting action presumably stems from a rather complicated process of anticipated contagion. Affective contagion occurs when mere contemplation of an emotionally charged situation causes the thinker to behave as if she believed the situation to obtain (or at least to be somewhat likely). In this case, the B-supporting subjects are presumably reluctant to engage in any sort of pretense concerning (their contribution to) A’s success, for fear that it might produce in them a belief that A has succeeded (a situation normally associated with A’s actually having succeeded), or for fear that it might weaken their actual commitment to B by producing in them a belief that A should succeed. (It is interesting to note that contagion seems more readily sparked by feared outcomes than by desired outcomes: note that while there is a sharp disparity between the willingness of actual B-supporters to pretend to engage in Abenefiting actions, there is no difference between the willingness of A-supporters and neutrals.) Finally, as the impatient reader no doubt has been eager to point out, we need not turn to laboratory experiments for examples of affective transmission. Sexual fantasy provides a rather striking example of the phenomenon: merely imagining a sexually arousing situation typically results in genuine sexual arousal. (As the existence of pornography reminds us, the effect is even more profound in cases of prop-based pretense.) Psychoanalysis provides another almost endless source of cases: Goethe’s earliest memory—of delightedly throwing crockery out the window and watching it smash on the streets below—is, according to Freud, behavior motivated by a fantasy of throwing his baby brother out the window, thereby ridding himself of his infant sibling rival (cf. Velleman 2000:266). Finally, the phenomenon that philosophers call the “paradox of fictional emotions”—that we seem, prima facie, to feel real emotions for characters we know to be fictional— suggests another general realm in which affective transmission occurs (see Gendler forthcoming). Cognitive contagion A second source of contagion can be found in the phenomenon of cognitive transmission; cases where mere contemplation of some emotionally neutral imaginary scenario causes the thinker to become (over-) sensitive to similar

134 TAMAR SZABÓ GENDLER

scenarios in the actual world. It is a well-known phenomenon—observed by anyone who has ever been pregnant or had a broken leg or bought a new car—that one’s attention to and evidential standards concerning the world are sharply affected by what is “on one’s mind”. When one is in a particular situation (e.g. pregnant, broken-legged, having just bought a new Subaru) one seems to observe an unexpectedly large number of others who are also in that situation. Three distinct explanations account for this phenomenon. First, one may be spending more time in settings that appeal to individuals with this condition, so one may indeed be encountering more people with broken legs or new Subarus than one had been previously. But this is not the full explanation. In addition, it is likely that—without consciously realizing it to be the case—one is more sensitive to the genuine markers of that condition, so that one attends to and conceptually processes more instances of pregnancy or new Subaru-hood than one would have otherwise. On top of this, one may well—again without realizing this to be the case—lower one’s standards of evidence for concluding that an individual is in that condition, so that one attributes the condition (perhaps falsely) in instances where, even had one attended to them previously, one would have withheld attribution. These various phenomena—raising to salience, affecting attention, altering evidential standards— are universal features of our cognitive behavior: “priming” and thinking-about affect both pattern-recognition and perceptual interpretation. In this light, it is not surprising that imagining p could have the sorts of carryover effects that it seems to, even in non-affect-laden cases. So, for example, if I spend the morning imagining that I am birdwatching, I will be more likely to attend to actual bird-encounters in the afternoon. Moreover, I may be ready to conclude that something is a bird on much thinner evidence than I would have had I not engaged in the pretense (e.g. hearing a rustle in the tree outside my study window may be sufficient to convince me that the sound has been made by a bird). And my imaginative engagement with birdwatching may even affect the patterns I perceive—I may actually see things differently as a result of what I have imagined. This phenomenon is familiar from psychological research on perception, and from the research on heuristics and biases. For example, the Availability Heuristic describes precisely such a tendency towards cognitive contagion: a tendency to make judgments concerning the likelihood or relative frequency of events or objects on the basis of the “availability” of such objects or events to memory, perception, or—Kahneman and Tversky explicitly note—imagination (Tversky and Kahneman 1972/1983:178). And research on attention suggests that perception is heavily dependent on the explicit or tacit intentions with which one approaches the perceptual scenario—intentions that may be equally well stimulated by imagining as by some other cognitive activity. In short, the evocation of perceptual and evaluative schemata is relatively indifferent to whether the evocation occurs as the result of something in the ambient environment, something in memory, or something brought to mind

PRETENSE AND BELIEF 135

merely as the result of imaginative rehearsal. In all three cases, the consequent availability of the object, event, or schema plays a central role in subsequent attention, perception, and even reasoning. Summary: contagion It seems clear that episodes of pretense typically exhibit features of contagion as well as features of quarantining. While we have a marked capacity to distinguish the imaginary from the real with regard to certain of our apparatuses for cognitive processing, others seem relatively indifferent to the question of whether the subject matter with which they are concerned was generated by the world or by the mind. Among these are the two families of cases discussed above: cases of affect-evoking imagination, and cases of schema-evoking imagination. In both families of cases, the traditional picture—that successful instances of pretense resemble belief-like states in terms of subject matter but differ from them in terms of motivational force—is misleading at least in terms of its second claim. (Recent research on eyewitness testimony suggests that such contagion occurs frequently in children and adults when imaginary episodes are misremembered as real (Ceci and Bruck 1993; Ceci and Friedman 2002; Loftus 1996).) Moreover, in neither case is this a consequence of some incidental or peripheral feature of our cognitive apparatus: cognitive and affective transmission result from features of the human mind as central as those that produce our capacity for quarantining (see Gendler forthcoming). Cognitive transmission is an inevitable by-product of fundamental processing features: the very features that make perception and information-processing possible make cognitive transmission inevitable. Without the use of schemata, attentional filters, and evidential standards adjustable on the basis of non-belief-based input, our finite cognitive and sensory apparatuses would be ineffective tools for making sense of the world around us. Affective transfer is a similarly inevitable by-product of similarly fundamental mechanisms. As research by Antonio Damasio and others has demonstrated, patients with damage to the ventromedial prefrontal cortex manifest a cluster of incapacities (Bechara et al. 1994; Damasio et al. 1991; Damasio 1995, 1999; LeDoux 1996). They lack autonomic responses to emotionally disturbing pictures (though they have no cognitive difficulty identifying such images, nor do they lack autonomic responses in general), and they reveal a marked tendency to engage in high-risk behavior (despite describing themselves as fully aware of its inadvisability). Together, these data seem to suggest that some sort of somatic realization of the potential consequences of a risky action is crucial to prudent decision-making: without it, the theoretical advantages of one or another course of action may be apparent, but these are not translated into action-guiding behavior. Without the capacity to feel something akin to real emotions in the case of merely imagined situations, we would be unable to engage in practical reasoning. What this means is that affective transmission is a fundamental feature

136 TAMAR SZABÓ GENDLER

of our cognitive architecture, and cannot be treated as straightforwardly and unqualifiedly deviant. Disparity The mirroring principle says that imagining and believing are attitudes with similar ranges of possible non-defective contents: the sorts of things that we are able to non-defectively imagine are the sorts of things that we could, in principle, nondefectively believe. But it seems clear that that non-defective imaginative content may differ from non-defective belief content in at least two ways: what I successfully imagine may be incomplete, in the sense that some of its features may remain permanently—even explicitly—unspecified and unspecifiable; and what I successfully imagine may be incoherent, in the sense that some of its features may be conceptually or logically incompatible. (I here skate over a number of important issues concerning the attribution of content.) The illusion that things are otherwise stems, I suspect, from intuitive reliance on a picture that treats imagining as just like belief, only off-line, and from a picture of prop-based pretense that treats principles of generation as complete, uniform mappings from one realm to another. Once we realize how complicated mapping relations are, even in simple children’s games of prop-based pretense, we can gain a sense for one of the mechanisms by which potentially imagined contents may come to differ from their belief-based counterparts, and hence how disparity might arise. Principles of generation As Kendall Walton has persuasively argued, games of make-believe are often governed by what he calls principles of generation—local conventions that govern how (certain) features of the actual world are to be mapped onto the imagined one. So, for instance, to pass the time on a long drive, we might agree—explicitly or tacitly—that cars “count as” lions and trucks as tigers. Once we have done so, certain fictional facts will obtain: if there is a car 50 feet from ours, it will be true in the fiction we have generated that there is a lion 50 feet from the lion we are riding—even if none of us notices it; if two trucks are traveling side-by-side, then two tigers are too. The rule-governedness of generative principles is part of what allows us to structure imaginative space in a way that we are able to make sense of its content. But even in simple cases of prop-based pretense, while some features are generative, others are not.5 The truck’s location may determine the location of the tiger, but other features of the truck—its carburetor, its mudflaps, its spare tire—may have no corresponding role to play in the realm of the pre-tense; and features of the pretense—the color of the tiger’s fur, the sharpness of its teeth, the length of its tail—may have no corresponding “base” in the realm of the prop. Consider, for example, how many different mapping schemes are employed even in an extremely simple pretend scenario, and how effortlessly the child moves among them.

PRETENSE AND BELIEF 137

The children watched as we introduced two animals, a monkey and a horse. We ‘fed’ the monkey with a yellow brick, explaining that he wanted some banana, and we ‘fed’ the horse with a red brick, explaining that he wanted some cake. Next, we introduced some more animals, telling the children whether the animal wanted banana or cake to eat. The two-yearolds almost invariably responded in accordance with the generativity principle. In ‘feeding’ the newly-introduced animals, none of them touched either the ‘banana’ or the ‘cake’ that had been given to the monkey and the horse. Instead, they spontaneously reached out and appropriately selected either another ‘banana’ or another ‘cake’ from two separate piles of red and yellow bricks available on the table. (Harris 2000:12) During the first part of the game, the pile of red bricks represents a pile of pieces of cake, and the child’s action in removing a brick from the pile represents selecting a piece of cake and carrying it to another location. Certain features of the bricks are understood to be relevant to play, others not. So, for instance, the location of the actual bricks indicates the location of the pretend cakes. The shape of the actual bricks may be taken to indicate the shape of the pretend cakes. But the color of the actual bricks may not indicate the color of the pretend cakes. Their lack of stickiness presumably does not indicate the lack of stickiness of the pretend cakes. Their density hopefully does not indicate the density of the pretend cakes. And the fact that they are made of painted wood surely does not indicate that the pretend cakes are composed of such ingredients. Think, moreover, about the complicated role that the color of the bricks plays in the imaginative exercise. That the bricks are colored red is important to the play— it is this that indicates that the bricks represent cakes rather than bananas. But there would be no problem with introducing a further stipulation in the game —that cakes at the top of the pile are vanilla and cakes at the bottom are chocolate. So to determine what sort of object a particular brick represents, appeal would need to be made to two sorts of properties: the color-properties of the brick indicates its category-membership—red bricks are cakes, yellow bricks are bananas—whereas the location-properties of one subset of the bricks (those red bricks lying in the initial pile) would indicate further of their features within the category—bricks at the top are vanilla, bricks at the bottom chocolate. Note how different these two sorts of mappings are. In the case of real-world items, visual and other sensory properties are indicators of features like banananess and chocolate-cakeness, whereas in this game, visual properties are indicators of the one (banana vs cake), whereas a property that could never criterially serve such a role in actual cases—location— distinguishes them along the other (chocolate vs vanilla). So far, we’ve only looked at the very first mapping—the one that makes the bricks in the red pile cakes. The next thing that happens is that the child moves one of the cakes to the “mouth” of the stuffed animal. Here, her actual-world

138 TAMAR SZABÓ GENDLER

motions serve to fix the pretend-world motions: the way she moves the brick from the pile to the “animal” is the way that the cake moves from the table to the horse. Note that unlike the brick, the child may well here represent herself both within and without the game: though she may be a zookeeper, or a horse-trainer, or a child on a nineteenth-century farm, she may also just be Sophie or Helena, engaged in an act of animal-feeding. So whereas the mapping that governed the relation between blocks in the real world and their imaginary-world counterparts took only a few of their elements—such as location and perhaps size and shape— the mapping that governs the relation between the child in the real world and her imaginary-world counterpart may take on nearly all of her features—at least at this stage. The next step in the story involves the child feeding the cake to the horse while the horse makes munching noises. Note that here, the mapping between the child’s actions and the events in the pretend world must simultaneously differentiate between the actions of her hands, and the actions of her voice. When the child holds the cake up to the stuffed animal’s mouth, the location of her hand indicates the location of the feeder’s hand in the story. But when the child makes the munching noises, the sound of her voice indicates the voice of the horse. In short, it seems that we are extremely flexible and adaptive about the principles of generation we use when we engage in exercises of prop-based pretense. Mapping may be partial, in the sense that only some of the features of the actual world entity are mapped into the pretend realm (see Fauconnier and Turner 1998); and mappings may be multiple, in the sense that features of many actual-world entities may be mapped onto a single entity in the pretend realm. Incompleteness and incoherence The result of partial mapping is that imaginary entities are potentially incomplete: they may explicitly lack determinates for their determinables; the result of multiple mapping is that imaginary entities are potentially incoherent: they may affirmatively bear incompatible properties. So, for example, I may imagine that it is between 50 and 75 miles from Lilliput to Brobdignag, with no further commitment that one rather than another of these distances is—even in the pretense—the distance between them; I may have a mental image of a spotted cow with somewhere between 10 and 20 spots, with no further commitment that one rather than another of these provides—even in the pretense—the genuine spot-count. Indeed, I may even commit myself to it being true that, in the pretense, no candidate determination correctly specifies the determinable in question. I might be committed to a view of the term “zillion”, for example, according to which a zillion is a one followed by a large finite number of zeros, but is determinately not equal to 10n, for any value of n. (Some find this last example rather unconvincing.) All of these, if successful, are examples of non-defective incomplete imagining.

PRETENSE AND BELIEF 139

In addition, non-defective imagining may be incoherent. I may imagine that seven and five both do and do not equal twelve (for a story to this effect, see Gendler 2000), or that there is a box that is both empty and not empty (for a story to this effect, see Priest 1999). The ease with which I am able to do so is a consequence of the ease with which “counting as” occurs in imagination: a number can “count as” the number 7 in a story, even if, in the story, that number in conjunction with 5 does not make 12; a box can “count as” empty in a story, even if it also contains something; the Scarecrow can “count as” a scarecrow in The Wizard of Oz, even if he is able to walk and talk. And when, in Mary Poppins, one of the Pleiades (Maia) comes down to do the Christmas shopping for herself and her six sisters, the little girl dressed in a sky-blue wisp of fabric “counts as” being a star, despite obvious discrepancies between their relatives sizes and chemical compositions (Travers, 1962:181–92). Without this sort of “cheapness”, the kinds of imaginative projects in which we engage—from childhood propbased pretense to sophisticated storytelling—would not be possible. Summary: disparity As with contagion and quarantining, episodes of pretense typically exhibit features of disparity as well as features of mirroring. While imaginary objects borrow some of their features from their real-world counterparts, they may differ in at least two ways: what is successfully imagined may be (recognized as) both incoherent and incomplete, whereas what is real may not. So, as before, the traditional picture— that successful instances of pretense resemble belief-like states in terms of subject matter but differ from them in terms of motivational force—is misleading, this time in terms of its first claim. And, as before, this is a consequence of features central to the nature of imagination itself. Acknowledgments For discussion of this material, suggestions concerning relevant literature, and/or comments on previous drafts, I am grateful to Tyler Burge, Greg Currie, John Hawthorne, Dom Lopes, Daniel Nolan, Shaun Nichols, Ted Sider, Zoltán Gendler Szabó, and David Velleman; Greg Currie and Shaun Nichols were especially helpful in suggesting empirical references, and John Hawthorne, Daniel Nolan, and Ted Sider pressed me on a number of important conceptual questions. Thanks also to audiences at the SUNY Buffalo Cognitive Science colloquium series, and at the Ohio State University Department of Philosophy, where some of this material was presented.

140 TAMAR SZABÓ GENDLER

Notes 1 In the remainder of the text, I will omit scare-quotes unless their omission would lead to confusion. 2 See, for example, Nichols and Stich, whose explicit goal is to provide an account of pre-tense that explains how it is that “the events that occurred in the context of the pretense have only a quite limited effect on the post-pretense cognitive state of the pretender” (Nichols and Stich 2000:120) (quarantining) and how it might be that “inference mechanisms treat the pretense representations in roughly the same way that the mechanisms treat real beliefs” (Nichols and Stich 2000:125) (mirroring). Given their caveats, my project can be seen as supplementary to theirs: what this chapter explores are aspects of the “limited effect” of the pretense on the pretender’s “post-pretense cognitive state” (contagion) and the ways in which the inference mechanisms that govern pretense are only “roughly” like those that govern belief (unproductivity). 3 In recent years, there has been intense debate among psychologists concerning exactly what this capacity for pretense amounts to, in light of the fact that children of this age are (a) generally incapable of solving standard (“Smarties-box”) falsebelief tasks; (b) fairly limited in their capacity to distinguish apparent from real identity in the case of visually deceptive objects; and (c) generally willing to attribute the behavior “pretending to be an X” to an individual unaware of the existence of Xs. In light of these data, Josef Perner has stressed that the capacity to treat one object as another need not involve the capacity to represent oneself as holding a particular attitude towards that object: the child may believe that the object before her is a blanket and retain that belief while acting as if (pretending that) the object before her is a pillow—without having any second-order thoughts concerning her own attitudes towards it. In particular, Perner argues, she need not believe that she believes that the object is a blanket, or represent her pretense-partner as bearing one or another mental attitude towards the object in question. Indeed, he contends, children of this age lack a differentiated notion of belief as opposed to pretend, making use rather of an undifferentiated concept that Perner terms “prelief ”. Others have challenged Perner’s position on various empirical and theoretical grounds. (Perner 1991; Perner et al. 1994; for related methodological and/or empirical discussions bearing on these issues, see Bruell and Woolley 1998; German and Leslie 2001; Harris 1994; Harris et al. 1994b; Harris 2000: chapter 3; Leslie 1987; Lillard 1993, 1994; and references therein.) 4 In order to do so, one might perform something like the following experiments. (1) Ask subjects to label the “cyanide” jar with a label reading “cyanide” and then to paste a second label over the initial label that reads “sugar”. (Perhaps doing the inverse with the other jar.) (2) Ask subjects to label the “cyanide” jar with a label reading “sugar” and the “sugar” jar with a label reading “cyanide”. (3) Ask subjects to label the “cyanide” jar with an “A” and the “sugar” jar with a “B”. (4) Ask subjects merely to imagine that one of the jars contains cyanide and the other sugar. And so on. 5 Sometimes when we engage in prop-oriented games of pretense, the main goal is to keep the mapping as steady and uniform as possible. So, for example, if I am trying to figure out what the room would look like if the couch were where the table is

PRETENSE AND BELIEF 141

now and the table were pushed back against the now-empty wall, I want to employ only two sorts of mappings into my imaginary realm: objects besides the table and the couch occupy exactly the same locations in the imaginary realm as they do in the actual world, and the couch and table are of exactly their real-world size, but occupying different locations. This is often the case when we employ mental imagery in the service of spatial problem-solving tasks.

8 LITERARY FICTION AND THE PHILOSOPHICAL VALUE OF DETAIL Eileen John

He recognized me as infallibly as I had recognized him; he appeared to know by instinct how a young American of an aesthetic turn would look when much divided between eagerness and modesty. (James 1944:419) Absent—that was what he was: so absent from everything most densely real and near to those about him that it sometimes startled him to find they still imagined he was there. (Wharton 1920:263) Finally he reaches over and takes the pennies from her hand. His nails graze her damp palm. (Morrison 1970:43) Next day, April 17, at eight o’clock the concierge buttonholed the doctor as he was going out. Some young scally-wags, he said, had dumped three dead rats in the hall. They’d obviously been caught in traps with very strong springs, as they were bleeding profusely. (Camus 1948:8) “Is your sister ill?” said he. Elinor answered in some distress that she was, and then talked of head-aches, low spirits, and over fatigues; and of every thing to which she could decently attribute her sister’s behavior. (Austen 1992:155) These passages are taken from works of literary fiction that have been claimed to be sources of philosophical knowledge (Barnes 1959; Ryle 1971; Wilson 1983; Hagberg 1994). I am interested here in how we ever get from sentences such as the ones quoted above to something recognizable as philosophical knowledge. I frame the question initially in the following terms: what is the philosophical value of detailed thought about imagined individuals?

EILEEN JOHN 143

One way to approach this question is to demonstrate the value of such thought by example. We can investigate how the detailed thoughts prompted by a particular literary work allow us to reach a particular philosophically substantive result. That is an important project, one taken on by the philosophers cited earlier and numerous others. Here I want to step back from that strategy, to see what can be said generally about the philosophical potential of the detailed thinking that literature tends to provoke. Thinking about how to explain the importance of richly detailed, complex thought tends to lead me into circular reasoning (“it’s a valuable kind of thinking because we really need to encounter the genuine complexity of particulars”). So the suggestions made below try to articulate some of the factors that underlie and make sense of the conviction that this is indeed a philosophically valuable kind of thought (see Beardsmore 1984; Novitz 1987; Nussbaum 1990; Hagberg 1994). Another way to approach this issue would be to start with philosophers’ own uses of imagined cases or “thought experiments”. In using a thought experiment, a philosopher sets up a scenario to be imagined, in order to prompt a judging response to that scenario. The judgment, it is hoped, will help advance the philosophers’ project, perhaps by doing one of the following things: supporting or serving as a counterexample to a philosophical position, revealing a new philosophical problem, or motivating the use of new concepts or distinctions. Here is Locke, for instance, urging us to distinguish men or human beings from persons: …for should the Soul of a Prince, carrying with it the consciousness of the Prince’s past Life, enter and inform the Body of a Cobler as soon as deserted by his own Soul, every one sees, he would be the same Person with the Prince, accountable only for the Prince’s Actions: But who would say it was the same Man? (Locke 1975:340) Or here is Putnam setting up a scenario to challenge traditional theories of meaning: [W]e shall suppose that somewhere there is a planet we shall call Twin Earth. Twin Earth is very much like Earth: in fact, people on Twin Earth even speak English. In fact, apart from the differences we shall specify in our science-fiction examples, the reader may suppose that Twin Earth is exactly like Earth. He may even suppose that he has a Doppelganger— an identical copy—on Twin Earth, if he wishes, although my stories will not depend on this. (Putnam 1973:700–1)

144 FICTION AND THE VALUE OF DETAIL

Perhaps we should view these imagined cases as lying on a continuum with certain works of literary fiction, and perhaps study of how these simpler cases on the continuum work can help us understand the more complex cases. It is not obvious that there is a theoretically satisfying continuity between what we do in such philosophers’ thought experiments and what we do philosophically in response to, say, a Henry James novel. The bases for judgment in the two kinds of cases appear radically different: schematic, unengaging scenarios on the one hand, and densely detailed, often imaginatively and affectively engaging scenarios on the other. A philosophers’ schematic thought experiment provides a pattern to follow, so that its details could easily be replaced by other details that would set up the same problem and prompt the same judgment. Our judgments in response to a literary work seem to emerge from and feel justified by the details of that particular work, and we probably do not have any idea of how to generate a parallel case. Martha Nussbaum contrasts the philosophers’ schematic examples, which “signal to the readers what they should notice and find relevant” and which hence offer “cooked” results, with the far more open-ended literary works (Nussbaum 1990:47). Onora O’Neill, discussing imagined cases as used in moral philosophy, presents the two kinds of cases as relatively disjoint in form and function, as she contrasts the theory-dependent use of philosophers’ “sparse sketches” with the use of “elaborate and extended” examples drawn from literary works to convey non-theory-dependent moral perception and reasoning (O’Neill 1986: 9–12). Nussbaum and O’Neill make quite different use of these similar observations; I will return to their views later on. If I were going to span the gulf between the two kinds of imagined cases, I would start by emphasizing that both ask us for judging responses: both seek some kind of collaboration from the audience, and in that way they share a sense of philosophical process as more dialogical, as not just the assertion of one thinker’s premises and conclusion. And, to reiterate the problem of this chapter, I do see both philosophers’ imagined cases and literary works as reflecting a need to work at some remove from the vocabulary and structure of theoretical discourse— though of course philosophers’ examples often do not stray far. At any rate, though I think there are links between philosophers’ brief, schematic forays into imaginative activity, and the philosophical activity that can be prompted by literary works, I do not think it would help to understand the schematic cases first. Maybe if we understand the extreme cases of highly detailed thought; some of what emerges will apply to the schematic cases as well, but I do not pursue that issue here. Very broadly, my strategy is to make suggestions about cognitive functions that seem especially well supported by detailed thought about individuals and to link those functions to philosophical needs and projects. One further question concerns how the literary aspect of this imaginative thought matters: should my focus be “literarily detailed thought about imagined individuals” or perhaps “detailed thought about literarily imagined individuals?” Beardsmore rightly emphasizes that it is a mistake to treat literary works as

EILEEN JOHN 145

examples for philosophy, on a par with real-life examples (or, I would add, on a par with philosophers’ schematic thought experiments), as if the “distinctively literary qualities of literature” played no philosophical role (Beardsmore 1984:60). I could say that the literary realm of imaginings is my de facto focus, since it seems possible that most of our detailed thought about imagined individuals happens to be prompted by literary works of fiction. But I hope that the suggestions I make go further than this. I think some of them reflect distinctively literary qualities of literature or literary practice more broadly construed, while some of them do not; I will note when aspects of “literariness” are specifically at work. An illustration Although the purpose of this chapter is not to defend the philosophical significance of a particular literary work, I will present one literary example in order to have a shared reference point for discussion. Since I use this case just as an illustration, I will sketch only enough of it for that purpose and will not do much to defend its philosophical role. I see Virginia Woolf’s novel To the Lighthouse as raising two questions about sympathy. One of the scenes that is relevant involves three characters, Mr and Mrs Ramsay and their young son James. Mrs Ramsay is knitting and reading aloud to James when Mr Ramsay approaches them. But his son hated him. He hated him for coming up to them, for stopping and looking down on them; he hated him for interrupting them…but most of all he hated the twang and twitter of his father’s emotion which, vibrating around them, disturbed the perfect simplicity and good sense of his relations with his mother. …There he stood, demanding sympathy. Mrs. Ramsay, who had been sitting loosely, folding her son in her arm, braced herself, and, half turning, seemed to raise herself with an effort, and at once to pour erect into the air a rain of energy, a column of spray… He wanted sympathy. He was a failure, he said. Mrs. Ramsay flashed her needles. Mr. Ramsay repeated, never taking his eyes from her face, that he was a failure… Charles Tansley thought him the greatest metaphysician of the time, she said. But he must have more than that. He must have sympathy. He must be assured that he too lived in the heart of life; was needed; not here only, but all over the world. Flashing her needles, confident, upright, she created drawing-room and kitchen, set them all aglow; bade him take his ease there, go in and out, enjoy himself. She laughed, she knitted. Standing between her knees, very stiff, James felt all her strength flaring up to be drunk and quenched by the beak of brass, the arid scimitar of the male, which smote mercilessly, again and again, demanding sympathy. He was a failure, he repeated. Well, look then, feel then. Flashing her needles, glancing round about her, out of the window, into the room, at

146 FICTION AND THE VALUE OF DETAIL

James himself, she assured him, beyond a shadow of a doubt, by her laugh, her poise, her competence…that it was real; the house was full, the garden blowing. (Woolf 1955:57–60) Clearly this story is about more than sympathy, but I will focus on that thread here. My basic judging response was to recognize Mrs Ramsay as having indeed provided sympathy to Mr Ramsay. This judgment raised questions because various other features of their interaction did not obviously fit with what I would call my “sympathy stereotype”, which contains both my conceptual understanding of sympathy and my rough-and-ready sense of how it shows up in human relations. The first question concerns the relation of sympathizer to sympathized. I took it to be some sort of constraint on feeling sympathy that we feel it for someone who is vulnerable while not, in that particular emotional exchange, being vulnerable ourselves. My stereotype presented sympathy as given from a perspective of relative safety or power, and perhaps this element of my stereotype counted for me as a conceptual constraint. The interaction between the Ramsays upended this conception. Her sympathy for him was not a gift, but was demanded with compelling force; it had her vulnerability to him built into it. I would categorize the philosophical result of this response to the characters as a conceptual clarification: it allowed me to understand more clearly the contingent status of one aspect of my sympathy stereotype. The second question concerns sympathy as a paradigm of an emotion that actually has some type of feel to it—a loosely specified but nonetheless characteristic feel (perhaps, a warm, to varying degrees sad, and relatively mushy feeling). In judging that Mrs Ramsay gave sympathy to Mr Ramsay, it was not clear to me that there was psychological “room” for her to feel what I took to be the characteristic feeling, in the midst of doing the work of shoring him up. I did imagine her to be feeling various things, but they had to do with what she was doing for him in giving him sympathy (which may have called for warmth but also for something more exuberant than I associated with sympathy). At any rate, I’m not sure precisely how to categorize this result—whether aspects of qualitative feel are conceptually and philosophically relevant to study of emotion depends on one’s theory of emotion. But it gave me a tentative new conception of sympathy as something active, something done or given to support the one in need, with the range of appropriate feeling left much more open to reflect the nature of the appropriate supportive activity. With this case in mind as a reference point, we can turn to considering the various cognitive roles of detailed thought about individuals.

EILEEN JOHN 147

Detailed, individualizing thought Let’s begin by considering the notions of detailed thought and thought about individuals. These often work closely together in generating cognitive outcomes, but they are distinct. As Ronald de Sousa puts it, they mark two kinds of contrast with generality, as more detail brings greater specificity of information, while being about an individual means making reference to a particular, a logically non-general entity (de Sousa 1997:180). We could have lots of detail about a general category (“Here is a step-by-step account of how eels digest food…”) and very little detail about an individual (“it was damp”). Works of literary fiction are most obviously sources of detailed thought, in the sense of providing informationally specific stimuli for thought. Consider a randomly chosen sentence from Nabokov’s Pnin: “gravely, comfortably, the gray-headed conductor sank into the opposite seat and consulted in silence a tattered book full of dog-eared insertions” (Nabokov 1953: 17). This is the dimension along which literary works are most clearly distinguishable from the schematic philosophers’ thought experiments. It is quite another matter to argue that works of fiction are sources of thought about individuals. They are fiction after all, and it is commonly held that the fictional characters and events that we seem to think about in response to fiction do not really exist. So perhaps there are no individuals to think about. The issues of cognitive role that I’m interested in do not, I believe, hinge on ontological and semantic accounts of fiction. We can have the sorts of thoughts that matter to this project within the scope of a pretense, or perhaps we can have them while just setting aside whatever ontological commitments about fiction we may have. At any rate, on these grounds I need to revise my focus, away from a notion of thought about individuals that requires commitment to the existence of the objects of thought. I want to peel off from that notion some features of cognitive role that can be applied as well to thought about imagined individuals. This approach, focusing on what I will call “individualizing thought”, allows me to work with a rather different conception of particularity or individuality than de Sousa sets up, since particularity in his sense is a yes-or-no matter of logical category, one not admitting of degrees. I think that our thinking in response to fiction is distinctive and cognitively valuable partly because it has great flexibility and variability in how “individualizing” it is: we can have a vivid, personally gripping experience with a fictional character, one that is strongly individualizing, and also (though perhaps not simultaneously) see that character more as a model of a certain type or as a symbol of some general problem. Detailed thought and individualizing thought mutually encourage each other in certain ways. Perhaps it is most clear how detailed thought can lead to individualizing thought, as specific details can build up a convincing sense of awareness of a unique individual—details can specify more and more fully what a particular individual would be like, or what it would be like to encounter that individual. There is no straightforward correlation between detailed thought and individualizing thought, however. This is in part because, as noted above,

148 FICTION AND THE VALUE OF DETAIL

accounts of general categories can have great specificity, but also because a multitude of details need not allow an individual to “gel” in our thoughts, while a few details might indeed allow for individualizing thought. This is a point at which it seems necessary to gesture at the literary qualities of literature. The details in a story I wrote could be quite copious in terms of information, and they might even be individuating, in the sense of specifying features that only one individual could have at a time (perhaps, precise geographical location, or parents and exact time of birth), but these details would not necessarily add up to something recognizable and plausible as, say, a particular person. This kind of failure occurs rather frequently, albeit to varying degrees, in works of fiction, which we might signal by saying that we just could not take a certain character seriously or could not see how the different aspects of the character hung together. When detailed thought does lead to individualizing thought, I would say it is often attributable to the elusive presence of the literary. The literary artist who succeeds at this particular task knows how to offer details that build on or resonate with each other. These details set up the sorts of questions and possibilities for answers that mark the way individuals are present to us. Individualizing thought may in turn influence detailed thought in a somewhat less direct way. If we take ourselves to be thinking about an individual, we normally expect certain kinds of detailed knowledge of that individual to be possible. For us to be considering an individual is for us to be thinking about something whose features could, theoretically, be specified quite fully (as opposed to the Platonic couch that has no specific color, size, matter, or location). So, while individualizing thought may not in itself be very detailed (“she wore orange”), it holds out the promise of specification and in that way may encourage us to seek out and build up our detailed thought about the individual. Individualizing thought can also encourage speculative thought about the detailed features of an individual, even in the absence of evidence. What can be said about the philosophical role of detailed, individualizing thought? Noel Carroll offers the principle that there is a proportionate relationship between the complexity of a philosophical issue and the elaborateness of a thought experiment that genuinely addresses that issue (Carroll 2001b:34–5). I am following up on Carroll’s idea, but without relying directly on the notion of the complexity of an issue. Instead, I try to link the elaborateness to loosely specified kinds of reasoning or thinking—philosophical “tools” that are used in many ways, on many different issues. This is my strategy in part because I do not know how to follow up on Carroll’s idea more directly; it would not be easy to account for what makes one issue more complex than another. My suggestions here are broken down into separate categories of reasoning or thought, but I recognize that in reality these functions are intertwined. Briefly, I suggest that detailed, individualizing thought can be important to all of the following: (1) modal, counterfactual, and causal reasoning; (2) affective response; (3) what I will call “context-establishing”; and (4) certain kinds of experimental and open-ended thinking. Let me give a sketch of what I mean by each of these

EILEEN JOHN 149

and why detailed, individualizing thought could be important to these functions. The kind of importance could vary. In some cases it might be a matter of psychological fact (we are the sort of creatures who happen to do better at certain kinds of thinking if we do it on a detailed, individual level), or it might be an evidential matter (certain kinds of thought are well-supported by evidence at that level). Modal, counterfactual, and causal reasoning are lumped together here because they share a very broad concern that involves, roughly, trying to track features through change or variation (maybe this broad similarity hints at the possibly quite deep relations between them, but I’m not assuming that). When we want to know if something is possible or necessary, we think about whether some features could change while others were held fixed. In counterfactual thought, we try to understand the consequences of varying certain facts—what would be the same and what would change? In causal reasoning we try to see what role certain features play within a transformation. More often than not, perhaps, we aim for general conclusions with these kinds of thinking (can x’s also be y’s? If there were no x’s, would there be any y’s? Do x’s cause y’s?). But even if our ultimate concerns are general, I hope it sounds plausible that we could benefit from approaching such questions by tracking the features or conditions of individuals. An individual gives us something to “attach” features to in a holistic way, and this can give us a manageable focus for scrutinizing the features for missing information, incoherence, or unclarity. Having “located” features in an individual, we then have a cognitive perspective from which to consider change and variation of conditions. We commonly conceive of individuals as having somewhat resilient identity through change (whereas it’s not clear, for instance, that abstract entities can undergo change), and it seems that some of these kinds of thinking make most sense if applied to resilient targets of thought. As I suggested earlier, I also think that part of how individuals are present to us, or how they have an identity for us, involves posing certain questions, primarily about past and future: we need to root an individual in its history and point it to some possible future(s). This is just to gesture at the idea that we take stock of individuals in a way that engages us in causal and modal reasoning (What has led it to this point? Where can it go from here?). The importance of having a detailed conception of an individual, for purposes of these kinds of reasoning, would depend on the particular kind of changes and conditions that were at stake. But it is not hard to envision possibilities, counterfactual situations, and causal networks that would call for appreciation of a fairly detailed set of circumstances. Certainly one common criticism of philosophers’ thought experiments is that they do not provide enough detail to support judgments about possibilities (Dennett 1984; Wilkes 1988). Wilkes favors attention to real-life cases rather than thought experiments in part because in real life “the quest for details that might or might not be relevant can be met” (Wilkes 1988:46). To return to the Woolf example, that scene involved some description of the son’s physical and mental reactions to his father’s presence: “by looking

150 FICTION AND THE VALUE OF DETAIL

fixedly at the page, he hoped to make him move on; by pointing his finger at a word, he hoped to recall his mother’s attention, which, he knew angrily, wavered instantly his father stopped. But, no. Nothing would make Mr Ramsay move on” (Woolf 1955:57). The physically manifest insistence of both father and son, and my acceptance of their confrontation as genuine, is a tiny piece of what the novel gives and of what I do with it that helps characterize the exchange of sympathy and helps set up my judgment about the possibility of vulnerable sympathizing. It is not true of course that precisely those details would be necessary for arriving at that judgment. Nonetheless, those details about how an interested third party reacts to the interaction between the Ramsays, and the plausibility of those details, can matter to whether the scene registers with a reader in the way that addresses the questions about sympathy. One last point here is that the role of imagined individuals in literary fiction seems particularly promising with respect to modal and counterfactual reasoning. The questions raised by these kinds of reasoning explicitly ask us to consider more than the actual world; it seems then that fiction’s freedom to work beyond the confines of the actual could aid in investigation of what can and must be or what would be. Aristotle, for instance, confidently links literary art to knowledge of possibility and necessity when he remarks that it is the poet’s business to relate “such things as might or could happen in accordance with probability or necessity” (Aristotle 1958:18).1 On the other hand, the literary nature of the imagined individuals raises problems, as this can be seen as reason to distrust their integrity and relevance to philosophical thought. I will come back to this type of objection later on. Second, it seems likely that detailed, individualizing thought has an important relation to affective response. By affective response I mean to refer to a mixed bag of responses that integrate felt experience with our cognitive, valuing, and desiring capacities. Emotions are the prime examples of affective responses, but more experientially, cognitively, and evaluatively diffuse responses such as moods (excitement, gloom, eagerness, cheerfulness, nervousness) are included as well. Now, even if our affective responses do not necessarily spring from encounters with individuals, it seems safe to say that our paradigms of affective response are rooted in individuals’ experiences of individual situations. These responses embody a particular individual’s perspective on the world, and they are in general quite sensitive to the detailed circumstances accessible from that perspective. Affective responses are often self-directed but they can have quite complex relations to others, as we have empathetic feelings and feelings for the actions and predicaments of others, feelings that reflect the relations between our own and another’s perspective, and so on. In sum, these responses commonly emerge from an individual’s perspective on social settings with complex, interacting sources of feeling. If drawing on our affective responses matters to philosophical activity, it seems that philosophers would at times need to rely on, or would benefit from, forms of detailed, individualizing thought. This kind of thinking provides a basis for the

EILEEN JOHN 151

engagement with individuals and their social settings that is the central “stuff” of affective response. Let me recognize here that there is a great deal of debate about how to describe the affective responses we have to merely imagined individuals— do these responses involve genuine emotion at all? One might use some of those worries to argue that even if philosophers need detailed, individualizing thought, they should not turn to fiction to get it. I am comfortable saying that, while affective responses to fiction should not be assumed to play the same functional role as responses to real life, they often play enough of the same role to be relevant to philosophy. In some cases the fictional source of thought and feeling could even be advantageous, if it allows us to take up positions and develop feelings that would be quite hard for us to experience when rooted in our actual circumstances. But what is the relevance of affective response to philosophical activity? For one thing, philosophers investigate psychological and moral concepts that directly involve our affective life. This includes philosophical study of the emotions, such as sympathy, as well as study of character concepts such as trustworthiness and dignity, which can involve affective impact and interaction. For those projects, it seems plausible that philosophers would need access to affectively engaging experience as evidence, counterevidence, and stimulus for theorizing. In the Woolf example mentioned above, I certainly needed my feelings in response to the characters to help me arrive at judgments about whether sympathy had been offered and under what conditions. Further support for the relevance of affective response to philosophical work comes from a good deal of contemporary work in moral philosophy. I will here just refer to the work of two moral philosophers who are, in different ways, making room in their moral theories both for detailed, individualizing thought and for the affective experience which that thinking can prompt. Lawrence Blum, following on ideas of Iris Murdoch, argues for the moral centrality of “loving attention” to particular individuals (Blum 1994). In contrast to traditions emphasizing the impartiality of moral reasoning and judgment, Blum’s view sees the moral agent as motivated by knowledge of and care for particular others. Affective responses to others are integrated into what counts as appropriate moral perception. Blum is thus arguing primarily for the need for detailed, individualizing thought and a reliance on morally relevant feelings in the course of living one’s life, rather than making a case for a specifically philosophical need for such thought and feeling. His own work, however, makes substantial use of constructed cases, involving imagined individuals, and while they are not literary works, they are more fully developed, with more human specificity, than most philosophers’ imagined examples. His practice, then, suggests that philosophers concerned with the nature of morality itself (whether or not morality requires an impartialist theory, for instance) will need evidence from the most basic level of moral activity. Barbara Herman aims to defend Kantian moral theory from charges that it ignores the particular. In that project, she argues that before a Kantian moral

152 FICTION AND THE VALUE OF DETAIL

agent can do what is required for moral judgment, namely, formulate maxims for particular actions and test them using the Categorical Imperative, the agent must be able to recognize the features in particular situations that have the potential to be morally relevant. We acquire this form of moral knowledge by learning what she calls “rules of moral salience” (Herman 1993:77). These rules are learned and applied in the experience of particular situations and in that way they work with detailed, individualizing thought about real life. The questions these rules need to address suggest that emotional response would be quite important to learning and applying them: “who is a moral agent or end-in-himself ?… What are the conditions of agency for ends-in-themselves? In what ways are such agents vulnerable?… What counts as force? deception?” (Herman 1993:86). The rules will furthermore need to be critiqued and improved, through better understanding of people’s needs, the full effects of our actions, and our parochial tendencies (Herman 1993:88). These questions and tasks all seem to be usefully approached in part by using the resources of specifically emotional response, given the way that emotions bring together beliefs, values, and desires in social context. Emotions are of course not inevitably trustworthy “detectors” of moral salience. Nonetheless, we frequently rely on emotion in noticing such things as exploitation, manipulation, force, and neglect. Feelings of embarrassment, of being condescended to, of being ignored, and empathetic versions of these and other feelings, contribute a great deal to making such moral matters salient to us. Herman’s view, like Blum’s, concerns most directly the “lived” level of moral experience and reasoning, rather than the level of moral theorizing. But articulating and critiquing implicit rules of moral salience does fall within the purview of moral theory, and that argues in favor of theorists drawing on detailed, individualizing, and affectively engaging thought. I hope it is suggestive that these two examples—of theories with quite different overarching commitments—each give a role to this kind of thought. The third type of cognitive function, “context-establishing”, is thinking that projects a larger context, perhaps a complete world, around what has been directly presented or specified. It may be that this function only matters when we are dealing with imagined scenarios—maybe we automatically get a context (in this case a complete world) when thinking about actual events. If so, then the power of detailed, individualizing thought with respect to context-establishing would only count as an advantage in relation to schematic imaginings. So, this cognitive function may return us to the contrast between literary works and philosophers’ thought experiments, by arguing for the virtues of the richer imaginative projects.2 At any rate, context-establishing with respect to an imagined scenario means coming to see the scenario as embedded in, shaped by, and contributing to a context beyond the specified content of the scenario. The nature and depth of the context established can vary greatly, and the impact of context on what is imagined can likewise vary greatly. For instance, the context for philosophers’ thought experiments that summon up ordinary, quite possibly actual circumstances may be something as broad as “the actual world with all physical

EILEEN JOHN 153

and psychological regularities held fixed.” We commonly use (most of) the features of the actual world to set the background context for what is imagined. With respect to literary works, the context will usually be more specific and historically rooted. The brief episode from Woolf’s novel sketched above makes sense in part because we place it within a larger social world in which male and female carry certain meanings. Mr and Mrs Ramsay do not themselves bear the whole burden of generating the clash of male and female that informs the exchange. Many of the details build up a sense of that social world: Mrs Ramsay creates the drawing-room and the kitchen, she knits, she holds her son in her arms, and Mr Ramsay meanwhile wants success with the abstractions of metaphysics and is saddled, at least figuratively, with sharp metal weapons. Similarly, the directly portrayed personal life of the Ramsay family belongs in a world that also includes as context Mr Ramsay’s professional life. Mrs Ramsay’s initial feeble attempt to meet his demand for sympathy, by mentioning Charles Tansley’s opinion of him, is a detail that gestures at that part of the larger world. And, although this episode does not emphasize it, the novel also relies on us taking a fairly specific structure of class relations as the context for how the Ramsays live (perhaps such things as the fact that Mrs Ramsay is knitting for a charitable purpose and the reference to the “drawingroom” hint at class). As this example illustrates, in a sense the context is just “there”, imagined as an objectively constraining framework of circumstances, laws, social practices, historical trends, or whatever happens to matter to that context. The context in this sense is something to rely on and to reckon with in understanding and responding to the scenario. But in another sense the context reflects what we choose to emphasize and what we see as a meaningful constraint on the scenario. The context thus shows something about what we take to be the possible dimensions along which the scenario can develop. We do not project a social class context for Putnam’s Twin Earth, and so we do not have to reckon with class issues in responding to that thought experiment, and it would not be fruitful to consider how it could develop in class terms.3 But most readers probably do give To the Lighthouse a meaningful class context, at least to some degree, and this opens up the possibility of asking class-related questions about it. We could perhaps choose to project a strictly physicalist, nonpsychological context for the novel, such that the only interesting dimension along which the story fit into a larger world involved spatiotemporal arrangements and interactions of physical substances (perhaps the middle, “nonhuman” section of the novel could be read as supporting that choice). The general idea is that context-establishing thought involves selection of constraints and emphases, which then influences further thought about the imagined scenario. Detailed, individualizing thought can establish contexts in very straightforward ways, but also in ways that make use of literary qualities. Certainly a few wellchosen details can “locate” a scenario historically and culturally. Consider the following opening lines of a novel: “we’ve got a ranch house. Daddy built it. Daddy says it’s called RANCH ‘cause it’s like houses out West which cowboys

154 FICTION AND THE VALUE OF DETAIL

sleep in. There’s a picture window in all ranch houses and if you’re in one of ‘em out West, you can look out and see the cattle eatin’ grass on the plains and the cowboys ridin’ around with lassos and tall hats. But we ain’t got nuthin’ like that here in Egypt, Maine” (Chute 1985:3). Time period, geographic and political location, class identity, and a strand of cultural mythology are all efficiently specified to some degree here. The use of a particular kind of colloquial, nonstandard English, which helps specify class identity, is a stylistic, specifically literary detail that contributes to setting up social class as possibly important to the context. The context-establishing power of detail in literary works is also of course influenced by the way we are trained to read (and write) literature. It is a central part of literary practice that details be used so as to point beyond themselves. We know the details have been carefully chosen, and one of the things they are chosen to do is to guide us to a larger context in which they can be located, made sense of, prioritized, connected to other details, and so on. This does not mean that details in literary works will always work in a contextestablishing way, but it means that readers usually try to get them to work that way and may be met halfway in the effort by authors. It may seem that detailed thought does more to establish context than individualizing thought, and perhaps that is true. Maybe projecting a context only requires envisioning kinds of things, in circumstances of certain types, and envisioning the possibility of individual members of those kinds. But it still could be true that individualizing thought, in which we imagine individuals having features relevant to the context and operating within that context, helps us establish contexts with more understanding and sophistication. In the Woolf example, suppose that we take a historically specific and problematic kind of relation between men and women as part of the context, and suppose that the possibility for changing that relation is one of the things to be understood in that context. It seems that individualizing thought about that relation could be extremely valuable for prompting more than shallow thought about what could change it. Assuming that the connection between context-establishing and detailed thought at least is granted, I still need to give a sense of how context-establishing can serve as a useful philosophical tool. What sorts of projects do philosophers pursue in which it could matter that we embed a scenario in a larger context or world? One initial thought is that context-establishing might matter in a practical way, by affecting the seriousness with which we respond to the scenario. Perhaps we take scenarios more seriously that do not appear out of nowhere in an isolated, imagined space, but that seem to have roots in a larger world, even if that world too is merely imagined. Seriousness of this sort could lead to scrutinizing the conditions of the imagined scenario in a more probing way, aiming for deeper sorts of consistency, or thinking harder about the consequences of the judgments one makes about the scenario. A more directly philosophical motivation for context-establishing relies on the idea that some judgments about a scenario only make sense if there is a larger

EILEEN JOHN 155

context in which facts about that scenario reverberate or in which participants in the scenario matter. Suppose, for instance, that morality is an essentially social phenomenon. It seems that, in a scenario that raised a moral issue but involved directly imagining only one individual, we might have to posit a social world of some kind for that individual. Or suppose that we are trying to give a normative evaluation of an imagined cultural practice that would only matter as a cultural practice if it had a history and a potential future in its community. Even if we only specify the practice in its present moment, it could be important to our judgment about it that we treat it as something with a temporally wider context. Or, very broadly speaking, it seems that making judgments about some scenarios could require conceiving of the consequences of events, decisions, or actions taken in the scenario as extending out causally into a larger world, rather than being contained within the boundaries of what is directly imagined. Consequentialist moral theories explicitly make this kind of extended causal context important, and it seems that many scenarios involving personhood and personal identity might need such a context as well, if it were important to take the imagined circumstances as having sustained consequences within a life. Finally, I want to discuss briefly a cognitive function that I will call “experimental” thought. This is thinking that allows us to experiment with what is important to an issue, without yet having settled on a theoretical account or perhaps even on the theoretical options. Linking this function to detailed, individualizing thought echoes Nussbaum’s description of literary works as more open-ended and less “cooked” than philosophers’ imagined examples, as well as O’Neill’s claim that literary works can be used by philosophers in a non-theorydependent way. It is not as if the details of literary fiction are theoretically unencumbered; often they depend on implicit or even rather explicit theoretical assumptions, such as the idea of competing physical and psychological explanations for behavior implied in the Jane Austen passage I quoted at the beginning. But the detailed, individualizing content of literary fiction commonly works together with typical practices of reading to encourage nontheoretical concerns and less theoretically committed thinking. The detail of fiction can achieve this in part by giving us a sheer curiosity for plot. I hope that even the brief passages quoted at the beginning had a bit of this effect. How does a day that begins with three bleeding rats in the hall continue? How does the contact of his nails grazing on her palm matter to either of them? Thinking at this level (what will happen next? why did it happen? what does it have to do with that other thing that happened?) displaces the theoretical level of knowledge from the position of priority it has within philosophy. When concerned with plot we certainly appeal to theoretical knowledge, to predict, explain, and evaluate, but we probably approach theories to some extent as opportunists, choosing whatever helps with understanding and appreciating the story. This opportunism can be valuable, if it leads us to see unsuspected theoretical virtues and inadequacies.

156 FICTION AND THE VALUE OF DETAIL

The detail of literary fiction in particular can be a challenge to theorizing because of its complexity, density, and suggestive quality. The James and Wharton passages quoted at the beginning of the chapter are single sentences, and not especially long ones, that describe very complex experiences of recognition. From a philosophers’ perspective, perhaps, these sentences recklessly mingle issues of epistemology, metaphysics, ethical character, social, and existential meaning. We might ask as philosophers, “how could anyone have evidence to justify such ambitious, integrating assertions?” But as readers and human beings, we probably think these sentences are fairly easy to understand and are perhaps plausible as assertions about these characters. The ambitiousness of the two sentences from Morrison’s novel, concerning the nails grazing the damp palm, is not discernible within those sentences, but in their context within the story, that moment of physical contact is charged with issues of personhood, freedom, and moral and aesthetic perception. It seems that philosophy would benefit from working with this kind of ambitiously dense and complex material because philosophers are supposed to be acute critics of our concepts and distinctions. Philosophers need to ask if we are using the best possible categories for understanding ourselves and the world, and if it turned out that philosophically favored categories could not keep up with or keep track of how we regularly think about human life—to the extent that we get evidence about that from literary fiction—that would be a valuable stimulus for philosophy. The value of suggestiveness is mingled with that of complexity and density, but has potentially distinct value. In the Woolf passage, we are told that Mrs Ramsay “assured him…that it was real; the house was full, the garden blowing.” The language used here is rather simple, but I think it is recognizably literary. It does not sound like what anyone would actually say when trying to comfort someone, and it does not appear to be intended to record what Mrs Ramsay said out loud. It is rather, as we say, suggestive of what she offers him and what he needs. But why does he need to hear that the garden is blowing or that “it is real” (and what is “it” exactly)? In reading the novel, I did not worry about those questions—I thought that string of assurances was just right. Although I probably could not convince someone who thought the language was not right, I think I do know what “it” is, and why it is not just utter banality to say that it is real, and why it helps in this situation to assert that the garden is blowing. Or at least I have a nice suggestive swirl of ideas, images, and feelings about all of that, perhaps not anything that would stand up as knowledge under epistemological scrutiny. Taking this as one illustration of what suggestiveness can be like, what could the philosophical value of this literary quality be? Returning to the idea of experimentation, a suggestive work encourages a certain kind of freedom and experimentation with how to fill out one’s suggestive “swirl” of thinking. As with density and complexity of detail, this could have challenging, helpfully disruptive effects on theoretical work. But a suggestive response may also be the appropriate kind of thinking for certain philosophical issues. Perhaps some objects of philosophical study (including, perhaps, personhood and the nature of art) are

EILEEN JOHN 157

evolving, so that the clarity of understanding we may achieve at certain points should be revisited regularly and confronted with the less orderly contents of new suggestive associations, newly mingled ideas, images, and feelings. Or perhaps it is not a matter of evolving over time, but a matter of inevitably “unfinished business”, as we try to reconcile within a single concept or theoretical approach an issue that refuses to be neatly resolved (perhaps, the nature of right action or of rationality). In that case, it seems that thought with qualities of suggestiveness could be appropriate for seeing more honestly the different directions in which we are pulled by our philosophical conceptions. These last comments about complexity, density, and suggestiveness are very incomplete efforts to point to special qualities of imaginative thought prompted by literary fiction. The four general categories laid out here—modal, counterfactual and causal reasoning, affective response, context-establishing, and experimental thinking—should also be taken as speculative and incomplete. I hope they are suggestive, at any rate. In conclusion, I consider briefly two broad objections to giving detailed, individualizing thought prompted by literary fiction the kind of cognitive and philosophical potential I have given it here. Objections: obscurity and artifice Works of literary fiction, as I have emphasized, commonly present us with a densely detailed account of events concerning imagined individuals. Considered as sources for philosophical activity, works of literary fiction can be criticized on these grounds for overwhelming us with detail, for not isolating any particular philosophical issue, and hence for allowing us to be distracted and misled by irrelevant features of the imagined circumstances. Roy Sorensen, for instance, rejects the idea of counting literary works as philosophical thought experiments because in a thought experiment, “the details should support, rather than engulf, the experimental intention,” but as stories approach the length of literary works “the proportion of theoretically irrelevant detail grows” (Sorensen 1992:223–4). Given what I have argued above, I would like to construe this as a practical problem, rather than as a reason to doubt the philosophical value of such detailed imaginings. The practical problem—that in reading a work we may never see our way clear to isolating and thinking about a philosophical question—is of course not necessarily a problem at all, from the perspective of literary experience. But it seems true that literary works give us a lot of freedom to do different sorts of things in response; the detail can steer us toward nonphilosophical concerns and pleasures as well as toward philosophical ones. So it seems to some degree to be a matter of chance whether a work of fiction functions in any of the ways outlined above. That makes it an unreliable source of philosophical activity, but still potentially valuable. Onora O’Neill’s critique of the use of literary examples in moral philosophy raises a number of different objections (O’Neill 1986). She sees at least some uses of these examples as having morally conservative tendencies, as emphasizing

158 FICTION AND THE VALUE OF DETAIL

private over public morality, and as entrenching a spectator’s passive perspective in moral reasoning. Here I will only address the further objection she raises concerning the artificial authority of literary works. O’Neill emphasizes the way in which the literary source “imposes a largely shared interpretation of examples. The only acceptable disagreements about the construal of literary examples are those for which there is warrant within the text.” Thus, “in respecting the integrity of literary examples, the depth and ubiquity of moral disagreement are obscured” (O’Neill 1986:14). The literary example brings with it its own structure of possibilities, values, and meaning, and in that way it creates an artificially limited context for moral thinking and theorizing. While O’Neill might grant that detailed, individualizing thought has value for philosophers, it seems that she would rather we sought it out strictly from real life. Let me grant first that it is indeed possible for the context established in interpreting and responding to a work of fiction to be distorted and misleading in relation to a moral problem or other philosophical issue. There is no guarantee that a literary work will help us frame a problem helpfully or investigate it fruitfully. However, first of all, O’Neill has given an overly discouraging account of our ability to think critically and freely about what a work of fiction presents us with. For certain kinds of interpretation, we are constrained by what appears in the text, and by what we assume about the author’s intentions, but much of the interesting thought that is prompted by fiction, some of which I have gestured at above, depends heavily on what we bring to the text. Our questions, reasoning, and feelings, our competence as concept-users, our capacities for suggestive thinking, all participate in what emerges from experience with the work. So, even if a literary text has a kind of integrity unto itself, I take it also to be part of our literary practice that we participate in making something more out of the text through our responses to it. Second, it would be rash not to be open to the philosophical talents of literary artists. Maybe some writers draw me into a context for moral thinking that is carefully constructed within their texts and that guides my thought very powerfully, but I think some of these writers are able to do this because they know philosophically important things and I learn from what they construct. Finally, if there is an implicit contrast here between artificially limited encounters with literary texts and potentially deep and challenging encounters with real life, it is not so clear how to defend the advantages of real life. We have limited, interpretationbound understandings of real moral problems as well. Even what we gain from participating as moral decision-makers does not provide us with definitive moral experience, but may entrench us in problematic habits of moral thinking. Perhaps the experimental aspects of our thinking in response to fiction would help address at least some of O’Neill’s worries. If we need to ensure against artificially narrow conceptions of philosophical problems, we might do much worse than to consult some ambitiously dense, complex, and suggestive works of fiction.

EILEEN JOHN 159

Acknowledgment I thank Cheshire Calhoun, Guy Dove, and Leigh Viner for very helpful comments. Notes 1 Aristotle appears to reject my kind of appeal to individualizing thought, however, since in his view the poet portrays what general types of people could or must do and shows this in part by choosing typical and repeatedly used names. I agree that fictional characters can function essentially as types (e.g. characters such as Everyman, Knowledge, and Discretion), but I think there is usually something more interesting going on. Again, I would emphasize that our responses to fictional characters incorporate many degrees of individualizing thought. For a very interesting reflection on Aristotle’s point see Nussbaum (1990:317–18). 2 I am not sure about whether our thinking about actual events automatically comes with a larger context attached. Does our causal contact with the world set the larger context for actual-world thought or is that larger context fixed in some way by the explicit content of our thought? Or maybe we have a standing intention to root most of our thoughts in an actual-world context. 3 Locke’s prince and cobbler in some sense do bring class issues into play, as he uses the detail of opposed social positions to mark extreme differences in personal identity. I assume Locke did not intend anyone to think the scenario had potential to develop in some way relevant to class and would not have counted class structure as part of its context. Could someone argue that class should be included in the context?

9 FICTION AND VALUE Roman Bonzon

How can an appreciation of fictional works bear upon an understanding of moral or ethical value? I will investigate this question by first refining it and then analyzing progressively more complex answers. This will, I hope, enable me to make my own account as persuasive as possible. Ethics and the imaginary There is a certain very familiar use of fiction in ethical or moral thinking. Two characteristic examples are to be found in Bernard Williams’s “Critique of Utilitarianism.” One concerns a chemist, George, who is reluctantly considering a job at a laboratory that pursues research into biochemical warfare. The other is a dilemma facing Jim, who must choose between killing one Indian and letting twenty Indians be killed (Williams 1973a:97–100). These fictions are intended, in part, to stir our intuitions concerning the relative weights we would assign to the competing demands of personal integrity and utility; and there are two of them because Williams wants to make room for the possibility of a divergence of answers to the question, “what should one do?” not envisaged either by utilitarianism or by any kind of nonconsequentialist theory. For these purposes the fictional nature of the examples is almost completely a matter of convenience: Williams could have posed instead problems from real life. The case of George is, if not commonplace, not rare either. Jim’s situation, though rarer, is nevertheless a credible one. It takes no great effort to imagine its actually happening. “Fiction” here means simply “imaginary” or “made up”. Arguing for the necessary role of fiction in thinking about ethical or moral values, Martha Nussbaum suggests that our actual lives do not afford us sufficient material for the tasks of ethical reflection: “we have never lived enough,” she says. “Our experience is, without fiction, too confined and too parochial” (Nussbaum 1990: 47). This is surely right. Consider the crucial use made by Judith Thomson of the imaginary case of the violinist in “A Defense of Abortion.” She first lays before us an argument for the conclusion that a human fetus may not be aborted, resting on the premise that a fetus’ right to life is stronger and more stringent than its mother’s right to decide what shall happen in and to her body. She then asks how the argument would be affected were we to substitute for the fetus a famous

ROMAN BONZON 161

unconscious violinist totally dependent upon the functioning of the kidneys of an involuntary donor. It would yield what she takes to be the clearly unacceptable conclusion that the violinist may not be unplugged, the problem apparently lying with the premise that the violinist’s right to life is stronger and more stringent than the donor’s right to decide what shall happen in and to her body (Thomson 1986:38–9). Fiction here expands the space for the operation of our moral intuitions, allowing us to think of the fetus in a stark new light. This use of fiction in ethical thought differs from Williams’s since it requires us to stretch our sense of possibilities further. Greater imaginative effort is needed to respond to this case, just as greater imaginative power was necessary to create it. Williams himself notes that Thomson’s example “is effective just because of its ghastly unreality” (Williams 1985:214, n. 3). His own examples, by contrast, are so plain that they are almost recipes for examples; the reader will have no trouble thinking up his own versions. This is not to say that imagination plays no significant role in the fictions of George and Jim. By his shaping of detail and shading of character and circumstance Williams is able to provide us with a more nuanced appreciation of the moral dilemmas involved. And indeed in the introductory section of his “Critique” he says “the aim [of these examples] is not just to offer or elicit moral intuitions against which utilitarianism can be tested. … rather, the aim is to lead into reflections which might show up in greater depth what would be involved in living with these [utilitarian] ideas” (Williams 1973a: 78). But Thomson’s example makes us see things that matter morally in an entirely new way, and her ingenuity plays no small part in this. Nevertheless, these two kinds of use bear a significant resemblance to each other: the fictions in both cases are wholly in service of some ethical point, and have no interest beyond their capacity to engage moral intuitions or stimulate philosophical reflection. The example in each case is meant to be exhausted by this aim since it was invented solely to achieve it; no further task of appreciation is either necessary or desirable. Moreover, the point each serves is supposed to be readily and unequivocally discernible in the fiction, whose use in ethical reflection would be spoiled by uncertainty or ambiguity. Sometimes, however, the fiction is richly woven and calls upon even greater imaginative engagement to understand its connection to ethical or moral thought. A famous example is given by Iris Murdoch in “The Idea of Perfection,” the first of the essays gathered under the title, The Sovereignty of Good. Developing her own view of the proper conception of moral thinking, she asks us to consider the following scenario: A mother, whom I shall call M, feels hostility to her daughter-in-law, whom I shall call D.M finds D quite a good-hearted girl, but while not exactly common yet certainly unpolished and lacking in dignity and refinement. D is inclined to be pert and familiar, insufficiently ceremonious, brusque, sometimes positively rude, always tiresomely juvenile. M does not like D’s accent or the way D dresses. M feels that her

162 FICTION AND VALUE

son has married beneath him. …the mother, who is a very ‘correct’ person, behaves beautifully to the girl throughout, not allowing her real opinion to appear in any way. …Thus much for M’s first thoughts about D. Time passes, and it could be that M settles down with a hardened sense of grievance and a fixed picture of D, imprisoned (if I may use a questionbegging word) by the cliché: my poor son has married a silly vulgar girl. However, the M of the example is an intelligent and well-intentioned person, capable of self-criticism, capable of giving careful and just attention to an object which confronts her. M tells herself: ‘I may be prejudiced and narrow-minded. I may be snobbish. I am certainly jealous. Let me look again’. Here I assume that M observes D or at least reflects deliberately about D, until gradually her vision of D alters. If we take D to be now absent or dead this can make it clear that the change is not in D’s behavior but in M’s mind. D is discovered to be not vulgar but refreshingly simple, not undignified but spontaneous, not noisy but gay, not tiresomely juvenile but delightfully youthful, and so on. And as I say, ex hypothesi M’s outward behavior, beautiful from the start, in no way alters. (Murdoch 1998:312–13; emphasis in original) The density of the example is required by the depth of the issue it serves. Murdoch is trying to get us to appreciate that the quality of a person’s vision of another, and not just the qualities of his character (Williams) or actions (utilitarians), can be the locus of ethical value, and her attempt involves our imaginative participation in a fiction that is detailed, open-ended, metaphorical, suggestive rather than explicit, and shaped already in ethical terms. She is interested in bringing the reader close to an active engagement with, almost an actual experience of, the moral point she is trying to make. Unlike the Thomson example, which lacks its characteristic resources and consequently its distinctive appeal, Murdoch’s approaches the condition of good literary fiction, which has, to use Nussbaum’s words, a “way of making the reader a participant and a friend” (Nussbaum 1990:46). It is plausible, particular, emotionally involving, realistic, and meant to feel if not familiar then not far-fetched. Our understanding and agreement are not so much coerced as confidently and quietly solicited. Unlike Williams’s examples, however, which are also plausible and lifelike, Murdoch’s fiction depends almost entirely for its effect upon an accumulation of subtle detail, a succession of narrative and perspectival shifts. Murdoch is concerned not so much with pointing up a critical though ignored moral consideration but with making salient an aspect of moral experience that the reader is able to recognize as part of his own. Nevertheless, what all of these examples have in common is more fundamental than any differences. Murdoch’s story, while coming close to being literary fiction, is not art. Its deployment of the powers of literature is governed by a nonartistic aim. Like Williams and Thomson, Murdoch here uses fiction to introduce, illustrate, or illuminate a standing ethical theme or idea; like theirs, her

ROMAN BONZON 163

fiction represents its author’s estimate of the best way of preparing for and making a moral point, which would fail in proportion as the point is not recoverable from the fiction or a different one is. The interest of all these stories, however interesting they may be in themselves, fully subserves the interest of the ethical point. And again this is because they were invented for the sole purpose of playing that role. Ethics and the imaginative Would the situation be drastically altered by the use of extant pieces of literature rather than fictional examples custom-made for the purpose? Consider their use in Colin McGinn’s Ethics, Evil, and Fiction. McGinn describes his method in terms of a distinction between two ways he thinks literary texts can serve as vehicles of moral thought, both of which he says are “handily exemplified” in the Bible: there is the “commandment” style of moral discourse starkly on view in the Ten Commandments, and there is the “parable” style typified by such New Testament fables as the Prodigal Son. According to him, philosophers have focused exclusively on commandment-style texts, and this has resulted in a crude and simplistic representation of moral life. His aim is to investigate those texts that embody the parable paradigm, claiming that this holds out the promise of a fuller picture. Unlike the commandment paradigm, which issues in “simple and unqualified” injunctions “meant to be followed mechanically,” the parable “operates by engaging the [reader’s] mastery of folk psychology and applying it in a dramatic or narrative context; the ethical lesson is meant to fall out of this activation of cooperating faculties” (McGinn 1997:172). The texts he exerts his ingenuity upon in this way—and they are ingenious interpretations—are The Picture of Dorian Gray and Frankenstein, and his main finding is that they give us detailed pictures of evil characters. This kind of resource, he suggests, no moral theory should fail to avail itself of. It makes no difference that McGinn claims to have first chosen to concentrate on some of his favorite novels and allowed his reading to give rise to the philosophical theme “instead of settling on a philosophical theme and then finding books (good or bad) that dealt with that theme in some way” (McGinn 1997:vi). The point is that the novels are treated, finally, as veins to be mined for philosophical nuggets, and are therefore made to bear the same relation to ethics as the examples Williams, Thomson, and Murdoch tailored for the purpose. Like commandments, after all, parables are essentially didactic, and this function is completely realized once the embedded insights are strip-mined from the soil of the text. That McGinn’s use of literary works to illuminate ethical or moral value does indeed cast light upon the philosophical point he is interested in should be conceded. Dorian Gray and Frankenstein do in fact yield the insights McGinn claims to find in them. This is because those insights form part of what might be called the prepositional or descriptive moral content of the novels, the set of moral truths

164 FICTION AND VALUE

directly stated, implied, presupposed, or made plausible or probable by the texts (complications arising from voice and point of view aside). But there is quite another use of literary fiction in ethical thought we need to consider. I have in mind Martha Nussbaum’s well-known discussion of The Golden Bowl (in Nussbaum 1990). Like McGinn, she is convinced that traditional moral philosophizing has impoverished itself by its failure to recognize and take full advantage of the resources of literature. However, she has a more sophisticated notion of how this failure is disclosed. McGinn thinks the failure is that of ignoring the different rhetorical possibilities of commandments and parables: the former are or are intended to be unambiguous injunctions, the latter call for more or less elaborate interpretation. Nussbaum would consider this account, in turn, too crude and oversimplified. What she thinks traditional moral philosophy ignores are the different formal possibilities created by literary styles that are in no sense didactic but rather exemplary or enactive, and the different realizations of moral content they open up. They relate to ethical ideas not by saying or telling (however deeply interpreted they need to be) but by showing and being. For Nussbaum, moral or ethical philosophy is the enterprise of investigating conceptions of the human good. Part of the inquiry will be conducted along traditional lines, by means of definitions, distinctions, arguments, and objections, and this purpose is best served by philosophical texts and, as McGinn argues, by literary works that can be made to yield an “ethical lesson” under interpretive pressure. However, she claims …there are candidates for moral truth which the plainness of traditional moral philosophy lacks the power to express, and which The Golden Bowl expresses wonderfully. Insofar as the goal of moral philosophy is to give us understanding of the human good through a scrutiny of alternative conceptions of this good, this text and others like it would then appear to be important parts of this philosophy. (Nussbaum 1990:142) She suggests that James’s novel “calls upon and also develops our ability to confront mystery with the cognitive engagement of both thought and feeling. To work through these sentences and these chapters is to become involved in an activity of exploration and unraveling that uses abilities, especially abilities of emotion and imagination, rarely tapped by philosophical texts,” which have “a good claim to be regarded as important parts of the moral assessment process” (Nussbaum 1990:143). Like McGinn, Nussbaum is alive to the resources made available to philosophical reflection by literary art. Unlike him, however, she does not see novels as mere seams of moral truths to be mined “by engaging the [reader’s] mastery of folk psychology and applying it in a dramatic or narrative context.” For her the “ethical lesson” does not “fall out of this activation of cooperating faculties.” Rather, proper attention to the literary work, specifically to its style,

ROMAN BONZON 165

demands the employment and cultivation of the very abilities of emotion and imagination necessary also for “the moral assessment process.” She is closer in this respect to Murdoch, who had earlier suggested that the quality of a person’s attention and consciousness is internally linked to his pursuit of the good. (Murdoch, in turn, cites the influence of Simone Weil (Murdoch 1998:327).) She goes beyond Murdoch’s use of fiction in the example from The Sovereignty of Good by giving us in The Golden Bowl a literary work that engages the powers of literature in demonstrating how this can be so. Let us call the kind of moral truth Nussbaum’s use of literary works serves to make available experiential moral content to contrast it with the descriptive or merely propositional content McGinn’s method yields. To arrive at the latter kind of content requires the capacity to work out, by reliance upon common background knowledge, folk psychology, and sheer ingenuity, plausible moral interpretations of entangled and enigmatic texts. The former, on the other hand, demands openness to the variety of formal and literary effects of the texts, as well as exercise of sensitivities, both cognitive and affective, to the imaginative possibilities they set in play. McGinn requires the reader to figure out the meaning of a work of literature for the purpose of moral illumination; his concern is of a piece with the kind of interest Williams and Thomson direct at fictions. Nussbaum, on the other hand, asks that the reader experience for himself the moral significance of the elements of a literary work; indeed, that the moral content of a literary work become part of the reader’s experience. Her demand, though deeper and wider ranging, is continuous with Murdoch’s. Valuing fiction Nevertheless, for all the grandiose claims they pronounce on behalf of the value of fiction for ethical thought, one cannot overcome the feeling that neither Colin McGinn nor Martha Nussbaum takes literary works seriously as works of art in relation to moral thinking, that they somehow see them in radically diminished form, merely as codes to be cracked either for their ethical import or the benefit of the exercise. This is comparatively easily shown in the case of McGinn, for whom such aesthetic qualities as brilliant irony, witty dialogue, formal symmetry, and subtle self-referential aestheticism serve merely as ornaments to the more serious task at hand: “I shall attempt,” he says, “to tease out the conceptual apparatus with which Wilde operates in the book, and indicate what lessons it draws” (McGinn 1997:123). Aesthetic qualities are here conceived and presented as having exclusively instrumental or accidental value. But, as I will argue below, to fail to take the aesthetic qualities of a work as valuable in themselves, as intrinsically valuable, is precisely to fail to treat the work as a work of art. McGinn conceals his true reductive valuation of literary works of art by concentrating on novels whose obviously moralistic intent limits their artistic claims. Dorian Gray and Frankenstein, both fundamentally didactic Gothic tales, are suitable candidates for

166 FICTION AND VALUE

the McGinn treatment just because they offer him, finally, no complicating aesthetic distractions to impede the full achievement of his purpose. This valuation of literary aesthetic qualities, in turn, stems from McGinn’s unacknowledged conception of literary works—of novels in particular—as mere repositories of ingeniously hidden and entangled propositional moral content, the extraction of which is the true task of the moral philosopher inclined to indulge his taste for literature. But if this is so, then he fails to do justice to the fact that the moral themes and ideas are being extracted from something that is a work of art before it is anything else. This is shown by the fact that what he gets from, say, Dorian Gray can be comprehended in a non-aesthetically significant summary of the book. For McGinn’s purposes, such a summary would be just as good as (perhaps be even better than!) the work itself. (McGinn’s own book contains such a summary, and would therefore count as superior to Wilde’s novel on these terms.) It would then simply be fortuitous that Dorian Gray exists to tempt and titillate the analytic abilities of the ethical theorist and provide, moreover, occasion for display of acute literary criticism. There would be no philosophical advantage in the existence of artistic talent over the philosopher’s own capacity, displayed by Williams, Thomson, and Murdoch, to generate imaginary situations to house his moral ideas. Furthermore, the actual propositional moral content of literary works of art tends to be boring, trivial, commonplace, fatuous, or banal (Macbeth tells of the evils of regicide; Dorian Gray suggests that ugliness of soul can be concealed, temporarily, by beauty of countenance). This is because the propositional moral content of a work of art, so far from being its point (as McGinn would have it be), is really only the basis upon which the author is enabled to achieve whatever other purposes he has (meditating upon the nature of the link between ambition and virtue in Macbeth, for instance, or between beauty and virtue in Dorian Gray). Concentration upon its propositional moral content will avert attention from the true significance of the work. Finally, McGinn’s way with works of literature leaves it completely mysterious why anyone who has once identified the moral of a novel or play would have any interest in returning to it. The only response open to McGinn would be to claim that revisiting a literary work gives us a keener appreciation of its aesthetic qualities; but then, plainly, he cannot have been taking those qualities seriously into account in his use of it for moral purposes. Nussbaum, on the other hand, certainly registers an awareness that works of art essentially are bearers of aesthetic qualities, and that our regard for them properly depends upon our appreciation of those qualities. But she does not seem to have a clear sense of the relation between the aesthetic qualities of a literary work and its moral content, and therefore leaves it mysterious why a great literary work of art can encompass rich and ramifying material for ethical reflection while a lesser work cannot. Her choice of The Golden Bowl disguises this, because the high aesthetic achievement of James’s novel is inseparable from its depth and density as a moral text. It is amenable to Nussbaum’s method of close attention to stylistic

ROMAN BONZON 167

features just because that method does reveal those aesthetic qualities that bear the weight of the book’s ethical fabric. What about great literary works of art whose “moral lessons” elude capture in McGinn’s net? What ethical resource exactly does Persuasion, War and Peace, Madame Bovary, let alone The Golden Bowl or Ulysses provide a reader like McGinn? And how will Nussbaum cope with lesser works like Barchester Towers, The Mayor of Casterbridge, Brideshead Revisited or even The Picture of Dorian Gray or Frankenstein? What if we were to take literary works seriously as works of art, with greater or lesser value as such? How then will they relate to ethical or moral thinking? The value of fiction My answer to this question is that the value of a literary work for moral reflection depends entirely upon its value as a work of art. To support this claim and suggest how the dependence is to be conceived I will first have to provide an account, at least in outline, of the value of a literary work as a work of art; and to this end I start with the familiar idea that to take literary works as having value as art is to take them as valuable in virtue of the value of a certain kind of experience they are uniquely able to afford (cf. Budd 1995:4–11; Levinson 1996b:11–12; Kieran 2001:215–17). This kind of experience is best characterized as aesthetic. Aesthetics has to do with a fusion of the cognitive and the affective. Aesthetic experience is the experience of perceiving pleasurably or painfully the sensible or intelligible qualities of an object. However, unlike certain views that similarly appeal to pleasure and perception in their account of aesthetic experience—the most influential probably being Hume’s (Hume 1985)—the present view denies that the pleasure or displeasure felt by the subject is merely contingently or causally connected with his perception of certain qualities of the object; rather, I claim that cognition and emotion inextricably determine the character of the experience. It is impossible, in aesthetic experience, to perceive the relevant qualities of the object without experiencing the associated pleasure or pain since it is that feeling that allows us to perceive those qualities. Likewise, it is impossible to experience the relevant pleasure or pain implicated in an aesthetic experience of an object without experiencing it as a response to the associated qualities of the object since it is perception of those qualities that allows us to have that feeling. What differentiates our aesthetic experiences are the aesthetic concepts that characterize them. Hence, one object is elegant, another is meretricious; one is shallow, another profound; another sentimental, yet another moving. These concepts relate an object to a kind of experience it makes possible that is defined by a unity of perception and feeling, an experience that is, in short, aesthetic. This account of aesthetic concepts should be sharply distinguished from the traditional ways of thinking about them that we have inherited from such aestheticians as Beardsley, Sibley, Dickie, and Kivy. In the first place, this account of our possession of aesthetic concepts as necessarily involving dispositions to experience pleasure or

168 FICTION AND VALUE

pain cuts across the usual way of classifying aesthetic concepts as emotive or expressive, behavioral, reactive or affective, formal or regional or gestalt concepts (think about sad, nervous, and loose, which are, respectively, expressive, behavioral, and formal concepts but not, on my account, aesthetic ones). Second, the perceptual aspect of aesthetic concepts should not be limited to the sensory: perception can also be of such intellectual qualities as aptness, tact, humor, gaiety, vulgarity, wit, crudeness, profundity, and frivolity. These are qualities that are available in our experience of objects and that do not depend upon operations of causal laws or inferences from aesthetic principles. We simply perceive them. Interest in concepts involving an unanalyzable compound of the cognitive and the affective was sparked (or perhaps fanned) by Bernard Williams’s discussion of “thick” ethical concepts in Ethics and the Limits of Philosophy (Williams 1985:129– 30, 141–2). Thick concepts embody a union of properties and attitudes, of description and evaluation. In ethics, the concepts that have been characterized in this way include the Aristotelian virtues and vices like bravery, benevolence, cowardice, generosity, and the like. As I have described them, aesthetic concepts similarly embody a union of qualities and feelings, more specifically, of the qualities of objects that can be perceived (in the broad sense) by a subject and the feelings of pleasure or pain that qualify and are qualified by such perception. Hence, so far as pleasure and pain are not value-neutral, such concepts too have inherent descriptive and evaluative aspects. With respect to such ethical concepts as bravery and treachery, Williams writes that “how we ‘go on’ from one application of [the] concept to another is a function of the kind of interest that the concept represents, and we should not assume that we could see how people ‘go on’ if we did not share the evaluative perspective in which this kind of concept has its point” (Williams 1985:141). Precisely the same thing may be said of such aesthetic concepts as elegance. What enables us to group objects of attention together as elegant is not their common possession of some perhaps complexly specified set of nonaesthetic qualities, but the appropriateness of our use of the concept of elegance—along with its inevitable connotation of pleasure—with respect to them. Correct applications of such concepts require participation in the aesthetic interests, cognitive repertoire, and evaluative practices of their users, a like-mindedness and like-heartedness with them; the concepts are simply not accessible from the outside. Subjects without our emotional economy, or lacking our aesthetic tastes or our cognitive capacities could not possibly grasp or deploy them (cf. McDowell 1981:§2; Wiggins 1987:§§8–9). To forestall a possible objection it should be pointed out that what is involved in traffic with thick concepts is not necessarily the actual assumption of the asso ciated attitude—though that is the normal or paradigmatic case—but the recognition that the attitude is appropriate. Prevailing factors may prevent or inhibit a subject from experiencing the approbation or disapprobation, the pleasure or pain, he nevertheless knows he would rightly feel upon perceiving the relevant qualities on other circumstances. As with other “response-dependent” concepts

ROMAN BONZON 169

what is being invoked here is the disposition to experience the relevant feeling, not its actual experience. (I owe this qualification to Dom Lopes.) Now the value of a work of art as a work of art is the value of the aesthetic experience it is able to afford. It is of course not merely possible but likely that the value of a work of art comes also from its capacity to enlarge our understanding of the world, to engage our sympathies toward others, or to develop our own powers of imagination and discrimination. But these are benefits that are conferred by a work of art merely as an instrument for cognitive, social, or psychological ends. Its only value as a work of art resides in its capacity to provide the kind of experience that, as we have just seen, is characterizable solely in terms of aesthetic concepts. Hence the value of a work of art as a work of art—ergo the value of a work of literature as a work of art—is commensurate with the feelings of pleasure and pain captured by the aesthetic concepts invocation of which is required by an adequate account of our experience of the work. On this view, works of art have only instrumental value as works of art; it is their qualities given in aesthetic experience and represented by aesthetic concepts that are intrinsically valuable. Aesthetic and moral value With this account of the value of literature as art in place, I can finally address the topic of this essay directly. How is an appreciation of literary fiction related to an understanding of moral value? My answer is that they are related insofar as ethical ideas, or the substance of moral reflection, are related to aesthetic qualities. This means that the moral content of a literary work of art will be relevant to it as a literary work of art only if it is relevant to some aesthetic concept our experience of the work requires us to employ. To defend this claim I will need to distinguish two kinds of experiential moral content a literary work of art can encompass. There is, first, the kind that operates through the causal bases of our experience. Working from a conception of the psychological, moral, political, sociological, historical, and so on makeup of the audience, the author is able to activate those factors to cause the audience to have the kind of experience he wants his work to produce without the audience necessarily being aware that it is being worked on in this way. Here is an example from The Portrait of a Lady. We feel Isabel Archer’s horrified moral revulsion when we learn that Madame Merle is Pansy’s mother, and that is why she is so eager for Isabel to encourage—or at any rate not discourage—Lord Warburton’s attentions to the girl. Our reaction does not depend upon our consciousness of how this information is being conveyed; it is prompted simply by our moral beliefs and our knowledge of the plot. However, there is also the kind of experiential moral content of literary works that operates, not just causally, but also by routing the audience’s response through its understanding, grasp or appreciation of the way the causal factors are being deployed in the design, coherence, and structure—in brief, the medium—of

170 FICTION AND VALUE

the work. Let me give a crude but concise example from the same work. Portrait of a Lady opens in a very carefully described setting—the lawn of Gardencourt during afternoon tea—with the arrival of a young lady whose suitability for marriage is the subject of lively comment even before her physical appearance upon the scene. It ends in twilight again at Gardencourt after some years and many changes, with Isabel’s decision to return to her husband in Rome. Now the moral significance of these structural components—and it is reinforced throughout—of the novel does not lie in the moral content they make available either propositionally (there is none) or merely causally (say, our sympathy for Isabel and the kind of life she can look forward to with her unspeakable husband). Rather, what they signify is the journey Isabel makes from innocence to experience and the moral awakening and deeper sense of moral responsibility she arrives at in the end, as a result—as James puts it in his preface—of her “affronting her destiny.” And this realization can only be fully had by feeling the moral journey in the literary design and its significance. I will call this second kind of experiential moral content aesthetic, since it is experiential moral content made available through—and only through—an aesthetic experience. The pleasure he takes at the sense of completion, of symmetry, of arrival and departure meshing with departure and return, made possible by the work gives the sensitive reader satisfaction in the moral truth whose recognition is made possible by it. I emphasize that this aesthetic experience is not a purely formal one, though of course it requires the formal elements to be in place. Beyond its purely structural role one has to recognize that Gardencourt in that opening scene is so described as to hint at Eden (Powers 1970:68). One has also to realize that independence and unhappiness can be inseparably united (better: that freedom and happiness are not necessarily conjoined) for strong and self-possessed natures like Isabel’s. But it is the aesthetic experience that affords us the moral insight, which we could not have unless we saw it as involved in our having the aesthetic satisfaction that we do have. Our moral perception and experience are brought about by an awareness of its content being as it were bodied forth in the novel’s aesthetic design. Thick aesthetic concepts work by giving us the means of understanding why our experience is a certain way, why a certain configuration of features presents itself as that configuration on the basis of a feeling of pleasure or displeasure we are disposed to take in them. The sense of completion, of a limited whole, of basis and variation, of stasis and change, that we get from the pattern and design of character and incident in Portrait of a Lady proves pleasurable, not for purely formal reasons, but because they signify—give us a satisfying picture of—the nature of moral change and growth, that is, a conception of the nature of moral change and growth we feel to be true. Our pleasure at the design of the novel comes from our awareness that that design is the way it is because it represents part of moral reality, but a part we would not notice had our aesthetic experience not demanded explication in its terms. We feel deep and lasting satisfaction and pleasure upon finishing James’s great work. Why? In no small part because of its

ROMAN BONZON 171

shape. We are guided by the words and the plot (the handling of the medium) to feel this shape, but its true significance is a moral one. Now Nussbaum emphasizes the experiential moral content available in literary works of art, and in this way marks an advance over McGinn’s painfully constricted view. But she fails to make the distinction I have just drawn between the causal and the aesthetic. She unsystematically calls attention to both, and then claims that it is the large undifferentiated mass of experiential moral content that allows us to use works of art for the purpose of gaining moral insight. But I claim that its causal experiential moral content forms no part of the basis of a work of art’s being a work of art. The reason for this is that it is available through nonartistic means: vivid descriptions, propaganda, sentimental romances, thrillers, and the like. All these work directly upon the audience’s emotions and opinions to secure the experience their authors want to induce. In fact, experiential moral content is more effectively communicated in these forms, since in them the audience is not belabored with any interpretive demands: it is encouraged to let its critical defenses down and simply react. (More cynically: it is not afforded the opportunity to entertain the possibility of another way of relating to the work.) Hence, to include it as Nussbaum does among the benefits conferred by a literary work of art as a work of art is to betray a failure to comprehend the real value of a literary work of art for moral reflection. Aesthetic experiential moral content is the only kind of moral content that literary works of art contain as works of art. Furthermore, the aesthetic experiential moral content of a particular artwork is only available through it, since the (total) set of aesthetic qualities of an artwork is unique to it (this follows from the conception of aesthetic qualities as representational as well as formal, emotive, and perceptual), whereas the propositional or causal experiential moral content of any artwork whatsoever can be had in nonart replacements. Unlike our reaction to the revelation of Pansy’s parentage, our understanding and appreciation of the significance of Isabel’s moral journey can be had through no other route than an actual reading of Portrait of a Lady. This places me in a position to say: (1) there can be no substitute for reading a literary work of art to get its aesthetic experiential moral content; it cannot be had through some other means (counterexample, possible-world scenario, or other fictional device); (2) the aesthetic experiential moral content of a literary work may well be its point (as in, arguably, Portrait of a Lady)—though it need not be, as in the case of many lesser literary achievements; (3) rereading is reexperiencing, which explains why we keep returning to great literary works. Unlike the reexperiencing of causal experiential moral content, the reexperiencing of the aesthetic experiential moral content of a great work of art always induces a sense of wonder and newness: it opens up our view of how life is as it enriches our conception of what art makes possible. At every rereading of Portrait of a Lady we are conscious of what we missed

172 FICTION AND VALUE

the previous time, and know that the next reading will prove even better than the present one. Literature and moral philosophy It has recently been argued by Noël Carroll and Berys Gaut that a grasp of the ethical content of works of literature can be relevant to an appreciation of their aesthetic value (Gaut 1998; Carroll 2001a). Their arguments have been intended to counteract the view, associated with formalists and others who see art as inhabiting an autonomous realm, that there is no place for moral considerations in the aesthetic evaluation of works of art. My concern in this essay has been to argue what is roughly speaking the converse position: that the aesthetic content of such works not only can be relevant to their value as ethical texts but in fact completely determines it. Like Carroll and Gaut I oppose the idea that there is a sharp divide between ethical concerns and aesthetic merit; unlike them, my argument is with those who think that works of art can be made to yield their ethical lessons without regard for their aesthetic value. If a work of fiction is a work of art, then its distinctive moral content is recoverable only as part of an attempt to resolve the aesthetic concepts applicable to it. Recall Martha Nussbaum’s claims on behalf of The Golden Bowl: it “calls upon and also develops our ability to confront mystery with the cognitive engagement of both thought and feeling. To work through these sentences and these chapters is to become involved in an activity of exploration and unraveling that uses abilities, especially abilities of emotion and imagination, rarely tapped by philosophical texts” (Nussbaum 1990: 143). In other words, the process of acquiring moral enlightenment from James’s novel is essentially an aesthetic task. More generally (to encompass McGinn’s use of literature as an aid to moral reflection), the task of understanding a literary work in an attempt to learn from it is an aesthetic task if the work is a work of art, that is, its moral content is not propositionally or causally detachable from it in the way philosophical, didactic, or sensationalistic works not merely make possible but positively demand. Once the moral content—the moral value—available in a literary work is only accessible aesthetically, through cognitive, imaginative, and affective engagement with its style, form, characterization, and meaning, its acquisition is necessarily mediated by aesthetic concepts. Sometimes, as is surely the case with the novels of Austen, Tolstoy, and James, what Nussbaum finds in the specific case of The Golden Bowl applies: the moral content they allow a discerning reader to comprehend is not available otherwise. These works enrich and enlarge the ground of moral thinking, enabling us the better to pursue the larger ethical project of investigating conceptions of the human good. But this is only because they are all supreme works of art, permeated by manifold intense, luminous, and distinct aesthetic qualities. Experience of these is what enables us as readers to broaden our ethical horizons and deepen our moral understanding. The novels are irreplaceable as ethical texts precisely because they are irreplaceable as art.

ROMAN BONZON 173

Consider, by way of contrast, literary works that are valueless or have little value as works of art. If they nevertheless have important ethical lessons to teach, this will be because they are explicitly intended to be didactic. Their aesthetic qualities are deliberately confined to those necessary to make the moral point. In fact, it would be misleading to speak here (when the ethical value so far outstrips the aesthetic) of works of art at all; they are simply convenient fictions. Recall the examples discussed by Williams, Thomson, and Murdoch. They all have aesthetic qualities that are, however, mobilized only for the sake of the moral or ethical ideas they are designed to serve. These fictions are not created primarily to bear their aesthetic qualities; they are not meant to be works of art. They are pedagogical texts, and our approach to understanding them, like their authors’ intention in creating them, is strictly controlled by an awareness of this function. This brings us to a claim Colin McGinn makes, that “moral considerations enter intimately into the construction of fictional works. The novelist must constantly treat of moral questions, and take some position on them” (McGinn 1997: vi). I think it is possible to be skeptical here when the fictional work is a work of art. While it is true that if a literary artist’s main purpose is to convey some moral truth, then moral considerations will enter intimately into the construction of fictional works, it is not the case that a literary artist’s main purpose is to convey some moral truth. And while it does not follow that moral truths will therefore not enter intimately in the construction of fictional works, it would be surprising if the pursuit of nonmoral purposes—specifically, aesthetic ones —nevertheless necessarily involved the appeal to and deployment of moral truths in the body of the works. This is not to deny that all literary works of art do convey moral ideas or ethical perspectives—how could they not? But the sense in which this is true does not require literary artists to have them firmly in mind before (or even after) the process of creation. (I will return to this in the final section.) Compare on this matter the views of Iris Murdoch, who is in a position to speak with some authority about the role of philosophical—and specifically, moral — considerations in the creation of literary works of art. In an interview with Brian Magee, she says that she sees “no ‘general role’ of philosophy in literature. …Of course writers are influenced by the ideas of their time…but the amount of philosophy they succeed in expressing is likely to be small. I think as soon as philosophy gets into a work of literature it becomes a plaything of the writer, and rightly so. There is no strictness about ideas and argument, the rules are different and truth is differently conveyed” (Murdoch 1998:18–19). Magee is unconvinced: after all, was War and Peace not composed in part in order to articulate the distinctive philosophy of history Tolstoy wished to advance? Does Tristram Shandy not self-consciously and avowedly bear the direct influence of Locke’s theory of ideas? This forces Murdoch to explicitness about a theme that is among the deepest of her concerns, and that underlies all her writings, fictional and philosophical: “in general I am reluctant to say that the deep structure of any good literary work could be a philosophical one. I think this is not just a verbal

174 FICTION AND VALUE

point. The unconscious mind is not a philosopher. For better or worse art goes deeper than philosophy. Ideas in art must suffer a sea change. …Of course some writers reflect much more overtly but as in the case of Dickens their reflections are aesthetically valuable in so far as they are connected, through character for instance, with substructures which are not abstract” (Murdoch 1998:21; emphasis in original). Literature and philosophy Let us take it as established that a merely philosophical (and not, or not also, aesthetic) use of literary works will fail to take them seriously as works of art, and that a literary use of philosophical ideas will tend to distort or oversimplify them. What then is the proper way of conceiving the relation between fiction and philosophy? More specifically, how, in light of the intense interest the question has generated, which has elicited a variety of incompatible answers, are we to conceive the relation between ethics and literature? One answer, of course, is that there is no systematic or general relation. This is the view that Murdoch seems to be proposing in response to Magee’s questions. In this final section of the essay, however, I will discuss another possible answer, one which Murdoch also, later in the conversation, seems to think might well contain the truth about these matters. I will first lay out the view, and then say something about how the two views are not, appearances to the contrary notwithstanding, incompatible. I will conclude with the difficulty of accepting either view, and mention what seems to be the only sensible resolution. Immediately after denying that “the deep structure of any good literary work could be a philosophical one” Murdoch says: “there is always something moral which goes down further than the ideas, the structures of good literary works are to do with erotic mysteries and deep dark struggles between good and evil.” Magee is not slow to follow up: “if the fiction writer is dealing with ‘something moral which goes down further than the ideas’ this must mean that fiction unavoidably involves the writer in presuppositions of not only a moralphilosophical but even of a moral-metaphysical sort” (Murdoch 1998:21). Murdoch eventually responds: It is important to remember that language itself is a moral medium, almost all uses of language convey value. This is one reason why we are almost always morally active. Life is soaked in the moral, literature is soaked in the moral. …So the novelist is revealing his values by any sort of writing which he may do. He is particularly bound to make moral judgements in so far as his subject matter is the behavior of human beings. …The author’s moral judgement is the air which the reader breathes. …The bad writer gives way to personal obsession and exalts some characters and demeans others without any concern for truth or justice, that is without any suitable aesthetic ‘explanation’. …The good writer is the just, intelligent judge. He justifies

ROMAN BONZON 175

his placing of his characters by some sort of work which he does in the book. …A great writer can combine form and character in a felicitous way (think how Shakespeare does it) so as to produce a large space in which the characters can exist freely and yet at the same time serve the purposes of the tale. A great work of art gives one a sense of space, as if one had been invited into some large hall of reflection. (Murdoch 1998:27–8; emphasis in original) What is being suggested here seems to be that human life and human reality are already moral: there is no fundamental distinction between fact and value. Insofar as a work of literature is about human life and human reality—and what work of literature is not?—it is itself already moral: there is no fundamental distinction between fiction and value, either. How does this square with the claim that there is no general role for philosophy —hence, for moral philosophy—in literature? If what is meant by “moral philosophy” is a distinctive content that is capable of interacting (either as data or as consequence) with a moral theory, then that has no necessary or definite place in literature, in the creation and appreciation of works of literary art. The aesthetic concepts that govern and underlie the activities of artists and audiences do not recognize pre-established constraints and limitations imposed by ethical concepts except as necessary material for the aesthetic substance of a literary work, and any attempt to force recognition is, as Murdoch says, apt to damage the work. However, if what is meant by “moral philosophy” is the human activity of learning about and interacting with reality itself, then literature and moral philosophy are simply aspects of the very same activity. It is not clear that philosophers like McGinn or Nussbaum will be satisfied with these views, which depend on two distinct conceptions of the moral. On the narrow conception, we deprive ourselves of the resources of great literary works to which Martha Nussbaum’s description of The Golden Bowl applies: “major or irreplaceable work[s] of moral philosophy, whose place could not be fully filled by texts which we are accustomed to call philosophical” (Nussbaum 1990:138). But the broad conception, under which not only The Golden Bowl but many other literary works of art as well (though not, perhaps, Frankenstein or Dorian Gray) would be major or irreplaceable works of moral philosophy simply by virtue of being major or irreplaceable literary works of art, is too broad: it deprives us of any distinctive discipline of moral philosophy. These are painful options, but I do not see that there is any other choice. Let us formally adopt Bernard Williams’s distinction between “moral philosophy” and “ethics”, according to which the former is a species of the latter. Ethics has to do with the considerations relevant to answering Socrates’s question, “how should one live?” while morality or moral philosophy limits itself to certain of those considerations, namely those that have to do with obligation (Williams 1985:6–7). Then we may say: literary works of art have no special relation to moral philosophy, since they respond to different interests and aims; literary

176 FICTION AND VALUE

works of art, however, are intimately and essentially bound up with ethics, since they help us to answer Socrates’s question, and the kind of help they give is not available elsewhere. I will end this essay with a return to the beginning, and with a new beginning. What exactly is the relation of aesthetics to ethics? How, in other words, does a literary work of art help us with how we should live? This, as Plato knew, is the deepest question about art. The answer he gives in The Republic has seemed to many to be: it does not; it harms us. However, a more imaginative—an aesthetic— reading of the whole of Book X and especially of the Myth of Er at the end, and not just of the notorious opening section about the banishment of poets, has convinced one commentator and at least one of his readers that this is not in fact Plato’s final answer (Sinaiko 1998). In Metaphysics as a Guide to Morals Iris Murdoch asks this question, too. What she says there, adumbrated in the remarks I have quoted and developed here, indicates that she thinks this may also be the deepest question about life itself. Acknowledgment To James Shelley, Brian Kierland, and especially Dom Lopes I am grateful for help with the ideas and their expression in this chapter.

10 IMAGINING THE TRUTH An account of tragic pleasure James Shelley

The problem of tragedy—sometimes called “the paradox of tragedy”—has been understood in various ways. I will understand it here simply as the problem of explaining why tragedy gives us the pleasure it does, given that it has the content it has. In what follows, I propose a series of constraints that I claim any adequate solution to the problem must satisfy, and then I develop a solution that satisfies those constraints. But I do not claim that the solution I develop uniquely satisfies the constraints I propose. I aim merely to narrow the field of contending solutions, and then to draw attention to an overlooked contender in that narrowed field. I I propose that an adequate solution to the problem must satisfy four constraints: (1) The solution must be of the pain-pleasure variety. In his Values of Art, Malcolm Budd distinguishes between pain-pleasure and no-pain solutions to the problem of tragedy: pain-pleasure solutions posit complex hedonic responses to tragedy— responses that combine pain and pleasure, the pleasure presumably overbalancing the pain; no-pain solutions posit wholly pleasurable responses to tragedy (Budd 1995:116–19). I am unsure what need there is for an argument on behalf of painpleasure solutions. It seems obvious that our experience of tragedy is not one of unalloyed pleasure. In fact, it is not clear that anyone has ever really disagreed. Budd controversially interprets Hume as giving a no-pain solution, and notes that Aristotle can be controversially interpreted as also giving one (Budd 1995:110– 16). Still, however plausible the claim that someone has given a no-pain solution, considerably less plausible will be the claim that they set out to do so. When Hume, for example, sets out to explain why tragedy pleases when it seems that it should not, he does not additionally set out to explain why tragedy does not pain when it seems it should. That tragedy does not pain is simply not a datum to be explained; if anything, it is an unforeseen consequence of explaining what is to be explained: tragic pleasure. This leads me to believe that if the no-pain theorist does exist, he likely begins in the pain-pleasure camp, and only defects once he finds that he cannot explain why tragedy pleases without denying that it pains.

178 JAMES SHELLEY

(2) The pain and pleasure must be internally related. Pain-pleasure solutions can be divided into two kinds: those asserting a mere concurrence of pain and pleasure, and those asserting a relation between them, whether causal or essential. Given that it is the content of tragedy that pains us, the search for a source of merely concurrent pleasure will likely focus on form, as it does, for example, in Santayana: The agreeableness of the presentation is…mixed with the horror of the thing; and the result is that while we are saddened by the truth, we are delighted by the vehicle that conveys it to us. The mixture of these emotions constitutes the peculiar flavour and poignancy of pathos. (Santayana 1988:138) More recently, Mark Packer has argued that certain solutions to the problem— Aristotle’s, in particular—fail because of the erroneous assumption “that a causal or intentional relation must hold between the positive and negative affects produced by tragic drama” (Packer 1989:211). Packer’s own solution locates the source of tragic pain in tragedy’s presentation of “individual fictional characters and their particular plights” (Packer 1989:216) and the source of tragic pleasure in the “organization, necessity and universality” that tragedy embodies (Packer 1989: 217). But I think an insuperable difficulty faces any such merely concurrent painpleasure solution. Insofar as the pain and pleasure of tragedy are merely concurrent, there is no reason to believe that the pleasure cannot be had independently of the pain, and so no reason to endure the pain to have the pleasure. Tragedy has no monopoly on agreeableness of presentation, nor on organization, necessity, and universality. So why submit ourselves to tragedy to experience the pleasures these afford? I think that Aristotle’s assumption is right: the pleasure and pain of tragedy must be so related that the pleasure cannot be prised from the pain. (3) The pleasure must be one of relief. Aristotle is right, too, about the particular kind of pleasure to be explained. For the problem of tragedy is not the problem of explaining how tragedy might yield some pleasure or other: no problem I know of has its solution in an explanation of how tragedy might yield gardenvariety formal or mimetic pleasure. The problem of tragedy is the problem of explaining how tragedy yields its proper pleasure, that is, tragic pleasure. Aristotle describes that pleasure as one of relief, and while I am not proposing that an adequate solution must describe tragic pleasure in precisely this way, I am proposing that the pleasure it explains must be so describable. At a minimum, this means that the pleasure must be backward-looking, since to take relief in one thing is always to take relief from some other. A further consequence is that the explanation of that pleasure must be backward-looking as well, since explaining a pleasure of relief requires reference both to the thing we take relief in and to the thing we take relief from. So a solution, for example, that describes tragic pleasure as a species of delight, and then concentrates merely on features inherent in tragedy in explaining such pleasure, both fastens on the wrong pleasure and

IMAGINING THE TRUTH 179

offers an explanation of a correspondingly wrong kind. But a solution, for example, that describes tragic pleasure as a pleasure of resignation, as Schopenhauer’s does; or as a pleasure of consolation or comfort, as Nietzsche’s does; or as a pleasure of liberation, as Freud’s does; and then in explanation of that pleasure appeals both to tragedy itself and to the troubling situation from which tragedy delivers us—as Aristotle’s, Schopenhauer’s, Nietzsche’s, and Freud’s do— such a solution both fixes on the right kind of pleasure and offers an explanation of the correspondingly right kind. (4) The solution must be consistent with—if not explanatory of—the high value traditionally accorded the greatest tragedies. Satisfying this final constraint may seem comparatively easy. But, if I understand him correctly, Budd thinks it difficult, perhaps impossible, that any pain-pleasure solution should do so. When I earlier noted Budd’s distinction between no-pain and pain-pleasure solutions, I did not note his distinction of a third category—no-pleasure solutions— nor that his own solution falls within it. A no-pleasure solution is not a solution that denies that tragedy pleases us; rather, it is a solution that denies that pleasure plays a prominent role in explaining why we value tragedy. A no-pleasure solution, then, is not a solution to the problem of tragedy as I have defined it, but a solution to a broader problem that Budd refers to as “the problem of the nature and value of tragic experience” (Budd 1995:116). In offering his own no-pleasure solution to this problem, I think that Budd provides one of the best explanations of the high esteem in which we traditionally hold the greatest tragedies. The high value we place on tragedy, he maintains, is in large part a function of the high value we place on the truth it makes available to us: …the reason a spectator finds it intrinsically rewarding to submit herself to a process that involves her suffering over imagined tragedy is that she values acknowledging truths about possibilities inherent in human life, no matter how unpleasant they may be—truths that in everyday life, whenever possible, she is liable to push to the fringes of her consciousness. (Budd 1995:202) I think that Budd is right to link our esteem for tragedy to its capacity to get us to own up to truths that we disown in everyday life. But why make this point in service of a no-pleasure solution? Budd never answers this question, but I suspect its answer lies in his inability to find a link between tragedy’s capacity for asserting its painful truths and its capacity for pleasing. After all, given that the truths in question are so unpleasant that in everyday life we push them to the fringes of consciousness, why should they please when forced on us by tragedy? All the same, I think there are reasons for preferring a solution that managed somehow to link the two capacities. One reason appeals to our long history of thinking that pleasure must figure prominently in an account of tragedy’s value. Budd, so far as I know, is among the first to maintain otherwise, and this raises the question why so many philosophers—some with sharply varying notions of the role pleasure

180 JAMES SHELLEY

plays in our everyday lives—have agreed that pleasure plays a central role in our valuing tragedy when it does not. Of course they could all be wrong, but their all being so would itself require explanation, and this is something Budd does not undertake. A second concern is that it is not clear how a no-pleasure solution can make sense of the way we look forward to attending performances of tragedy. We believe that trips to the dentist are valuable: we are willing to part with money and to brave pain for the good dentistry provides. But we do not look forward to such trips. Why, if Budd’s solution were correct, would our attitude toward attending tragedy be any different? II I will take a step toward offering my own solution by considering Freud’s essay “Creative Writers and Day-Dreaming.” Ritchie Robertson has recently dismissed this essay as “notoriously inadequate” because it “does nothing to explain the power of great art, although or because it deals with painful material, to give us a feeling of satisfaction, harmony, and uplift” (Robertson 1999:xxi). Whether Robertson is right depends on how strong a claim he is making. If the claim is simply that Freud ultimately fails to explain how art that deals with painful material might (among other things) please us, it is true enough. But if the claim is that Freud says nothing of use in explaining how art dealing with painful material might please us, then it is false, as I hope to show. Freud claims that we take pleasure in works of narrative fiction because they, like daydreams, present as fulfilled one or more of our unfulfilled desires. This may make it seem as if pleasure in narrative fiction just is pleasure in the contemplation of pleasant content—pleasure in the contemplation of a pleasant fictional world. But while Freud does not deny that we take such pleasure (he thinks that the pleasure we take in narrative fiction “probably arises from the confluence of many sources” [Freud 1959:153]), his emphasis is less on the content of the fictional world than on the troubling situation from which contemplation of that world releases us. Wish-fulfillments are essentially “correction[s] of unsatisfying reality” (Freud 1959:146), and “our actual enjoyment of an imaginative work proceeds from a liberation of tensions in our minds” (Freud 1959:153)—tensions between wishes or desires that the world be one way and the reality that it is some other. So narrative fiction relieves us from a world that is in tension with desire by replacing it with a world that is not. Clearly this account cannot be made to apply to tragedy without serious accommodation, for the obvious reason that tragedy appears to be the very opposite of wish-fulfillment: in the fictional worlds tragedy portrays it is not desire but fear that appears to find fulfillment. Freud’s strategy is to explain away this appearance. If we do not recognize tragedies as wish-fulfillments, this is because we have disowned the wishes they fulfill: the wishes that find fulfillment in tragedy tend to be sufficiently abhorrent to consciousness to have been banned from it. So—to take the obvious example—we take pleasure in Oedipus the King

IMAGINING THE TRUTH 181

because it liberates us from a world in which our unconscious oedipal desires are unsatisfied by giving us instead a world in which they are satisfied. However resourceful this accommodation, I think that the fear-fulfilling appearance of tragedy cannot be explained away, at least not the way Freud attempts here. A glance at the argument of “Creative Writers and DayDreaming” will bring this out. At its center is a comparison between the content of daydreams and the content of popular narrative fiction. That daydreams are wish-fulfillments is obvious given their content. So to the degree that we find an overlap of content in works of popular fiction, we have reason to believe that they too are wish-fulfillments. And we do tend to find similar content in such works: we find “a hero who is the centre of interest, for whom the writer tries to win our sympathy by every possible means”; we find that “the women in the novel invariably fall in love with the hero”; we find that “the other characters in the story are sharply divided into good and bad,” such that “the ‘good’ ones are the [hero’s] helpers, while the ‘bad’ ones are [his] enemies and rivals” (Freud 1959:150). And, finally, there is the “one feature above all that cannot fail to strike us about the creations of these story-writers”—that the hero seems to have been placed “under the protection of a special Providence”: If, at the end of one chapter of my story, I leave the hero unconscious and bleeding from severe wounds, I am sure to find him at the beginning of the next being carefully nursed and on the way to recovery; and if the first volume closes with the ship he is on going down in a storm at sea, I am certain, at the opening of the second volume, to read of his miraculous rescue—a rescue without which the story could not proceed. The feeling of security with which I follow the hero throughout his perilous adventures is the same as the feeling with which a hero in real life throws himself into the water to save a drowning man or exposes himself to the enemy’s fire in order to storm a battery. It is the true heroic feeling, which one of our best writers has expressed in the inimitable phrase: ‘Nothing can happen to me!’ It seems to me, however, that through this revealing characteristic of invulnerability we can immediately recognize His Majesty the Ego, the hero alike of every daydream and of every story. (Freud 1959:149–50) I have no complaint with the argument so far. I think it is a good argument to the conclusion that some works of narrative fiction are wish-fulfillments. But Freud wants to broaden this conclusion to encompass all imaginative writings. He concedes that he is “perfectly aware that very many imaginative writings are far removed from the model of the naïve day-dream,” but yet finds that he cannot “suppress the suspicion that even the most extreme deviations from that model could be linked with it through an uninterrupted series of transitional cases” (Freud 1959:150). I do not doubt Freud’s suspicion, but you can grant it without granting anything interesting about works deviating from the model of

182 JAMES SHELLEY

the naïve daydream: the most the truth of Freud’s suspicion could show is that the line blurs between works we have reason to believe are wish-fulfillments and works for which we have no such reason. But the problem is not merely that Freud has failed to give us reason to believe that all works of narrative fiction are wish-fulfillments; the problem is that he has given us good reason to believe that some are not. For the content characteristic of tragedy does not merely deviate from the content characteristic of wish-fulfillment: the content characteristic of tragedy crucially opposes the content characteristic of wish-fulfillment. Consider the feature of wish-fulfillment that Freud makes so much of. If it counts as evidence toward a work’s being a wish-fulfillment that its hero seems to have been placed “under the protection of a special Providence,” that he is sure to evade disaster no matter how he flirts with it, that, in short, “nothing can happen to him,” then it must count as evidence against a work’s being a wish-fulfillment that its hero seems to have been placed—perhaps has been placed—under some divine condemnation, that he is sure to succumb to disaster no matter how he strives to avoid it, that there is, in short, nothing that cannot happen to him. If tragedy appears to be the opposite of wish-fulfillment, then this is because it is. Why think, then, that “Creative Writers and Day-Dreaming” is of use in explaining how tragedy might please us? I think it is of use precisely because tragedy is the opposite of wish-fulfillment. My hypothesis is that if you take Freud’s account of the pleasure of wish-fulfillment and invert it at the right point, you will have the basis of a solution to the problem of tragedy that will satisfy the four constraints proposed above. Freud, if I have understood him, believes that wish-fulfillment pleases us by resolving a tension between desire and reality. But a tension between desire and reality is, in principle, resolvable in either of two directions. Wish-fulfillment pleases us by resolving the tension in favor of desire: it liberates us from a world in which desire has been thwarted by reality by replacing it with a situation in which desire prevails. Tragedy pleases us, I conjecture, by resolving that same tension, though in reality’s favor: it liberates us from a world in which reality has been thwarted by desire by replacing it with a world in which reality prevails. III I will try to lend both clarity and plausibility to this conjecture by explaining what this talk about desire and reality thwarting one another comes to, conceding from the start that the explanation will proceed at a regrettably abstract level. Both desire and reality constrain belief. To say that reality constrains belief is simply to say that belief is constrained by what is the case, or, more precisely, by what is the case according to the best available evidence (I shall ignore this qualification in the future, as nothing here rides on it). But the fact of selfdeception implies that we cannot explain all belief by appeal to this constraint: in cases of self-deception we believe the world to be a certain way not because it is but because we desire it to so be (Mele 1987:125). That reality constrains belief I

IMAGINING THE TRUTH 183

take to be uncontroversial; that desire additionally does, I take to be no more controversial than the existence of self-deception. I will say that beliefs are troubled to the degree that these two constraints oppose one another. Troubled beliefs present a continuum, running from those in which the truth-constraint clearly prevails to those in which the desire-constraint clearly does. I will call belief of the former kind truth-prevalent and belief of the latter kind desire-prevalent (insofar as they are false, which in the normal case they are, desireprevalent beliefs are self-deceptive). I see no reason to believe that troubled beliefs divide without remainder into these two kinds: between them we will presumably find a series of cases in which neither constraint clearly prevails. Troubled beliefs, then, are beliefs that we are under some constraint or pressure not to have. We resist having them, with a resistance proportional to the strength of the non-prevailing or thwarted constraint. The pleasures proper to wishfulfillment and tragedy, I propose, are the pleasures of being relieved of such pressure—the pleasures of being able to give in to that which, in everyday life, we have been resisting. I propose that the pleasure proper to wish-fulfillment is the pleasure of giving into the pressure exerted by desires on truth-prevalent beliefs. Wish-fulfillment relieves us from the pressures of thwarted desires by replacing a world in which the truth prevails over those desires with a world in which it does not. I propose that the pleasure proper to tragedy is the pleasure of giving into the pressure exerted by reality on a desire-prevalent belief. Tragedy relieves us from the pressures of thwarted truths by replacing a world in which desire prevails over those truths with a world in which this does not happen. With this in mind I want to return to Budd’s remark about the value of tragedy, cited in the opening section of this chapter: …the reason a spectator finds it intrinsically rewarding to submit herself to a process that involves her suffering over imagined tragedy is that she values acknowledging truths about possibilities inherent in human life, no matter how unpleasant they may be—truths that in everyday life, whenever possible, she is liable to push to the fringes of her consciousness. (Budd 1995:202) Budd, we recall, makes this remark in service of his no-pleasure solution to the problem of the value of tragedy, convinced, it seems, that tragedy’s alleged capacity for pleasing cannot be linked with its capacity for bringing to consciousness “truths about possibilities inherent in human life” that are so unpleasant that we have pushed them from it. I have now asserted a link between the two. We take pleasure in the truths tragedy forces into consciousness precisely because we have been forcing those truths from consciousness. It is because we have been resisting such truths in everyday life that we take such relief in being able to give in to them in tragedy. There is, however, this asymmetry between tragedy and wish-fulfillment. The reason that truth does not prevail over desire in wish-fulfilling fictional worlds is

184 JAMES SHELLEY

that desire faces no opposition there. The truths that thwart desire in the everyday world are replaced in wish-fulfilling fictional worlds by fictional truths that are in concert with desire. The reason that desire does not prevail over truth in tragic fictional worlds, however, is not that truth is unopposed there. The desires that thwart truth in the everyday world are not replaced in tragic fictional worlds by fictional desires that are in concert with the truth. Desire, unlike truth, holds constant across the fictional/nonfictional divide. So in tragic worlds truth prevails though not unopposed—it prevails over the very desires that prevailed over it in the everyday world. So whereas the move from engagement in the everyday world to engagement in a wish-fulfilling world is a move from a troubling situation to a non-troubling one, the move from engagement in the everyday world to engagement in tragedy is a move from one troubling situation to another. Thus whereas the relief of wish-fulfillment may be likened to the relief of temporarily laying down a burden, the relief of tragedy is more like the relief of temporarily shifting a burden, say, from one set of muscles to another. This difference between tragedy and wish-fulfillment may help to explain the comparative ease with which we depart from everyday contexts to wish-fulfilling as opposed to tragic contexts. It may also help to explain the comparative popularity of wish-fulfillment over tragedy. But while there is this asymmetry, and it seems to favor wish-fulfillment, it is balanced by a second asymmetry, one that seems to favor tragedy. Although wishfulfillment substitutes non-troubling fictional truths for troubling real truths, we do not actually get to believe those substitute fictional truths, nor (more importantly) do we actually get to disbelieve the real truths they have replaced. Wish-fulfillment replaces—or perhaps, more accurately, displaces—a troubling belief with an opposing, non-troubling make-belief. So we are not actually but merely imaginatively liberated from that troubling belief. And though the relief afforded by this imaginative liberation is real enough, it is not comparable, I think, to the relief afforded by actually being liberated from troubled belief. Tragedy, as noted, replaces a troubling falsehood with a no less troubling truth, but it is a truth that we actually get to believe, and the falsehood it replaces is a falsehood that we actually get to disbelieve. Tragedy—so long as its spell lasts—replaces belief not with make-belief, but with belief. This difference between tragedy and wish-fulfillment may help to explain the comparative power of the former. Both asymmetries point to a question regarding tragedy corresponding to which there is no question regarding wish-fulfillment: how does tragedy secure the prevalence of truths over the very desires that, in everyday life, prevail over those truths? How, in other words, does tragedy secure our belief in truths that we are unable to believe in everyday life? This question deserves a much more complicated answer than the one I am about to give, in spite of the fact that the answer I am about to give begins with the admission that things are more complicated than I have let on. Just now I said that in tragic worlds truth prevails over the very desires that prevailed over it in the everyday world. But this cannot be true. Tragedies are works of fiction.

IMAGINING THE TRUTH 185

Tragic worlds—the worlds tragedies prescribe that we imagine—are fictional worlds, and so are composed of fictional, not real, truths (Walton 1990:41–3). So it cannot be true that real truths prevail over real desires in fictional worlds, since there are no real truths in fictional worlds. I also said, just now, that tragedy replaces belief not with make-belief, but with belief. But this too now appears suspect, for similar reasons. Tragic worlds consist of fictional truths, and fictional truths are objects not of belief but of make-belief (Walton 1990:39–41). It thus seems that it cannot be true that tragedy replaces belief with belief. When I said the things that now appear suspect, if not patently false, I was speaking loosely. Speaking strictly, in tragic worlds it is not a real truth but a fictional one that prevails over our desires, and we do not, of course, believe but make-believe that fictional truth. But the fictional truth that prevails over our desires is the same in content as a truth over which those very desires have been prevailing in everyday life: we make-believe that p, where p is a true proposition that our desires have been preventing us from believing. These considerations, I suggest, will lie at the heart of a complete answer to the complicated question at hand. For however resistant we are to believing certain propositions, we are surely less resistant to imagining them to be true. And all that tragedy ostensibly asks of us is that we imagine, and where is the harm in that? We know that nothing that happens in a performance or reading of a tragedy is real. We know, in particular, that the calamities that engulf tragic heroes will not spill across the fictional/nonfictional divide to engulf us. What we do not suspect is that tragedy has linked our assenting to the fictional truth of certain propositions to our assenting to their truth. Suppose that a particular tragedy prescribes that we assent to the fictional truth of the proposition that a good man, acting in accordance with his best judgment, acts in a way that leads, unforeseeably though with astonishing ease, to his own destruction. In assenting to this fictional truth, we assent to the fictional truth of the proposition that a good man, acting in accordance with his best judgment, may act in a way that thus leads to his own destruction. At the same time we see that there is nothing fiction-specific about the relation between the man, his actions, and the destruction they unforeseeably bring. For we see the man performing the actions, and the actions leading to the destruction, all in accordance with the standards of probability and necessity at work in our own world (Aristotle 1987:40). And though there may be fictionspecific elements to the man and to his destruction, they are not fiction-specific in ways that prevent our seeing the man as genuinely human and our seeing his destruction as a genuine—and terribly near—human possibility. And so we assent, in spite of ourselves, to the truth of the proposition that such a man, acting in accordance with his best judgment, may act in a way that thus leads to his own destruction. In this way, tragedy may be said to smuggle assent to tragic propositions across the fiction/reality border. We find ourselves believing something that until now we had been unable to believe. And, if the central thesis of this chapter is correct, we find ourselves pleased to believe it.

Part III SENSORY IMAGINATION

11 SEEING TWICE OVER Christopher Williams

It is possible to characterize the imagination by means of two contrasting models. On the one hand, we can regard the exercise of the imagination as a type of conception, and so come to think of imaginability as equivalent to conceivability. On the other, we can regard the imagination as having a closer relation to perception. This second characterization surfaces in use of the colloquial term “the mind’s eye”, and the naturalness with which we use that term, and other language that is like it, when we talk about the experience of art suggests that the perception model plays an important role in our thinking about art. Yet latter-day philosophers have tended to find the perception model unattractive—and not just in the special context of the arts, but everywhere, on general grounds having to do with the empiricist account of the mind. The modern antipathy to the imagination-as-perception view is traceable to Ryle, who thought that the imagination-as-perception view had the absurd consequence that people must have duplicated sets of seeings (Ryle 1949). According to him, a child who plays with a doll, imagining that the doll smiles, would see an unsmiling doll in public space and an unattached smile in a private picture gallery, as it were. But if the perception view is to have this consequence we have to suppose that imaginative perception and ordinary sense perception are somehow in competition with each other, that imaginative perception has to be adjacent to, or has to exclude, sense perception. Now it is absurd to think that people see double in the way that Ryle indicates, but there is nothing in the perception view as such that calls for seeing double. The smile that the child sees is on the doll, though it requires an imaginative turn for the child (or anyone else) to see it. This imaginative exercise may be said to be a seeing, but because it is possible to see the doll unimaginatively as well, without the smile, we could say that the imaginative seer does see the doll twice over—once without the assistance of the imagination, and once with. The doubling here, however, must not be interpreted as a pair of successive perceptions, lest we find ourselves with the duplicated set that Ryle rightly found incredible. Rejecting the imagination-as-perception view altogether can make it difficult to recognize some obvious truths for what they are. Hume—Ryle’s leading target — correctly observed that “pity depends, in a great measure, on the contiguity,

188 CHRISTOPHER WILLIAMS

and even sight of the object; which is a proof, that ’tis deriv’d from the imagination” (Hume 1978:370, my emphasis). Some exercises of imagination are not just causally dependent on sense perception, but are present in sense perception itself. Hume’s pity provides one example, but aesthetic experience provides a range of general examples. If we consider aesthetic contexts, we find especially favorable conditions for encouragement of the thought that a less ordinary type of perception can operate within the more ordinary, all the while remaining distinguishable from it. Successful artistic achievements at once furnish us with a focus for sense perception and yield imaginative satisfactions. In this essay, I want to show how these two perceptual exercises are related in the type of attention we bring to bear on aesthetic objects. This attention is an activity that the older aestheticians’ notion of “the aesthetic attitude” acknowledged, however imperfectly, and I hope to retrieve what was right about that notion, which seems to me artificially out of fashion. To suppose that psychical distance—to take the version of the aesthetic attitude that I plan to revisit—may enter into a due appreciation of a work is to make a supposition that some persons would regard as resting on an exploded psychology. Yet no such stigma attaches to the idea that imaginative resistance figures centrally in our aesthetic responses, and it has been noted that imaginative resistance has been a relatively unexplored topic of inquiry (Moran 1994:95; Gendler 2000:55). But if my remarks about the role of imagination in aesthetic experience are correct, it may emerge that psychical distance and imaginative resistance are much the same concept. In what follows, I begin by exploring the distinction between the two models of imagination, conceptual and perceptual. I then suggest that insofar as we seek to align the imagination as it is deployed in aesthetic experience with one model or the other, the alignment should be with the perceptual. We will then be in a better position to rehabilitate what we can of the aesthetic attitude and its underlying psychology. Sometimes it is possible to think of the exercise of the imagination in a valueneutral fashion, as merely something that we do. But it is also possible to think of the exercise evaluatively: sometimes a person does something valuable by exercising her imagination, and we want to commend her for the exercise. If our interest is to evaluate, one ready way to express a positive assessment consists in calling a person imaginative. Although “imaginative” can be given a value-neutral sense too, the link between imaginativeness and assessment is much stronger than the link between imagination and assessment. It is through the concept of the imaginative, in its common assessment sense, that I want to develop the idea that there are two types of imagination. The reason is that there are two types of imaginativeness. When we say that a person is imaginative, one thing we can mean is that the person has a remarkable facility in conceiving counterfactual states of affairs, or at least states of affairs that are not yet known to be factual or concerning whose factuality the person has no reasonable grounds for belief. We could call

SEEING TWICE OVER 189

these states of affairs “possibilities”. Imaginativeness, so understood as having to do with the conception of possibilities, is found in varying degrees in people, and a highly imaginative person will be marked by the degree to which her imaginings of possibilities are unusual, or by the extent to which those imaginings depart from, or are opposed to, the known facts. This is a very rough sketch, but it will suffice for the purpose of generating examples. To begin with an elementary case: a young child may be reluctant to go to bed because of a frightening story he has heard or watched, and his reluctance may be traced to his imagination. He is afraid (he says) of the goblin that is lurking in his toy chest. The child’s imagination is already running wild, but it may run wilder still: if the toy chest is opened and no goblin is discovered, the more imaginative child could explain that the goblin failed to appear because this particular goblin takes special pains to avoid detection whenever the adults are inconveniently in the room. But if it improbably transpired that a goblin was lurking in the toy chest after all, this discovery would not diminish the child’s imaginativeness. We might be disposed to say that the child’s goblin fantasy had a prophetic, anticipatory quality, in view of the surprising facts, but the fantasy itself was still the product of his imagination; and no belief is reasonably founded on its very free exercise. Imaginativeness of the sort the child exemplifies can be far more sophisticated. At the other end of the spectrum, Locke certainly evinced that he was an imaginative thinker by theorizing that Socrates and the mayor of Queenborough could share the same soul and yet be distinct persons; and the personal-identity theorists who are Locke’s successors have often shown themselves to be considerably more imaginative than Locke himself. By juxtaposing these elementary and sophisticated cases as I have done, I do not mean to cast an invidious shadow over the sophisticated member of the pair. That is, my intention is not to pronounce on the value of Lockean thoughtexperiments by assimilating them to childish fantasies. But there is a real issue to discuss here concerning the difference between imaginings that qualify as imaginative and those that do not, an issue that the juxtaposition may blur. Although we might praise the child for an elaborate bedtime-circumventing narrative, were it not for the stake we may have in getting him off to bed, it seems to me that we would, on further reflection, decline to attribute imaginativeness to him, quite apart from practical exigencies. The child, we may assume, is not disposed to test the goblin hypothesis, and he may not be able to distinguish actual goblins from hypothetical (whatever that might mean). If his imaginings are merely loose reveries, we might say that the child is imaginative in the value-free sense; but the child has to give us more in order to be imaginative in the preferred assessment sense. Imaginings have to be constrained by reasonable procedures for assessing belief-candidates, and have to be integrated with our other beliefs, if they are to count as properly imaginative. Whether Locke and his successors are properly imaginative likewise depends on considerations that are identifiable

190 CHRISTOPHER WILLIAMS

separately from the mere exercise of imagination that produces the exotic scenarios for their thought-experiments. Yet viewed from within, there is an important core similarity between the exercises of imagination in the two cases, the child’s and the Lockeans’. We do not call a person imaginative (in any sense) who is simply able to think about the world as the world exists, no matter how encyclopedically accurate the conception. We do call a person imaginative who is able to frame and entertain possibilities (though, once again, if we know just this we do not know enough to say whether an elevated compliment is appropriate). In any event, it would be useful, for the sake of the discussion, to have a synonym for “imaginative” in the spectrum of cases defined by children’s artlessness and philosophers’ sophistication. I would propose “ingenious” as a defensible choice, for in the class of imaginings these cases represent, it will generally be true that we can substitute “ingenious” for “imaginative” without loss of meaning and with a reproduction of the same evaluative and nonevaluative options. For the next class of cases whose imaginativeness I wish to characterize, an acceptable synonym for “imaginative” is more elusive. The cases that interest me chiefly are artistic, and I think that they are different from the preceding cases, and should not be assimilated to them. The artistic examples to which I will confine myself are literary; and doubtless this policy calls for justification, in view of the claim that sensory and imaginative elements occur together in aesthetic experience. When we read a novel, we do not rely on our senses for enjoying the work in the same way that we do when we look at a painting. For this reason, Peter Kivy has held that it is unhelpful to make generalizations from painting to literature, and that a literary work, in contrast to works in other art forms, has to be understood, not appreciated (Kivy 1997). The difficulty of identifying a perceptual component in a literary work is easily exaggerated, however, and one reason for thinking this is that, since some distinction between appreciation and understanding has to survive even in the case of literary works, it is not evident how we would capture the idea of appreciation without appeal to at least a perception-modeled component. And on the assumption that such a component is needed, we are brought back to the point of using literary examples. If aesthetic enjoyments are typically imaginative in a way that differs from the imaginativeness on display in my first set of examples, then showing how the difference emerges even for literary enjoyments—a seemingly inauspicious setting for the argument—should strengthen the feeling that there is indeed a difference. We sometimes hear it said that Tolstoy or Flaubert or Proust or Mark Twain are imaginative writers, which is equivalent, here, to saying that they are writers of artistic distinction. Their claim to merit is necessarily connected to the exercise of their imaginative abilities, but the first picture of imagination I have sketched provides an extremely meager explanation of what the imaginativeness of these writers consists in. Anna Karenina is a significant artistic achievement, and yet the plot of Tolstoy’s novel hardly required of its author an ability to conceive

SEEING TWICE OVER 191

recondite possibilities: at one admittedly plain level of description, the novel is the story of a Russian noblewoman who has an extramarital affair and eventually comes to grief. Tolstoy is doubtless less imaginative than (say) Melville with respect to what literary critics have called invention, which better illustrates the first type of imagination (since a story about an obsessive sea captain and the whale he pursues is at least mildly ingenious). But Tolstoy is an imaginative writer nonetheless, and his work, in the simplicity and familiarity of its story line, is not unrepresentative of the work that serious novelists do. If Tolstoy’s imagination is not expressed primarily in invention, we need to understand how this is so. A gifted imaginative writer delineates characters, circumstances, manners and opinions, or cultural moments in a peculiarly vivid, engrossing way. The language we have for describing the manner of delineation is imprecise, but we can and do say such things as these about Tolstoy’s characters: that Tolstoy makes them come alive for us, that they have a lifelike quality, that they are three-dimensional, that they are not cardboard spokespersons for the author’s religious or political enthusiasms, or that we see them in the round. It might be held that such talk is merely casual and philosophically uninteresting, but such a judgment would be premature, inasmuch as it is not really possible to formulate a favorable critical verdict on a novel—at least a novel that does not self-consciously flout our expectations of psychological engagement—without resorting to this style of discourse. If we take the practice of criticism at face value, then we have to explain the difference between the gifted imaginative writer and the writer whose work fails to elicit a positive evaluation. We could say that Tolstoy has an ability to envisage his characters and their situations, and that envisaging is the ability to present the characters and situations to the readers as if they were ordinary perceivable items that attract, for whatever reason, our sustained attention. The imaginative achievement of Anna Karenina lies in Tolstoy’s use of the telling verbal detail to attain just this presentation. If Tolstoy is not particularly ingenious, and if ingenuity as such is not required for artistic achievement, Tolstoy remains a perceptive writer, and a successful novel does need to demonstrate perceptiveness (though not invariably about character psychology). Thus, “perceptive” is a synonym for “imaginative” in the second type of case; and with this gloss it is even a more challenging project to try to understand imaginativeness independently of assessment. Still, if we insisted on finessing a technical, value-free meaning for “perceptive”, we would be left, perhaps, with the picture of a person who simply visualizes things that do not attract sustained attention. As with the imaginativeness that concerns the conception of possibilities, something more than the bare exercise of a capacity is needed for the exercise to be valuable. (But a difference between them is that true perceptiveness, in contrast to true ingenuity, can be verified by careful inspection of the imaginative exercise itself: a fact about perceptiveness that aesthetic formalism attempted to recognize, however clumsily.) On the proposal I am offering, then, an imaginative person will be either ingenious or perceptive. And although a person can be either ingenious or

192 CHRISTOPHER WILLIAMS

perceptive, we do not think ingenuity and perceptiveness are the same thing. And hence the reason for caring about acceptable synonyms, which are otherwise dispensable, is to make explicit a real difference between two kinds of imaginativeness. So far, for both kinds, I have concentrated on the creator, the person originally responsible for the imaginative exercise, but concentrating on the creator is merely an expository help. It is obvious that a person other than the original imaginer can entertain the fearful child’s thought, or Locke’s, and so be said to imagine the same thing that they did. We would probably not call the noncreator ingenious, but she is using her imagination to understand the ingenious conception, and is therefore derivatively ingenious. Similarly for Tolstoy’s reader. We would not call him perceptive (unless he were a critic who tellingly uncovered the features of the text that account for a reader’s satisfaction), but he would at least come to perceive things because of Tolstoy’s writing, and by relying on his imagination would be derivatively perceptive. Answering to the two kinds of imaginativeness are two kinds of imagination, which may be labeled “conceptual” and “perceptual”. Although it is possible to have a mixed imaginative project that incorporates both kinds of imagining, the two kinds are distinguishable in principle. But the terminology of perception might seem alarming. While it is unproblematic to say that a writer is perceptive, or that a reader has been made perceptive, it is not obviously unproblematic to say that the writer or reader has perceptions. The general root of the discomfort here is that the paradigm of perception, sense perception, does not strictly apply to the type of case we have considered. Sense perception requires a bodily organ, and any given perception is causally dependent on, and covaries with, states of the perceived object. But because the perceptive exercise is modeled on sense perception, we could speak of the exercise as “quasi-perceptual” if we did demand that a perception must conform strictly to the requirements of sense perception. That demand seems gratuitous to me, but the clarification should remove any lingering cause for alarm. I shall conclude this preliminary account of imagination by making an observation regarding two further differences between the two kinds. One difference involves an inference that one imaginative exercise licenses, but not the other; and a second difference involves the object of the exercise. In the conceptual type of exercise, if we say that a person imagines that she is -ing, it is typically the case that the person is not -ing or does not believe that she is -ing (unless the person is under a misapprehension about her mental action or the origin of her belief). This inference is assumed, for example, by a recent writer who says that in Triumph of the Will Leni Riefenstahl “doesn’t just ask us to imagine finding the Führer and his message appealing, but actually to find them so” (Devereaux 1998:241). The other feature to notice about conceptual imagining is the readiness with which we can represent a person’s imaginative exercise by using the imagine-that locution. To imagine a goblin in the room is to imagine that there is a goblin, in the example as it was originally presented. More

SEEING TWICE OVER 193

generally, it is to imagine that p, where the that-clause takes an object that is represented sententially. It is then an innocuous step to reconstruct statements of the form “S imagines that p” as having the shape of “S fictionally believes that p” or “it is fictional for S that p”. For perceptual imaginings, there is no required inference to the unreality of the imagined item or to nonbelief in the imagined item. To imagine the world of Tolstoy’s novel is to imagine an order of things that is continuous with the order we inhabit in our workaday lives, but we do not take, or do not have to take, a doxastic stand on the status of the extension that appears in the novel. To say that its characters are true to life is to indicate nothing more than a certain idealized neutrality about their status in our beliefs. (This part of the suggestion revives a portion of an idea that Collingwood had when he distinguished imagination from make-believe, as he understood them.) As for the object of the imagining, it is not the case that we must represent perceptual imaginings as imaginings-that-p. A plain line to take is that the object of a perceptual imagining is something whose properties are selectively available to perception, not a proposition. And this, after all, is what we should expect if some imaginative exercises are perceptual; as a consequence, perceptual imaginings are much less readily represented by the imagine-that locution. If we can recover anything from the notion of the aesthetic attitude by connecting that notion to the imagination, it will be the perceptual imagination that supplies the connection. But skepticism about this overall design may be aimed at one or more of its parts. First of all, there may be a suspicion that in distinguishing the perceptual from the conceptual imagination I have done the job too well: if the difference between them is as great as it appears to be, it may seem odd that we would refer to both of them as the imagination. And then there may be another suspicion (partly in the light of the first) that if aesthetic experience is to be explained by appeal to the imagination, it is the conceptual kind rather than the perceptual that does the work. For although the aesthetic attitude, so called, might be affiliated with the perceptual activity, we still need a reason to think that engagements with art, in particular, are not also affiliated— and perhaps primarily so—with the conception of possibilities. As for the first topic, we should concede that there is no single mental operation that can be called the working of the imagination. It is nevertheless striking that when philosophers have thought about (the) imagination, it has been natural for them to emphasize, at different points in their treatment, mental operations that are very unlike each other. To return to Hume, who helped himself most liberally to (the) imagination in explaining the mind’s actions, we find this pattern. In his more official theoretical statements on imagination, Hume stresses the arbitrary freedom of a faculty that rearranges our ideas, a freedom that, in the limit, produces the fabulous creatures of poems and romances. At other moments, when he merely invokes imagination, a much more conservativelooking mental operation appears. We learn, for instance, that the imagination enables us to think about the world beyond our fragmentary experience of it; and

194 CHRISTOPHER WILLIAMS

the operation here seems to be an enhancement of perception that is constrained, as perception is, by the way the world presents itself to us. Both the freedom and the constraints are attributed to imagination, seemingly without much selfconsciousness on Hume’s part that the attributions appear to be made to different things. Some of Hume’s ingenuousness in this area is doubtless due to a disposition not to draw sharp boundaries between perception and conception. But if we do draw the boundaries on his behalf, the issue about the disposition to regard two different things as the imagination remains, and the need for an explanation is magnified. Now it seems to me that there is an explanation, and that the notion of an unseen object is the key. In the case of the (possibly) ingenious child, the goblin is nonexistent, hence unseen, and the child’s conception of the goblin requires that the child apprehend an object that is not present to him. In the case of the perceptive novelist, the world imagined is an extension of our world, and the extension is not literally seen by us, just as the objects that lie beyond our view are unseen (while still belonging to our visual field, which is limitless). The novelist likewise apprehends something that is not present to him. If imagination is fundamentally directed at the unseen, so to speak, then both conceptual and perceptual activities should have imaginative potential. But here the topic of the relationship of aesthetic experience to imagination becomes newly pressing. Why suppose that there is a special relation to the perceptual kind? A brief answer is that, as the etymology attests, the idea of the aesthetic is inseparable from the idea of the perceptual, and therefore to the extent that we imagine things in aesthetic contexts we must also perceive them, and vice versa. This is a promising if abstract answer, but more needs to be said if it is to be generally serviceable. Literary works, after all, are typically works of fiction, and this would not seem to be coincidental. If literary works are not coincidentally fictional, then it might be thought that the imaginative satisfactions we obtain from them result from the fictional element. Against this line of thought, however, it may be held that the fictional element is irrelevant, or nearly so, to an explanation of the satisfaction, even though it is no coincidence that literary works are typically fictions. I think that this counter-reply is correct, but to see why we may need to have a greater awareness of the ways in which we use the concept of “fiction”. In philosophical parlance, the fiction concept applies to nonexistent entities, as well as to our thoughts about them. On the basis of these applications, the concept can be appropriated for prose compositions that describe fictional entities and that are meant to be understood as describing fictional entities. And so, to illustrate the point crudely: Anna Karenina (the character) is a fictional entity, and as a result Anna Karenina (the novel) is a specimen of fictional discourse. These usages are apt to seem unremarkable if we are making ontological distinctions; but they are remarkable if we approach them from the standpoint of the arts. When we think about the arts, the concept of fiction does not apply, in the first instance, to nonexistent individuals and the thoughts that represent them. Instead, the concept

SEEING TWICE OVER 195

applies to genres within particular artistic media, which are invariably paired against nonfiction genres within the same medium. Some media are hospitable to the distinction between fiction and nonfiction, but not all of them are, and we should try to account for the difference. We do not divide paintings, for instance, into fiction and nonfiction kinds, even though there are paintings that depict the real George Washington as well as paintings that depict the nonexistent fall of Icarus. This practice is important, and not merely sociologically, because the reasons that we have for looking at paintings, and the knowledge that we have of the way that paintings are made, obviate the need to sort paintings into fiction and nonfiction kinds. Paintings are manifestly artifacts and they provide a certain type of contemplative pleasure, and these palpable facts about them fail to motivate the introduction of the fiction/ nonfiction classificatory scheme. Consequently, it would be strange to insist that we must deploy the concept of fiction when we wish to explain our aesthetic responses to paintings, despite the fact (and it is a fact) that fictional entities can enter into the content of a painting. Since it is otiose, on aesthetic grounds, to deploy the fiction concept in a given medium unless we also have a basis for deploying the nonfiction concept, we should not suppose that the concept becomes relevant to the explanation of individual works in a medium when the concept is irrelevant to the explanation of the medium itself. And we have no basis for drawing the fiction/nonfiction distinction for painting—as well as for sculpture, music, dance, and architecture. These reflections about some of the artistic media seem obvious, but their obviousness may be lost unless we maintain a viable contrast between perceptual and conceptual imaginings. Without the right contrast in place, there may be an inclination to locate fictional elements in every exercise of the imagination, even those that involve the media we just considered, and to claim, for example, that it is fictional that we see George Washington in a painting (Walton 1990 takes this approach). It is of course not true to claim that we see Washington when we look at a painting. But it is true—and more pertinent—to claim that we see a picture of Washington; and this claim gives us no reason to add that because a picture is not a person it is fictional that we perceive a person in the picture. In any event, isolating the aesthetic pleasure remains to be done, and it seems that anything we want to say about the aesthetic satisfactions of paintings, insofar as the imagination has a bearing on them, can be said without our having to broach the fiction concept. If we turn now to the artistic media in which we distinguish fiction from nonfiction, there are two important classes. One of these is verbal. We make the distinction for the literary arts (though, revealingly, our stance toward poetry is at times more closely akin to our stance toward painting). The other is photographic, whether still or moving. Although we do not use the words “fiction” and “non-fiction” to describe a photograph or a film, the concepts are in evidence. We do speak of “documentary” and “nondocumentary” films; and we mark a difference between photographs that result from processes that have

196 CHRISTOPHER WILLIAMS

been creatively manipulated so as to produce effects that are not amenable to being described as fact-reporting (or fact-recording) and the photographs that commonly appear, or once appeared, in the newspapers or the family album. The two kinds of photographs, or films, do not correspond exactly to the distinction between art and nonart photography, or between artistic films and nonartistic, but the artistic instances will at least to a great extent sort themselves together under the broadly nondocumentary rubric and the nonartistic under the documentary. The same is true for writing, or (if poetry is an exception) for discursive writing. As soon as we explain why this pattern holds, an explanation of the aesthetic irrelevance of the fiction concept even for those media in which we draw the fiction/nonfiction distinction will be within reach. A still photographic picture is commonly used for fact-reporting (or factrecording), and the mechanical processes involved in its production make a photograph well suited to report or record the facts. Yet because artists can exploit photographic technology for different ends, it becomes intelligible to make distinctions. Discursive writing is analogous, though its fact-representing format is not pictorial. Words are the medium in which we think, and much of our thinking is concerned with stating facts, concealing facts, debating facts, or reasoning about facts. But words of course also furnish a medium in which artistic accomplishment is possible, and accordingly it is important to demarcate fictional and nonfictional uses of language. (The duality in our use of language may also provide an explanation of the deep Platonic suspicion of the literary arts: confusions between reasoning and imagining are likely if the medium for both coincides.) Even though instances of artistically discursive writing may deal prominently or intentionally with real people and real events, and even though some writing that is not artistically motivated may yield imaginative satisfactions, fiction and non-fiction are categories that it is useful for us to respect in discursive writing. This point is the inverse of the one I made about these categories in connection with painting. To the extent that we can make sense of art works as aesthetically interesting without classifying the works as fictional or nonfictional, we have reason for thinking that our imaginative response to Anna Karenina is no different. When we do resort to the dual classification scheme, we are only registering the fact that the medium in question has some artistic employment in addition to a standing nonartistic métier. This is the general rule in the arts. An objector might say that we need to understand literature, film, and photography as genuine exceptions to this general rule, and that imaginative engagements with works in these media should be understood differently, in a way that is more congenial to the claims of conceptual imagining. But even this qualified attempt to defend the primacy of conceptual imagining does not seem well made. Given the reasons for drawing the fiction/nonfiction distinction in the first place, it is plausible to maintain that Anna’s status as a nonexistent is not coincidental to the narrative. Our knowing that the story of Anna Karenina did not occur, or at any rate did not occur as Tolstoy told it (if the story had been a

SEEING TWICE OVER 197

reworking of actual events), helps to free the imagination from irrelevant preoccupations that the knowledge that the story is conceived as a true history—a story responsive to canons of historical accuracy—would tend to pose. We would want to relate the history of Anna to the larger order of fact that contains the facts of her history as a particular part. We would have questions about Tolstoy’s evidence, and whether Tolstoy had dealt with it appropriately. A recounting of the Anna story, for some possible historical Anna, might not have been able to deliver the same satisfactions without the taking of poetic license. (This last remark is a variation on Aristotle’s dictum concerning the merits of poetry over history, without Aristotle’s value judgment.) Such preoccupations are distractions, and they occur because the story of Anna, like any story, is told in words. No corresponding distractions occur when we view paintings of Russian noblewomen. And distractions are all that they appear to be. A writer (or reader) who thinks that the story he is writing (or reading) is a true history may have difficulty imagining aspects of the narrative, owing to his beliefs about the facts. But the story need not traffic in fictional characters or events in order to supply imaginative satisfactions as such. To forge a more visible connection between the perceptual imagination and the idea of the aesthetic attitude, I want to return to Hume one more time. In a little-discussed passage from Book Two of the Treatise, Hume has this to say about the imagination: ’Tis obvious, that the imagination can never totally forget the points of space and time, in which we are existent; but receives such frequent advertisements of them from the passions and senses, that however it may turn its attention to foreign and remote objects, it is necessitated every moment to reflect on the present. …When we reflect, therefore, on any object distant from ourselves, we are oblig’d not only to reach it at first by passing thro’ all the intermediate space betwixt ourselves and the object, but also to renew our progress every moment; being every moment recall’d to the consideration of ourselves and our present situation. (Hume 1978:427–8) That which Hume initially seeks to explain by means of this imaginative activity does not really require the postulation of that activity. His basic goal is to explain why spatially removed objects are easier for us to imagine, and to have concerns about, than temporally removed objects. The answer that he really gives is that space comprises a set of coexistent parts that are “capable of being at once present to the sight or feeling” (Hume 1978:429), whereas the moments of time are not susceptible of simultaneous presentation. This difference between space and time in our potential experience of each would alone suffice to account for differences in imaginative facility.

198 CHRISTOPHER WILLIAMS

But even though the explanation can go through without the psychological mechanism Hume invokes, there does seem to be something to the mechanism. The project of establishing contact with remote objects involves an exercise of the perceptual imagination, and presumably we are to suppose that the imaginer occupies a succession of standpoints en route to a nearer perceptual apprehension of the object, while undergoing no bodily change of standpoint, which accounts for the resistance within the imagination. The succession of standpoints can be conceived as virtual movements on the imaginer’s part, and it is in fact natural to think of the imagination as racing from one locale to another, with the world as a whole—an indefinitely extensive perceptual field—for the backdrop. This remark about the whole of the world, which is present to a person’s imagination but absent to sense perception, should not be taken to mean that the imagination constructs a world that exists enduringly in our absence out of fragmentary sensory materials assembled in our presence. The world-constructing activity is a task that Hume elsewhere assigns to the imagination, in his epistemology, and P.F.Strawson has criticized Hume for thinking that the task was one for which the imagination was needed. Strawson argues, on Kantian grounds, that perception presupposes an enduring world, and so the problem to which the imagination’s activity is a proposed solution would be impossible to solve if the imagination had to solve it (Strawson 1974:49). We can grant the Kantian point as much as we please. But the activity Hume describes in Book Two is quite doable in the light of that point, since relying on the imagination to establish contact with a remote object differs from relying on it to establish, as it were, remoteness. Hume’s mechanism does give us a general way of describing differential imaginability. The less able we are to envisage the intervening space or time between us and some item, the more difficulty we shall have in sustaining our imaginative focus on the item. A poor grasp of the intervening points could be due to faulty or seriously incomplete information about the contents of the interval, but it could also be that the item that a person tries to imagine is so unlike the world with which the person is familiar that there is not enough space for a virtual movement of imagination. Bizarre possibilities are often difficult to imagine because we cannot convincingly locate them vis-à-vis our own situations (which are typically psychological, but bodily relationships provide obvious metaphors for psychological conditions). Subsequently, Hume exploits his mechanism to explain the admiration we can have for remote objects, and this move perhaps reveals something of greater importance. He says that, when they are present to us, a great traveler is admired because of his distant journeys, and a Greek medal because of its antiquity (Hume 1978:433). If we assume that Hume has got the facts about admiration right, we may notice that the admired objects themselves are not perceptually distant from us; and this reveals a difference in his two treatments of remote objects, the unadmired and the admired. The best way to understand the imagination-assisted admiration is to suppose that imagination is necessarily operative in the perception

SEEING TWICE OVER 199

of the admired object if it is operative at all. It might not be operative, since it is perfectly possible to behold the traveler or the medal without calling to mind their relationships to non-present realities. And so, not everyone’s sense perceptions need be informed by these further perceptions. How exactly imagination informs a sense perception is a very difficult matter to explain, and (as Strawson suggested) may be inexplicable. One thing that is clear, at any rate, is that we cannot avail ourselves of the most straightforward explanation in this area, which is that we see the relationships in the present object, in the sense that we may see faces in clouds. But there is enough in these materials to show how we need not picture imagination as operating merely next to sense perception, nor as simply contributing to sense perception a conceptual structure that it must have if the sensory inputs are to be other than merely blind. The second half of this last claim marks out an option that we might take if we find just the option marked out in the first half unacceptable. If we do not acknowledge a distinctively perceptual imagination, while rejecting the idea that imagination and perception are serially exercised, we will suppose that the imagination is present in sense perceptions merely if they are conceptualized (see Walton 1990:295). To make this true point is to fall short of saying that certain exercises of imagination cannot occur without a sense perception, and making it would not help us to explain why some people, possessing imaginativeness to a greater degree, can perceive more than other perceivers do. Imaginativeness, where perceptual imagination is concerned, is reflected in how people look at things. The way in which an imaginative person looks at things was once expressed by the idea of an “aesthetic attitude” that we have when we appreciate works of art or natural objects that we are viewing as if they were works of art. Specifically, I want to consider the classical version of the idea that Edward Bullough presented in his account of psychical distance. The title of Bullough’s paper indicated a twofold role for psychical distance—as a psychological “factor in art” and as an “aesthetic principle” that separates the beautiful from the merely agreeable—but my remarks will be restricted to the first of these two roles. Bullough is worth a reexamination because the doctrine of psychical distance has been unjustly criticized. Psychical distance has sometimes been treated as if it were an illusory phenomenon—a psychological phantom—or, if not illusory, it has been treated as an idiosyncratic disposition that perhaps acquires a false glimmer of importance in virtue of being tailor-made for another doctrine, aesthetic formalism, which is widely assumed to be disreputable. But general skepticism about the perceptual imagination as such is the main motivator of skepticism about psychical distance. And the discontents with formalism, too, are at least partly traceable to an anti-perceptual proclivity in more recent theorizing. A better understanding of both perception and imagination will enable us to do justice to what Bullough saw. The “aesthetic attitude” is not wholly unproblematic, however, and one obstacle to a just appreciation of Bullough’s account concerns the very idea of the

200 CHRISTOPHER WILLIAMS

aesthetic attitude. Aestheticians have voiced dissatisfaction with the idea, but the real difficulty with it has not been satisfactorily identified. Critics of the aesthetic attitude have targeted the allegedly distinctive aesthetic ingredient in the proposed attitude. Yet the underlying difficulty is not with that ingredient, but with the more general notion of an attitude, which seldom receives scrutiny in its own right. An attitude is a stance toward things that a person can adopt, or not adopt, in at least a semi-voluntary fashion. The item known as the pro-attitude, familiar to us from moral philosophy, is an especially transparent example of a freely adoptable stance; but attitudes, as ordinarily understood, trail similar baggage. It makes sense, after all, to encourage a person to change her attitude if her attitude is unhelpful, and we do not offer comparable encouragements for changes of belief. An attitude is also a stance whose adoption cannot be explicated by reasons, except in a pragmatic, strategic sense: for instance, the job of living will go better if a person changes her unhelpful attitude, and that can give the person a reason to make the change (and a reason for others to exhort her to change). If we then bring an attitude to bear on artworks or natural objects, it will seem— correctly seem—that we are changing a merely subjective orientation. And it will further seem that a peculiarly aesthetic value in what we perceive might be willed into being. As an explanation of art or the aesthetic, any approach that posits a mere attitude is highly unappealing. Bullough occasionally uses the word “attitude”. But to assume that he is an attitude-theorist, with the unappealing trimmings, would be incorrect because he does not consistently describe his chosen phenomenon as an attitude. When in his central expository example Bullough speaks of the difference between the person who is terrified by the sea fog and the person who appreciates the weird grotesquerie of the sights, he says that it is a difference in outlook, not attitude; and I do not think that we should just take it for granted that an attitude is meant, or must be meant. The language Bullough uses when he talks about psychical distance in general psychological terms is somewhat casual, even tentative, but he explicitly keeps the sense-perceptual metaphor in focus. If we cash the metaphor, an outlook is a perceptual standpoint; and attitudes, as I have just characterized them, are not perceptual in a meaningful respect. We could insist, perhaps, that any perceptual element in Bullough’s exposition is an unfortunate residue of bad ways of thinking, and that he should properly, and charitably, be regarded as describing a mere attitude. But if my argument has been on the right track, the attitude construal is precisely the wrong one to put on Bullough’s words, and such a construal will in fact yield the uncharitable interpretation. It would be best, on Bullough’s behalf, to eschew the attitude idiom altogether in order to explain aesthetic response. Taking our cue from Bullough’s injunction to “direct our attention to the features ‘objectively’ constituting the phenomenon” (Bullough 1957:94), it would be a happier idiom to speak instead of a certain kind of attention that more imaginative people bring to bear on their sense perceptions and that less imaginative people do not. Yet lest it be thought that I am tendentiously caviling about a word, we should notice that Bullough’s

SEEING TWICE OVER 201

fog-at-sea example does not have to be construed as involving attitude-shifts at all. There is nothing in the example itself that promotes the suggestion that it is possible for a person to oscillate willy-nilly between surveying the sea fearfully and surveying the sea rapturously. Indeed, the example suggests a different interpretation. The fearful perception informs the rapturous, and any appearance of oscillation can be explained by the fitfulness with which we occupy the standpoint of the more imaginative and less accessible second perception. With the attitude detour set to rights, we should reconsider the psychological phenomenon at work in the aesthetically satisfying—or, as we might say, aestheticized—response to the sea fog. And here we encounter another obstacle to an appreciation of Bullough. For the satisfying response contains two elements, and not one, as Bullough’s exposition would lead us to suppose. We have, first of all, a capacity for detachment: an ability that a person may have to overlook the practical implications of a given situation. The capacity for detachment is what Bullough has in mind when he commands the reader to abstract the danger and unpleasantness from the experience of the sea fog. Yet standing back, as it were, from the physical threat that the fog poses is not sufficient to explain the “intense relish and enjoyment” of the aestheticized response, since practical detachment may merely enable a person to be a more discriminating perceiver of the various ordinary sensory features that are detectible in a foggy scene. (If I am suitably detached, I will be better able to take in the more minute variations of density in the fog over time.) Although it may be satisfying for me to make finer sense-perceptual discriminations, the satisfaction that Bullough describes certainly goes well beyond this. And so we come to the second element in the aestheticized response. Bullough says that if we respond to the sea fog in this way, we see the sea fog as if the fog were a shape-distorting and shape-concealing veil of milk, and as if the objects that the veil occludes could be touched by merely putting out our hands, and as if the sea were hypocritically denying any suggestion of danger. This description of the sea could virtually succeed as a description of a painting that depicts a seascene, and it seems hardly accidental that it could so succeed. If a person responds to a night on the sea or to a depicted nighttime sea-scene in such a way that Bullough’s extended description of the response is revealingly apposite, the same sort of capacity is being exercised in either case. This capacity is the second element, which may be called aesthetic sensitivity, because it involves receptiveness to an imaginative perception. That Bullough’s account of psychical distance includes this new element of aesthetic sensitivity becomes more evident in Bullough’s later remarks on colorappreciation. According to Bullough, most people will explain their likings for certain colors in terms of simple opposing qualities such as warmth or coldness, while a more discerning minority resorts to a vocabulary that is more readily appropriate to persons, and will explain their likings, in turn, by saying that the colors are energetic, affectionate, reserved, treacherous, and the like. The most intuitive way to characterize the perceptual responses of the minority is to

202 CHRISTOPHER WILLIAMS

say that they exhibit an aesthetic sensitivity that the majority lacks. The two types of response to the sea fog that Bullough considers parallel these two types of response to colors. Psychical distance is properly analyzed, then, as a combination of detachment from a practical involvement and aesthetic sensitivity to an object of occurrent perception. What may have led to Bullough’s failure to make the combination explicit is the role that the imagination plays in both detachment and aesthetic sensitivity. To stand back from a physical threat without literally escaping from the threat must of course be an imaginative activity, just as the detection of perceptual qualities that invite metaphorical description involves such an activity. A perceiver with little imaginative ability would be neither detached nor sensitive, and because imagination is necessary for attributions of both detachment and sensitivity it could be possible to think, mistakenly, that a simple item is being attributed. Detachment is properly regarded as an enabling condition for aesthetic sensitivity. A person who is terrified does not have the presence of mind to observe the weird grotesquerie surrounding him, does not have the opportunity to exercise his imagination in the course of perceiving the sea fog more concretely —an exercise that requires a degree of attention that is easily eroded if there are distractions. By not being able to sustain the right degree of attention, the terrified person merely perceives the sea fog and its accompanying dangers. Thus far in this reconstruction of Bullough’s account there has been no unavoidable occasion to use the crucial term “distance”. But at this point it is not difficult to see why the term might prove congenial to a theorist who is interested in the psychology of aesthetic response. Because appreciation demands a measure of imagination that not everyone, or not anyone at every time, can bring to bear on the object of attention, it is natural to regard the imaginatively perceived object as being less accessible to perceptual interaction than the object that is merely perceived in a fearful manner, and hence natural to picture the imaginatively perceived object as more distant. Bullough does not always put the best face on the consequences that he draws from his account of distance, and he accordingly exposes it to irrelevant criticism. This is yet another obstacle for us, and there are two improvements to make regarding Bullough’s exposition of the consequences. The first concerns his treatment of the further topics of under- and over-distance. The problem is that under- and over-distance do not differ by the absence or presence of a single type of property. When we appreciate the foggy sea aesthetically, we put the fog “out of gear” (in Bullough’s notorious phrase) with our practical interests, while the jealous husband who watches Othello fails to put Shakespeare’s play out of gear, and his failure reveals a lack of distance on his part. Now the out-of-gear talk is not very illuminating on its own, to be sure, and we would need to hear more about it than Bullough actually tells us. That difficulty is superficial at any rate, but the problem I have in mind does not depend on our finding this talk unacceptable as such. Rather, the problem is that if a lack of distance on X

SEEING TWICE OVER 203

involves not putting X out of gear, then it is quite unclear what a surfeit of distance, so to speak, could possibly amount to. Once we are out of gear with the fog or the play for aesthetic purposes, it is not readily apparent how we could become more out of gear. Yet that dubiously intelligible result is what the notion of over-distance tends to suggest. If we examine Bullough’s examples of over-distance, we shall see how little help the initially drawn contrast between the two fog-surveyors gives us. He turns away from differences between spectators and turns toward differences among objects, and one example of over-distance is provided by “idealistic Art,” which is art “springing from abstract conceptions, expressing allegorical meanings, or illustrating general truths.” Bullough’s explanation of over-distance in the context of idealistic Art is that its abstractions make it impossible for the appreciator to stand in a suitably personal relationship to them: They appeal to everybody and therefore to none. An axiom of Euclid belongs to nobody, just because it compels everyone’s assent; general conceptions like Patriotism, Friendship, Love, Hope, Life, Death, concern as much Dick, Tom and Harry as myself. (Bullough 1957:103) Owing to this indiscriminate appeal, an abstraction either leaves us cold or invites us to pay no attention to the work that suffers from it. The phenomenon that Bullough identifies by means of the term “overdistance” is real enough, but it would be better served if he had more forthrightly come to grips with the role of the perceptual imagination. The real explanation for our being unmoved by abstractions and generalities is that they offer us no quasi-perceptual satisfactions. The example of Euclid’s axioms is rather ill-chosen, for some have found it unproblematic indeed to suppose that Euclid “looked on beauty bare,” that geometrical demonstrations meet the perceptual quasiperceptual conditions required for aesthetic satisfaction. Nevertheless, the idea behind the example, that an abstraction “belongs to nobody,” is true in the sense that there is no perceptual standpoint from which it is possible to appreciate something that would not lend itself even to geometrical mode of perceptual presentation. A complete abstraction is an object of thought, and not—or not enough—an object of appreciation. We can thus equate under-distance with fairly unimaginative perception and over-distance with not having a suitable object for an imaginative perception to take in. These identifications would thereby remove any urge to think that there is a single continuum of response that involves under- and over-distance at the termini of the continuum. The other improvement of Bullough’s exposition to add concerns the term “distanced” (and in turn “under-distanced” and “over-distanced”), which Bullough tends to use offhandedly and uncritically, as do his readers. Certainly, it is a linguistic convenience to say that a person has a distanced appreciation; and

204 CHRISTOPHER WILLIAMS

the language need not involve us in mistakes. But if we indulge the habit of using this shorthand, it may become easy to assume that we are distanced as a result of some distancing activity of ours, and we may then arrive, irresistibly, at the verb “to distance” as a back-formation. If we think that this verb names an action that we can perform, we will certainly give aid and comfort to the construal of psychical distance as an attitude. Psychical distance will seem to be adjectivally related to the distancing action that precipitates it, and because we may perform the action merely at will, psychical distance will again seem to return to the territory of attitudes, and consequently be of marginal interest to aesthetic theory. But the perceptual imagination gives us the wherewithal to avoid this dead end. The difference between the two viewings of the foggy sea is the difference between an unimaginative sense perception and a sense perception that is imaginatively assisted. So put, the contrast is much the same as that which obtains between a viewing of a Greek medal that fails to appreciate the resonances of its antiquity and a viewing that is alive to those resonances. There is nothing that we do, beyond relying or not relying on our imagination; and so no special mental activity, one that we might perform within our imagination, needs to be postulated. If we invoke a more sophisticated descendant of Hume’s notional spacetraversing mechanism in order to explain the interest that the ancient medal has for us, we shall be able to show why the metaphor of distance so aptly captures the variable degree of perceptual accessibility in the first place. An imaginatively perceived medal, or foggy sea, is an object that is present to the senses, and such an object is of course not distant from a viewer in a literally spatial sense. But as a visible fragment of something that is not present to the senses—namely the circumstances of its origin (in the case of the medal), or the circumstances of a non-endangering situation (in the case of the foggy sea)—an object can seem to exist at a remove from our unassisted perceptual standpoint. It makes sense to regard such items as “distant” from us, and because of the psychological resistance involved in forcefully imagining them, it also makes sense to suppose that a traversal of notional distance occurs when we do succeed in imagining them. If the preceding gloss on Bullough’s thought is acceptable, it would be more strictly accurate to speak of a “distant” object of appreciation (i.e., an appreciated object that is separated from us by distance) than to speak of a “distanced” appreciation. This terminological substitution still leaves us with a metaphor— metaphors are probably unavoidable here—but at least it is less misleading to assign the image of distance to a space that the mind can notionally navigate in the course of performing other actions than to assign it to the actions themselves. The neo-Humean interpretation of psychical distance I have given outlines an account of the imagination in aesthetic contexts. More needs to be said about art specifically, and although I cannot explore the details here, I shall conclude with a few observations on that topic. At the beginning of this essay I touched on the thought that the imagination could not be perceptual because if it were, the phenomenology of imaginative

SEEING TWICE OVER 205

experience must be intolerably different from what we take that experience to be. The type of accusation that this view makes against a perceptual view of the imagination also appears in a criticism of psychical distance that we have not so far considered—namely the claim that Bullough’s doctrine must distort our understanding of literary works in which the author, or a protagonist, directly addresses the audience and by so doing violates the distance between the audience and the world of the work (see Dickie 1964). This is not a worrisome criticism, since we can plausibly regard a direct address to the audience as a moment in which the audience members have an opportunity to play themselves (in much the same way that a celebrity can make an appearance in a film in propria persona). But the larger observation to make on this occasion is that it is a misunderstanding to suppose that the perceptual view of imagination can distort the phenomena in some such fashion as this. Properly speaking, the account I have sketched should have no implications regarding the phenomenological content of imaginative experience: the account merely seeks to explain how that content is possible, whatever the content is. I also noted that psychical distance seems to be made tailor-made for aesthetic formalism. There is no elegant comment to make on that aspersion, because “formalism” is not a simple, easily identifiable thesis. But to the extent that psychical distance does play an essential role in our understanding of art, my account of it does nothing to reinforce formalism in its most extreme formulation, which denies that the content of a representational work enters into a genuinely aesthetic appreciation of a work. But the account does reinforce at least one idea that is often associated with formalism, or more broadly, with the position that there is a distinctively aesthetic category of appreciation: the idea that direct perceptual acquaintance with an aesthetic object is crucial to appreciation, and that no substitute for such acquaintance is allowable. The account reinforces at least this idea because the imaginative exercise, of the perceptual variety, cannot occur apart from a distinguishable perceptual exercise, of the sensory variety. To grasp the aesthetic qualities of the sea fog, which call for imaginativeness on our part, we must be attending to the sea fog, which is also an occurrent object of sense perception. And with this reflection we return to the two types of imagination. We are now in a position to see a deeper reason why the perceptual imagination is more intimately related to our experience of art than the conceptual. Conceptual imagining does not require that a person enjoy any perceptual episode in the course of the imagining, and this fact about it can either make imagination appear to be an activity that is merely causally related to perception or else reduce the imaginative element in perception to Kantian concept-ladenness. On this picture, according to which we see but once if we see at all, when we appreciate a painting we engage in an imaginative project that could in principle occur without the act of looking. That perception is fundamental to aesthetic experience is as obvious a truth as Hume’s proof that pity depends on imagination,

206 CHRISTOPHER WILLIAMS

and the alternative picture I have been exploring here reflects an effort to respect that truth. Acknowledgments I wish to thank Berys Gaut and Dominic McIver Lopes for their comments on an earlier version of this chapter. A remark by James Conant inspired some of my remarks on fiction.

12 OUT OF SIGHT, OUT OF MIND Dominic McIver Lopes

Conceptions of imagination are typically the result of a triangulation that places it in relation to conceptions of perception, art, and fiction, and much depends on the weight accorded to each of these. For centuries, imagination was taken to be sensory imagining. Many philosophers held perception always to involve the operation of imagination in this sense. Some held all thought to be imagistic. Happily, such views have long been laid to rest. In recent philosophy, imagination is taken to be a prepositional attitude, specifically the attitude properly taken towards fictional propositions. Imagining is imagining that p. Some have argued that even our understanding of what pictures represent lies in our imagining fictions. But this view of imagination goes too far in deprecating the sensory imagination: such phenomena as visualizing a familiar face and “hearing” a piece of music in one’s head cannot be adequately understood if we take imagination to be a prepositional attitude. We therefore need an account of sensory imagining alongside accounts of purely propositional imagining. I argue that sensory imagining involves experiential states of the kind involved in perceiving. These states contribute to the proper appreciation of works of art. Vision and visualization To visually imagine (or “visualize”) a scene is in part to have an experience as of the scene. It is this that it shares with seeing the scene. But visualizing a scene is not merely having a visual experience as of the scene in the absence of seeing the scene. One question to ask is how visualizing should be distinguished from seeing. Another requests some account of the nature of visual experience such that it may figure both in visualizing and in straightforward seeing. A third question asks what makes any experience a visual experience and hence what makes any sensory imagining an instance of visualizing. Before proceeding, it will be helpful to develop a rough conception of the distinguishing characteristics of experiences, as against other kinds of representational mental states. First, experiences are necessarily occurrent: they always have a precise duration and occupy a precise location in the stream of conscious mental events. This is not the case with beliefs or desires, for instance, which are quite often dispositional. (Experiences but not beliefs are the stuff of

208 OUT OF SIGHT, OUT OF MIND

episodic memory: I can remember that I experienced something without recalling the experience itself but I cannot remember having a belief without recalling the belief itself.) Second, experiences necessarily have a phenomenal character (beliefs need not). Finally, the content of experience is nonconceptual: the properties it represents need not be ones of which the perceiver possesses concepts. A perceiver may have experiences as of shapes, colors, chords, and timbres for which she possesses no concepts. The contents of belief, by contrast, are conceptual since a property represented in belief is necessarily one of which the believer possesses a concept. Immanence and transparency The hypothesis that visualizing and seeing both involve visual experiences should not be read as conflating visualizing with seeing. Indeed, it is consistent with the intuition that visualizing and seeing differ phenomenologically. To see this, consider a well-known error of some early epistemology. A visual experience is a mental representation with visual content and phenomenology (however these are to be characterized). Some philosophers added that the vehicles of visual experiences possess, in much the way that pictures do, the properties that they represent. Experiences were thought, as the point is usually put, to be pictures in the head. This additional claim performed two tasks. It was intended to explain how some mental representations could have imagistic content: a visual experience represents a blue triangle because the vehicle of the experience has the properties of being blue and triangular. But it also figured as part of an account of the justification of perceptual belief: I am justified in believing that I see a blue triangle when I can introspect a mental object that possesses the properties of being blue and triangular. Hence the views that we introspect the properties of the mental vehicles of experiences and that experiences represent the properties they do because their vehicles literally have those properties. Both views are false, of course, and there is hardly need to rehearse the reasons why. However, there is a fact about visualizing that explains part of the intuitive appeal of the early conception of seeing and that is useful in distinguishing visualizing from seeing. When it comes to visualizing, we are subject to what Sartre calls the illusion of immanence (Sartre 1948). It appears, when we visualize a scene, that we introspect the properties of some mental object, the vehicle of the visualizing, where those properties match those of the imagined scene. In visualizing it is as if we introspect mental objects that have the visual properties of objects in the nonmental world. Thus when I visualize a chartreuse gorilla it seems that I introspect a mental object that is the vehicle of my visualizing and that is chartreuse and gorilla-like in appearance. Properties visualized seem to be “stuck on” the vehicle of experience. Of course, properties visualized are also experienced as belonging to the object imagined. The interesting fact about visualizing is that properties visualized are experienced both as properties of the object imagined and as

DOMINIC MCIVER LOPES 209

properties of a mental object. This is a fact about the phenomenology of visualizing. In some early philosophy the susceptibility of visualizing to the illusion of immanence is erroneously extended to perception. In so far as ordinary seeing is not subject to this illusion, the fact that visualizing is subject to the illusion of immanence distinguishes it from seeing. An oftremarked feature of seeing, as of perceptual experiences generally, is its transparency. An experience is transparent provided that the properties comprising its content are experienced only as properties of the scene represented, not as properties of the vehicle of the experience. Transparency, so defined, is incompatible with the illusion of immanence, for that illusion consists in experiencing represented properties as properties of the vehicle of the experience. In seeing, represented properties seem to be “stuck on” the objects represented, not on the vehicle of the experience itself. Of course, we are owed an explanation of why seeing is transparent while visualizing is subject to the illusion of immanence. It is doubtful that we know enough about vision, visualization, and their respective roles in the cognitive economy to be able to produce one at the moment. One possible starting point is provided by another, somewhat surprising, difference between vision and visualization. Seeing sometimes allows for aspect switches—one may see the duck-rabbit figure as a duck or as a rabbit. The switch from one aspect to the other is often spontaneous. In addition, what aspect one sees is not due to what one believes: one may believe the figure represents a duck and yet only see the rabbit aspect. Instead, the aspect one sees appears to be a matter of attention, a mechanism that organizes salient experienced properties into a unified whole. To see the parts of the duck-rabbit figure as a rabbit instead of a duck, one must see them as put together in a new way. But whereas seeing allows for aspect switches, the psychologists Deborah Chambers and Daniel Reisberg demonstrated that visualizing does not (Chambers and Reisberg 1985). Chambers and Reisberg briefly showed their subjects an ambiguous picture, such as the duck-rabbit. When later asked to visualize the figure, all reported visualizing either a duck or a rabbit. Yet none was able to discern the other aspect in imagination, even when told what animal to look for and even though, when asked to draw their image, they were easily able to discern the missing aspect in their drawings. Chambers and Reisberg claimed to have shown that it is not possible to make discoveries through mental images. However, there is ample evidence that, while they cannot discern alternative aspects in images, subjects can make many other sorts of discoveries with images (see Finke 1990; Peterson et al. 1992). A better explanation of Chamber’s and Reisberg’s findings is that mental images come with attention “built in”. Images of the duck-rabbit as a duck and as a rabbit represent the same outline shape and textures, but attention foregrounds and organizes these properties differently in each case. Thus if one visualizes an object as F and wishes to visualize it as G, it is necessary to form a new image. One may

210 OUT OF SIGHT, OUT OF MIND

make discoveries in or reinterpretations of an image only if no shift of attention in this sense is required. That seeing is both transparent and also allows for aspect switches is likely to be no accident. One hypothesis is that since attention is “built in” to visualizing, the content of visualizing depends on the will in a way that the content of seeing does not—nothing is visualized without the direction of attention. Thus experience is not transparent unless its content is independent of willful attention. Seeing is transparent because there is a sense in which its content is independent of the perceiver. Sometimes two mysteries can be solved once put into proper relation to one another. Berkeley’s Puzzle It is one thing to say that visualizing requires having an experience; it is another to say that the experience is one of the same kind we enjoy when seeing. An account of visualizing, if my hypothesis is correct, depends on our conception of visual experience as it occurs in ordinary seeing. But Bernard Williams has warned that we are in danger of being driven headlong under the power of this hypothesis into a view of Berkeley’s (Williams 1973b). Berkeley’s Puzzle issues from the following argument. (1) Visual experience necessarily represents scenes as from a point of view. (2) Any experience that is viewpointed in this way is personal—it represents the experiencer as observing the scene from the viewpoint so as to endow her with a first-person way of thinking of the scene. It seems to follow, on the hypothesis that visualizing involves having a visual experience, that visualizing is also personal, necessarily representing what is visualized as observed by the visualizer so as to endow her with a first-person way of thinking of the scene. The puzzle is then that a visual imagining of a scene (e.g. the big bang) as one not seen by anyone is incoherent, though we seem to perform this sort of task not infrequently and indeed with little if any sense of dissonance. This provides leverage for an objection to my hypothesis. If it is possible to visualize a scene as unseen without incoherence, and if (1) and (2) are true, as they appear at first glance to be, then the hypothesis that visualizing requires visual experience as it also occurs in seeing must be false. Williams gives two reasons to reject the Berkeleian argument. The first depends upon and introduces an important distinction. Visualizing requires a visual experience. However, the content of visual imagination need not be determined in whole by the visual experience, which constitutes only part of the imaginative project that counts as an act of visualizing. One might visualize a tiger with forty-five stripes by means of an experience of a tiger with an indeterminate number of stripes. Alternatively, one may visualize a scene by means of an experience whose content is determinate in respects of which the project of visualizing is indeterminate. In Williams’s example, a man visualizing a bath and having recently visited the Bonnard show, may find himself unable to visualize a bath without a woman in it (Williams 1973b:33). The woman figures in a

DOMINIC MCIVER LOPES 211

description of what he visually experiences but not in a description of what he, in a larger sense, visualizes. The content of the experience may, as it were, outstrip that of the imaginative project, so that a single experience may serve several imaginative projects. In such cases what is imagined derives partly from prepositional imagining. Thus a person may visualize a tree as not seen by anyone without having a visual experience representing it as unseen. That the tree is seen is an element of the content of the experience visualizing requires, but not of the project of visualizing something unseen. In defense of Berkeley’s argument, Christopher Peacocke maintains that its conclusion still follows provided it is interpreted as requiring that the unseen tree’s being seen is part of the content of the visualizing that derives from propositional imaginings (Peacocke 1985:29). Prepositional imaginings are needed to distinguish imaginings that differ despite involving type-identical experiences. Imagining being at the helm of a yacht, imagining the experience of being at the helm of a yacht, and imagining what it would be like were a brain surgeon causing you to have an experience as of being at the helm of a yacht may all invoke the same visual experience. What distinguish them are the conditions one propositionally imagines to hold—that the experience is perceptual, that it is an illusion, or that it is left open. The Berkeleian argument requires that visualizing an unseen tree be construed as (1) having an experience as of a tree, (2) propositionally imagining that one perceives the tree, and (3) propositionally imagining that nobody perceives the tree. The incoherence of the conjunction of (2) and (3) cannot be dissolved by an appeal to the distinction between (1) and (3). There are cases, of course, where the visualizing should not be construed in this way. It may be indeterminate in the context of some visual imagining that one is perceiving the visualized object; it may be determinate that one is not. But when one propositionally imagines that one is perceiving and that the object of perception is unperceived, the imagining is incoherent. While Williams’s first objection grants the premises of the Berkeleian argument, Williams also disputes the second premise, namely that experience which represents things from a perspective must represent that perspective as occupied. Williams draws an analogy to seeing in theater. Spectators see Othello in front of a palace from a certain point of view—not from a perspective in the stalls but from, say, the front of the palace—yet they are not at any specifiable distance from the palace. This is because they are not in the scene represented. Although they see the scene, they do not do so in the way they see the props and actors making up the scene nor in the way dramatic characters (fictionally) see the scene and other characters. Visualizing is from a point of view, but Williams concludes “there can be no reason at all for insisting that the point of view is one within the world of what is visualized. We can, then, even visualize the unseen” (Williams 1973b:37). Arguments from analogy get nowhere when the basis of the analogy is in dispute. Rejecting Williams’s analogy, Peacocke simply insists that it is in the nature of visualizing that it present a scene as experienced “from the inside”

212 OUT OF SIGHT, OUT OF MIND

(Peacocke 1985:29). That this is not a condition upon theatrical representation is, Peacocke urges, irrelevant. The lesson is that Williams should have introduced his discussion of theater not by way of analogy but by way of counterexample. His reason for treating theater as nothing more than the basis of an analogy is that seeing in a play is not a case of visualizing—it is a case of seeing some things (props and actors) that represent other things (scenery and characters). However, what is at issue is the nature of visual experience, and not just as it occurs in ordinary seeing. The question we must ask is whether visual experience, being viewpointed, necessarily represents that viewpoint as occupied (so as to endow the occupant with a first-person way of thinking of the scene). We have the answer if what is seen in a play is viewpointed yet does not represent the viewpoint as occupied by the spectator. What reason can there be to deny that my experience of Othello or Kane is not a visual experience? Representational seeing is an apparent counterexample to the claim that viewpointed experience necessarily represents the viewpoint as occupied. An argument is required to justify setting the counterexample aside and to show that the claim is true. A corollary of such an argument would be that representational seeing involves a kind of visual experience that does not figure in ordinary seeing. Representational seeing and visualizing reveal something about the nature of visual experience generally: they reveal what is overlooked by an exclusive regard for ordinary seeing. Taking this point seriously, we may also cast doubt on the first premise of the Berkeleian argument—the claim that visual experience is necessarily viewpointed, representing a scene as viewed from a single, relatively determinate viewpoint. The usual argument for this claim is that a crucial function of vision is to provide egocentric information about the perceiver’s location in space. Vision evolved because this information is essential to survival. But experience that does not represent an environment from a single, relatively determinate viewpoint cannot provide information about the perceiver’s location in that environment. The problem with this argument is that there is no reason to think that visual experience does not occur when one of its functions is not served, even one for which it was designed. Indeed, there are cases in which we normally say we see objects though we are unable to locate them in relation to us—for instance, in trick mirrors or closed-circuit television systems. There are also cases of persons who suffer from optic ataxia (or Balint’s syndrome) as a result of damage to the occipito-parietal region of the cerebral cortex. Since the occipito-parietal region processes the “where” stream of vision and the occipito-temporal region the “what” stream, these individuals are not able to locate the objects they see in space, although they are able to recognize, name, and describe them (De Renzi 1982). What reason is there to deny these persons can see? That they cannot see well, or fully, does not show that they cannot see at all. If seeing need not represent the person seeing as occupying a viewpoint within the scene that she sees, neither need the visual experiences involved in seeing.

DOMINIC MCIVER LOPES 213

The same may be said of visual experience in general. Remembering that we should presume that there is something to be learned about visual experience generally from representational seeing, it is notable that pictures which have multiple or indeterminate viewpoints are not uncommon. Cubist paintings are often thought to represent a set of objects as if seen simultaneously from multiple viewpoints and pictures in the axonometric projection system characteristic of Chinese scroll painting do not have a determinate viewpoint; nevertheless, we see in these pictures the scenes that they represent. Since seeing in pictures need not replicate every feature of seeing, an independent argument is needed to show that the visual experience in virtue of which we see things must possess a single, relatively determinate viewpoint and hence that the experience in virtue of which we see things in pictures is not the kind of visual experience in virtue of which we see things ordinarily. What goes for seeing-in also goes for visualizing. The fragmented and multiple viewpoints of cubist images and the indeterminate viewpoints of Chinese scroll painting may surely have taken their first form in the visual imaginations of their makers and may surely be visualized by their fans. Why should the contours of experience in seeing-in and visualizing not tell us about contours of visual experience more generally? Since the intuitive connections between perspective and vision are extremely strong, it may help to consider a nonvisual case. Ordinary hearing always represents how things sound from a single, relatively determinate location in space. Hearing is directionally sensitive and sounds are always heard as at determinate distances (although a loud, distant sound cannot be distinguished from a closer but fainter one). These facts about hearing no doubt reflect the importance to an organism’s survival of its being able to orient itself in relation to sound-making objects— sound is an indication of movement, friendly or hostile, sometimes edible. Should we conclude that auditory experience necessarily represents the hearer as occupying a single, relatively determinate location with respect to the auditory scene? The perspectival content of sound experiences is often lost when sound is represented, as in mono recordings, yet many have surely heard Martin Luther King’s voice via his recorded speeches. (Has nobody I have spoken to on the telephone heard me? If they have not, then in what sense have I spoken to them?) Likewise, although imagined sounds are only rarely if ever perspectival, it is natural to suppose that they involve having auditory experiences. I can imagine the sound of the kettle drum burst that opens Beethoven’s Ninth Symphony without imagining it as coming from the back of the orchestra. The hypothesis that visualizing requires having visual experiences does not entangle us in Berkeley’s puzzle. Visualizing is not essentially perspectival and personal and neither is visual experience in general. Having reached this conclusion, we must still say something about what makes visual imagination specifically visual.

214 OUT OF SIGHT, OUT OF MIND

Modes of sensory imagination Just as discussions of perception typically default to discussions of visual perception, discussions of sensory imagination typically default to discussions of visualization. But we must not forget that the sensory imagination operates in the same variety of modes as does perception. We are obliged to provide an account of what, if anything, distinguishes the different modes of sensory imagination. This obligation does not rest on the assumption that sensory imagination can take place in only one modality at a time. A great deal of sensory imagining is multimodal, the imaginative equivalent of cinema in its combination of sounds with visual images of movement. Tactile properties, odors, and tastes can also be added to the imaginative mix. It would be odd to claim that multisensory imagining involves synchronizing separate imaginings in different modalities. In imagining a lecture of Michael Dummett, I do not on the one hand auditorally imagine his speech and on the other hand visualize his lips moving, and then bring the one in step with the other. I imagine the visual and auditory properties of the lecture as a unified whole. One might think nevertheless that there is reason to distinguish various modes of sensory imagining. Some properties (e.g. colors and chords) can only be imagined in one modality, but other properties (e.g. shapes, directions, and textures) can be imagined in more than one modality. What is remarkable is that imagining such a property in one modality differs phenomenologically from imagining that property in another modality. Just as what it is like to see a cube differs from what it is like to identify a cube by touch, so it is one thing to visualize a cube and another to imagine it haptically. This can be explained if sensory imaginings come by way of distinct channels, albeit ones that can feed into a single multimodal experience. How, then, are we to understand the fact that sensory imagining is channeled by mode? One answer is that visual, auditory, and tactile imaginings involve modality-specific experiences. Call this the “specificity thesis”. It seems hard to deny that the thesis is true. Surely what makes visualizing different from auditory imagining is that the former depends upon a specifically visual experience and the latter on a specifically auditory one. Of course, what plausibility the thesis has it borrows from having substituted a new question for an old one. Instead of asking how to classify imaginings by modality, we are now asking how to classify experiences into modalities. This question has traditionally received two answers. One is that experiences in the different sense modalities differ in respect of their intrinsic, nonrepresentational, consciously accessed properties (qualia). If this answer is correct, we have come to the end of our inquiry: visual imagining is distinctively visual because it involves experiences having visual qualia. Moreover, multimodal imagining involves an experience with a mix of qualia characteristic of different senses. But let us appeal to qualia only in the last resort.

DOMINIC MCIVER LOPES 215

An alternative is to claim that sense experiences belong to modalities in respect of the kinds of properties they represent. If these kinds of properties are to individuate the modalities, then it must be the case that some kinds of properties represented by experiences in any given modality are not represented by experiences in any other modality. We are committed to the “content specificity thesis”: each mode of sensory imagining involves a type of sensory experience that necessarily represents some property type that is not represented by experiences involved in any other mode of imagining. Call the modality-specific types of properties that are necessarily represented by imaginings in that modality the modality’s “specific imaginables”. The content specificity thesis says that modes of sensory imagining are individuated by their specific imaginables. The thesis presents us with the challenge of identifying the specific imaginables, at least of the more prominent modes of sensory imagination. It is hard to see how this challenge could be met if we must individuate the full range of possible senses (including such senses as echolocation) alongside the more familiar set of five. Setting this problem aside, the challenge remains formidable enough, for not all the specific sensibles of a perceptual modality are specific imaginables of the corresponding mode of sensory imagining: specific sensibles of a perceptual modality need not be represented by imaginings in that modality. Colors, for instance, are specific sensibles of vision (only vision represents them) but it seems that one can visualize a scene as colorless or as indeterminate with respect to color. Since colors need not be visualized, they are not specific imaginables of visualizing and cannot be used to distinguish it from other modes of imagining. Likewise, it is plausible to suppose that we see objects only by seeing the light they emanate, transmit, or reflect. Yet it is possible to imagine a scene without imagining it as either luminous or illuminated. Unless instructed otherwise, most people asked to visualize an object will have an image that is indeterminate as regards the lighting—they do not visualize the object as in the dark but neither do they visualize it as in the light. We should be wary of promoting features characteristically represented by experience as it occurs in perception to the status of essential ingredients of experience tout court. Nor should we underestimate the demands this lesson imposes. The only types of properties that seem necessarily to be visualized are spatial ones, but spatial properties are also represented by visual and tactile imagination. Perhaps the content specificity thesis is needlessly strong. Why require that every imagining in a modality represent a specific imaginable? Why not require only that visualizing, for example, involve experiences of a type some of whose tokens represent imaginables specific to vision? This weaker notion of specific imaginables does not entail that every visualizing represent color or illumination. It only requires that every visualizing involve a type of experience some tokens of which represent color or illumination. The difficulty with this proposal is that it begs the question of which modality any given imagining belongs to. Why not say that a sensory imagining of a cube that is indeterminate as to color and

216 OUT OF SIGHT, OUT OF MIND

illumination is a tactile imagining? As we shall see, this sets a problem that is by no means trivial. Following a suggestion of M.G.F.Martin, Robert Hopkins has ingeniously proposed that to ascertain the specific imaginables of visual and tactile imagination we must consider how spatial properties (specifically, perspectival properties) are represented in vision and touch (Hopkins 1998:172–80). The thought is that whereas in vision objects are always experienced as separate from the point of view from which they are observed, objects are always experienced in touch as in the same place as some “point of touch”. Touch therefore implicates an experience of bodily space: in touch, points in bodily space are experienced as identical to some points in environmental space. It is important to note that not every spatial property need be represented in touch by representing the colocation of body and object. Touch can represent the object, only part of which I am now in contact with, as the fender of my automobile. Hopkins’s claim is that in touch “some part of the perceived object is always presented as in the same place as some part of the subject’s body” (Hopkins 1998:179, my emphasis). The suggestion is therefore that tactile imagination necessarily involves an experience of some part of the imagined object as in the same place as some part of the imaginer’s body. By contrast, visual imagination necessarily involves an experience of objects as separate from a point of observation. Represented colocation of the object perceived with the perceiver’s body is surely a specific sensible of touch—experiences in no other sense modality have this kind of content. However, we have seen that not all specific sensibles of a modality are specific imaginables of the modality and also, more generally, that what is characteristically represented by experience in perception may not be represented by experience in the corresponding mode of imagination. Is represented colocation essential to tactile imagination? There is reason to think not. Consider the imaginative capacities of congenitally blind persons, whom we may suppose to have tactile imaginings. What reason is there to suppose that the contents of the tactile imaginings mentally manipulated by the blind always include colocation at a point of touch? We know that blind people possess the same spatial imagination skills as sighted people (Marmor and Zaback 1976; Carpenter and Eisenberg 1978; Kerr 1983; Zimler and Keenan 1983). In one standard imagery experiment, subjects are asked to imagine an aerial view of an island with various items (a pond, a tree, a hut, a well) located at various points. Are blind people incapable of imagining the island unless it is part of the content of the image that they share a point of touch with part of the island? Do they then imagine that they are capable of touching objects dozens of meters away from the point at which they imagine them observed? We have three options. One is to reject the supposition that those incapable of vision are incapable of visual imaginings. The empirical evidence may be explained on the hypothesis that although blind people cannot see, they can visualize. This would mean that a blind person can visualize spatial and textural properties in virtue of having encountered them in touch. To do so, she need

DOMINIC MCIVER LOPES 217

only imagine them from a separate viewpoint rather than from a point of touch with her own body. This is hard to swallow. It certainly implies that people are very often mistaken about what kinds of experiences they are able to enjoy. One may also wonder how, if this is the case, it is possible to account for the visual character of visualizing and seeing. Still, one might embrace it as a consequence of the view on offer rather than an objection to it. For those who find this position too unpalatable, one alternative is to reject the claim that tactile experience necessarily represents parts of objects as colocated with a point of bodily touch. This option enables us to explain the empirical data concerning the imagery skills of blind people as a product of tactile experience. But there is a third option, namely to allow that some spatial experiences are amodal—they are indeterminately tactile or visual. The content specificity thesis is therefore false: some sensory imaginings possess no specific imaginables in virtue of which they belong to one and only one sense modality. If this is correct, we may recuperate the thesis that we visualize only when our images represent properties that are specific sensibles of vision and we have tactile imaginings only when our images represent properties numbered among the specific sensibles of touch. The virtue of this proposal is that the specific sensibles of a modality also count as its specific imaginables. The drawback is that it means a great deal of imagining turns out to be amodal, even though most people take their imaginings almost always to be in one or more of the modalities. Is it possible to decide between these options? The choice between the second and third options is particularly troubling. It is always a good idea to try to avoid inputing to people widespread error about the consciously accessible features of their experiences. So can we determine the modality of an imagining independently of determining its content? There might be a way to do so empirically. It has been shown that imagining an object interferes with the ability to perceive it and perceiving an object interferes with the ability to imagine it. Moreover, these interference effects are more pronounced when the imagining and the perception are in the same modality (Segal and Fusella 1970; Hampson and Duffy 1984; Craver-Lemley and Reeves 1992). Seeing strongly interferes with visualizing and touch with tactile imagining. The accepted explanation of this is that one mechanism interferes with another to a greater degree when both mechanisms involve experiences of the same type. If this explanation is correct, interference effects can be used to probe the conditions in which any given imagining involves experiences in any given modality. We do not know that the content specificity thesis is false but we are not yet prepared to see that it is true. More must be known about what people are doing (or what they take themselves to be doing) when they have imaginings in the various sense modalities. However, if it turns out that the content specificity thesis is false, must we then accept that modes of sensory imagination are individuated by qualia? There is another possibility. We have seen that while sensory imagining requires having an experience, it may also incorporate an element of prepositional imagining. It may be that although experiences involved

218 OUT OF SIGHT, OUT OF MIND

in some sensory imaginings are amodal, the sensory imaginings are endowed with a modality in virtue of a prepositional imagining. Among the conditions that one might propositionally imagine to hold is that one is visualizing. A blind person may amodally imagine an island from the air in the sense that the content of her experience is neither specifically tactile nor specifically visual, and yet she may rightly be said to imagine it haptically if she imagines that her experience is a tactile one, or to visualize it if she imagines that her experience is a visual one. This solution has the virtue of not implying that people are always mistaken when they take themselves to be visualizing and do so by means of an amodal experience. A stronger claim has been made, namely that visualizing requires imagining that one sees (Walton 1990). This claim is objectionable on at least one count. Imagining that p entails the capacities to grasp that p and indeed also to believe that p. But being able to visualize comes just with being able to see and no more requires the rather more sophisticated capacities to believe or grasp that one is seeing than seeing requires the capacity to believe or grasp that one is seeing. It is likely that very young children can visualize before they acquire mental concepts, including concepts that apply to modes of awareness of the world. It would be strange to discover that small children could not duplicate the results of the simpler mental rotation tasks that have famously shown human adults to employ mental images. This objection does not tell against the proposal that imagining in a particular sense modality sometimes depends on imagining that one is experiencing in that sense modality. The proposal simply entails that we can only attribute amodal imaginings to small children and others lacking concepts of the sense modalities. This is a verdict we should be able to accept, especially since the creatures in question, lacking modality concepts, are blind to their predicament. Again, whether this proposal has anything to explain depends on whether the experiences involved in imagining are ever amodal, and this is an empirical question that has not yet been answered. If sensory imaginings involve experiential states of the kind involved in perceiving, then we should be wary of translating our conception of experience in perception to an account of experience as it occurs in imagination (or in representational art). The experiences we have when we see may be viewpointed and those we have when we touch things may always involve a point of touch, but it neither follows that the experiences we have when we visualize are viewpointed nor that the experiences involved in tactile imagination include a point of touch. We may have something to learn about experience from its involvement in sensory imagination. Delivering on this promise, however, depends on a better acquaintance than we now have with the empirical contours of sensory imagination.

DOMINIC MCIVER LOPES 219

Sensory imagination in the arts The operation of sensory imagination in the arts may prove an important source of data. Until recently few philosophers have had much to say about imagination in the interpretation and appreciation of art (Kant is the notable exception). If imagination was discussed at all in connection with art, the emphasis was on the creative imagination. This neglect is due in part to the traditional conception of imagination as a faculty that operates in all perception, not just perception of artworks. Attitude theories of imagination, especially those developed by Kendall Walton and Greg Currie, are to be credited for having reversed this state of affairs. If imagination is a prepositional attitude properly taken towards fictions, then it deserves a central place in philosophical accounts of art—and nothing said so far casts doubt on this. But while traditional theories of imagination place it in close relation to perception and in distant relation to art, attitude theories represent a swing of the pendulum to the opposite extreme. Thus for Walton all art is fiction and depends on propositional imaginings (Walton 1990:2–3). I argue that Walton’s account of the fictionality of pictures and the role of propositional imagination in appreciating them is inadequate. This argument, if successful, indicates how to proceed in developing an account of the sensory imagination in appreciating all the arts. Imagination and depiction In Mimesis as Make-Believe, Kendall Walton is clear that not all imaginings are propositional. One may imagine that bears fly kites or one may more simply imagine bears flying kites (or kite-flying bears), and it is also possible to imagine doing things, as when one imagines being a bear flying kites. Yet Walton’s account of the visual imaginings he takes to be prompted by pictures casts them as propositional imaginings. It implies that the attribution of these propositional imaginings can explain what is explained by the attribution of sensory imaginings. Walton’s is an account of pictorial imagining as propositional imagining. According to Walton, pictures are designed to function as games in pictorial games of make-believe. What makes the games specifically pictorial is that they enjoin us to imagine, of our looking at a picture that it is an instance of looking at what the picture depicts. Bonnard’s Nude in a Bath and Small Dog is designed to function as a prop in a pictorial game of make-believe that prescribes, because of the way the picture’s surface is marked, that the viewer imagine, of her seeing the surface, that it is a case of seeing a woman in a bath. This view has three interesting features. First, if a picture prescribes an imagining that one is seeing its subject, the picture is a fiction—it is fictional when I look at the Vermeer that I am seeing a woman holding a balance. Second, one’s experience of seeing the picture surface is penetrated by the pictorial imagining that the seeing is seeing whatever the picture represents. The seeing of the picture surface and the imagining that one sees the scene it depicts are “bound

220 OUT OF SIGHT, OUT OF MIND

together as a single phenomenological whole” (Walton 1990:295). Third, the imaginings prescribed by pictures are relatively rich, allowing “for the fictional performance of a large variety of visual actions, by virtue of actually performing visual actions vis à vis the work” (Walton 1990:296). For instance, one may fictionally scrutinize the woman in a bath by actually scrutinizing the Bonnard. There is therefore more to visual imagining as it figures in pictorial games of make-believe than merely imagining that one is seeing (contra Moran 1994:88–9). In pictorial imaginings, experience of the picture surface assumes the role that is played by an imaginative experience in traditional accounts of sensory imagining. That is, in pictorial imagining, the propositional imagination harnesses an occurrent visual experience in order to shape its content, borrow its phenomenology, and sustain a rich variety of imagined visual actions. When I look at the Bonnard, I not only imagine that I see a woman bathing but this imagining gets its phenomenology from my seeing the Bonnard’s surface, and the content of my seeing the Bonnard’s surface is penetrated by the content of the imagining that I see a woman bathing. This is why I may fictionally scrutinize a woman bathing by scrutinizing a flat, marked surface. Is it the case that the attribution of pictorial imaginings to picture-viewers can explain what is explained by the attribution of sensory imaginings to pictureviewers? Do pictorial imaginings, as Walton has construed them, possess the three features that are characteristic of experience? Are they necessarily occurrent, do they have a rich phenomenology, and is their content nonconceptual?—The answer is obviously affirmative on counts one and two. Since pictorial imagining is seeing the picture surface in a way that is penetrated by certain prepositional imaginings, it necessarily possesses a visual phenomenology and is necessarily occurrent, having a precise duration. But it is not obvious that Walton’s account of pictorial imagining construes its content as nonconceptual. What we see in pictures is nonconceptual—we need not possess a concept of a property in order to see the property represented in the picture. I may see a particular three-dimensional shape in a depicted scene without possessing a concept of that shape. But it cannot be that the shape is represented by my experience of a shape on the picture surface, for the picture surface neither has the shape in question nor do I see it as having the shape (I see it as flat). Nor again can it be that I imagine that I see the shape, as long as imagining is a propositional attitude whose content is conceptual. In other words, the nonconceptual content of my imagining of the shape cannot be identified with the nonconceptual content of my experience of the picture surface, for the shape in question is not part of that content; but neither can it be identified with the content of a propositional imagining, for that is conceptual. It is not the case, as Walton’s view requires, that the divide between nonconceptual and conceptual contents coincides with the divide between the contents of experience of the picture surface and the contents of the propositional imagining that gives it shape. If I have an experience as of an F when I look at a picture that is G but I do not

DOMINIC MCIVER LOPES 221

possess a concept of Fs, then my having the F-experience cannot be attributed to my imagining of my seeing the G that it is a case of seeing an F. Walton seeks to accommodate the thought that it is one thing to imagine that one has a visual experience with such and such content but it is another to imagine having the experience in a way that essentially involves having the experience, or one like it. Since, as Walton sees, pictorial imagination is a case of the latter phenomenon, he proposes that pictorial imagination consists of propositional imagining integrated with ordinary experience of pictures as physical objects. But ordinary experiences of pictures as flat, colored surfaces cannot bear or come to assume the kind of contents that this proposal requires. Something more is needed. Attitude theories of imagination provide for a unified account of representation in the arts. All pictures, for instance, are fictions and require the exercise of imagination for their interpretation. This will not do; not all pictures are fictions and interpreting pictures is not a matter of propositional imagining. We may wonder, then, whether picture-appreciation involves non-propositional visualizing. Surprisingly, there are reasons to think not. Visualizing and seeing differ in the following respect: the experiences involved in visualizing are susceptible to the illusion of immanence, whereas those involved in seeing are transparent and hence are like seeing. Nude in the Bath and Small Dog has certain properties and it also represents a number of other objects as having various other properties but, Fauvist aspirations aside, these properties appear to us to be stuck on the objects concerned (the woman, the bath, the dog, and the picture itself); none appear to be properties of mental objects. Seeing what a picture represents is like seeing the properties of its surface: both represent properties wholly as properties of external objects. What accounts for the difference between seeing in pictures and ordinary seeing is the twofoldness of the former. Experiences of many pictures in most circumstances amalgamate features of the picture surfaces with features of the objects they represent. Awareness of one aspect does not quash awareness of the other: the two aspects blend into one “twofold” experience. In some cases, such as trompe l’oeil pictures, twofoldness is either absent or suppressed. These pictures have the same phenomenology as seeing—they are illusionistic. Indeed, this is additional proof that seeing in pictures is not susceptible to the illusion of immanence and so does not consist in sensory imagining. None of this shows that imagination plays no role in pictorial interpretation. Understanding pictures of objects that do not exist or events that have not occurred may depend on propositional imaginings. Things stand with depiction as they stand with language. Propositional imagination explains fictional representation in both media of communication. But it nevertheless remains the case that a pluralistic conception of imagination that embraces sensory imagination as independent of propositional imagination ironically limits the role of imagination in the interpretation of pictures.

222 OUT OF SIGHT, OUT OF MIND

Imagination across the arts This limitation may be the price we must pay for an expansion of the role of imagination in other areas. A full appreciation of some works of art may involve sensory imagining. Some poetry, for instance, is designed to stimulate visual imaginings, as is some music. Should we welcome the proposed expansion? The consensus position is an emphatic no. Two kinds of objections are typically raised. The first is that sensory imagining is not required to grasp the content of art works. One can come to fully understand what a painting or novel or symphony represents (if anything) without engaging in any sensory imaginings it might prompt. To think otherwise is to espouse what Peter Kivy has called the cinematic model of interpretation, which is to be found in Locke and also the phenomenologists (Kivy 1997). There are several reasons to doubt that sensory imagery is required for the interpretation of literary works, to take one art form. Not every reader has sensory imaginings while reading. Since these may be either less determinate or else more determinate than literary descriptions, having them cannot fully explain our ability to grasp the descriptions in question. An additional worry is that if having an imagining and its degree of determinacy are optional then one’s interpretation of the work is not subject to the author’s control and, as Kivy puts it, “reading a novel should be under the control of the author to some large degree” (Kivy 1997:74). Thankfully, the present account of sensory imagination in the arts is immune to this objection. Visualizing, for instance, is not necessary to the interpretation of pictures—and each of the worries just expressed lends additional support to this. Arguments meant to show that sensory imagination is not required for interpreting art works are hardly new—I.A.Richards deployed each of the arguments mentioned above in an attack on sensory imagining that established orthodoxy in twentieth-century literary criticism. However, Peter Kivy has introduced a second objection (Kivy 1997:74–6). Noting that some literary descriptions have great power to arouse the sensory imagination, he nevertheless argues that the descriptions in question do not have the purpose of stimulating the sensory imagination (just as this chapter does not have the purpose of provoking its reader to visualize bathing nudes). This objection hinges not upon the interpretation of literary works but rather the aesthetic appreciation of them. It entails that having sensory imaginings while reading adds little of value to our experience of the work as a work of art. The arguments for the first objection do double duty here. Since the imaginings accompanying reading are optional, sometimes relatively indeterminate and sometimes too determinate, they are not subject to authorial control, so they can add nothing of value to her work. This assumes that whatever value accrues to a work of art as a work of art it must have intentionally—and this is a tendentious assumption. It seems that a great deal of what we legitimately value in many works of art is not entirely subject to the artist’s control. Jazz numbers are to be admired

DOMINIC MCIVER LOPES 223

in part for their providing material for unanticipated and revealing improvizations. Granting the tendentious assumption, however, the account of sensory imagination offered here contains the resources to sidestep the argument. We have seen that sensory imaginings are elements of larger imaginative projects in which they are constrained by prepositional imaginings. The sensory imaginings that one undergoes while reading a description are constrained by the relevant prepositional imaginings. Of course, there remains room for individual variation, but the same is true of literary interpretation. Ellen Esrock, summarizing an extensive review of the empirical research on reading and imagining, lists some of the purposes to which the evocation of visual imagery by literature can be put (Esrock 1994:188–201). It enhances memory of story elements. It increases comprehension of spatial descriptions. It helps convey a sense of the concreteness of fictional worlds. It evokes the psychodynamics of vision, provoking, for instance, voyeuristic interests. It helps craft a point of view upon the narrative for the reader to occupy. It establishes formal contrasts. It assists in thematizing the act of imagining itself. Note that these achievements do not require fine-grained control over when visualizing occurs or with what degree of determinacy. They are also sometimes key achievements to the success of a literary work as a work of art. Sensory imagination can also have value when it makes experiences of artworks multimodal. Although visual imagination can only be elicited by works the grasp of which does not occur through seeing, it may add a visual element to works of literature or music. In this way sensory imagination helps our experience of works of art approach the condition of perception, in which we often enjoy experiences in many sense modes simultaneously. This can be a merit in a work of art, though its absence may not be a demerit—cinema does not allow much in the way of sensory imagining unless it is tactile, gustatory, or olfactory. Imaginings in each of the sense modalities have their specific virtues. In visualizing, connections among features of a situation, or gaps and inconsistencies, can leap to the eye—indeed, they might be invisible if the situation were described. In auditory imagination nuances of an emotional state can be conveyed in a way that a description or a picture can only approximate. Whatever value might accrue to a painting because of its visual appeal might also accrue to a poem because of its appeal to visual imagination. This means that a visual aesthetics is not limited to the visual arts but must include an account of visualizing as it occurs outside painting, sculpture, and film. It also means that an account of appreciation of works in each of the art media cannot limit itself to properties of the experiences by means of which the work is perceived. Music is an auditory medium but it does not follow that its effects are exclusively auditory, since works of music can be designed to induce nonauditory sensory imaginings. If this is right then we may wonder how nonvisual works can prompt visual imaginings, how nonauditory works can prompt auditory imaginings, and the like. This is a problem philosophers have hitherto attempted to address in a comprehensive way.

224 OUT OF SIGHT, OUT OF MIND

Attitude theories have made imagination a central concept of philosophy of art by linking it to fiction. They have knit art to the attitudes, that part of cognition that trades in prepositional contents. A notion of sensory imagination as involving quasi-perceptual experience may be indispensable to an understanding of artworks as objects of perception whose value lies in our perceiving them. Acknowledgment I am grateful to Robert Hopkins, Eileen John, Matthew Kieran, Mohan Matthen, and Patrick Rysiew for their comments.

13 THE IMAGED, THE IMAGINED, AND THE IMAGINARY David Davies

In this chapter, I want to examine the role that should be ascribed to the imagination (the “imagined”) in an adequate account of the experience, understanding, and enjoyment of fictions (the “imaginary”) that are presented to receivers in the cinematic medium (the “imaged”). Strictly speaking, I have sacrificed precision for epigram in my title, for not only is the realm of the imaged larger than the realm of the cinematic, and not only is the cinematic a mixed medium that presents its fictions through image and sound, but it is also questionable whether one should identify what is fictional with what is imaginary —as I have argued elsewhere, there is nothing incoherent in the idea of making fictional what one knows to have actually occurred (Davies 2000). Readers familiar with recent debates in the philosophy of film will recognize my subject as bearing upon the analysis of what has been termed “film experience” in a broad sense, and will also rightly anticipate that one of my concerns will be Gregory Currie’s comprehensive “cognitivist” model of film experience presented in his Image and Mind and defended in subsequent writings. Currie challenges the assumption that narrative cinema has the capacity to elicit in viewers a certain kind of “illusion” that they are actually watching the events portrayed, and that it is in virtue of its illusory powers that cinema is able to engender the sorts of experiences that viewers standardly report. Currie’s account is attractive because it offers a detailed model of film experience that both identifies the cognitive capacities that the latter requires and argues for their place in a more general theory of human cognition. But a number of critics have claimed that it fails to accommodate genuine elements of film experience that proponents of the “illusion” thesis have sought to capture in their theories. Richard Allen, for example, in his book Projecting Illusion, attempted to do justice to what was right in the “illusion” thesis, proposing that film experience involves what he termed “projective illusion”. Projective illusion, according to Allen, involves a loss of “medium awareness” and an “imagined seeing” of the events cinematically represented. Dominic Lopes has also argued, against Currie, that film experience involves a kind of “perceptual illusion” that is a form of “imagined seeing” (Lopes 1998). My approach in this chapter is broadly sympathetic towards Currie’s critics. I think they are correct in identifying a dimension of film experience that is missing

226 DAVID DAVIES

in Currie’s account. But I think we need to set out more clearly the overall structure of Currie’s model of film experience in order to determine how the claims of his critics bear upon this model, and also to determine what options are open to us if we grant those claims. I shall sketch Currie’s model in the next section, and examine his critics’ objections to that model in the following two sections. I want to draw a more general moral from these reflections, however, concerning the import of the sorts of criticisms raised by Allen and Lopes against Currie. I think we need to reassess the plausible idea that our philosophical concerns about film experience can be resolved by the sort of cognitivist model proposed by Currie, while also resisting the idea that such concerns are properly dissolved by Wittgensteinian therapy (e.g. Allen 1997). At the end of this chapter, I shall suggest an alternative strategy for responding to these concerns, one that provides us with a basis for saying the kinds of things that writers like Allen and Lopes rightly want to say. Currie’s model of film experience We may distinguish four elements in Currie’s account of film experience, corresponding to four conceptually distinguishable stages in such experience: Stage I. A theory of image recognition. Currie claims that recognizing what a cinematic image represents—recognizing that it is an x-representation (e.g. a black-horse-representation)—involves an exercise of the same sub-personal recognitional capacities that subserve recognition of the thing represented (e.g. a black horse) (Currie 1995a:80ff). Just as my ability to recognize horses involves the triggering of an appropriate subset of the feature-detectors associated with my sub-personal concept of “horse”, so it is in virtue of the triggering of an analogous subset of these feature-detectors that I recognize something as a representation of a horse rather than as a representation of, for example, a cow. This establishes a sense in which what Currie terms the “Likeness” thesis—that is, the thesis that the experience of looking at a cinematic representation of x is like the experience of watching x (Currie 1995a:20)—is true. But, since the truth of the Likeness thesis is established by facts about the sub-personal capacities that make both kinds of experience possible, it does not follow that there is a likeness of experienced content between cinematic experience and non-cinematic experience of a sort that might support the claim that viewers are subject to some kind of illusion when they watch a cinematic representation of x. There is, we might say, an objective likeness between an x-experience and a cinematic-representation-of-xexperience, but this does not entail a particular kind of subjective likeness between these experiences. Stage II. A theory of the content of film experience. This is not much developed by Currie, save implicitly, but he clearly seems committed to the view that, in watching a film, I standardly apprehend what I am watching as a representation. Thus, when the cinematic image causes my x-detectors to fire at the sub-personal level, my experience is one that I would describe as an x-

THE IMAGED, THE IMAGINED, THE IMAGINARY 227

representation-experience, not as an x-experience. To hold, of some experiential context, that “experience represents the world as being a certain way, when in fact it is not that way and the subject does not believe it to be that way” (Currie 1995a:29) is to endorse a “perceptual illusionist” account of that context. While admitting that he does not have a conclusive argument against the “perceptual illusionist” thesis that, in film experience, looking at an x-representation is subjectively like looking at an x, he nonetheless explicitly rejects such a view (Currie 1995a:30). This rejection is also implicit in his argument against the thesis that moviewatching involves a perceptual illusion of movement within the image (Currie 1995a:34–42). Currie argues that I am not subject to such an illusion because the movement I experience is a genuine property of the representation. If this is to rebut the illusionist account of movement, then what I am aware of (opaquely construed) must be a representation, to which I ascribe (correctly) the relevant properties of movement. In other words, we do not standardly suffer from perceptual illusion in perceiving movement in the image since what we perceive is a “moving image”, and that is exactly what is really there (where a “moving image” is a representation in which movement is perceived by a normally constituted observer). If the content of experience in standard conditions of viewing a film is to be described in terms of x-representations rather than x’s, then there can be perceptual illusion only if we experience the representations as having properties they lack. Since we do not do such a thing, there is no illusion. Stage III. A theory of the determination of what is true in a cinematic story. On the basis of my engagement with what is cinematically presented, I form beliefs about the content of the representations that I am observing. I determine what is true in the story N by referring these contentful representations to a maker who is presenting them in this order for the purpose of telling a story. I then infer from the representations presented that the storyteller is telling a story in which it is true that p. This is intended to parallel the manner in which I interpret literary fictions: in each case, interpretation is a matter of referring “evidence” (the literal meaning of the text, the “appearance” meaning of the image) to a maker whose causally efficacious storytelling intentions explain why the evidence is the way it is (Currie 1995a:239ff). In arriving at beliefs about what is true in the story being told by means of the cinematic representations, I also engage in various exercises of what Currie terms “secondary imagining”. In the first place, such imaginative activity helps me to determine what is implicitly true in the story concerning the thoughts and feelings of the characters. I am able to imaginatively infer this, given what they are represented as doing or undergoing. Second, I may need to use my imagination to determine what the maker should be taken to be doing by presenting to me this sequence of cinematic images, given that I take the purpose of such presentation to be the narration of an intelligible story to a viewer like myself.

228 DAVID DAVIES

Stage IV. A theory of our imaginative engagement with cinematic fiction. Those beliefs about what is “true-in-N” that are the output from stage III of the model are then given as input to a “mental simulator”, which simulates acquiring those things taken to be true-in-N as beliefs about the actual world. This, for Currie, is a matter of “impersonal” imagining—that is, imagining that p is true. He argues against what he terms the “Imagined Observer Hypothesis”, which takes our imaginative engagement with cinematic fiction to be a matter of “personal imagining”— imagining that I am actually witnessing p. This distinction between “personal” and “impersonal” imagining is to be understood in the following manner: just as, in reading a fictional text, the content of my imagining is not “I am reading a text in which are described events E, and E are real”, but, for events E described in this text I am reading, “E are real”, so, in viewing a fictional film, the content of my imagining is not “I am viewing a representation of events E, and E are real”, but, for events E represented by the images I am viewing, “E are real”. Currie maintains that film experience, so conceived, involves neither cognitive nor perceptual illusion. Cognitive illusionism, here, as a thesis about our experience of fictional films, is the claim that such experience standardly involves the formation of false beliefs about the world. Most obviously, it might be argued that, in watching a fictional cinematic representation that p, I form the false belief that p, or that I am observing that p, or that I am watching a representation of real events. Some forms of cognitive illusionism are inconsistent with Currie’s “theory of the content of film experience” (“stage II” above), according to which our conscious appropriation of the film is as an x-representation. This seems to rule out the acquisition, on the basis of film experience, of the false belief that I am observing the represented events rather than a representation of those events. Other forms of cognitive illusion are incompatible with Currie’s account of how we determine what is true in a cinematic narrative that we believe to be fictional (“stage III” above). For example, if we mistook the fictional representation for a representation of actual events, it is likely that we would also infer false beliefs about the world based on the content of the experienced representation. But, in determining that content, we would not refer the representations to a storyteller, but would infer directly to states of the world taken to be causally responsible for the cinematic image. Cognitive illusion of this sort seems possible only when we are genuinely confused as to the status of a narrative, rather than being a feature of our engagement with cinematic representations with which we engage as fictions. A variant of cognitive illusionism is what we might term “pretense illusionism”, the idea that, while film experience does not lead us to form false beliefs about the world, it does lead us to imagine certain things that are false. Obviously, this is harmless if the things it leads us to imagine are those things we take to be true in the story. But Currie’s target, here, is the thesis that we imagine we are watching real events. This thesis, however, is to be rejected on the basis of Currie’s theory of our “imaginative engagement with fiction” (“stage IV” above).

THE IMAGED, THE IMAGINED, THE IMAGINARY 229

For to say that we imagine we are watching real events is just to say that our imaginative engagement with fiction involves “personal imagining”. Perceptual illusions, as noted above, are illusions that deceive the senses—that cause us to have sensory experiences that fail to correspond to those states of the perceptible world responsible for our having such experiences—but that do not lead us to make false inferences and thereby come to have false beliefs. Standard cases of such illusions arise in respect of trompe l’oeil paintings and the MüllerLyer illusion. While, when we are ignorant of the illusory nature of such entities, we may suffer from cognitive illusion, we may be subject to perceptual illusion even when we know that what we are looking at is a canvas, and will be so subject even when we know that the lines in the Müller-Lyer drawing are really equal in length. As noted above, currie’s “theory of the content of film experience” (“stage II” above) seems to rule out certain kinds of perceptual illusions. For example, if, in experiencing a cinematic representation of x, I am aware of it as an x-representation rather than as an x, then it seems that I cannot suffer from the perceptual illusion that I am looking at an x when I view a cinematic representation of an x. To summarize Currie’s response to various kinds of illusionist theses, what we experience, or judge ourselves to be experiencing, or judge to be before us, are representations or images of x’s, not x’s themselves. This is why there is no illusion in normal film experience, according to Currie. We neither have nor judge ourselves to be having x-experiences (by stage II), nor do we believe we are experiencing representations of real x’s (by stage III), nor do we imagine we are watching real x’s (by stage IV). Since we experience what is occurring on the screen as a representation of moving x’s (by stage II), and since there really is motion in the image, we do not suffer from perceptual illusion either. Projective illusion Allen maintains that standard film experience does involve a form of illusion (Allen 1995). He terms the distinctive kind of illusion associated with film experience “projective illusion”, where this is taken to involve a kind of “imagined seeing”. Projective illusion is characterized in terms of the features it shares or fails to share with other, more familiar, types of illusion, such as trompe l’oeil painting, mirror images, and the Müller-Lyer illusion. Allen proposes four features in terms of which illusions can be compared and contrasted: First, projective illusion involves a kind of sensory illusion where the senses are deceived but where we may lack the corresponding belief (Allen 1995:97ff). In the case of sensory illusions, Allen claims, we “think” of things as such-and-such without necessarily believing that they are such-and-such. Allen’s “sensory illusion” resembles Currie’s “perceptual illusion” in that it relates to subjective features of our perceptual experience, but the former, unlike the latter, is defined without reference to whether we are subject to cognitive (or as Allen terms it “epistemic”) illusion. He also talks about “entertaining” a prepositional content, but not

230 DAVID DAVIES

entertaining it assertively. Where we have sensory illusion, entertaining a thought about x is a matter of “imagined seeing”—imagining that we actually perceive x, according to Allen. In the case of protective illusion, we “entertain in thought” that we are witnessing the portrayed events: we imagine seeing those events. But we do not form false beliefs about the actual world, because, while we imagine the portrayed events to be fully realized before us, we do not think them as realized in the space of the actual world but in a “virtual reality”. Relatedly, Allen distinguishes between illusions that involve a loss of “medium awareness” and those that do not. Trompe l’oeil, when not recognized as such, involves loss of medium awareness, where this is a matter of taking an xrepresentation to be an x (Allen 1995:82). Protective illusion also involves loss of medium awareness. I do not experience the cinematic image as a representation, but as “a fully realized world” of experience (Allen 1995:88). Allen distinguishes between illusions that are “context dependent”—where we are subject to the illusion only when we have certain beliefs about what we are looking at—and illusions that are context independent (1995:100). Some sensory illusions are context independent—for example, the Müller-Lyer illusion and trompe l’oeil. We are (or may be) subject to the sensory illusion whether or not we know that the lines are equal in length, or that we are looking at a painting. Other sensory illusions, such as mirror illusions, are highly context dependent. We are subject to the sensory illusion only as long as we do not believe we are looking at a mirror. Thus, in such cases, we are subject to sensory illusion only as long as we are subject to epistemic illusion. Projective illusion is context dependent, in that the illusion can be broken by my knowledge that I am watching a film (as with mirrors). Unlike mirror illusions, however, I do not suffer from epistemic illusion even when I am under the sensory illusion, because it is part of the sensory illusion that the fully realized reality I am experiencing is not part of the actual world. Also, unlike mirror illusions, the illusion can occur in spite of my knowledge that I am watching a film. Finally, Allen wants to relate the different kinds of sensory illusions to the operation of “seeing as” (Allen 1995:101ff). In some cases, whether we are subject to an illusion is to some extent within our control, because the illusion involves seeing something one way rather than another, and we are able to influence the way in which we see that thing. In the case of the duck-rabbit drawing, for example, I can see this as either a duck-picture or a rabbit-picture, and I have some control over which way I see it (I know how to elicit one way of seeing it rather than the other). To see it as a rabbit-picture is to think of it as a rabbit-picture, and similarly for duck-picture, Allen claims. He maintains that this also applies to (recognized) trompe l’oeil, but not to the Müller-Lyer illusion (because I cannot see it as being the way it is, whereas I can see the trompe l’oeil either as a representation or as what it represents). Projective illusion is voluntary, and can be characterized in terms of seeing as. The distinction to be drawn here is between: (a) as one subject to projective illusion, viewing the cinematic image as a fully realized world, and (b) as a medium-aware spectator, viewing the

THE IMAGED, THE IMAGINED, THE IMAGINARY 231

cinematic image as a succession of shots of various pro-filmic events. However, projective illusion differs from standard cases of aspect seeing such as the duckrabbit drawing in its relation to volition. In standard cases of aspect seeing, we can simply choose to see things one way rather than the other, whereas in the case of projective illusion we can only allow ourselves to lose medium awareness (Allen 1995:106). The phenomenon Allen characterizes as “projective illusion” brings out a distinction between two ways in which we can engage with a cinematic fiction that seems to be borne out by reflection on our film experience. There is a sense in which some viewings incorporate an awareness of the representational nature of what is being experienced and some viewings do not. Furthermore, our capacity to view a film “as a fully realized world” seems to be context dependent in the sense characterized by Allen. In certain films, representational techniques are either intentionally or unintentionally obtrusive, making it difficult for viewers to “get involved” in the narrative. Where stylization is deliberately employed for such purposes, we have Brechtian “distantiation”. More generally, any departure from standard “representational style” is likely to at least compete with, if not undermine, our viewing of the film as a “fully realized world”, Allen himself notes that “to some extent the experience of projective illusion is simply a function of audience familiarity with the conventions of classical narration” (Allen 1995:116). However, any attempt to bring Allen’s claims about “projective illusion” to bear in a critical way on Currie’s model of film experience faces serious obstacles. One problem, to which I shall return later, is the role played by the notion of “medium awareness” in Allen’s account. In what sense of “medium awareness” can Currie be said to claim that medium awareness is not lost in film experience, and in what sense of “medium awareness” does Allen claim that medium awareness is lost in film experience? A more pressing problem, however, is to clarify to which element in Currie’s analysis of film experience the phenomenon of “projective illusion” might be taken to pose a challenge? If projective illusion involves “imagining seeing”, then it looks like a challenge to stage IV in Currie’s account— his championing of “impersonal imagining” over “personal imagining”. However, Allen often characterizes projective illusion as a matter of how we “experience the film”: “you perceive a fully realized though fictional world that has all the perceptual immediacy of our own; you experience the film as a projective illusion” (Allen 1995:107). Furthermore, projective illusion is something that happens to us when we allow ourselves to lose medium awareness. This makes it look more like a challenge to stage II in Currie’s account—his analysis of the content of film experience. As we have seen, Allen attempts to elucidate the phenomenon he terms “projective illusion” in terms of “imagined seeing”. But is “imagined seeing” to be taken as a distinctive kind of imagining or as a distinctive kind of seeing? If the former, then “projective illusion” is a challenge to stage IV of Currie’s model, as a version of the “Imagined Observer” account of our imaginative engagement with cinematically presented fictions. But, as Currie points out in his critique of personal

232 DAVID DAVIES

imagining, the idea that the film viewer “imagines seeing” the represented events in this sense is problematic. Suppose I ask you to imagine that you are observing the events represented in a film such as Jurassic Park. Then it is surely a condition for your succeeding that, internal to your imagining, you do affirm the reality of what you are observing, where this seems to involve taking what you are observing to be a real event in the world you yourself inhabit. But then, internal to your imagining, it seems you must also be imagining various other things concerning the relationships that obtain between the represented events and yourself, in virtue of your imagined location in the represented world. Even if we impose less stringent closure conditions on imagination than belief, it seems that I can only imagine seeing the represented events in Jurassic Park if it is also part of my imagining that I stand in certain other spatial and consequential relationships to the represented events. But, the objection goes, such imaginings are surely not usually part of my engagement with a cinematic fiction. In this way, imagining observing something is different from simply “entertaining” a possible state of affairs, since you are asked to imagine observing the events, and this requires that you imagine believing the events to be really taking place. If I ask you to entertain a hypothesis, as in the form “suppose, for a minute, that there are once again dinosaurs walking the earth under the following conditions”, on the other hand, I am not asking you to imagine believing that such is the case. But Allen’s “imagined seeing” might also be construed as a distinctive kind of seeing—or, more accurately, a distinctive kind of visual experience—in which case “projective illusion” represents a challenge to stage II in Currie’s account of film experience. Our experiencing of the projected image is describable as “imagining that one observes that p”, or “entertaining the thought that p”, and what differentiates projective illusion from normal seeing is that in the latter case we have, in addition, a mental act of endorsing or asserting the contents of that experience. There is no corresponding mental act in the case of imagining seeing —merely the absence of such a mental act. But, if imagining seeing is just another way of talking about the unasserted content of visual experience, then “imagining seeing” refers merely to the visual nature of the content, not to an observational relation in which we take ourselves to stand to the things imagined. In this case, there seems to be no reason why my “imagining seeing” that p commits me to any further “imaginings” relating me spatially or consequentially to the represented events. However, on this second construal of “imagined seeing”, Allen’s characterization of projective illusion in terms of “imagination” seems unmotivated, for imagination seems to involve the exercise of some sort of mental capacity. This criticism also applies to Allen’s talk of “imagined seeing” as a general feature of sensory illusion, where the latter includes such phenomena as the Müller-Lyer illusion. There is something very odd in the claim that, when the two lines of equal length continue to look to be of different lengths even after I have empirically verified their equality, my visual experience involves an exercise of “imagined seeing”. What it surely involves is a visual experience whose

THE IMAGED, THE IMAGINED, THE IMAGINARY 233

content I fail to endorse as representative of the real world—it is a case of the world “looking as if” it is a way I know it not to be. The experiential dimension of engagement with cinema Dom Lopes has developed a more subtle version of the charge that Currie’s model fails to do justice to the phenomenology of film experience (Lopes 1998). Lopes, like Allen, argues that film experience involves a kind of perceptual illusion, in that viewers have an experience “as of” the thing represented. His characterization of the nature of that illusion is, I think, much more promising than Allen’s rather confused notion of “protective illusion”. However, as Currie himself notes, Lopes misrepresents Currie as at least open to the idea that film experience involves a form of perceptual illusion, and, as a result, wrongly assumes that one might integrate the features of film experience deemed to be missing in Currie’s account into the sort of cognitivist model that Currie favors. Lopes argues that our experience of films involves two kinds of experience: first, there is what he terms “screen experience”—experiences of “patterns of light and sound”—and second there is what he terms “cinematic experience”— experiences “as of the scenes and events those patterns portray” (Lopes 1998:343). It is crucial, he further maintains, that we provide some account of cinematic experience that can be applied to both documentary and fictional film, given that “there is no intrinsic or phenomenological difference between experiences as of actual objects and events in movies and experiences as of fictional ones” (Lopes 1998:345). Lopes argues that Currie’s claim that represented motion is real rather than illusory only counters an illusionistic account of screen experience, and does not undermine the idea that cinematic experience involves perceptual illusion. Indeed, he further argues, Currie grants that such a view of cinematic experience is consistent with his “triggering” account of how viewers are able to recognize cinematic representations (stage I in Currie’s model of film experience). According to Lopes, “if I visually recognize something as Stallone, then it will seem to me that I am seeing him; likewise if a film represents Stallone by engaging my capacity to recognize him, then in watching the film, I will have an experience as of Stallone” (Lopes 1998:346). As he puts it later, talking of Dr Strangelove, “a viewer has an experience as of Sellars just because this Sellarsrecognition ability is triggered” (Lopes 1998:350). But, as we saw in the presentation of Currie’s position, Currie explicitly resists such a connection between objective likeness characterizable at the level of sub-personal mechanisms that subserve visual recognition, and the sort of subjective likeness that Lopes alludes to in his talk of having an experience “as of” the thing represented. Indeed, Currie himself cites this misrepresentation of his position in his response to Lopes: “Lopes asserts that I concede the role of perceptual illusion. Partly true— but the illusion I admit is sub-personal. It is an illusion to which the viewer’s visual object recognition system is subject, not one to which the viewer is subject” (Currie 1998:365). If “experience as of” is a matter of how things seem

234 DAVID DAVIES

to a viewer, then the former cannot be cashed in terms of the sort of sub-personal illusion admitted by Currie. Lopes’s principal charge against Currie is that the latter’s account of film experience fails to accommodate an essential experiential dimension of the latter. Currie, it is claimed, identifies cinematic experience with “impersonal perceptual imaginings”, identifying imagination in such contexts as a matter of simulated beliefs rather than as a matter of some kind of visualization. To this, Lopes objects that “cinematic experience cannot be a matter of imagination if imagination is simulated belief, for experience cannot be simulated by anything belief-like” (Lopes 1998:349). Of course, as Lopes recognizes, Currie’s model of film experience does not deny that there is an experiential dimension to our engagement with cinematically presented narratives—this dimension is represented, at stage II, as an awareness of a sequence of representations through which a narrative is being conveyed to us. The dispute arises at stage IV, where our imaginative engagement with a film is taken by Currie to be “impersonal” and belief-like in a sense that, so Lopes maintains, wrongly excludes experiential content from the imaginings. Lopes proposes to remedy this lacuna by incorporating, at stage IV, the sort of “perceptual illusionism”—“experience as of”—for which he has argued in discussing stage II of Currie’s model. (I remind the reader that the analysis of Currie’s model into stages I–IV is my own, and that I am recasting Lopes’s argument in terms of that analysis. I should also like to express my gratitude to Lopes for helping me to arrive at a clearer conception of his views and for correcting earlier misreadings.) Currie himself misrepresents this disagreement, taking it to be solely a matter of how we characterize the imagining that we must posit at stage IV of the model of film experience in the case of films taken to be fictional (Currie 1998:363–4). He takes the issue to be whether it is fruitful to think of this as a matter of “imagining seeing”. The alternatives, for Currie, are: (a) his “impersonal perceptual imaginings”, where there is no “imagined seeing”, but, rather, what is imagined is propositionally characterizable and derived from pictorially represented information contained in the representations acknowledged at stage II, (b) the “personal perceptual imaginings” that he rejects, where the viewer’s “imagining seeing” is glossed as her imagining that she is herself observing the presented events, and (c) Lopes’s proposal, taken to be an alternative way of taking talk of “imagining seeing” where this is impersonal but ineliminably perceptual in content —what Currie describes as “imagining E in a seeing sort of way.” Currie assumes that his overall position would be undamaged were he to concede to Lopes that the latter is a better way of thinking about the content of the imaginings through which we engage with fictional cinematic narratives. But, once we relate Lopes’s proposal to his construal of stage II of Currie’s account, we can see that the proposed revision of stage IV is less innocent than Currie believes. For the proposal is presumably that the content of our imagining when we imagine “in a seeing sort of way” is just the content of our experience of the film (stage II of Currie’s model). In that case, the content of that experience cannot be characterizable in terms of an

THE IMAGED, THE IMAGINED, THE IMAGINARY 235

awareness of representations, but must be characterized as an awareness of representeds, to be understood in terms of “experiences as of” what is represented. The deeper challenge that this poses to Currie’s model emerges if we ask, as we have thus far refrained from doing, how we are to think of the “perceptually illusory” experiencings that Lopes characterizes as “experiences as of” what is represented. He seems to think that these can be incorporated into Currie’s cognitivist model of film experience with little difficulty. Lopes wants to deny that “experience as of” depends upon my more general beliefs as to what I am doing— more specifically, my belief that I am watching a movie, which, if “experience as of” were cognitively penetrable, would be incompatible with my having an experience as of watching Spock on the bridge of the Enterprise when watching Star Trek. He therefore assumes that cinematic experience—“experience as of”— belongs at Currie’s “sub-personal level of experience” where designated ranges of feature detectors fire. But there are two serious problems with this view. First, as noted earlier, Currie wants to insist that the sorts of “objective likenesses” at the sub-personal level that explain our capacity to recognize representations do not in any way entail the sorts of “subjective likenesses” to which Lopes must appeal if his talk of “experience as of” is to capture how things “seem” to viewers. Second, the “intentionality” of ranges of feature detectors is defined extensionally—it is a kind of natural meaning that depends upon the things in the world that are correlated with the firing of some appropriate subset of feature detectors in the relevant sensory modules. But if the contents of our imaginings are to be the contents of our cinematic experiences, the latter must be characterizable in terms of fictional entities and states of affairs such as Spock, the Enterprise, and the attack of the Klingon fleet—none of which can count as the meaning of the firing of an appropriate set of feature detectors. This then raises the deeper question alluded to earlier—how are we to understand Lopes’s talk of cinematic experience in terms of “experience as of”? What place is there for such a notion in the sort of cognitivist model proposed by Currie. It is, in fact, a necessary feature of the sort of model which Currie endorses that the sort of content in terms of which Lopes wishes to characterize cinematic experience only enters at the level of judgments, beliefs, and other intentional operations upon the content of experience. He spells this out very clearly in his response to Lopes, distinguishing his position from Lopes’s as follows: If we take the case of seeing a cinematic image of Cary Grant while watching Bringing up Baby, Lopes will say that I have the illusion as of seeing a confused and exasperated paleontologist, while I will say that I have the non-illusory experience of seeing (the representation of) Cary Grant carrying on in a certain way, and am led very naturally and thoughtlessly by this to imagine a confused and exasperated paleontologist. That, I think, is the essence of what divides us. (Currie 1998:362–3)

236 DAVID DAVIES

Crucially, for Currie, perceptual experience is to be understood in terms of the operations of sub-personal perceptual modules that are triggered by extensionally individuated cues in the internal or external environment. But the contents of what Lopes terms “experience as of” are individuated intensionally, and depend on the conceptualizing of those extensionally individuated cues in a particular way. Conceptualization, on the cognitivist model of film experience to which Currie subscribes, can enter the picture only through propositionally characterizable operations of the understanding such as judging, believing, and imagining. Thus Currie can only acknowledge the phenomenon identified by Lopes as “experience as of” as the result of exercises of the imagination, as the preceding citation makes clear. A similarly revealing passage in Currie’s response to Lopes relates to such phenomena as the Müller-Lyer illusion. The latter, as a genuine perceptual illusion, is importantly different from the sort of cinematic experience that Lopes wishes to characterize as a form of perceptual illusion, according to Currie: When I see the two lines in the Müller-Lyer set up, or look into an Ames room through an appropriately placed pin hole, I have experiences as of lines of different lengths, or of people of strikingly different sizes. And these experiences put pressure on my beliefs. I know the lines are the same length, that the people are roughly the same size, but what is disturbing about the experience is that I cannot reconcile it with my belief. I do not think that cinematic experience creates the same kind of tension: we do not sit there struggling to maintain our beliefs in the face of a contradictory experience. Rather, the cinematic experience is one which encourages (very powerfully and usually without deliberation or other cognitive effort on our part) imagining that Spock is on the bridge. (Currie 1998:362) What is it, for Currie, to have “experience as of lines of different length”? If this is to stand in a relationship of tension with our belief that the lines are the same lengths, it must itself involve a propositional content of the form “that the lines appear to be of different length”. The operations of our sub-personal perceptual system cause us to make a perceptual judgment with this content, a judgment that conflicts with our standing belief that the lines are equal in length. The perceptual judgment is one we are caused to make in virtue of the fact that the Müller-Lyer figure “tricks” the perceptual system: the “lines of different length” feature detectors are triggered by lines of equal length, inviting us to judge, contrary to our standing belief, that the lines are of different length. Since “experience as of” must, on Currie’s model, involve a conceptual operation elicited by experience, and since there is none of the tension one would expect if that conceptual operation were one of judging that Spock is on the bridge, given our belief that we are watching a fictional representation, the appropriate account of “experience as of”, for Currie, must be in terms of imagination.

THE IMAGED, THE IMAGINED, THE IMAGINARY 237

Modes of medium awareness The real debate, I think, is much deeper than it might appear. To make room for the idea that the phenomena to which both Lopes and Allen allude are genuine features of film experience, we must challenge the very model of perceptual experience to which Currie is appealing in the preceding passages— and this is tantamount to saying that we must question whether the notion of “experience” required for an account of film experience is the cognitivist one upon which Currie is drawing. To see what this model must leave out in our cinematic experience (and, I would argue, in many other experiential domains central to our talk about the arts), it will be useful to return to the notion of “medium awareness”, which, it will be recalled, was employed by Allen in his attempts to characterize “projective illusion”. While Currie does not himself talk about “medium awareness”, he is clearly committed to the latter in one sense by his contention that we experience a film as a sequence of cinematic representations, which grounds his arguments against illusionism. In response to the claim that film viewers believe (falsely) that they are watching the actual events cinematically represented, Currie (and others) maintain that this claim is contradicted by the manner in which viewers behave. If I really believed myself to be observing from close up a herd of predatory dinosaurs, I would not remain seated in the cinema, but would attempt to flee or hide. What I believe, rather, is that I am watching a cinematic representation of a herd of predatory dinosaurs. This argument against cognitive illusionism draws upon a more general thesis about the purposes served by, and the constraints upon, the ascription of intentional states such as belief. Belief-ascriptions serve our interest in providing intentional explanations of the manner in which agents behave, and in anticipating how they will behave, in given circumstances. On one prominent view of these matters, it is in terms of differences in belief that we make sense of differences in behavior in response to identical or similar stimuli (Davidson 1974). Given this link between belief ascription and the explanation of behavior, a film viewer’s failure to act in the way in which persons believing themselves to be confronted by x would be expected to act is prima facie grounds for denying that film viewers believe themselves to be confronted by x. So the cognitive illusionist’s claim—that film viewers are subject to the illusion that they are observing what is cinematically represented—should be rejected. If the issue of “medium awareness” is an issue about what the viewer believes in this sense—does she believe she is watching an x-representation or that she is watching an x?— then we have good reason to hold that the viewer does not lose medium awareness. Let us term medium awareness in this sense—having the kind of belief that one is watching a representation that can be cited in explanation of one’s failure to respond actively (even if one responds passionately!) to what is portrayed in the cinema—MA1. The example in terms of which Allen introduces the notion of medium awareness—unrecognized trompe l’oeil painting—suggests that what he has in

238 DAVID DAVIES

mind is MAL For, in the case of (successful) trompe l’oeil, what I lose is indeed MA1. I falsely believe that I am looking at x, when I am actually looking at an xrepresentation. But there are good reasons to think that, when Allen maintains that projective illusion requires a loss of medium awareness, he does not mean a loss of MA1. Loss of MA1 will normally entail epistemic illusion— Currie’s cognitive illusion—since one will form various beliefs about one’s environment based on what is portrayed, and, unless what is portrayed tracks what is truly the case in one’s environment, these beliefs will be mostly false. Thus, in addition to the original false belief that one is watching x rather than an x-representation, one will come to have many other false beliefs as a result of one’s engagement with the cinematic image. But Allen denies that projective illusion normally involves any kind of epistemic illusion. This is one respect in which projective illusion differs from (successful) trompe l’oeil. Projective illusion, for Allen, involves imagining that one is seeing the represented events, but not believing them to be part of the real world: the “fully realized world” that one perceives is “virtual”. Furthermore, the “medium awareness” whose loss is a condition for projective illusion, for Allen, is something we can allow ourselves to lose, but this is surely not an appropriate way of describing MA1. For one thing, we do not have the degree of voluntary control over our beliefs required for the kind of loss of medium awareness associated with projective illusion. So it seems that loss of “medium awareness” of the sort posited by Allen as a condition for projective illusion must be compatible with retaining MA1. How, then, is Allen’s “loss of medium awareness” to be understood? The terms in which he describes what is going on in cases of unrecognized trompe l’oeil seem to anticipate his later attempts to bring together projective illusion and “seeing as”. He distinguishes between (a) seeing an x-representation as a pictorial representation, and (b) seeing an x-representation as if it were x. This seems to mirror the distinction, in the case of projective illusion, between (a) viewing a film as a succession of shots of pro-filmic events and (b) viewing the cinema as a fully realized world. In the case of projective illusion, as we have seen, talk of “seeing as” is to be cashed out in terms of how we think about things, where this in turn is cashed out, for sensory illusions, in terms of our “imagining seeing” things. “Thoughts”, here, are described as things that have propositional content but that are not entertained assertively. This suggests that for Allen, medium awareness in the case of cinema is a matter not of believing that one is watching a representation in the sense of MA1, but of “thinking about” what one is watching as a representation. It is then a condition for the occurrence of projective illusion that one lose medium awareness so construed—that one not think about what one is watching as a representation—in which case one may (but need not) come to “think about” what is represented in the manner of projective illusion. The latter is to be understood as an imagined seeing of what is represented, without any corresponding assertion that this is the way the world actually is. While, as Allen puts it, projective illusion involves perceiving “a fully realized though fictional

THE IMAGED, THE IMAGINED, THE IMAGINARY 239

world that has all the perceptual immediacy of our own,” we do not believe, as in the case of trompe l’oeil, that we are perceiving reality. But the terminology is puzzling here. What is it to “think about” a cinematic representation as a cinematic representation, where this is to be distinguished from having the belief, in the MA1 sense, that the representation is a representation? In particular, what is it to think about a film as a representation when one already believes that it is a representation? Presumably what we want to capture here is the distinction, noted above, between two ways in which we can be engaged by a film as we watch it—the distinction between admiring camera angles and admiring Juliet’s poise. In other words, to “think about” a film as a representation is to make certain kinds of judgments about what one is experiencing as part of one’s ongoing engagement with the film: it is one kind of mental process that can occur while we are watching the film. Thus talk about what we are “thinking about” in watching the film is talk about our occurrent mental states, as opposed to dispositional mental states such as the belief that one is watching an xrepresentation rather than an x. For example, when I admire the camera angle, or wonder how a particular sequence was shot, or whether a given sequence is a flashback, I am clearly “medium-aware” in this sense. But when I admire Juliet’s poise and beauty, or wonder how the murder victim was shot, or whether Kane’s childhood experiences explain his puzzling behavior, it seems that I am not “medium-aware” in this sense. Whether I am currently thinking about camera angles or Juliet, I continue to believe that I am watching a representation as long as I continue to be disposed to behave in the sorts of ways Currie describes in arguing against cognitive illusionism. Thus I retain MA1. But how we “think about” what we are watching can change many times in the course of watching a film: we can engage the film in both representational and narrative terms. We may introduce the term “MA2” to characterize medium awareness understood as a feature of how we (occurrently) think about a film. It seems to be MA2 that must be lost, according to Allen, if projective illusion is to occur. Since MA2 is a matter of how we occurrently think about the film, it is something we can allow ourselves to lose (unlike MA1), to the extent that we can decide to concentrate on other facets or aspects of the film. But further refinements are necessary. For Allen characterizes the distinction between “medium aware viewing” and “projective illusion” in terms of aspectseeing and “seeing-as”, and, as a consequence, maintains that loss of medium awareness is a necessary condition for projective illusion to occur. We could no more simultaneously apprehend a film as a representation and as a fully realized world than we can see the famous figure simultaneously as both a duck-picture and a rabbit-picture. But is it really the case that we can only apprehend a film either as xrepresentation or as x, and never as both simultaneously? Certainly, it is often the case that we seem to switch from “thinking about” a film in one way to thinking about it in the other. This switch frequently takes place when stylization in film narrative leads me to think about how a film represents what it represents. For

240 DAVID DAVIES

example, in viewing the closing sequence of Antonioni’s The Passenger, I move from an awareness of the main protagonist being visited by threatening looking strangers in an isolated hut to an awareness that the final shot involves a very long single take (five minutes plus) with the camera tracking in an approximate 360° curve. But what of films that are highly stylized throughout? For example, in Peter Wollen and Laura Mulvey’s Riddles of the Sphinx, every sequence in the film is shot in such a fashion. (Wollen, interestingly, wrote the screenplay for The Passenger.) Must one cease to apprehend this stylistic feature of the film in order to engage with what is represented as a “fully realized world”? Or is it the case that, as the viewer grasps the overall stylistic properties of the film, the capacity to operate simultaneously at both the “representational” and “objectual” levels increases, so that she is able to both (a) retain occurrent medium awareness, yet (b) experientially appropriate what she is watching in terms of x=what is represented rather than in terms of its being an x-representation? There are, I think, at least two good reasons to reject an appeal to aspect seeing in a model of film experience. First, Wollheim has famously argued against the idea that our critical engagement with representational paintings is a matter of “seeing-as”, replacing this notion with that of “seeing-in” where the latter respects the “twofoldness” of such engagement—our taking account at a given time of both the represented content and the manner in which that content is articulated in a medium (Wollheim 1980). Second, there seems to be evidence from the other kinds of illusions discussed by Allen that it is possible simultaneously to apprehend something both as a representation and as what is represented, and thus, apparently, to “think about” something at a given time in ways inconsistent with the thesis that what is going on here is some kind of “aspect seeing”. Consider a recognized trompe l’oeil. I can think about a painting as a painting, yet still be subject to the perceptual illusion, and thus see it as an x rather than as an x-representation. For example, the ceiling cornerpieces in the Accademia in Venice, which one knows are paintings from having inspected them from other angles, obstinately continue to look like architectural embellishments when viewed from below. I will describe my encounter with such a painting in the following way: I realize that this is a painting, and I can admire the skill with which it is done, but I still see the structure projecting towards me. Or consider various kinds of “op-art”—for example, the painting Brushstroke in the recent Sensation exhibition in London. This is a flat canvas that nonetheless looks curved, and no amount of concentration on “how the trick is done” can make the canvas look flat. I can certainly “think about” this painting as an artifact— indeed, this is just what I am doing when I am trying to work out how the trick is being performed—but I nonetheless see it as what it represents (a curved surface) rather than as what it is (a flat canvas). Similarly, with the Müller-Lyer illusion, we may not only believe (dispositionally) that the two lines are equal in length, but also be thinking (occurrently) about them in that way, but we are still unable to get them to look the same length.

THE IMAGED, THE IMAGINED, THE IMAGINARY 241

These examples are offered, in the first instance, to motivate the question: why should it not be possible to retain MA2 while apprehending the film as a “fully realized world” of the sort posited in Allen’s projective illusion? But they also raise a more basic question, as to what is involved in an adequate characterization of a sensory illusion. For consider: how should we describe the difference between the following situations: (a) I look at the trompe l’oeil, realizing it is a trompe l’oeil, but I still see it as what it represents; (b) I look at the same trompe l’oeil, still realizing that it is a trompe l’oeil, but now I see it as a representation— I can see “how its done”. The difference, it seems, is not a matter of how we “think about” the representation. Rather, what changes is what we see, or the manner in which we experience the picture, for I am already thinking about the picture as a representation in both cases. Certainly, to the extent that I see the trompe l’oeil as a painting, I will think about it differently. But my coming to see it that way does not seem to consist in a change in the way I think about it. This suggests that we need to distinguish between how we think about what we observe and how we perceptually experience what we observe in order to give an adequate account of the phenomenon Allen describes as projective illusion. Allen, as we know, wants to insist that sensory illusion, of which projective illusion is a species, is a matter of “thinking of things as being such-and-such” but not entertaining assertively the proposition that things are such-and-such. But, if we allow the contents of perceptual experience to be propositionally characterizable, so that we can be said to experience the world as being such-and-such, where these ways of experiencing the world are individuated intensionally, then we have something that fits Allen’s description: for we can characterize how things appear to us in prepositional terms while denying that the propositional contents are entertained assertively in experience. This is what the above examples (the Müller-Lyer illusion and the recognized trompe l’oeil) demonstrate. I think about the ceiling of the Accademia as a painting, but I perceptually experience it as an architectural relief: in so doing, I entertain assertively the proposition that it is a painting but not the proposition that it is a relief. If this is right, it suggests that we may want to distinguish a third kind of medium awareness, where “awareness” is a term that refers to our conscious perceptual experience—how we see or hear things, for example. In this sense of medium awareness (term it MA3), to be medium aware is to see something as an x-representation, rather than seeing it as an x. In the case of the trompe l’oeil described above, the move is from not being MA3 to being MA3. This allows us to describe cases where I think of something as a representation but am still subject to sensory illusion as cases where I am both MA1 and MA2 but fail to be MA3. The necessary condition for perceiving a film as a fully realized world is loss of MA3, which will usually require loss of MA2: but we may be capable of so perceiving a film even while retaining MA2, as I hypothesized in the Riddles of the Sphinx example. I think this also promises to clarify Lopes’s talk of cinematic experience in terms of “experience as of”. For the latter, as characterized by Lopes, seems to

242 DAVID DAVIES

refer to experiences that are individuated in terms of their conceptualized content. It is in this sense that the difference in experience between someone immersed in a movie and someone who walks in halfway through and observes the same scenes is not capturable in terms of something “belief-like”—it is no more a difference in how we think about something (our occurrent doxastic states) than it is a difference in what we believe about that thing (our dispositional mental states). Towards a non-cognitivist model of film experience If this is plausible, it opens up a couple of avenues of inquiry. First, we can look again at how the phenomenon that Allen labels “projective illusion” and that Lopes treats as a form of perceptual illusion is to be characterized. We may be able to replace talk of “entertaining in thought”, “imagined seeing”, and “experience as of” with a more adequate account in terms of the content of perceptual experience. But second, and crucially, we must provide an account of perceptual experience consistent with the idea that the content of such experience can vary in the way required by our analysis. One challenge that must be met is the following. It might be argued that, to the extent that it makes sense to ascribe fine-grained content to perceptual experience, such content derives from the way we think about something neutral given in experience. We must draw a distinction between perceptual experience strictly speaking and how we think about what we perceive as we perceive it. On this view, our experience of the world involves conceptualizing what is given in perception, where this is a matter of thinking about the given in a particular manner. In this case, it might be argued, the difference between MA3 and the absence of MA3 is a difference in the manner in which we think about what is perceived, not a difference in the “content of perceptual experience” more narrowly conceived. If we characterize the “neutral” content of experience so conceived as CE, the idea would be that: (a) MA3 is a matter of thinking of CE as an x-representation, (b) loss of MA3 is a matter of thinking of CE as an x. In this case, MA3 would just be MA2 applied to a neutral CE, and it would make sense to talk of the content of perceptual experience only if we used “perceptual experience” in the wider sense that incorporates thought about CE in the narrow sense. There are difficulties with this initially plausible view, however. First, it does not seem to fit all of the cases that lend themselves to analysis in terms of a difference in the content of perceptual experience. For example, in the case of the jazz enthusiast and the amateur, it seems that they hear the music differently, or that they hear different things in listening to the music. It seems wrong to say that they hear the same thing, but think about it differently, if we are talking about the content of experience rather than the nature of the stimulation that occasions such experience. The same would apply to A and B watching the same movie, where A joins B halfway through.

THE IMAGED, THE IMAGINED, THE IMAGINARY 243

Second, as the above discussion of recognized trompe l’oeil paintings indicates, the proposed view will have difficulty with cases most naturally described in the following manner: I think of what I am experiencing as an x-representation but experience what I am experiencing as an x. In the case of a trompe l’oeil painting known to be such, I think of what I am perceiving as a trompe l’oeil painting of an x, but I experience it as an x. Clearly, in such a case, it cannot be in virtue of thinking of what I am perceiving as an x-representation that my perceptual experience possesses its fine-grained content. But it is precisely in virtue of thinking of what I perceive in this way that I am MA2. So, if the finegrained content of perceptual experience derives from how we think about such experience, “thinking about” must be distinct from the sort of medium awareness at issue in claims about MA2. These considerations by themselves are unlikely to convince anyone who is initially skeptical about the idea that perceptual experience possesses fine-grained content independently of how we think about some neutral experiential material. The real argument against the proposed view of perceptual experience must be more general. We find one such argument—an argument to the effect that such a concept of experience is incapable of featuring in a coherent story about how thinking about the world is possible—in John McDowell’s Mind and World. McDowell argues that we need to think of perceptual experience as conceptualized, and also as reaching to the facts themselves, if we are to satisfy the requirements of “minimal empiricism”, that experience be capable of serving as a tribunal against which our thoughts of the world can be measured. The problem with McDowell’s account, however, is that, as presented, it seems to commit us to the idea that the world consists of facts that are individuated in the manner of Fregean senses (e.g. Dodd 1995 and Pietroski 1996). The challenge is to come up with a way of understanding the McDowellian account that avoids this unfortunate consequence. I have suggested elsewhere how McDowell should be interpreted in order to avoid such objections (Davies 1999). The key notion, I have argued, is one that closely resembles, at least in its label, the notion in terms of which Lopes tries to elucidate cinematic experience. Experience must be thought of as having both an extensional and an intensional dimension. Viewed extensionally, it is always experience of a shared objective world of objects, events, and processes, that admit of different right (and indeed wrong) descriptions. Viewed intensionally, it is always of that objective world as being a certain way. Perceptual experience is thus, in Strawson’s pregnant phrase, a subjective glimpse of an objective world (Strawson 1966). If we can avail ourselves of such a model of perceptual experience, as falling within what McDowell terms the “space of concepts”, then we can appeal to this model in characterizing the sort of cinematic experience that can furnish us with the content for our “visual imagining” in our imaginative engagement with cinematic fictions. We still have to argue, however, that our perceptual engagement with cinema is characteristically not MA3—that is, that its conceptualized content is correctly

244 DAVID DAVIES

regarded in terms of “experience as of x”, rather than as “experience as of an xrepresentation”, in Lopes’s terms. Currie might respond by pointing to the need to integrate into one’s account of film experience certain processes of extranarrative reasoning. The challenge might be expressed as follows: how can the proposed account explain Currie’s stage III, where we infer what is true in the story given what is represented in the image, by referring representations to a hypothetical narrator who is using the representations to tell a story? The further question, here, is how we are to relate an account of our reasoning about a story to an account of our imaginative engagement with cinematic fiction (Currie’s stage IV). In fact, this instantiates a more general difficulty for one who endorses a “thick” conception of the content of perceptual experience. For the flow of such experience seems to incorporate what we would normally take to be inference or reasoning. The question is where the McDowellian is to draw the line between perceptual experience and reasoning on the basis of perceptual experience. What must obviously be avoided by the McDowellian is the assumption that, whenever we wish to ascribe different beliefs to A and B in a given stimulus situation, we must analyze this in terms of identical perceptual experience and different inferences based on that shared experience. For this will lead us to a “thin” conception of the contents of perceptual experience of the sort endorsed in traditional empiricist accounts. I obviously cannot address these questions here. But I think that a defense of the accounts of cinematic experience pressed against Currie by Allen and Lopes requires that they be addressed eventually. To put this point another way, I do not think such an account of cinematic experience can be adequately defended either by seeking to incorporate it into the sort of cognitivist model of perception endorsed by Currie (Lopes’s strategy), or by arguing that our seeing the sorts of fictionalized states of affairs that we imagine to be the case in our engagements with fictional films is established by merely pointing to the fact that this is the way we talk about such things in ordinary parlance (Allen’s more recent strategy). In this chapter, I have defended the spirit of the objections leveled against Currie’s model of film experience by critics such as Allen and Lopes, while suggesting that we can take proper account of these objections only if we relate them more carefully to the different elements in Currie’s model. More specifically, I have defended Lopes’s idea that our experiential engagement with cinematic narratives must be characterized in terms of what he calls “experience as of”. I have further argued, however, that this element in film experience cannot be accommodated by Currie’s cognitivist model in the manner suggested by Lopes. If our imaginative engagement with cinematic fictions involves a form of visual rather than propositional imagining, the contents of such imagining should be seen as the contents of our perceptual engagement with what is presented on the screen, in which case we must reject Currie’s model of such engagement (stage II). Furthermore, so I have argued, Currie’s cognitivist model of our perceptual experience of film does not allow for the sort of conceptualized experience that is

THE IMAGED, THE IMAGINED, THE IMAGINARY 245

involved in “cinematic experience” on Lopes’s analysis. An adequate model of film experience, then, must draw upon the richer models of perceptual experience proposed by writers such as McDowell. Acknowledgment I wish to thank Karen Bardsley, Berys Gaut, Rob Hopkins, and especially Dom Lopes, for very helpful comments on earlier drafts of this chapter.

14 FILM AND THE TRANSCENDENTAL IMAGINATION Kant and Hitchcock’s The Lady Vanishes Melissa Zinkin

Cinema…is able to record time in outward and visible signs, recognizable to the feelings. And so time becomes the very foundation of cinema as sound is in music, color in painting, character in drama. (Andrey Tarkovsky, Sculpting in Time) Recently, in philosophical analyses of film, there has been much discussion of the imagination. Philosophers, such as Richard Allen, Gregory Curry and Murray Smith, all make use of a theory of the imagination to explain, in varying ways, our relation to the objects depicted in moving pictures projected onto a screen. These philosophical approaches tend to refer to findings from cognitive psychology and are often presented as alternatives to psychoanalytic approaches to film in which the film is supposed to be an “imaginary signifier” that acts as the viewer’s unconscious. The imagination does seem to be a handy term for explaining how the mind works when we watch a film since it connotes a faculty we use when we relate to objects not as they are in themselves, but as images that provoke our fancy and require an aesthetic stance. However, although references to the imagination as what enables us to put ourselves in the place of another (Smith 1997), or as what enables us to think that we are seeing something real (Allen 1997), or as the unconscious faculty that is put into play when experiencing a film, do in fact articulate functions of a particular form of cognition, these uses of the imagination are not specific to watching films. Our capability to put ourselves in the place of another or to have our unconscious put into play is what makes fiction in general effective, and our ability to imagine seeing something is also used in painting and photography. There is, however, a theory of the imagination that can explain the particular way the mind works with regard to the moving image. This is Kant’s theory of the transcendental imagination in the Critique of Pure Reason. Unlike the theories used by contemporary philosophers of film, Kant’s theory has the distinction of describing the imagination as the faculty that determines time. Even more strictly than music, film requires the control of time. A certain film will

FILM AND THE TRANSCENDENTAL IMAGINATION 247

always take the same time, and although what a viewer sees in a shot can vary, how long the viewer looks at the shot cannot. Nor can the order in which the shots are seen. A film exists only in a fixed form of time. This is different from other art forms, all of which can be coherently presented in different amounts of time. Since what is essential to film is its control of time and this is one of the factors that distinguish it from other art forms, Kant’s theory of the imagination is well suited to explain what is specific to the experience produced by film. I believe that Kant’s theory of the imagination can be used as an alternative to both psychological and psychoanalytic theories of film. Such a way of understanding film is desirable since it enables us to describe the objective properties of cinema that create meaning regardless of any psychology the viewer may or may not possess. Kant’s theory of the imagination can do this because it considers the imagination to be a transcendental faculty of the mind. By “transcendental”, Kant means an a priori faculty of the mind that makes cognition possible. For Kant, if experience is to be possible at all, we must possess those transcendental faculties that provide the basis for this possibility. What is transcendental is therefore a necessary constituent of experience. It is contrasted with what is empirical or psychological in our experience and hence merely contingent. Kant articulates this distinction in the Critique of Pure Reason. He writes, The imagination is…a faculty for determining the sensibility a priori… in accordance with the categories…. Now insofar as the imagination is spontaneity, I also occasionally call it the productive imagination, and thereby distinguish it from the reproductive imagination, whose synthesis is subject solely to empirical laws, namely those of association, and therefore contributes nothing to the explanation of the possibility of cognition a priori, and on that account belongs not in transcendental philosophy but in psychology. (Kant 1997:B152) The “productive”, or transcendental imagination for Kant is thus what makes possible our sensible experience in the first place. It does this by creating, in accordance with the a priori rules, or categories, of thought, the original temporal forms that structure experience. It is not the faculty we use when we happen to make certain associations between the various sensations we receive. That faculty is what Kant calls the empirical, or reproductive, imagination. In this chapter, I argue that by looking at film with regard to Kant’s theory of the transcendental imagination we can explain our meaningful experience of film as what is made possible by specific temporal forms. In order to do this, I will have to show that unlike in the Critique of Pure Reason, where the imagination is guided by the “categories”, which are a priori rules for the construction of the objects of our knowledge, in the case of film, the imagination is guided by different rules that make possible our experience of the particular objects of film.

248 MELISSA ZINKIN

In what follows, I first explain Kant’s theory of the imagination and how it accounts for the possibility of film as a form of art. I then use Alfred Hitchcock’s film, The Lady Vanishes (1938), as an example of the imagination at work in constructing a film by focusing on some of the temporal forms that are present in the film; music, suspense and metaphor. I conclude with some brief comments concerning the role of the transcendental imagination with regard to historical time and the possibility of interpreting film with regard to its historical context. Kant’s theory of the imagination Suppose you are a film editor and you find on your editing table two separate frames. One is of a ship upstream. The other is of the same ship downstream. In what order do you splice these frames together? This, of course, is Kant’s famous example from the Critique of Pure Reason where he argues against Hume’s skepticism about the existence of causal relations. For the skeptical Hume, “all events seem entirely loose and separate”; as if they were pieces of film randomly spliced together. “They seem conjoined, but never connected” (Hume 1993:49). Kant wants to argue that there are necessary connections and causal relations in our experience, and that Hume is wrong. Kant’s point is that for it to be possible to have any experience at all, for things even to seem conjoined to us, they must first be organized by forms of thought that are necessary and a priori. So Hume, in Kant’s view, just by saying that he has experience of events in the first place, has in fact already presupposed the necessary connection in these events. Experience must also take place in time. For Kant, representations must first be arranged in time, or given a temporal form, in order to constitute a coherent experience. It is the imagination in its transcendental function that provides us with the original forms of time that make possible a unified experience out of the haphazard impressions our senses receive. Kant explains the workings of the imagination in the following passage: Now it is obvious that if I draw a line in thought, or think of the time from one noon to the next, or even want to represent a certain number to myself, I must necessarily first grasp one of these manifold representations after another in my thoughts. But if I were always to lose the preceding representations (the first parts of the line, the preceding parts of time, or the successively represented units) from my thoughts and not reproduce them when I proceed to the following ones, then no whole representation and none of the previously mentioned thoughts, not even the purest and most fundamental representations of space and time, could ever arise. (Kant 1997:A102) For Kant, our sense impressions of the outside world are isolated perceptions, appearing to the mind one after the other. It is as if our mind were a camera taking successive snapshots. Kant argues that if there is no additional

FILM AND THE TRANSCENDENTAL IMAGINATION 249

mental faculty to retain these snapshots and reproduce them next to the current one, for all we would know, we would be seeing the same thing over and over; the same point, for example, and never see that it was part of a line. We would have just one representation before the mind and not be able to think something about it and consider it as a part of a coherent object or event. Kant describes the imagination as the “faculty for representing an object even without its presence in intuition” (Kant 1997:B151). It is thus what enables us to recall or reproduce a representation and place it next to the present one so that we can construct a whole object out of many successively apprehended perceptions. In order for it to be possible to make these coherent objects, the imagination, in its transcendental function, must first create a figure or “schema” of time in which representations must be placed in order to be thought of as connected in the first place. It does this in accordance with the basic rules, or a priori categories, of thought. Take, for example, Kant’s schema of time for the category of causality. What must time be like, in order for someone to be able to make the judgment, “if I do one thing, then something else will happen”? Kant’s answer is that time must be viewed as successive. So succession is the schema of time according to the category of causality. Or, what must time be like in order to make the judgment that something is a substance that remains when its contingent attributes change? Time must be viewed as persisting. Thus persistence is the schema of time according to the category of substance (Kant 1997:A137/ B176–A147/B187). Kant’s point against Hume is that if we are to perceive a succession of events, if, for example, we are to order our representations successively in time so that the ship downstream comes after the ship upstream, this can only be done under the guidance of the rule of cause and effect. Thus, against Hume, there must be a necessary connection, or rule, governing the events we experience or else we would not be able to experience them as temporal events, or experience them at all, in the first place. The imagination is thus what creates the unified temporal forms in which experience is possible. A Kantian theory of film I will now argue that films can be seen as products of the imagination, understood in the Kantian sense—as a series of representations that are organized in time according to rules. By looking at film in this way, I want to show how temporal forms make possible the meaningful images that occur in film. The imagination that I am referring to is not the particular imagination of the viewer who watches, or receives, the film. Indeed, Kant calls the transcendental imagination “productive”, since it is what produces a unified appearance, or image, in a specific schema of time (Kant 1997:A123). Nor is it the particular, subjective, imagination of whomever makes the film. Along with Kant, I am not referring to a form of the imagination that is part of our subjective psychology, what, in the commonplace sense of the word, creates our whimsical fantasies. I therefore do not mean what Dr Hartz, in The Lady Vanishes, means by the word when he says

250 MELISSA ZINKIN

that Iris must be imagining things. Instead, it is a universal ability that works according to rules in order to make any temporal experience possible. But, someone might object: “when the imagination is guided by rules, for Kant, this is in order to create coherent objects of knowledge, and hence avoid Humean skepticism. But here we are not considering film as an object of knowledge, but as a form of art. According to Kant, only when the imagination is free from guidance by rules and is at play, does it have an aesthetic relation to an object. It is thus incorrect to refer to rules of the imagination when trying to explain our aesthetic relation to an object” (Kant 2000:241). My response to this is that what I want to show, with regard to film, is that there can in fact be rules to this play that are different from the rules of thought that Kant explicates in the Critique of Pure Reason. In this way, my understanding of film with regard to Kant’s theory of the imagination is consistent with Kant’s view of the imagination in both the first and third Critiques. In fact, in the third Critique, Kant never says that the imagination acts in an irregular or arbitrary manner when it is involved in producing a judgment of taste. He states only that the imagination is not subject to the particular rules of the understanding, as it is in the first Critique where the categories determine how the imagination is to schematize time. Thus Kant writes that, in a judgment of taste, the imagination is “free and yet lawful by itself” and “lawful without a law” (Kant 2000:241). However, despite the fact that (or rather because) Kant’s theory of aesthetic judgment does not include a theory of time, it is worthwhile to ask what schemas of time the imagination would create in its freedom from the laws of the understanding. The answer to this question, I suggest, would be a Kantian theory of film; a theory of the forms of time created by the imagination according to rules other than those articulated in Kant’s theory of knowledge. What follows should therefore be understood not only as an application of Kant’s philosophy to the study of film, but also as an interpretation of Kant that addresses the question of the nature of time in his aesthetic theory. In the Critique of Pure Reason, Kant’s notion of experience is restricted by a particular conception of time. This is time as what can be represented by a continuous straight line (Kant 1997:B154, A169/B212). For Kant, therefore, time cannot be suspended, nor can different moments of time, each of which are defined by the distinct representations they contain, overlap (cf. Kant 1997:A99). But perhaps, with regard to the experience presented in a cinematic work of art, there are a priori concepts other than the categories that guide the imagination to create other images of time in addition to those that Kant lists in the Critique of Pure Reason. These could be categories that make open, interrogative judgments or metaphorical judgments possible, for example. The form of time in which interrogative judgments would be possible would be one in which time is suspended; the “line” of time is broken and a moment is left hanging in isolation. The form of time in which a metaphorical judgment would be possible, such as one that states that the wheels of a train are the wheels of thought, would be one

FILM AND THE TRANSCENDENTAL IMAGINATION 251

in which two moments of time overlap in order to construct a new object different from what is represented in each moment. Such filmic schemas of time are revealed during those moments in a film when the Kantian cognitive form of the imagination is in a crisis with regard to the regular line of time and the standard rules of thought. The imagination is unable to follow the normative rules of thought in order to construct the figure of a recognizable real object. Free from these rules, however, the imagination comes up with forms of its own. As with many of Hitchcock’s films, The Lady Vanishes highlights and comments on problems specific to the medium of film itself. I believe that this film can be understood to be commenting on the very problems that can arise with regard to the cognitive function of the imagination. The resolution of such problems is indicated within the form of the film medium itself. In my discussion of The Lady Vanishes, I will therefore refer to the drama that is presented in the film to illustrate the philosophical problems related to the imagination. It should be kept in mind, however, that I am using this film as an example of general processes that are at work in all movies. These are the processes by which imaginative forms of time make it possible for film to have meaning in the first place. I should also note that, although discussions of time in film are usually analyses of film narrative, this is not my concern here. A narrative is “a chain of events in cause-effect relationship occurring in time and space” (Bordwell and Thompson 1990:55). The study of narrative in film involves figuring out how the images presented on the screen in a certain order work to tell a story whose content often must be inferred from these images. In this way, the study of narrative is about time because it concerns how the events that happen on screen in a certain order and length of time can represent or refer to events that happen in a different order and duration of time. The study of narrative in film therefore considers film with regard to how it represents time. In my discussion of time in film, however, I am not concerned with the representation of temporal order and duration, but with how particular forms of time in film can create objects or events that have certain aesthetic properties that produce meaning such as the abstract or explicit meanings, symbolic or implicit meanings, or symptomatic or repressed meanings, with which an event is attributed (Bordwell 1989:8–9). To the extent that most films are narrative films, the forms of time that I am discussing will occur in narrative films. Nevertheless, I am not discussing the forms of time that make the chain of events in a narrative possible, but those that make these events capable of providing us with aesthetic experience. The following analysis of The Lady Vanishes will not be quite a “transcendental deduction” of the categories that ground temporal forms in film. But it will show that underlying both our experience of suspense and the jarring and intensified experience of unusual metaphoric images in film, there are particular temporal forms that make these experiences possible.

252 MELISSA ZINKIN

The Lady Vanishes The Lady Vanishes is the story of Iris Henderson who meets Miss Froy, an old governess and music teacher, at a Tyrolean inn and shares a train with her on the way to London. When Iris wakes up from a nap, her companion has vanished. No one on the train seems to have witnessed Miss Froy. A brain surgeon on the train, Dr Hartz, tries to convince Iris that she is “imagining things” due to having been knocked on the head at the train station by a falling flowerpot. Finally, Iris receives support from the charming yet flip Gilbert, a musicologist, and together they discover that Miss Froy is concealed under bandages as Dr Hartz’s patient. Miss Froy reveals to Iris and Gilbert that she is a spy trying to smuggle back to London the secret clause of a treaty encoded in a piece of music. She entrusts this tune to Gilbert in case anything should happen to her, but ultimately she makes it safely back to London. This film is about memory and time. The main action of the film, which takes place on a train, is framed by scenes that involve music, especially the tune that is the secret code. In these scenes, music, which according to Kant expresses not concepts, but “aesthetic ideas of a coherent whole of an unutterable fullness of thought” (Kant 2000:329), presents a challenge to the Kantian view of imagination and memory. For Kant, the transcendental imagination has to do with placing mental representations in a temporal order according to rules, not with rhythm and harmony, the variations in duration in which representations occur in time. How, then, does one commit something to memory for which there is no original concept, or code, by means of which one can recall it again for later use? At the beginning of the film, we see Miss Froy at her window in the Tyrolean inn trying to memorize the tune played below by tapping her hand at regular intervals and thus submitting the tune to the measure of time, encoding each moment with a phrase of notes and abstracting the general rhythm of the melody, so that the imagination will have a rule by which to recall this piece of music. But even this is difficult, since other music is being played at the same time in the room above. In The Lady Vanishes, not only is there a code in the music within the film, but the code, or rule of the film itself, is music. Indeed, the connection between time and music is illustrated at the beginning of the film by the singing cuckoo clock. By placing the code of the secret treaty in music, the film indicates that it is possible to have rules embedded in the play of time. This can be seen in film’s capacity to play with the regular structure of time and still have these irregular forms of time have meaning and not be mere arbitrary fantasy. Such rhythms are produced by the varying lengths of time in which images are shown on the screen as well as the inter-cutting or repetition of images. Such rhythm does not necessarily relate the events in a story, but it does indicate the degree of tension and significance with which shots are to be attributed. In the final scene of The Lady Vanishes, in the government office in London, Gilbert is unable to remember the tune because he presently has another in his mind—the wedding march, since he and Iris have just become engaged. Here,

FILM AND THE TRANSCENDENTAL IMAGINATION 253

the problem concerns how to recall something when something else is already occupying one’s mental attention. This is a problem that Kant’s theory of the imagination is unable to handle. For Kant, the imagination serves to combine different representations into a unified object. But two songs cannot be so combined, one thinks of one or the other. For Kant, the imagination cannot superimpose one form of unity upon another—especially one temporal structure, such as a song, upon another song. Hence it is impossible for Gilbert to recall one tune with another already in his mind. Gilbert panics, his imagination is in crisis, but the tension is resolved when the tune is heard in the background (mirroring the beginning of the film in which the music plays over the opening credits) and Gilbert discovers that Miss Froy has made it safely back to London and that it is she who is playing the tune in the next room. This tune carries over to the final credits of the film, indicating that film, as a form that can play with time as can music, can resolve crises that occur when the imagination works with a straight line of time. Bounded by the scenes with music is the train ride during which most of the film takes place. The train is like the strip of film and the train of thought. These require respectively tracks, gates, or the intuition of time for their movement. Just as, in Kant’s philosophy, our experience is limited by the linear form of time, the movement of a train is limited by the tracks on which it runs, and a film is limited by the strip of celluloid that runs through the projector. In The Lady Vanishes, the train is made to represent the destructiveness of the passage of time and the fragility of memory. As what hurtles towards its end without stopping or taking the time to recollect anything it might have left behind, the train becomes the vehicle that makes it possible for Miss Froy to vanish into the past, and persist only in Iris’s memory. As what must be kept on time, the train is what ensures that Iris’s recollection of Miss Froy not be verified; the British cricket fans lie and say they never saw Miss Froy in order to prevent a delay of the train. And, as the passage of time in general, the train represents what makes Miss Froy forgettable to the passengers on the train, her gray old age. How does the film The Lady Vanishes work within the limits set by the line of time on which each moment must ever recede into the past? In fact, it is only on the train that Iris’s problems occur. Miss Froy vanishes. Is Iris’s memory of Miss Froy of a real object, or of a subjective fantasy? How is it possible to decide one way or the other? Indeed, it is hard to know whether an isolated representation of something is real or not when it is not connected with anything else in one’s experience. One can only know whether what one remembers is real if it can be connected with something else. According to Kant, this connection is achieved when the imagination connects representations according to a rule. For Iris, her experience of Miss Froy is isolated—especially since no one whom she asks seems to have shared her experience. This puts her in a state of great agitation as she strives to connect her memory of Miss Froy with something present—hopefully Miss Froy herself. Here, again, the film reveals a crisis of the imagination. Just as the trees that the train passes by recede irretrievably into the background, and the

254 MELISSA ZINKIN

earlier scenes of the film are irretrievably wound back up on the film reel, what was once present on the train is now gone. Without the imagination to recall this image and guarantee its persistence in time, and hence the substantiality of Miss Froy, her reality is in doubt and is assumed by others to be a product of Iris’s fantasy or particular subjective imagining. Iris becomes an easy subject for the dubious psychoanalytic theories of the evil Dr Hartz. Similarly, since there is no real object that persists “behind” the fleeting objects projected onto the screen, the way one recalls past shots is subject to the play of the imagination. A film must therefore have its own rules for making its images memorable so that they persist as significant for the film even when they are no longer present on the screen. Hitchcock excelled at this with his creation of MacGuffins. These are ordinary objects endowed with vital importance within the film often by means of visual tricks, such as illuminating the glass of milk in Suspicion, but also by means of having the camera actually linger just a second longer on such objects, as with the money in the newspaper in Psycho or even the playing of the tune in The Lady Vanishes (Truffaut 1984:138, 143). In Iris’s search for evidence that would confirm the objectivity of Miss Froy, her imagination gropes to find some representation that could be connected with what she remembers. It is this searching of the imagination that ultimately makes the film suspenseful. But the film becomes suspenseful only to the viewers. The characters in the film are merely in a state of agitation. This points to the difference between the story narrated by the film and the temporal forms created by the film. It is only when the images in the film are displayed in a particular way to the viewer that they become suspenseful. I believe that The Lady Vanishes becomes suspenseful at a very specific moment. This is when a representation of Miss Froy appears on the screen and we are waiting for Iris and Gilbert to become conscious of it. This representation is Miss Froy’s name, which she had previously spelled out for Iris in the condensation on the train window. Seeing this now would surely prove that Iris’s memory is correct. What makes this scene suspenseful is the typically Hitchcockian element of suspense. The audience knows something that the characters in the film do not know, but that they would want to know, once such a fact became available. The experience of suspense, which takes the form of the future interrogative judgment, “will they see the writing”? requires that this object of suspense be taken out of, or suspended from, the regular line of time, and hover just above it. However, the question is not merely “will they see the writing”? but “will they see it in time”? before their meal is finished and they leave the dining carriage. The suspenseful object is thus determined as something that exists in a rapidly contracting form of time. Suspense is therefore the “bending” of time by the imagination as it strives to connect an event that it represents as occurring in the future with one that is represented as present. The suspension of the object is due to its being seized from its context of significance that is in the future and made to hover over the present, thus contracting the line of time. Even if the suspenseful object is one of

FILM AND THE TRANSCENDENTAL IMAGINATION 255

dread, like a bomb under the table that is about to go off, but which one hopes will be deactivated, suspense still takes the same form of an imagined future event. Noël Carroll has defined suspense, as I have, with regard to a film’s setting up a question whose answer is anticipated. But in order to distinguish suspense from mere anticipation, Carroll explains that the suspenseful situation has psychological urgency, and this is because the outcome of the anticipated event has moral value. According to Carroll, suspense occurs either when evil is the likely outcome of a situation or when what is morally good is the unlikely outcome of a situation (Carroll 1996a:101). Perhaps this is right, although there are quite a few counterexamples to Carroll’s theory (cf. Knight and McKnight 1999:107–23). My aim, however, is to emphasize the temporal form of suspense, which Carroll does not, but, incidentally, Hitchcock did, as the necessary condition that makes possible any psychological attitude towards the film that might also result. Suspense is what happens when the imagination tries to create a form of time in which an event can happen sooner, by imagining a future object or event and trying to connect it with the present one. In the case of the bomb that is expected to go off, suspense is felt not with regard to the bomb’s explosion, but with regard to the actions that would prevent the explosion, which are hoped to happen very soon. What is the schema of time created by the imagination that makes it possible for such a suspended object to be thought? This would have to be a form of time that makes the future hover over the present. In such a form of time two different moments would exist next to each other, but not be continuous. Not only does Hitchcock achieve this temporal form by means of his presentation of shots; he also represents this form with a spatial image within a shot by having the word “Froy” appear from above as Gilbert pulls down the window. This word, written in the condensation of the window, then hovers between Gilbert and Iris as they flirt. For the whole scene, this object, even, and especially, when it is not pictured on the screen, is associated with the anticipation that both Gilbert and Iris will notice it. The climax, when Iris finally notices the writing, is a highly compressed moment of suspense. The word disappears and appears again on the window, as if it is not written on the train and moving along with it, but were impressed on it from outside. This represents the attempt to impose onto the present a time outside of the regular flow of time. The train and the straight line of time prevail, however, and the writing disappears when the temperature changes as the train moves through the tunnel. The imagination has failed to present to Iris the object of her memory. Nevertheless, hope does arrive in the next scene when the cook throws garbage from the train. The tea carton that Miss Froy had previously given to a waiter, and that Iris has mentioned to Gilbert, falls back upon the window. The smudge that remains signifies that a connection has actually been made, and Gilbert is convinced that Iris is telling the truth. Another form of time that occurs in film and that indicates that the imagination is working free from the rules that Kant presents in the Critique of Pure Reason is what I will call film metaphor. A metaphor can be generally defined as the transference of one concept onto another in order to create a new concept or to

256 MELISSA ZINKIN

transform one by means of the other. Along with Noël Carroll, I understand a “film metaphor” to be what occurs when disparate elements “are visually incorporated or amalgamated into one spatially bounded homogeneous entity. Elements are fused into a composite, but nevertheless self-identifiable, construct thereby visually indicating that these elements are elements of the same entity” (Carroll 1996a:213). If we see film as the product of the imagination in the Kantian sense by means of which representations are placed in a temporal order according to rules in order to produce a coherent form of experience, then filmic metaphors are those in which a representation is superimposed upon another in order to alter experience. The form of time here is one in which the past overlaps with the present. After Iris has accepted defeat in her attempt to prove the existence of Miss Froy and slumps in her seat in resignation, the train compartment becomes suffused with the remembered image of Miss Froy. The image of Miss Froy merges with that of each passenger in the compartment. Miss Froy is the contents of the train. Indeed, she is also the content of the film. She is an image isolated on the track of time and hence left behind—as the train in the film itself is left behind when Dr Hartz severs the compartments. Like film itself, which does not necessarily represent any real object, the concept of Miss Froy within the film cannot be connected with any single real object. Hence, her image can be projected onto anything. The crisis of the imagination that occurs when what I am calling a filmic metaphor is produced, happens when the imagination tries to recall something for which there is no complete schema, or rule, for recollection. Indeed, when Iris does produce a schema of Miss Froy by stating her physical characteristics; her oatmeal suit, her blue scarf, and so on, the image that is presented to her is still not the real Miss Froy, but Miss Kummel instead. In such situations where there is no complete rule by means of which we can recall what we want (or even really know what we want to recall in the first place), the imagination can only think of this particular thing as something else, by connecting it with something it in some way resembles. The overlapping, or metaphor, that results, is then also unique. It is the only way this memory can be articulated. The rule for this attachment of one image to another is thus not that of the concept of the object. Instead it is the reason for why the object is memorable in the first place (cf. Kant 2000:352). Miss Froy is memorable precisely because the memory of her is so dubious and fleeting. She is thus recalled as part of the train. As the schema for suspense is the temporal form of a future suspended over the present, so the schema of film metaphor is the present suffused with the past. Historical time Another form of time that is presented in film is historical time. Indeed, often the reason why a film is memorable is its relation to historical events. What makes something memorable is how it is marked by the form of time in which it first becomes a possible object of thought. For Kant, the rules for constructing an

FILM AND THE TRANSCENDENTAL IMAGINATION 257

object in time are indeed the same rules for remembering the object. We remember empirical objects because we have attained concepts of them by means of the basic categories of thought, which are the rules that guide the transcendental imagination in constructing an object in time. I have argued that images that appear in film are memorable because they are made to stand out from, or disrupt, the regular flow of time that one experiences when watching a film. But moments in a film, or even the entire film itself, can also be memorable with respect to how it appears with regard to the background of the historical moment in which it was made. Again, I will use The Lady Vanishes as my example. The Lady Vanishes was made in 1937 and released in 1938. 1937 was the year in which Chamberlain met with Hitler in Munich and agreed that Germany would occupy the Sudetenland. The film in which Iris struggles to get help for the vanished Miss Froy can easily be interpreted as an argument for British involvement in international affairs. The film begins with an avalanche, which alludes to Chamberlain’s speech in 1937 in which he warned that an “incautious move or loud exclamation may lead to an avalanche” (Fuchser 1982:81). The “pacifist” in the film has the ominous name of Todhunter (death hunter in German). Headlines referring to the cricket match read “England on the brink”, and the connotation is clear. The Lady Vanishes can thus be interpreted as implicitly being about the mood of England at a specific historical moment. The mood is one of suspense with regard to what Hitler will do next and hope that intervention will not be necessary. This hope is determined primarily by the memory of the British people of the devastation of the First World War—and it is this memory that affected Chamberlain’s decisions and were the basis for his policy of appeasement. Is there a form of historical time that is produced by the transcendental imagination? This would be a form that determines the possibility not of the experience of objects, but of historical events. As Yirmiahu Yovel puts it, for Kant, “history is the domain in which human action is supposed to create a progressive synthesis between the moral demands of reason and the actual world of experience” (Yovel 1980:6). This particular Kantian view considers history as the time in which the progress of human morality is enacted. But if, for Kant, reason is outside of time and hence outside of the world of experience, how can reason, which gives us the “highest end” of humanity, morality, actually be connected with empirical temporal experience? In the Critique of Judgment, Kant argues that taste enables us to make the transition from the sensible world to the moral world of reason. This is so because in judging what is beautiful, the imagination is free from the rules of the understanding and hence from the rules that limit it to images of what can be empirically known. In this way, the mind becomes capable of thinking of its purpose beyond the world of sensible nature; the morally good. Kant writes, “taste as it were makes possible the transition from sensible charm to the habitual moral interest without too violent a leap by representing the imagination even in its freedom as purposively determinable for the understanding and teaching us to find a free satisfaction in the objects even

258 MELISSA ZINKIN

without any sensible charm” (Kant 2000:354). But what is the temporal form that the imagination creates when it is free from the rules of understanding and is not used for the cognition of natural empirical objects? This would be a form of time that could indicate our relationship to the “highest good”, that is, our place in history. Although Kant himself hardly mentions time in his Critique of Judgment, I suggest we understand the aesthetic form of time, not as the regular line that Kant describes in his Critique of Pure Reason, but as it is presented in film, where moments can overlap, repeat, occur rapidly or slowly, and hang in suspense. In film, then, the imagination can mediate between reason and time. The rules that guide the imagination here are not those that make knowledge possible, but those concerning the moral relationship of a society to a historical situation, be it one of hope or an obsession with memory. In other words, the way that a film deals with time to create forms of possible experience is also what makes it possible to interpret films as historical documents, revealing the relationship to its own time of a particular society. One can think here of the postmodern use of time in Jim Jarmush’s Mystery Train (note the title) or the use of time in Tarkovky’s film about exile, Nostalgia. Indeed, it is through film, I believe, that the imagination can present us with a piece of history. And it is by understanding film as a product of the imagination in the Kantian sense, as containing temporal forms constructed according to rules, that it can first become an object that can be interpreted as having any psychological significance. The rules that guide the imagination in film, however, do not have quite the same status as Kant’s a priori forms of cognition. Since the object they construct is not the object of natural science with which Kant was concerned, but is instead a cultural product, the rules the imagination uses are not transcendental in the strong metaphysical sense of what must hold true in all places and times. Instead, with regard to film, they are the rules that make possible the particular experience of each culture. In this way, the imagination is a universal and necessary condition for the possibility of experience in general, but the rules it uses and hence the forms of time it constructs can differ from culture to culture. By looking at cinema with regard to Kant’s theory of the imagination, it is possible to avoid the psychological arguments intrinsic to both the analytic cognitive and psychoanalytic approaches to cinema. The empirical psychological approach has led Gregory Currie, to write, “we need an account of what is sometimes called the proper functioning of the imagination: a biologically oriented theory that explains the adaptive benefit we gain from having the capacity to imagine” (Currie 1995a:142). To my mind, this scientific approach is a good way to kill whatever is special about film and the aesthetic experience it provokes. Noël Carroll has argued that cognitive theories are better explanations of film than are psychoanalytic theories because their explanations presume that the viewer is responding to a film in a rational way and that there is a rational explanation for the way we experience the cinema (Carroll 1996b:65). According to Carroll, it is only when a rational explanation cannot be provided that we need a theory to explain what now presents itself as irrational behavior; this is when we

FILM AND THE TRANSCENDENTAL IMAGINATION 259

turn to psychoanalytic theory. Yet, it is also possible that what has to do with art and our aesthetic response to it is precisely what cannot be explained rationally, or at least with regard to our conscious beliefs and desires. In this way, Carroll is begging the question of whether cognitive theories really are best suited to explaining our response to the cinema. Psychoanalytic approaches to film, however, are often too ungrounded. Although many of these interpretations are fascinating, they rely on the assumption that the particular aesthetic element in film is based on its relation to the unconscious and the fantasies it produces. Since it is hard to prove that there really is any unconscious faculty that governs the production of these fantasies, psychoanalytic theories, as theories of film, tend to lack credibility. Kant’s theory of the imagination, I believe, can provide the basis for an approach to film that gives us general rules for explaining what makes film meaningful as an art form and also helps us to understand the particular meaning (or range of meanings) of a film with regard to its historical context. There are also historical reasons that justify using Kant’s theory of the imagination to understand the cinema, and Hitchcock’s The Lady Vanishes in particular. Even though Kant died in 1804, his philosophy was very much alive at the beginning of the twentieth century when film itself was coming into its own as a form of art. In the 1930s what we now call the split between analytic, or Anglo-American, and continental traditions in philosophy, had just begun to solidify. This was the division between the logical-scientific view of philosophy represented by Carnap and the existential-historical view represented by Heidegger (see Friedman 2000:129ff). At the center of this divide was the issue of the correct interpretation of Kant, especially with regard to his distinction between sensible and intellectual faculties—precisely the distinction that, according to Kant, was to be mediated by the faculty of the imagination. Carnap, and the analytic tradition that followed, focused exclusively on Kant’s view of logic and his analysis of the intellectual faculty. They dismissed the role of Kant’s sensible faculties and consequently the imagination that was supposed to bridge the gap between the understanding and sensible intuition. Heidegger, on the other hand, saw Kant’s philosophy as essentially that of the imagination, and hence denied the delicate balance between sensibility, thought and the imagination that Kant articulated. At issue, generally stated, was whether philosophy was to be a discipline that would ensure the objectivity of the sciences or whether it was to be a discipline that could articulate the subject’s experience of the world. My suggestion is that since philosophy in England in the 1930s was defined exclusively with concerns of logic and science, the study of the imagination had to find its home elsewhere. This was in the newly emerging art form of film that was produced by directors such as Hitchcock. Indeed, now that analytic philosophers are currently beginning to address film, they reveal their roots by their neglect of the Kantian theory of the imagination.

15 THE FUNERARY SADNESS OF MAHLER’S MUSIC Saam Trivedi

We often say of musical passages and works that they are sad, or that they possess other emotions and other mental states. For instance, the opening passages of Mahler’s Seventh Symphony, with the soft orchestral chords and tenor horn solo, would, I guess, be described by most competent listeners as sad, even funerary. But how can it be true that something inanimate such as music is sad? This I take it is, essentially, the problem of musical expressiveness. (Compare Kivy 1989: 6; Davies 1994: x; Matravers 1998:3, 102. Matravers suggests the problem involves finding the link between applying emotion-terms to artworks and applying them to people; such a link can be found in my view for we imagine in various ways that sad music is sad, in part because its structural and other properties resemble sad people in many ways.) And, right away, we can set aside the claim seemingly made by some that it is literally true to say that some musical passages are sad, no matter how this literalist claim is further fleshed out, whether in terms of music resembling our expressive behavior, or in some other terms (Davies 1994:162–3; Carroll 1999:95–105). For it seems odd to think we can literally attribute emotions and other mental states such as moods and feelings to inanimate things such as music. To say music literally is sad seems to imply that music literally or really has mental states such as sadness, which is just false. In what follows, I explore the possible solution to the problem of musical expressiveness that listeners imagine in various not always highly conscious ways that sad music is sad; that it is true imaginarily (or imagined it is true) that some music is sad. If I am right, then the music is not really or even metaphorically sad but rather it is only imagined to be (the kind of thing that is) sad, so that while vibrations in the air may be primary qualities or objects, and sounds secondary qualities or objects, expressiveness may be more like an imagined or a projected or perhaps even a tertiary property of the music (see Scruton 1997:93–4, 160–2, 169– 70), one that is somehow dependent (either in terms of causation or supervenience or emergence or in some other way) on the music’s structural properties in a way that remains to be clarified (see Davies 1994:256). But first I indicate very briefly why I am not persuaded by two better-established and currently quite popular alternatives to the imagination-based solution I favor.

THE FUNERARY SADNESS OF MAHLER’S MUSIC 261

Alternative views: metaphorism and arousalism One reason why Anglophone analytic philosophers and aestheticians may largely have neglected imagination when it comes to musical expressiveness is the popularity of metaphorism amongst those heavily influenced by the linguistic turn in twentieth-century philosophy (see, e.g. Goodman 1976; Scruton 1997; also compare Matravers 2000:104). By metaphorism, I mean the view that when we say that a piece or a passage of music is sad, we are merely using a metaphor and saying only that the music is metaphorically sad. (I leave aside Goodman’s notion of expression as metaphorical exemplification, which has been extensively criticized in Davies 1994, Matravers 1998, and Carroll 1999, amongst many other places.) Now, metaphors can be paraphrased so that the point of the metaphor can be grasped as expressed, without loss of meaning, by an adequate paraphrase. In particular, it is by means of an adequate paraphrase that one can explain the point of a metaphor to someone who has not grasped it. Donald Davidson has famously denied this view, claiming that metaphorical sentences taken as a whole have only a literal meaning, not a metaphorical meaning that can be paraphrased (Davidson 1984). He claims that taken literally, metaphors are usually false or else trivially true. I cannot adequately discuss Davidson’s view here, though I will mention one problem for that view. Davidson is right, I think, if he means that the individual words that constitute a metaphor do not have a special, metaphorical meaning but rather are used in their ordinary sense and in an imaginative way that points to similarities. For instance, in the metaphor “the spiteful sun”, the word “spiteful” does not acquire a special, metaphorical meaning that it does not have in nonfigurative contexts. However, I disagree with Davidson if he means the metaphor or the metaphorical sentence as a whole does not also have a metaphorical meaning in addition to its literal meaning. The problem for Davidson is that he cannot allow for the possibility that someone might grasp the literal meaning of a metaphor and yet be in the dark about the point of the metaphor, that is, its metaphorical meaning (see Moran 1989; Levinson 2001:10–11). For in Davidson’s view, there is no such thing as metaphorical meaning that one can fail to grasp until that is revealed to one by correct paraphrase. Consider as an example, the metaphorical expression “he is drowned in sorrow”. One possible literal meaning of this, an obviously false one, is that he sank in a large body of liquid called “sorrow”. Someone who does not know English too well might grasp this literal meaning of the metaphor, and yet not have grasped the point of the metaphor. That point of the metaphor is its metaphorical meaning given by the paraphrase that he is extremely sad, so sad that his sadness can be imagined to be like a deep ocean that he is immersed in and unable to come out of, as though he were drowned in an ocean. With that very brief digression into Davidson on metaphor behind us, let us return now to the alleged metaphor “the music is sad”, assuming, contra Davidson, that metaphors can indeed be paraphrased, even if it may be hard to

262 SAAM TRIVEDI

paraphrase some metaphors. Now if the locution “the music is sad” is indeed a metaphor, then we should be able to paraphrase its point, without loss of meaning, to someone who has not understood the point of the metaphor. There are (at least) three possible candidates for adequate paraphrase here: (i) in technical terms of musical analysis, so that to say that the music is sad is to say that the music is in a minor key, uses diminished chords, and so on; (ii) in nontechnical, audible terms used by laypersons so that to say that the music is sad is to say that the music is slow in tempo, low in pitch and volume, and so on; and (iii) a combination of (i) and (ii) so that to say that the music is sad is to say that it is slow, low, in a minor key, and so on. Neither of these three possible candidates for paraphrasing the allegedly metaphorical locution “the music is sad” does so without loss of meaning, leading to doubts whether the locution is metaphorical to begin with. While the locution “the music is sad” says something about the emotion-terms associated with the music, all three possible paraphrases say something very different about the music in talking about the music not in emotional terms but in technical and nontechnical, audible terms. The metaphorist might respond here that this only shows that there is more to the paraphrase of metaphorical expressive judgments than is contained in (i), (ii), and (iii) above; whatever is missing will have emotional terms and can simply be added to the paraphrase, making it complete. But, in reply, it is not clear that there are any other candidates here for possible paraphrase given that, by the law of excluded middle, technical, and nontechnical paraphrases (plus their combination) seem to exhaust all possibilities here. It really is incumbent on the metaphorist to suggest here other possible candidates for paraphrase, and for now at least there are no signs of this challenge being met. Moreover, it is not clear just what this missing element (that can be added on to (i), (ii), or (iii) above) is, nor is it clear that we can simply posit that the alleged missing element will have emotional terms, especially without knowing what it is in the first place. It must be doubted, then, whether expressive locutions such as “the music is sad” are metaphorical, given that they cannot be paraphrased adequately without loss of meaning. There is not enough space here to raise additional worries for metaphorism, and so with this very brief critique of metaphorism behind us, I turn to other possible solutions to the problem of musical expressiveness. Let us now consider arousalism, the popularity of which may also have contributed in part to the topic of imagination and musical expressiveness not being explored fully. At its simplest, arousalism claims that music is expressive of the emotion it arouses in the listener so that to say “the music is sad” is to say that the music arouses or evokes sadness in listeners. Here are some general worries for such simple arousalist theories. First, and perhaps most importantly, these theories conflate musical expressiveness and musical arousal. Expressiveness is a property of the music itself, whereas the aroused response, as an emotional effect that music has upon listeners, is something that belongs to listeners. While expressiveness is often integral to the aesthetic character of a musical work so that we usually praise

THE FUNERARY SADNESS OF MAHLER’S MUSIC 263

a musical work aesthetically for being expressive, arousal is usually accidental and usually not an aesthetic plus; unless we are talking about marches or funeral music or other music for special, rousing occasions (compare Scruton 1997:145). Expressiveness and arousal are distinct as concepts and as phenomena, even if they are often coextensive in that we are often aroused by expressive music. Second, just as we can often perceive other people as sad without necessarily being aroused emotionally to sadness or any other emotion or mental state, likewise a musical passage or piece may be heard as expressive of a certain emotion and yet not arouse any emotion even in appropriately backgrounded, attentive listeners who are not bored or tired or excessively familiar with the music. This may be because the music may not do a very good job of being expressive. Third, arousalism reverses the order of things (see Davies 1994: 199; Matravers 1998: 117) in claiming that sad music is sad because it (sometimes) arouses sadness in listeners, rather than claiming that sad music arouses sadness in us because it is (imagined to be) sad and we imaginatively empathize or sympathize or identify with it (or perhaps with an imagined musical persona in it), thus being aroused to sadness. There is, I would urge, a partial analogy between (hearing) sad music and (seeing) sad persons, given which, this implausible reversal of the order of explanation is like claiming that sad people are sad because they sometimes arouse sadness in us, rather than claiming that sad people arouse sadness in us because they are sad and we empathize or sympathize or identify with them, thus being aroused to sadness. Indeed, one might wonder if, unlike the alternative explanation just sketched, arousalists can explain at all why we are often aroused by music, but I shall not pursue the point here. The worries raised so far for arousalism arise for what one might call strong arousalism. In addition to strong arousalism, however, there are “weak” or moderate arousalist theories, which might be thought to escape at least some of these worries (see Trivedi 2000 for discussion of the moderate arousalism in Ridley 1995). One kind of moderate arousalism claims that expressive judgments are about the capacity of artworks to cause certain sorts of experiences involving certain emotional effects or arousal (Matravers 1998). Sad artworks are sad because part of the experience they cause in arousing feelings is similar to the characteristic experience we would have were we to see sad people; music is sad if it has the capacity to be experienced as sad through arousal. A second component of this brand of moderate arousalism is the claim that to say that music is sad (or expressive of sadness) is to say that it is appropriate to feel sad when we hear it; that music expresses that particular emotion to which our aroused feeling would be an appropriate response when, as qualified listeners, we hear the music under normal conditions. This brand of moderate arousalism faces some problems, in addition to the problems discussed earlier of conflating arousal and expressiveness, and reversing the direction of explanation. For starters, it is not clear that to say that a musical work or passage is sad is to say (just) that it has the capacity to cause sadness in

264 SAAM TRIVEDI

listeners and thus be experienced as sad; even though it may well have that capacity. This is analogous to the case of sad persons: to say someone is sad is not to say that they have the capacity to make others sad and thus be seen as sad; though they often have this capacity. Rather, it means that they possess a certain mental state, whether or not they make others sad or are seen as such. Analogously — and this is only a partial analogy—I would urge that to say the music is sad is to say that it is imagined in various ways to have sadness, in part because its structural and other properties resemble sad people in various ways. Second, there are qualms to be raised about the claim that to say music is sad is to say it is appropriate to feel sad when we listen to it. To begin with, pursuing the analogy between sad music and sad people, just as to say someone is sad is not to say that it is appropriate to feel sad when we see them (though it may often be appropriate to feel sad when we see them) but rather that they possess a certain mental state, similarly I submit that to say music is sad is not to claim that it is appropriate to feel sad when we hear the music (though it may often be that) but rather that the music is imagined in some way and at some level to be sad in ways resembling sad people. Moreover, even if it is always appropriate to feel sad (as opposed to it being appropriate sometimes to feel pity or some such emotion instead of sadness) when we hear sad music, we must dig deeper here and ask a fundamental question: why is it appropriate to feel sad when we hear sad music? Analogously—to return to the partial analogy between sad music and sad people— one can ask why it is appropriate to feel sad when we see sad people. The natural answer here, I suggest, is that it is appropriate to feel sad when we see sad people because they are sad, and it is very natural and human to empathize or sympathize or identify with them, thus being aroused to feeling sad. Likewise, one can claim that it is appropriate to feel sad when we hear sad music because it is sad, and it is very natural and human to be aroused to sadness musically when we imaginatively empathize or sympathize or identify with the music itself as animated (in the sense I explain later) or with a musical persona in it. But now this will lead to the further and basic question how something inanimate such as music can itself be said to be sad, a question that this brand of moderate arousalism does not answer in claiming that to say music is sad is to say it is appropriate to feel sad when we hear it; instead, perhaps all we are told is that it is appropriate to feel sad when we hear sad music. In contrast, we are given an answer to this basic question of musical expressiveness when we are told—as I will suggest later—that amongst other things we often animate the music at some level, imagining that it is the sort of thing that is sad.1 Imagination So much for metaphorism and arousalism as alternatives to imagination-based theories of musical expressiveness. Before exploring the issue of how imagination applies to musical expressiveness, we need to explore, very briefly, imagination itself.

THE FUNERARY SADNESS OF MAHLER’S MUSIC 265

What is it to imagine something? What is involved in imagination? Is there one, uniform sense of imagination involved in various imaginative activities? Or are there many different senses of imagination? If there are various senses of imagining, do they all apply to musical expressiveness? If so, do they all apply in the same way or to the same degree, or are there important differences in how they are involved in hearing music as expressive? I cannot address all these questions adequately in what follows, and no doubt there are also other questions that can be raised. But I hope at least to make a start here at addressing some of these questions that are seldom raised and addressed in the work done so far on musical expressiveness by various philosophers. The first notion of imagination relevant to art that comes readily to mind is that imagination involves forming mental images in the sense of visualizing scenes or events or things not present. Such a notion of imagination is to be found in Aristotle, who claimed that imagination involves images being presented to us, and that the soul can never think without a mental image. It is also the notion of imagination favored by the Romantic tradition. Indeed, the very etymology of the word “imagination” (from “imago”, meaning images or pictures) suggests a close link with mental images. Moreover, as Roger Scruton points out (Scruton 1992: 212–3), images have a subjective, phenomenological character; they are like perceptions in having a sensory component; and to describe them we must refer to a perceptual experience associated with them, in terms of what it would be like to have this perception. However, we do not always imagine in the sense of forming mental images when we experience art. In particular, when reading fiction, though we may intermittently and occasionally form images, we are not always visualizing every sentence uttered by one character to another, or visualizing every metaphor. Instead, sometimes when we read fiction, we may imagine in that we suppose something is said by one fictional character to another without actually picturing in our minds how these fictional characters look, what they wear, where they live, and so on. Indeed, there are ways of speaking in ordinary language where when we ask someone to imagine something is true, we ask them to suppose or think it is true without necessarily forming mental pictures. For example, to imagine that the universe is infinite is to suppose or think that this is the case, and it is not clear that one is being asked to form a mental picture of an infinite universe, if one can do so in the first place. There are in fact many activities that are imaginative, as Ryle stressed, and thus many different senses of imagination (Ryle 1949). In addition to imaging, these may include pretending; supposing or assuming or entertaining that some proposition is true without actually believing or affirming it (Scruton 1992:214; 1997: 88–92); thinking of a possibility; acting; impersonating; fancying; having false beliefs or delusions (such as when someone has delusions and imagines she is Queen Victoria); dreaming and daydreaming; making-believe; and so on (New 1999:69– 73; Thomas 2000). These various ways of imagining can have a long or a short duration; be intermittent; be voluntary (in the sense of being under our control)

266 SAAM TRIVEDI

or not so; they can be engaged in while we are involved in other activities (e.g. daydreaming in class that you are a Distinguished Professor); they can be spontaneous or otherwise; and they can be engaged in consciously or otherwise (Walton 1990:13–21). Is there a defining quality common to all and only these various kinds of imaginings? Even if it is true that all varieties of imagining share a family resemblance in involving some sort of contrast with reality (New 1999:71–2), clearly this is at best a necessary but not sufficient condition. For all kinds of things, such as false propositions themselves, involve a contrast with reality without themselves being forms of imagining thereby. Perhaps imagination involves new perspectives, new ways of seeing things, literally or otherwise (Hamlyn 1994:362); or conceiving, framing, and entertaining possibilities beyond known facts. Imagination and musical expressiveness Not all these notions of imagination need be involved when we experience art, even if many are. In particular, to turn at last to imagination and musical expressiveness, not all notions of imagination need apply to cases where we hear music as expressive. There are, I suggest, a variety of ways in which imagination is involved when we hear music as expressive, and I sketch some of these below (see also Trivedi 2001). There may also be other kinds of imaginings involved in hearing music as expressive in addition to the ones I note later, and that is a task for further inquiry on this issue to explore. Note that the imaginings involved in hearing music as expressive are constrained and guided by the features of the music, and not totally independent of them (Walton 1990). This means that not everything is up to the imaginer, and thus allows for the possibility of shared, imagination-involving experiences when it comes to hearing music as expressive. There are various kinds of imaginings involved when we hear music as sad and imagine it is sad. One kind of imagining involved when we hear music as expressive involves imagining that someone’s mental states are being expressed musically (Levinson 1996a; see also Budd 1995; Ridley 1995; Stecker 2001). That is to say, attentive, musically sensitive, appropriately backgrounded listeners sometimes imagine that an indeterminate musical persona’s mental states are being manifested through the gestures, development, and other properties of the music: while hearing the music, it is felt as if someone (in the music) is crying or wailing or laughing, and so on. What kind of imagining is this? Of the varieties of imaginings mentioned earlier, I suggest that imagining a musical persona involves imaging, not imaging of the visual kind but of the auditory kind, assuming that there are instances of imaging corresponding not just to sight but also to the other senses (Martin 1995). It is not a very specific or precise auditory imaging, which is why the persona is indeterminate. And describing this imaging would involve referring to an associated perceptual experience, in terms of what it would be like to hear someone crying or wailing or laughing, and so on. Moreover, the causal

THE FUNERARY SADNESS OF MAHLER’S MUSIC 267

story that in part allows this kind of imaging, I suggest, is one that involves resemblances between how the music sounds, on the one hand, and our vocal and bodily expressive behavior as well as the affective feel of the emotions themselves (see Kivy 1989; Davies 1994; Budd 1995). For example, when we imagine that someone is crying and expressing their sadness musically, this is in part because we hear the sound of the music resembling human (or some other creature, humanlike or not) crying or wailing. And the sound of the music may resemble human crying due to timbre, pitch, articulation, dynamics, even tempo sometimes, and so on. For instance, I have sometimes, but not always, heard the opening clarinet glissando in Gershwin’s Rhapsody in Blue in terms of an imagined musical persona crying or wailing, due in part to its timbre, pitch, articulation, musical shape or contour, and dynamics resembling human crying. At other times, however, I have heard the same passage in terms of an imagined musical persona yawning loudly and waking and rising from bed, due in part to similar musical features, especially musical shape or contour, resembling human yawning and getting up. A different kind of imagining when hearing music as expressive involves an imaginative identification of the auditory experience of hearing the music with the experience of feeling the emotion we hear the music as expressing (Walton 1988, 1994; Budd 1989, 1995). We imagine of our experience of hearing the music that it is an experience of our feeling the emotion the music is expressive of. We may imagine in such cases that the music is expressive of our own emotion, as we may identify imaginatively with the music and imagine, for a while at least, that we are the music (see Scruton 1997:363–4). Note that this need not be an experience of real or actual arousal, as has sometimes been thought (Matravers 1998:137–40), but one that involves at best an imagined arousal. What kind of imagining is involved here? I suggest that here we may either have a case of our supposing that one kind of experience is another kind of experience, that there is a unity; or perhaps we fancy that one kind of experience is another kind of experience. Perhaps something similar happens often when we see movies and plays and empathize or sympathize or identify with the characters portrayed, imagining often that it is our own mental states that are being expressed, that the fictional character’s experience of emotion and its expression is our own experience. It is doubtful, though, that when seeing plays and movies, we imagine of our perceptual experience that it is an experience of our feeling the mental state expressed and perceived. That sort of imaginative identification of experiences may be unique to music, perhaps due to its abstract nature, and due to our hearing expressiveness in it, and getting caught up in it, readily and immediately, and in a way that might well be more ready and immediate than what is involved in experiencing many other art forms; as Lopes suggests, perhaps nuances of emotional states can be conveyed in auditory imagination in ways that the visual and literary arts can only approximate to (Lopes, Chapter 12, in this volume). This is not to deny that other kinds of imaginative identification of experiences may also happen when experiencing other arts, for example,

268 SAAM TRIVEDI

imagining of our perceptual experience of seeing a picture that it is an experience of seeing what is depicted (Walton 1990). On other occasions, we may hear music as sad, and this may involve imagining more simply in that we may entertain or suppose that the music is sad without actually believing this, given that we know that music is inanimate and cannot really or literally be sad. Yet another kind of imagining may be involved, at least for a while, in hearing music that is very intensely expressive, as is the case with some musical passages in the Beethoven symphonies and also in his late quartets. Here, at least whilst in the grip of hearing the music as intensely expressive, we may have false beliefs, or undergo illusions that we cannot help but have, that the music we hear as sad is itself the very thing that is sad. In such cases, we may animate the music and make it come alive for us, endowing it with, or projecting onto it, life and lifelike qualities and imagine, amongst other things, that it is the sort of thing that is sad (see also Kivy 1989:57–9; Scruton 1997:96; Trivedi 2001). We may here be committing what is sometimes called the pathetic fallacy (Flew 1979:263), which strictly speaking is not a fallacy but rather the mistake of attributing mental states and dispositions to inanimate things. In such cases, we may be doing something similar to what we do when we see comic strips animating inanimate things such as cars, trees, the sun, and the like, depicting them as smiling or crying or angry. Within the world of the comic strip, during our experience of the strip, we imagine that the car (or other inanimate object) itself is alive and possessive of mental states, and that it itself is smiling on a nice sunny day, going at 80 miles on open, clear highways; and we imagine that the car itself is glum when it has a flat on a cold, rainy day, and is depicted as being gloomy. Something similar happens when little children play with dolls, imagining that these inanimate threedimensional representations of human beings are themselves happy or hurt or possessive of some mental state. Something similar also happened, I suggest, when our ancestors in ancient Greece, Egypt, India, Japan, and elsewhere animated aspects of nature and the elements (the sun, the moon, the wind, fire, the earth, the ocean, and so on) as various pagan and animistic gods possessive of life and mental states (Scruton 1997: 158, 355–6). Animating the music itself as sad or joyous (or whatever) may thus have connections to other aspects of our imaginative lives that may also involve imagination in the sense of our briefly having false beliefs or illusions. Another instance where we may animate the music is this: some jazz passages, as played on the piano, may be heard as if the music itself is exploring and wandering, perhaps questioning things, in search of something that it itself does not quite know. Note that this kind of imagining, animating the music itself, is phenomenologically and otherwise stronger than the kind of imagining where we may entertain or suppose a proposition without believing it insofar as the former kind of imagining may involve actually believing, at some level of our consciousness and no matter how briefly, that the music itself is sad, and thus animating may involve what David Davies (Chapter 13, this volume) calls cognitive as opposed to perceptual illusions; note also that I am not

THE FUNERARY SADNESS OF MAHLER’S MUSIC 269

describing animating the music in terms of hearing as or some such notion that derives from discussions of pictorial representation and ultimately, via Wollheim, from Wittgenstein on aspect perception. No doubt there are also other kinds of imaginings involved in hearing music, and in hearing it as expressive. Here are some more instances, which may involve imagining either in the sense of supposing or entertaining propositions without believing them, or they may involve auditory imaging, or else they may involve fancying. Rapidly ascending orchestral passages may be heard as if something is soaring into the sky (perhaps a bird); for example, one may hear as such the passage in the Vivace section of the first movement of Beethoven’s Seventh Symphony where twice the strings rise an octave rapidly, to and from the dominant, just before recapitulating the famous theme with its distinctive, repeated, dotted rhythmic figure. Rapidly descending Impressionist piano passages may be heard as if water is cascading, or flowing swiftly in a brook. I cannot stress adequately the importance of noting that not all these ways of imagining need be highly foregrounded or conscious when we hear music as expressive (Kivy 1989:173; Matravers 1998:154). In other words, we may be engaged in some sorts of imaginings without noticing or being conscious at the time that this is what we are doing; daydreaming and dreaming are perhaps the clearest instances of this. This qualification is important for it blocks possible phenomenological objections to the effect that one never imagines music as sad (or as expressive of other mental states) in any of the ways mentioned above. Listeners should, I suggest, listen very carefully and repeatedly, again and again, to all kinds of music, across styles, eras, cultures, and so on, and scrutinize their own experiences of musical expressiveness to find the imaginings involved therein; if they can play musical instruments, that might be even better when it comes to discerning these imaginings. To borrow and vary somewhat the Wittgensteinian prescription to “look and see”, one might say: “Hear and listen!” Moreover, there may also be certain inhibiting factors that may prevent these kinds of imaginings as well as our being conscious of them when we hear music as expressive, and indeed they may prevent hearing the music as expressive in the first place. These inhibiting factors may include inattentiveness (see the claim in Williams, Chapter 11, this volume that more imaginative people bring a certain kind of attention to bear on their sense perception that less imaginative people do not); fatigue; boredom; excessive familiarity with the music; not being sufficiently sensitive, musically speaking; not being an appropriately backgrounded listener who is minimally familiar with the basic musical idiom and style of the musical passage or piece in question (e.g. I do not have sufficient grounding in Chinese opera to hear it correctly as expressive of certain mental states, and thus very often I cannot hear it as expressive nor engage in the imaginings involved in hearing it as expressive); and so on. Furthermore, people’s musical sensitivities, powers of imagination, and capacities of imagery vary a great deal (Hamlyn 1994:363; Matravers 1998:193), and this may also help explain why some claim that they are

270 SAAM TRIVEDI

never aware of imagining sad music is sad (or joyous music joyous), in any of the various ways of imagining mentioned above. I also want to stress here briefly that the various kinds of imaginings involved in hearing music as expressive may be made possible by the structural features of the music, its gestures, its musical development, and so on. And, we should add to this causal story about what makes music expressive, and allows us to hear it as such, that various resemblances between music and our vocal and dynamic expressive behavior and also between music and the affective feel of emotions (Kivy 1989; Davies 1994; Budd 1995) may also allow us to hear music as expressive and imagine in various ways that it is so. Moreover, given that these imaginings themselves differ in ways spelt out above, these causal factors need not always allow hearing musical expressiveness in the same way. I now turn to addressing some possible worries about the proposal sketched above. One possible concern about the proposal I have sketched may be that it describes the subjective, idiosyncratic experiences of just one listener, myself. In response, I must point out that there is an emerging consensus amongst philosophers writing about musical expressiveness that we do engage in many of the imaginings described above when we hear music as expressive. Claims about hearing music as expressive in terms of an imagined, indeterminate musical persona have been made or at least granted by other philosophers before me (see Budd 1995; Ridley 1995; Levinson 1996a; Stecker 2001); claims about our imaginatively identifying our auditory experience of the music with our feeling the emotions expressed have also been made before (see Walton 1988, 1994b; Budd 1989, 1995); and claims about our animating the music have also been made earlier (see Kivy 1989). To be sure, this emerging intersubjective agreement does not in itself show that philosophers who make such claims are infallible as a group in their collective judgment. But still, insofar as this emerging consensus may constitute something like a Humean joint verdict of true critics, we may have some reason to take these philosophers’ various claims about imagination and musical expressiveness seriously and not dismiss them outright, despite occasional disagreements amongst them about the specifics involved. Yet another concern about the proposal sketched above may be that to be a theory of expressiveness, a theory of musical expressiveness must link somehow with our ordinary concept of expressiveness, which involves the outward manifestation of inner mental states. This link must obtain because musical expressiveness is, after all, a kind of expressiveness. In response, at least some of the imaginings mentioned earlier do involve the idea that it is imagined that someone or something’s mental states are being manifested outwardly and expressed through the music (see also Trivedi 2001). Who or what this someone or something is may vary: sometimes it may involve an imagined, indeterminate musical persona; sometimes it may involve animating the music itself so that it is felt as if the mental state expressed musically belongs to the music itself; and sometimes we may imagine that the music is expressive of our own mental states as we identify with the music and imagine we are the music or that our

THE FUNERARY SADNESS OF MAHLER’S MUSIC 271

experience of hearing the music is an experience of our feeling the mental states expressed. Thus, I submit, the proposal sketched above maintains some link with our ordinary concept of expressiveness in the ways just specified, even if not all the different kinds of imaginings sketched above as being involved in hearing musical expressiveness relate easily to the ordinary notion of expressiveness; this might well show that perhaps some of these involve experiences of something other than musical expressiveness but I shall not pursue the point here. A third worry about the proposal sketched above may be this. How do we know when we are right in hearing a musical piece or passage as expressive of one mental state rather than other? How does my proposal make sense of the fact that it is correct to hear the second movement of the Eroica Symphony as expressive of grief and incorrect to hear it as expressive of glee? In reply, I suggest that we can justify hearing music as expressive of a given mental state either in terms of the music and its character resembling our vocal and dynamic expressive behavior and the feel of mental states themselves; or in terms of a consensus amongst competent (but fallible) listeners (those who are musically sensitive, attentive, appropriately backgrounded in musical terms, imaginative, and so on) as to the expressive nature of the music; or in terms of a combination of resemblance and consensus. I submit that more often than not, we do have a consensus about the expressive nature of musical works and passages; most people would describe the second movement of the Eroica Symphony in terms of some negative emotion such as grief or heavyhearted resoluteness rather than in terms of some positive emotion such as glee. Thus, my proposal can allow for normativity when it comes to judgments of expressiveness. I also want to note here, before concluding, that there may be a link via resemblance between imagination and musical representation in the few cases where music may be said to be representational. Sonic resemblances (in terms of timbre, dynamics, and so on) between the rolling crescendos in the fourth movement of Beethoven’s Pastoral Symphony, for example, and the sound of thunder may allow us to imagine, in the sense of entertaining without believing, that the musical sounds we hear are (or that they represent) the sounds of a storm. To conclude, with the more traditional solutions to the problem of musical expressiveness—metaphorism, arousalism, expression theories, resemblance-based theories (for a critique of the last, see Ridley 1995; Trivedi 2001), and so on—all facing difficulties, it is high time alternative solutions were adequately explored. I have mentioned problems for metaphorism and arousalism earlier, and I will briefly add that, amongst many other problems that I cannot detail here (but see Davies 1994; Carroll 1999), expression theories face the problem that the composer or performer need not feel the mental state that the musical work or passage is expressive of; and resemblance-based theories face the problem that they do not relate to the ordinary concept of expressiveness in dealing instead with resemblances, which may give us the causal story behind expressiveness but do not suffice for expressiveness. One alternative to these older, problematic theories is to claim, as I have suggested, that we imagine in various ways that sad

272 SAAM TRIVEDI

music is sad (and we imagine joyous music is joyous, and so on). And this may be done in either of the many ways mentioned above, though there may also be other ways of imagining involved in hearing musical expressiveness. One might disagree with the specific ways in which I have fleshed out our imagining, for example, that sad music is sad. Even so, it is, I think, undeniable that imagination is involved in hearing music and in hearing it as expressive. It would be folly to underestimate imagination and its powers when it comes to the arts (as analytic philosophers and aestheticians largely did till some years back), even, or perhaps especially, when it comes to musical expressiveness. For the role played by imagination in hearing music as expressive may explain in part why we find such experiences valuable. We may imagine the music is sad (or joyous or possessive of some mental state), and empathize or sympathize or identify with the music itself, as animated, or with its persona, and this may often arouse emotion in us. Amongst other rewards, this may provide us with an emotional release; it may give us a rewarding sense of emotional communion as we feel we are not alone in being capable of having certain mental states that the music itself or its persona imaginarily also has, and that may be aroused in other listeners too; it may give us a reassuring sense of relief that we are human and sensitive enough to perceive emotion in music and respond to it emotionally; and it may be rewarding in giving our powers of imagination a certain degree of freedom to adventure, play, and explore when we hear music as expressive. Exploring these and other rewards of listening to music with imagination in hearing it as expressive is, however, a topic for a different occasion. Acknowledgments For helpful feedback on various things in this chapter, I thank John Brown, Matthew Kieran, David Kolb, Jerrold Levinson especially, Dominic Lopes, Nirmalangshu Mukherji, Daniel Nathan, Diane Raymond, Allen Stairs, Fred Suppe, and audiences at Bates College and at Simmons College. Note 1 Matravers approvingly quotes Bouwsma’s famous claim that “sadness is to the music rather like the redness to the apple, than it is like the burp to the cider” (Bouwsma 1950: 98). But I submit that my view is closer to Bouwsma’s claim than is Matravers’s arousalism, for just as redness is a mind-dependent property of the apple, likewise musical sadness (and musical expressiveness in general) is an imaginationdependent and thus broadly mind-dependent property of the music, if I am right. On a related point, I disagree with Matravers: “part of the analysis of red is the claim that the observer ‘experiences x as red’; the analogue in the aesthetic case is that a work ‘arouses a feeling’ in the observer” (Matravers 1998:188). I submit that the aesthetic analogue for expressiveness is not arousal but rather competent listeners in normal circumstances readily and immediately “experiencing music (or the artwork) as sad”,

THE FUNERARY SADNESS OF MAHLER’S MUSIC 273

which is different from arousal and need not involve arousal (as Matravers allows) even if the two are often coextensive.

16 SCULPTURE AND SPACE Robert Hopkins

This chapter is about the aesthetics of sculpture. That is, it asks what aesthetic satisfactions sculpture has to offer. There are two directions from which we might approach this topic. We might concentrate on what sculpture has in common with the other arts, thereby explaining what it is about sculpture that makes it an art at all. Or we might direct our attention to what, if anything, is distinctive about sculpture. I will adopt the latter approach. Not that I assume that sculpture is aesthetically distinctive. It is hard to see what, prior to investigation, could justify this view. My motive is rather that, if we do not look for sculpture’s distinctive aesthetic interest, we are unlikely to find any that is there. Since I think that philosophical reflection on the arts both feeds on and itself nourishes critical engagement with them, failing to look for anything distinctive about sculpture not only threatens to impoverish philosophy; it might also limit our engagement with the art. There are other philosophical questions we might raise about sculpture. It will help to distinguish them from the one that will be our main concern. An issue that will concern us a little is the nature of sculptural representation. At least a good deal of sculpture represents. What is the form of representation involved, and how does it relate to representation by other things, such as pictures and language? Another question will hardly concern us at all: how should sculpture be defined? The answer might seem obvious: it is representation by threedimensional (3-D) figures. However, confidence in this quick reply is equally rapidly undermined. Think of abstract sculpture, and representations in three dimensions that are usually given their own categories, such as models and maquettes. There are interconnections between these questions. One might, for example, attempt to define sculpture in terms of the form of representation it exhibits (having somehow finessed the issue of abstract sculpture). Or one might use an account of sculptural representation to drive sculptural aesthetics, perhaps deriving from the former aesthetically revealing conclusions about the range of what can be represented sculpturally, or offering sculptural representation itself as able to engage our aesthetic interest in ways in which other forms of representation cannot (see Lessing 1962). Nonetheless, the questions are distinct, and should be treated separately.

ROBERT HOPKINS 275

Sculptural representation It will introduce some useful ideas if we begin with the issue of sculptural representation. What one says about this depends in part on where one thinks the crucial contrasts lie. All acknowledge some important differences between representation by sculpture and representation in language. Even Goodman, who sought to make room for the idea that both are, at root, a matter of convention, devoted considerable ingenuity to stating the difference between them (Goodman 1976). In contrast, few have attempted to differentiate sculptural representation from representation by pictures. Goodman himself clearly intended what he said about sculpture to apply equally to pictures. Indeed, it is the latter that provide his ostensible subject matter almost throughout (the exception is Goodman 1976: chapter 1, sect. 4.) In this, if in nothing else, he was followed by Flint Schier (Schier 1986). Schier’s central idea is that some representation engages the same processing capacities as what it represents. This defines a general notion of iconic representation, intended to cover, not just pictures and sculptures, but other representations too, such as mime and some aspects of theater. So those of very different outlooks have concentrated on distinguishing sculpture from language, at the cost of classifying sculpture with pictures. Yet prima facie there is a difference between the way in which pictures and sculptures represent. This thought looks especially appealing in the context of a particular approach to the topic. In discussing pictures, Richard Wollheim has argued that the form of representation they exhibit might be constituted by the distinctive experience to which it gives rise (Wollheim 1968, 1987). This idea is every bit as plausible when applied to sculpture. The thought is that to grasp the content of a sculpture is to see it in a special way. One’s experience of the sculpture is permeated by certain thoughts. These do not merely accompany the experience, they determine its phenomenology. They are thoughts of the object represented—a horse, say. Now, it is true of seeing a horse in the flesh that it is an experience permeated by thoughts of horses. For that is just what it is to be an experience with the content that a horse is before one. But it need not be part of the present approach that sculptures generate the illusion that their objects are present. When we see a horse sculpture, there is no horse before us, we do not take one to be there, and our experience does not have the phenomenology of seeing a horse. Rather, it presents us with a crafted lump of marble, bronze or whatever. But, although we see nothing but marble or bronze to be before us, we experience that material as organized in a distinctive way. It is organized by thoughts of the sculpture’s object: a prancing Arab stallion with flowing mane, and so forth. More needs to be said about this special experience, and in particular about its phenomenology. The way to do this is to specify the experience’s structure, to say in what way thoughts of the absent horse permeate one’s experience of the marble before one. Such a specification should replace unsatisfactorily vague talk of thought “permeating” experience with something more precise. There are various accounts we might adopt, and some are discussed later. The point now is

276 SCULPTURE AND SPACE

that one factor bearing on their plausibility is how far they are able to distinguish our experience of pictures and of sculpture. The general description in the last paragraph applies equally to the pictorial case, as is not surprising given the source of its basic idea. The only more detailed account considered so far, the illusion view, also fails to distinguish the two. For by assimilating our experience of both sorts of representation to our experience of things in the flesh, it denies itself the resources to distinguish between them. But this is surely a failing. It is simply not plausible that we experience pictures and sculptures in the same way, and hence important, if the representation each involves is defined by reference to those experiences, to capture the experiential difference. Of those sympathetic to the experiential approach to pictorial and sculptural representation, few have explicitly discussed the latter. (This reflects more general neglect of the topic of sculptural representation, and indeed of the other two questions described in the previous section.) However, it is easy enough to see whether their accounts are able to distinguish the two. Wollheim himself is pessimistic about how far the forementioned rough outline can be filled in with an account of the experience’s phenomenology and structure (Wollheim 1987: 46–7). Nonetheless, he does offer something more, and what he says seems unlikely to yield a natural way to differentiate the sculptural from the pictorial (Vance 1995). He talks about the experience of pictures having two “folds” or aspects. One is constituted by our awareness of the properties of the object, here a painted surface, before us; the other by our awareness of its representing something absent. Much the same clearly applies in the sculptural case. And Wollheim’s pessimism about the prospects, in the case of pictures, for saying more serves only to banish further the possibility of distinguishing the two. Kendall Walton is more sanguine about the chances of analyzing these experiences (Walton 1990). In his view, pictorial experience is constituted by our imagining certain things. We imagine, of our looking at the marks on the picture’s surface, that it is our looking at the things represented, and we do so in such a way that the imagining, which is rich and vivid, permeates the perception, thus giving it its distinctive phenomenology. There are questions about the coherence and the plausibility of this account (see Hopkins 1998:20–2). But the point now is that, however attractive the view, it does not naturally yield a complementary, though distinct, account of our experience of sculptures. True, Walton does have something to offer on this score; but his comments are very brief. They reduce to the idea that sculptures are more likely than pictures to license imagining that the represented object is actually located in gallery space (Walton 1990:63); and that there is a wider range of actions for sculptures than for paintings for which it is natural to imagine performing that action on the represented object in virtue of performing it on the representation—his examples include caressing, and various perceptual actions and movements (Walton 1990: 227, 296). These differences seem too peripheral to provide a satisfying distinction between our experience of pictures and of sculpture. It is not plausible that, before sculptures, we always imagine in the particular ways Walton describes, and

ROBERT HOPKINS 277

not plausible that the mere possibility of so imagining constitutes (rather than reflects) an ever-present difference between the two art forms. Yet, if there is a difference between the way we see pictures and the way we see sculpture, it is surely present all the time. However, at least one account of pictorial experience does naturally capture its difference from the sculptural analogue. If we construe both as experiences of resemblance, we can distinguish them in terms of the respects in which resemblance is experienced. It is not a straightforward matter to state the resemblance relevant to the pictorial case, but not our present concern to do so (see Hopkins 1998). It’s enough to note that the overwhelmingly plausible candidate for sculptural resemblance, resemblance in 3-D shape, is precisely not a respect in which pictures are experienced as resembling their objects. Now, the view that sculptures are seen as resembling what they represent in terms of 3-D shape is, for all its intuitive appeal, in need of considerable defense. Here is not the place to undertake that (see Hopkins 1994, 1998). For our purposes, it suffices to have some sense of the possible approaches to sculptural representation, and some grasp of the experiential approach in particular. We will make use later of the idea that sculpture is experienced in the light of thoughts about what is represented, without those thoughts engendering any illusion about what is present. We will also return to the question of imagination’s involvement with sculpture. Both themes arise as we address the aesthetics of sculpture. The aesthetic question: some preliminary thoughts Whatever the right account of sculptural representation, it is unlikely by itself to yield a satisfactory aesthetics of sculptural art. Certainly none of the above accounts reveal, unaided, why things exhibiting the form of representation they describe are of interest. To do this, they need supplementing, at the minimum with some claim about the aesthetic interest of representation in general, and ideally with some explanation that engages with the details of the particular form of representation they describe. Moreover, if, as we do, one seeks to distinguish, in aesthetic terms, between sculpture and the other arts, some of the above accounts are clearly impotent. They fail to distinguish sculptural representation from at least one other kind, that by pictures; and thus cannot yield an account of what is aesthetically distinctive about sculpture, as opposed to some wider class of visual arts. Some would not find this a fault. It is plausible that, from an aesthetic perspective, sculpture is more closely akin to drawing and painting than it is to the literary arts. So we might, as Lessing notably did, go so far as to treat sculpture and painting as one, developing an aesthetic common to both by contrasting their charms with those of literature (Lessing 1962). His distinguished example notwithstanding, we will continue to take the aesthetic question in key part to be what sculpture has to offer that painting and drawing do not.

278 SCULPTURE AND SPACE

Where might an answer lie? If we divide candidate features into three crude categories, one emerges as key. Some of sculpture’s aesthetically engaging features are clearly common, not merely to all visual art, but to a good deal of art tout court. Expressiveness, beauty, the ability to explore ideas by embodying the universal in the representation of a concrete particular: these are found, if not in all arts, then at least in many. At the other end of the scale, some of sculpture’s features are, while undeniably its own, too elemental to provide a comprehensible basis for our aesthetic interest in it as sculpture. Several examples come to mind. Sculpture is the only art, if pottery does not count, at the heart of which lies the formation of distinctive 3-D shapes. But while a responsiveness to the appeal or otherwise of shapes is certainly aesthetic, the engagement it offers is at too low a level to provide, by itself, a plausible basis for sculpture’s claims to be an art. A parallel point applies to the charms of sculpture’s distinctive materials: stone, metal, wood, and clay. (Though see, in this connection Adrian Stokes’s extraordinarily powerful rhapsody on the beauty and significance of limestone and marble (Stokes 1934: part 1).) Yet again, although there might be much of aesthetic interest in a distinction between two fundamental ways of making sculpture, carving versus molding (see Carpenter 1960 on “glyptic” vs “plastic” sculptural art); it is hard not to think that our attention, at least, should focus on the differing consequences of these operations, rather than the operations themselves. At any rate, I will concentrate on the middle ground, on features exploiting, rather than simply constituted by, sculpture’s most basic resources; but not so high-flown as to be formed elsewhere from quite different ingredients. A natural thought is that the aesthetic difference between painting and sculpture stems from the senses they engage. Sculpture, being 3-D, demands to be touched as well as looked at; painting can only be appreciated visually. Many have been attracted by this contrast. Herbert Read made it the centerpiece of his account of sculpture (Read 1961), and the position has its advocates still (Vance 1995). However, there have also been those who deny vigorously any such difference between the two arts. Rhys Carpenter is quite explicit: …sculpture is a visual and not a tactile art, because it is made for the eyes to contemplate and not for the fingers to feel. Moreover, just as it reaches us through the eyes and not through the finger tips, so it is created visually, no matter how the sculptor may use his hands to produce his work. … sculptured form cannot be apprehended tactiley or evaluated by its tactual fidelity. (Carpenter 1960:34) And the denial is implicit in the influential work of Adolph Von Hildebrand, who saw sculpture’s role as that of providing a series of 2-D silhouettes (Von Hildebrand 1932). (Hildebrand, it should be conceded, thereby assimilates the aesthetics of sculpture to that of painting. However, the assimilation is only partial. For, while a painting offers only one such silhouette, a sculpture can offer

ROBERT HOPKINS 279

many, varying with the point from which it is seen. Interestingly, Hildebrand considered this difference to be to the disadvantage of sculpture. Apparently Cellini, on very similar grounds, drew the opposite conclusion.) As the presence of this debate perhaps suggests, the issues here are messy (see Hopkins 2002). We can sidestep them. For, as Hildebrand’s view shows, what really matters here is not which senses are engaged, but what their engagement gives one access to. A satisfying sculptural aesthetic needs, in the end, to concentrate on this last. In the absence of a description of what it is one is accessing, appeals to the engagement of a particular sense reduce to the existential claim that there is something of aesthetic interest, which is thereby made available to us. Thus, although we will briefly return to the theme of the different senses, we will do so only once we have some sense of what it is that we are thereby able to appreciate. Different spaces A quite different starting point is the thought that sculpture is distinctively related to the space in which it lies, that it interacts with that space as pictorial art does not. Thus Hegel: …a sculpture…remain[s] essentially connected with its surroundings. Neither a statue nor a group, still less a relief, can be fashioned without considering the place where the work of art is to be put. A sculptor should not first complete his work and only afterwards look around to see whither it is to be taken: on the contrary, his very conception of the work must be connected with specific external surroundings and their spatial form and their locality…halls, staircases, gardens, public squares, gates, single columns, triumphal arches, etc., are likewise animated and, as it were, peopled by works of sculpture, and, quite independently of this wider environment, each statue demands a pedestal of its own to mark its position and terrain. (Hegel 1974:702) And, extending the point, David F.Martin: …the space around a sculpture, although not a part of its material body, is still an essential part of the perceptible structure of that sculpture. And the perceptual forces in that surrounding space impact on our bodies directly, giving to that space a translucency, a thickness, that is largely missing from the space in front of a painting. With a painting the space between us and the canvas is, ideally, an intangible bridge to the painting, for the most part not explicitly entering into our awareness of the painting. (Martin 1976:282) There is certainly something attractive about these thoughts. Sculpture interacts with its space: it matters, in appreciating a sculpture, what sort of space it is in. This

280 SCULPTURE AND SPACE

is true of pictures too, of course. If we hang a painting too high on the wall, it can looked “cramped” by the ceiling. Likewise, if we place a statue in too small a space, it can look suffocated by it. But there is a difference. In the case of the picture, the sense of crampedness would persist even if one had not yet made out what the picture represents, or if a roughly similarly colored and sized canvas, though one not representing anything, were put there. In the case of sculpture, at least sometimes the effects of fitting with or failing to fit the space is dependent on its representing what it does. If one were to fail to see its content, or if one substituted a nonrepresentational object of a roughly similar shape and size, the effect of cramping would not necessarily persist. This suggests a difference between the two, but how are we to articulate it more precisely? We might summarize the claims of the last paragraph by saying that in the pictorial case what looks cramped is the representation, the picture; whereas in the sculpture’s case it is the thing represented that seems crushed by its surroundings. And this suggests the following account of the difference between the two. The space of the sculpture is the space around the representation itself, what we might, without prejudice as to its nature or location, call “gallery space”. The space of the painting is distinct from gallery space, the space depicted in a picture is a separate realm from the space of the depiction and the viewer. However, as it stands this will not do. Both picture and sculpture, the representations, exist in the space in which we perceive them. And neither the objects depicted nor those sculpturally represented exist in that space. A sculpture of a horse introduces no more horses into the space in which it is seen than does a picture of one. Where, then, is the difference between the two? We might try the thought that, while the sculpted horse is not present in gallery space, it at least seems to be. But this is just illusionism, a view we have already rejected. Besides, even if we could construe our experience of sculptures as somehow involving the apparent presence of their objects, we could as easily do the same for pictorial experience. Another response would be to note that pictures often represent spaces in a more full-blooded sense than sculptures. For a picture may show a range of objects arranged within a volume of containing space, while a sculpture presents nothing more than, say, a prancing horse, without surroundings or companions. But not only does this seem a contingent feature of some sculptures and some pictures, since sculptural groups are possible and context-free horses can be depicted; it also fails to connect with the issue in hand. Any sculpture and any picture represents a space, in representing at least one object, and the spatial relations between its parts. Our question is whether that represented space is differently related, in the two cases, to gallery space. And on that question the reply is silent. A final, and forlorn, stab at answering that question would be to claim that, while pictures always represent other spaces, sculptures represent the gallery space itself, and this is the sense in which the space of the sculpture is the space in which it resides. There are difficulties in interpreting this claim, but, more importantly, it seems a picture can represent gallery space, as when a photograph of a concert hall hangs in the hall itself; but such a picture is not related to its

ROBERT HOPKINS 281

surroundings in anything like the way in which sculpture is, and which we seek to capture. To make progress, we need to distinguish two senses in which spaces may be the same, or different. The first is more metaphysical. Two spaces differ in this sense if they do not form parts of a continuum. The space represented in a picture may be different from gallery space in that it is not part of the spatial continuum of which gallery space is a part. An example is the space depicted in Bellini’s Sacred Allegory. There is no spatially continuous route, however circuitous, from the gallery to the space represented in the painting, if only because that space is not actual. Of course, matters are more complicated if, as in one of Bellotto’s cityscapes of Verona, what is represented is actual space. But we can prescind from these complications: the Bellini sort of example provides the clearest possible case in which, in one sense, picture space and gallery space are not the same. The problem, of course, is that the space represented by a sculpture (e.g. the space occupied by the (represented) arm of one of Degas’s sculpted dancers) is also different from gallery space in this sense. The other sense of “same space” is more everyday. In this sense, the space outside a window is different from the space within the room, in that, though equally parts of one spatial continuum, the two constitute different parts of it; and, moreover, parts presented to us as clearly different, with different natural boundaries, organizing contours, focal points, and the like. I suggest this second sense provides the only reasonable way to construe talk of pictorial space being different from that of the gallery, while sculptural space is the same. Setting aside the metaphysical issue, pictorial space is different from gallery space in just the sense in which the space outside the gallery’s window is: it is experienced as a discrete spatial unit, with its own organizing features. Not so for the space the sculpture presents. But what exactly is our positive account of the sculptural case? We cannot say that the sculpted object is experienced as lying within the perceived spatial unit that is gallery space, on pain of falling back into the error of illusionism. And it is not enough to say that the sculpture itself is experienced as lying therein—that is equally true of the picture, the representing marks themselves. Martin, in the earlier quotation, claims “the space around a sculpture, although not a part of its material body, is still an essential part of the perceptible structure of that sculpture.” But what does this mean? Sculpture and organization: Langer The answer lies in the most sophisticated account of sculpture in the literature, that sketched with tantalizing economy by Susanne Langer in Feeling and Form. Her way to frame the general approach within which we have been operating is to say that sculpture creates, compared with painting and, she adds, architecture, a distinctive form of “virtual space” (Langer 1953:86). I take this to mean that our experience of sculpture needs characterizing as having a distinctive spatial content.

282 SCULPTURE AND SPACE

That content is distinctive in presenting us with a separate space in the everyday sense described above, a discrete perceptual unit, organized in a particular way (Langer 1953:88). What is that way? Langer’s answer has two parts. First, she notes that quite generally we experience our surroundings as organized around our possible movements and actions: …the kinetic realm of tangible volumes, or things, and free air spaces between them, is organized in each person’s actual experience as his environment, i.e. a space whereof he is the centre; his body and the range of its free motion, its breathing space and the reach of its limbs, are his own kinetic volume, the point of orientation from which he plots the world of tangible reality—objects, distances, motions, shape and size and mass. (Langer 1953:90) Second, we are able to see the space around a sculpture as organized around its kinetic possibilities: A piece of sculpture is a center of three-dimensional space. It is a virtual kinetic volume, which dominates a surrounding space, and this environment derives all proportions and relations from it, as the actual environment does from one’s self. (Langer 1953:91) As she summarizes in a slogan: “sculpture is literally the image of kinetic volume in sensory space” (Langer 1953:92). These are the ingredients for a solution to our problem, that is of completing our account of sculpture’s and painting’s differing relations to surrounding space. But Langer’s ideas need careful handling, and some adapting, if they are to be of use. A central question is whether for Langer the space we experience as organized by the sculpture is gallery space, as it must be for her suggestions to bear on our problem. There are at least hints that she thinks not. One is her insistence on talk of sculpture’s creating “virtual” space, another her phrasing when keen to reconcile her account with the persistence of our sense of our own kinetic potential (Langer 1953: cf. 92). But whatever Langer’s actual view, the crucial claim, from our point of view, is certainly open to her. She should say that, just as we see the marble that makes up a statue as organized in a particular way, organized by the thought of whatever is represented; so we see the space actually surrounding a sculpture as organized in a particular way, organized by our sense of the potential for movement and action of that represented item. Neither experience involves illusion: they are never of a kind to mislead us about the nature of our surroundings. Rather, the experiences have the very broad structure outlined earlier: perception itself is transformed by the organizing thoughts, though not so as to yield an experience that in any way fails to be veridical. (For all that, very broadly speaking, they share a structure, the experience here is

ROBERT HOPKINS 283

distinct from that considered in the discussion of sculptural representation. For any experience essential to sculptural representation will be experience of the sculpted material as organized by thoughts of the represented object itself; but here the experience is of the surrounding space, and the thoughts brought to bear are of the sculpted object’s kinetic potentialities.) So, what is special about sculpture is that the experiences it supports include experiences of the gallery space as organized in a distinctive way. For paintings, in contrast, the parallel phenomenon stops at the boundary of the marked surface— the marks are perceptually transformed, the surrounding space is not. Below we consider the merits of this account as the core of sculptural aesthetics. Before doing that let us tidy up one or two other issues. First, Langer too becomes embroiled in questions about sculpture’s relations to sight and touch: Here we have the primary illusion, virtual space, created in a mode quite different from that of painting, which is scene, the field of direct vision. Sculpture creates an equally visual space, but not a space of direct vision; for volume is really given originally to touch, both haptic touch and contact limiting bodily movement, and the business of sculpture is to translate its data into entirely visual terms, i.e. to make tactual space visible. (Langer 1953:89–90) This is not quite right. The phenomenon she makes central is indeed at least partly “tactual”, since the organizing principle in our perception of the space around the sculpture is the sculpted object’s potential for movement and action, and these do essentially involve “haptic touch” and limiting contact with the body. But it is not clear that sculpture “makes tactual space visible” as painting does not. For painting can certainly evoke an environment as organized kinetically The differences between the two lie elsewhere, and are twofold. First, in painting the environment seen as so organized is not that actually surrounding the picture, but that depicted within it. Second, the center around which it is organized will lie at the point of view from which the scene is depicted, a point the actual viewer imaginatively occupies. In the sculptural case, in contrast, the viewer does not see gallery space as organized around the sculpted object by imagining herself in that object’s shoes: her own actual point of view remains the only relevant one. From that point of view, she experiences the space around the sculpture as shaped by the sculpted object’s potential to move and act in various ways. If Langer fails to see that these are the only important differences, it may be because she has not taken some of her own lessons to heart. If she fully embraced the thoughts offered in the first of the earlier quotations from her book, thoughts so reminiscent of Merleau-Ponty, she would see that all visual experience is experience of “kinetic volume”, that is, is permeated by a sense of possible movement and action. And this includes pictorial experience, the experience in which we grasp the content of pictures.

284 SCULPTURE AND SPACE

However, Langer’s view does have one consequence for the proper mode of appreciating sculpture. For touching the sculpture itself will hinder us from perceiving its surroundings as appropriately organized: “handling the figure, no matter what it gives us, is always a mere interlude in our perception of the form. We have to step back, and see it unmolested by our hands, that break into the sphere of its spatial influence” (Langer 1953:92). Although Langer does not say why this should be, it is easy to think of reasons. For one thing, to touch the sculpture is to be too near to the center of the space around it to experience that space as appropriately organized. For that experience of organization is essentially visual, however informed by other senses and proprioception, and from up close one cannot take in visually enough of that space at one go. For another, to explore the sculpture by touch is to reinforce one’s sense of one’s own actual kinetic possibilities, and this may, as a matter of psychology if not of logic, necessarily reduce one’s ability to see the space as constructed with another object at its kinetic center. Organization and the imagination Langer’s account leaves certain questions unanswered. Most obviously, it would be good to know more about the experience she makes central. She tells us something about the thoughts that permeate it, but nothing about its structure, about how those thoughts enter it. I know at least part of what I would like to offer her here, but there are complications. My suggestion is that the experience Langer has described is essentially an imaginative one. It is constitutive of that experience that the viewer is imagining certain things of the sculpted object. One framework to which we might appeal in making such a claim is that developed by Kendall Walton (Walton 1990). He suggests that a good deal of our engagement with representations, of any kind, is a matter of our imagining certain things as a result of recognizing that the world is in fact a certain way. The actual properties of the representation establish that it is to be imagined that there is such-and-such object or state of affairs; and, at least sometimes, our actual relations —be they spatial, perceptual, or psychological—to the representation establish that something further is to be imagined about our relation to the represented thing or state. Indeed, we have already seen one application of this scheme, in Walton’s account of pictorial experience. As we then noted, Walton does in passing discuss sculpture, and in particular its differences from painting. Previously we said that his appeal to the likelihood of imagining the sculpted object to be present, and the range of actions we might imagine performing on the item do not suffice to capture what is distinctive about sculptural experience, and the representation it defines. Note that they also plainly fall far short of committing him to anything like Langer’s view. Nonetheless, his theory does provide a congenial context within which Langerian claims could be made. We would merely need to say that the experience she has pinpointed is one of imagining the represented object to perform certain actions, or perhaps to

ROBERT HOPKINS 285

have a certain potential for action, in gallery space; imaginings licensed by properties of the sculpture itself and of the space in which it stands. What is attractive about some such appeal to imagining is that it accommodates the sense that seeing the gallery space as organized around the sculpted object is an experiential matter, but one extending in a somewhat optional way the basic experience of seeing the represented thing in the sculpture. Certainly, one could have the latter experience without the former, and one way to make the step to the more complex experience might be to imagine certain things of the represented object. However, it does not follow that imaginings are constitutive of the experience Langer has described: perhaps they are just a means to that end. And it certainly is not plausible to suggest that seeing the gallery space as organized around the object involves visualizing, or (pace Langer) any form of tactile or haptic imagining, of the represented object’s movements. Such deliberate imaginative episodes seem at best to accompany, and at worse to interfere with, the Langer experience. It may seem to ease the problem if we tinker with what is imagined— not actual movement or action, but merely the potential for it. But then it is unclear quite what it is to imagine the sculpted object having the potential for those things, if not to imagine (in some way) its doing them. What we are looking for is a way to make room for some thoughts of the object’s potentialities, and the problem is that any clear sense in which imagining might do this is also one in which it is not so plausible after all that imagining with that content goes on. Perhaps this will seem unfair. It is only implausible in the first place that Langer’s experience involves visualizing, or other forms of sensory imagining, if it is assumed that such imagining will be both deliberate and at the forefront of consciousness. But can it not be involved in some way that is both more basic and less prominent? The appeal of this line lies partly in generality. There are three experiences before us. One is a form of seeing ordinary things—the experience of the environment as organized around one’s own kinetic potential. Another is the experience in which a sculpture’s content is grasped. The third is Langer’s experience of gallery space as organized around the represented object. I have suggested that imagination plays a role in the third, and earlier implicitly rejected any role for it in the second. Walton explicitly gives it a role in the second, and certainly could give it a role in the third. But what of the first? We might see that too as involving imagination. For what could constitute experiencing things as organized around one’s own potentialities, if not some lived sense of how one might move and act upon them? And why not then see the imagination, operating in the background, as providing just that sense? This would enable us to give a single sort of account of all three experiences, and, whatever one thinks about the second, surely the first and third are intimately related. The problem with this line is that it seriously weakens our hold on what is meant by “the imagination”. It is no more obvious that the experience of things as organized around one’s capacities involves imagining than it was that Langer’s experience does. To insist that it does have a role, just one that is both

286 SCULPTURE AND SPACE

fundamental and in the background, looks unhelpful. For sure, that insistence fits into a long and distinguished tradition. For those who have argued that we do experience things as organized around our kinetic potential have certainly thought that this was a, perhaps the, fundamental form of perceptual experience (Gibson 1950, 1986; Merleau-Ponty 1962). And there is a long history, running back through figures of the stature of Hume and Kant, of seeing perception itself as fundamentally involving something called the “imagination”. The difficulty is that, as P.F.Strawson noted some time ago (1974), it is quite unclear that “imagination” is univocal, between its use in such contexts and its use to describe mental activities which are deliberate and at the forefront of consciousness. The proposed defense of imagination’s role seems fated to run into this quagmire of conflicting uses and missing definitions. Thus I have made relatively little progress in explicating Langer’s phenomenon. It remains unclear quite what thoughts about the represented object must be deployed. And it remains unclear in the context of which attitude (our candidate has been imagining) they are instigated. My hope is that nonetheless our grasp on Langer’s account is sufficiently tight for us to assess it. Langer assessed What are the strengths and weaknesses of Langer’s view? (1) First, Langer is well placed to explain why the subject matter of so much sculpture has been severely limited. People and animals form almost the sole subject matter of the greater part of what has been called “sculpture”. Even the twentieth century, with its radical reconceiving of sculptural art, saw masters such as Henry Moore defending the view that the human form is sculpture’s proper subject. For Langer, this should be perfectly comprehensible. If sculpture is “the image of kinetic volume in sensory space”, one would expect it to concentrate on representing whatever can form the center of such kinetic volumes, and that, since it is the larger creatures that dominate our experience of actual movement and action, means people and certain animals. However, it may seem that to this strength there corresponds a weakness. Will the account not strain to accommodate certain sculptural works, particularly more abstract ones? For, to put the point crudely, if nothing is represented, or nothing definite enough to have “kinetic potentialities”, how can the sculpture organize surrounding space in the way described? Partly in recognition of this problem, Langer characterizes what a sculpture needs to support the key experience as “living” or “vital” form, the sort of unity, the fittedness of part to part and part to function, which organisms exhibit: There is nothing actually organic about a work of sculpture. Even carved wood is dead matter. Only its form is the form of life, and the space it makes visible is vitalized as it would be by organic activity at its center. It is virtual kinetic volume, created by—and with—the semblance of living form.

ROBERT HOPKINS 287

(Langer 1953:89) This solution requires clarifying. The thorniest question is over what “living form” amounts to. But a related obscurity concerns what it is supposed to be a property of. Is it the sculpture itself that exhibits living form, or some object the sculpture represents? If Langer takes the former option, the proposal at least promises to apply to all the current problem cases. For Langer can claim that all abstract sculptures exhibit living form without having to suppose that they represent anything, as their abstractness apparently precludes. The cost of this move is that she is forced to offer a different account of our experience of these sculptures from that, at least in our elaboration of it, she is offering for sculpture of other kinds. For there we have her claiming that the gallery space is experienced as organized around the kinetic potential of the sculpted object, while here she must say that it is experienced as organized around the potential of the sculpture itself. The alternative avoids this fracture. Even apparently abstract works represent, and it is the represented object that displays living form. Such sculpture is thus not abstract in the sense of representing nothing. Rather, its abstractness amounts to the fact that what is represented is relatively imprecise: not a horse or man of a certain specifiable type, but merely something shaped in certain only loosely specifiable ways, perhaps with certain broad capacities for movement, and the like. In effect, this line treats abstract sculpture just as Richard Wollheim proposes to treat abstract painting (Wollheim 1987:62). One might worry that the move will be empty. What is the force of claiming that the sculpture has a content, albeit one limited to shapes and other elemental properties, rather than simply talking about the sculpture itself, in terms of its having those properties? But the worry is misplaced, provided there is reason to distinguish between two sets of properties, those of the sculpture itself, and that at least partially distinct set that composes (for what alternative bearer is there?) what it represents. For instance, with at least some “abstract” sculptures, appreciating them surely requires a sense of solidity, for all that the sculpture itself is manifestly hollow. Rather, the difficulty the maneuver faces is one of incompleteness. For even if many abstract sculptures can be accommodated in this way, why think that all can? Might not some be truly abstract, not merely not representing anything very specific, but not representing anything at all? If there are such cases, then at least some of the original counterexamples to Langer’s view stand. Perhaps either option is feasible. I will not attempt to adjudicate between them. In the following I discuss how far Langer can afford to concede recalcitrant cases. If, as I will argue, she has some leeway, then perhaps at least the second defense here does not come at too high a price. For now, let us continue to review the possible cases. Are there other examples of sculpture to which Langer’s account fails to apply? (2) One possible sort of case is provided by very small sculptures, such as Cellini’s saltcellars in the form of animals. What is the space they are experienced as

288 SCULPTURE AND SPACE

organizing? However, the difficulty here is not that these pieces fail to organize their surroundings, so much as that their doing so leads to faintly absurd consequences. For there is something peculiar in the idea of the other condiments falling with reach of the saltcellar lion’s pounce, while the little creature is itself dominated by the vase of flowers close by. If so, the case already fits the essentials of Langer’s view, even if the experience she makes central would here be of dubious aesthetic worth. But the example merits further discussion, if only because it introduces two themes of more general theoretical interest. The first is that the case highlights something implicit all along, that is, that Langer’s experience depends not merely on the sculpture, but also on the surrounding space. It must be such that it can, or can fruitfully, be seen as organized around the sculpted object. The point brings with it various complexities, to do with the role of intention in aesthetic effect, given that at least some sculptures are sufficiently portable to be displayed in a range of environments, many of which would not have been known to the sculptor, and some of which might even have been beyond anything he could have anticipated. But these complexities need not delay us, since they arise for many aesthetic effects, and not just Langer’s. We do better to note that the fact that Langer’s phenomenon depends as much on the sculpture’s context as on the sculpture itself allows us to build a bridge to some of the cases on the periphery of sculpture, as naturally conceived. For some 3-D models, such as tiny replica cars, might also organize the right surrounding space in something like the way Langer describes. It is a further question how special the surroundings must be for this to occur, and whether its occurrence would be aesthetically interesting; but at least their disposition to produce this effect, in the right circumstances, provides a link between these peripheral cases and the core of sculptural art. Perhaps the difference between the two is in part a matter of the range of gallery spaces each is disposed to organize around itself, traditional sculpture dominating a wide range of spaces, these peripheral cases requiring very specific circumstances for the effect to obtain. The second theme is that of scale, of how the size of the sculpture relates to that of the sculpted object. In general, of course, the size of a representation, even a broadly mimetic representation, does not directly determine the size of object represented—small portraits do not depict small people. The same is true of sculpture, even on a view on which sculptural representation depends on our experience of it. The saltcellar does not represent a lion the size of the lump of silver; one does not take it to represent such a lion; and one does not even see in the lump a lion of that size. And this, even though one sees the size of the block quite clearly: it is just one of several properties of the sculpture set aside as irrelevant to its content. (The relative size of head to body matters, the absolute size of either does not.) However, in certain respects this ability selectively to ignore certain features of the piece is weakened if we view sculpture as Langer suggests we should. For it is difficult to see the space around the sculpture as organized around the represented

ROBERT HOPKINS 289

object’s kinetic potential, without concentrating on that space in such a way as to be aware of the region of it occupied by the sculpture itself. It does not follow, certainly not as a matter of logic, that one must come to see the space as formed around a sculpted object of just the sculpture’s proportions. But, at least as a matter of psychology, I think it hard for this not to be the result. This, of course, does not undermine the main point above. There is no pressure here for the sculpture to represent a lion of the saltcellar’s size. For there is no suggestion that Langer’s experience is the one relevant to sculpture’s content. And even if the upshot is that the content-determining experience is corrupted, so that it does incorporate the lion’s small size, in general experiences can determine content without every aspect of the experience determining a correlative aspect of content (Hopkins 1998: chapter 6). But it does indicate some of the delicacy of the relation between sculpture and surrounding space, if Langer’s experience is to occur in an aesthetically satisfying way. (3) Other possible counterexamples suggest themselves. Might sculptured groups not exhibit a self-containedness that prevent them from interacting with their surroundings in the right way? Might, as it were, each not draw the kinetic fire of the others, so that our sense of their possible movement and action does not involve their interacting with the gallery space at all? I am skeptical. There are certainly examples in which the interrelations between the various depicted figures are key—Giambologna’s Samson and the Philistine, for instance. But such cases do not negate Langer’s phenomenon, so much as background it. One still experiences the space around the sculpture in the light of the room needed for Samson to swing his weapon, or the space into which the hapless Philistine might spring, for all that one’s attention is concentrated on the intense spiral of the two interlocked men. Other cases do seem to efface Langer’s effect more completely. Here we might think of any one of Ghilberti’s panels for the doors of the Florence baptistery. But they do so, not merely through the interrelations of sculpted figures, but through setting them in a clearly defined sculpted space, a classical receding plane. And in doing so, they precisely lie nearer the boundary with art properly thought of as pictorial. For the effect of the sculpted surrounding space is precisely to create a sense of a different space from that of the gallery, in the sense traced above. Perhaps it seems question-begging to dismiss these examples as only marginally sculptural. But there are other considerations motivating this, such as their combining full-blown sculptural representation with low relief and even at times engraving. In these respects too, they seem to occupy a region between sculpture and painting, and that is how, in general, tradition has received them. If this marginal status is essential to their frustrating our attempts to experience them in the way Langer has described, they seem as readily to provide a confirmation of her view as a threat to it. (The Ghilberti panels in fact involve several other factors antipathetic to Langer’s phenomenon. The most important is the small scale of the figures, relative to the vast space of the piazza which they would have to organize. Another is the quality of montage produced by so many distinct

290 SCULPTURE AND SPACE

sculpted scenes lying side by side. In effect, this puts the organizational power of one panel in competition with that of the others, thereby further neutralizing any influence each might have.) (4) However, there is a more central class of counterexample: sculpted portrait busts. It seems hard to marginalize these—they have a central place in our tradition of sculpted art. Yet it is far from clear that Langer’s claims apply to them. For what is the “kinetic potential” of the represented head, or head and shoulders, such that gallery space is seen as shaped by its possible movements? Of course, it is true that we experience these sculptures as interacting with their surroundings. For we ourselves interact with them in a way somewhat akin to our interaction with other people. Hence it is not mere ease of inspection that encourages us to position such busts at head height. But it is not clear that this interaction is the sort Langer has described. For one thing, it does not clearly involve action, at least not action in some sense in which the environment might be appear receptive or resistant to such—any action here involved would seem to be more psychological in nature. And for another, it needs spelling out how any such interaction differs from that with a picture portrait. So Langer faces a challenge. She must either show that busts are after all experienced in the way she has described, or show that bust portraiture is more marginal than it seems, or concede that busts do indeed constitute a central class of examples to which her claims do not apply. I think there is some hope for her with respect to the first two responses here. But let us suppose that these fail, and that she must adopt the third. How serious is this concession? Indeed, how important in general is it that Langer’s claims fit all sculpture? (5) The answer is that it is not very important at all. For Langer need not have ambitions that are thereby frustrated. One such ambition would be that of defining sculpture. My suspicion is that the notion of sculpture is subject to too many pressures—from the etiology of the works in question, the form of representation they exhibit, and indeed the aesthetic satisfactions they offer—for the class to be genuinely unified. That, indeed, creates some room for a revisionist definition, and as such I think Langer’s would be as well placed as any other, for all that it might have to reject bust portraiture as genuinely sculptural. But her claims are of interest independent of any such application of them, and she would do well not to burden herself with extra argumentative commitments. Her ambitions are better seen as purely aesthetic, and in particular as capturing what is distinctive about sculpture, aesthetically speaking. She can seek to do this without claiming that her answer exhausts what there is to value in sculpture—and hence without having to make the highly implausible claim that bust portraiture is without aesthetic value. And she can even, I think, avoid claiming that her account states what it is for sculpture to be valuable as sculpture. For it might be, as indeed seems plausible, definitive of sculpture as an art that it offer, not only its own distinctive satisfactions, but others common to many other arts as well. It should be enough for Langer that the feature she has described is (a) one of

ROBERT HOPKINS 291

substantial aesthetic interest; and (b) unique to sculptural art. So let me end by considering whether her account meets these conditions. I begin with the second, with whether the pictorial arts in particular could ever offer what Langer claims sculpture does. (6) One sort of case must be conceded immediately. It seems there at least could be a perfect trompe l’oeil painting. There is no reason why such a painting could not depict, and thus be mistaken for, a sculpture. And then it seems very likely that such a painting could induce precisely the experience Langer has described. For even if that experience requires one to engage with the sculpture from a range of positions, one can imagine a cleverly positioned series of perfectly illusory pictures providing just such a range. However, the concession here is trivial. For it does not force Langer to accept that the satisfactions of sculpture could be reproduced by a picture seen as such. And surely in general the aesthetics of an art form is a matter of what it offers, when appreciated for what it is. More substantial concessions are harder to force. It is easy to think of cases in which there is some interaction between depicted space and gallery space. Sometimes, as in Masaccio’s Holy Trinity or Titian’s Pesaro chapel, depicted space is a continuation of the actual space in which the painted surface lies. Sometimes the way depicted light falls within the depicted scene is dictated by the actual sources of light in the surrounding room. One can even imagine a case in which the way gallery space appears organized is affected by one’s perception of the depicted scene. The columns in a church might be so arranged that one could see them as grouped in either of two ways. A picture depicting the continuation of the nave might show the continuing space in such a way as to compel one to apply one grouping rather than the other to the continuation, thereby determining which way gallery space itself appears. But although here gallery space is experienced as organized a certain way, and it is so as a result of grasping the picture’s content, the organizing principle is not the potential for action of any depicted item. Nor can I think of any example in which this, the crucial condition, is met. Thus, at least until some such example is produced, I am inclined to think that Langer has indeed identified something that many sculptures do offer, and no picture could. I doubt that the impossibility goes deep. No doubt it turns on facts about our psychology, and perhaps our physiology. But it goes quite deep enough for aesthetics. For we make sculptural and pictorial art for ourselves as we are, not for possible creatures with other perceptual and processing capacities. And it is a conclusion of genuine significance if, as we are, we can find in many sculptures something no picture could offer us. (7) However, one serious challenge to Langer’s view may remain. This is to show how the feature she describes is aesthetically significant. From one perspective, this demand seems unreasonable. There is a form of vertigo familiar in aesthetics, induced by seeing that to any account of the value of something a further question always arises, that of why the terms in which its value has been explicated are themselves not just as questionably of value as the original item. If there is in general no cure for this vertigo, it is unfair to challenge Langer to

292 SCULPTURE AND SPACE

provide a foundation for sculpture’s value so steady as to banish it. But from another perspective the demand is justified. For consider the pictorial phenomena just described, those coming closest to the Langer phenomenon. These seem rather trivial achievements, somewhat as illusionistic triumphs sometimes are. The sculptural phenomenon is one to which they are only distant approximations. But where, in the differences between it and them, does its value reside, such that, though they are trivial, it is not? From this perspective, Langer needs to tell us how the sculptural form of “virtual space” amounts to a serious achievement, why, that is, we should care about it. And, from this viewpoint, if she cannot do that, her phenomenon, for all that it is distinctive to sculpture, is a mere curiosity of that art. Acknowledgments Thanks are due to audiences at the Open University and the Imagination and the Arts conference for stimulating discussion. I am grateful to Rosalind Hursthouse for the example of Cellini’s saltcellars, and for helpful discussion.

Part IV AFTERTHOUGHTS

17 THE CAPACITIES THAT ENABLE US TO PRODUCE AND CONSUME ART Gregory Currie

Kendall Walton once identified a tendency he called “have theory, will travel”: the relentless and increasingly implausible application of one’s favored system to all the problems you can lay your hands on. I try to avoid this tendency here, especially since the editors asked me to draw attention to un- or under-explored issues in imagination and the arts. By way of compromise I spend just the first section advertizing an approach that I think is in danger of being misunderstood. After that the grip of theory is relaxed. I offer some suggestions for taking forwards some of the discussion in these pages; whatever their defects, they are not based on narrowly doctrinal assumptions about what imagination is. No theory about imagination and the arts can be credible if it is wholly a priori, and a number of authors in this volume make significant contact with empirical work on various aspects of mind. A good deal of this chapter focuses on the natural history of the imagination; in the final section I reflect, in a disappointingly inconclusive way, on how the art of prehistory can help us here. I hope to present a more developed view of this last problem in a later piece (Currie, in preparation). Perhaps the problems I discuss are a ragbag. They do seem to me to have this much in common: that solving them will help us to understand the cognitive capacities that make art possible. Imagining events and imagining being a character A number of authors discuss the relations between imagination and the theory of mental simulation (ST), which was developed as a way of explaining how we understand the mental states of other people (see Davies and Stone 1995a, b; Currie and Ravenscroft 2002). ST has no canonical formulation, but it can reasonably be taken as making two, separable, claims. The first is that mind reading tasks are undertaken by imaginatively projecting ourselves into the situation of the agent whom we wish to understand; we take on, in imagination, the beliefs and desires of that agent. The second claim is one about what it is to do this— how we take on, in imagination, these beliefs and desires. According to ST, to take on, in imagination, the belief that p (Bp) and the desire that q (Dq) is to take

THE CAPACITIES TO PRODUCE AND CONSUME ART 295

on—that is, really take on—states that are like Bp and Dq. Call the states we really take on the imagined belief that p (IBp) and the imagined desire that q (IDq). Other common descriptions of these states include “pretend” and “off-line” beliefs/ desires. In what respects are IBp like Bp, and IDq like Dq? First in respect of inferential role; having IBp and IDq enables me to reason, theoretically and practically, as I would if I had Bp and Dq. Second, in respect of affective power; if believing that it will rain and desiring that it not rain make me feel disappointed or worried, then having imaginative counterparts of these states should make me feel disappointed or worried also. (Imaginative counterparts of beliefs and desires will have these capacities to mimic beliefs and desires only in certain circumstances and even then only approximately.) This explains how it is that putting ourselves in another’s shoes can help us see how that person will respond; we can reason from the same premises he reasons from, make the same decisions he makes, and feel the same emotional urgings he feels. As I say, these claims are separable: you can accept the point about the role of imagination in mind reading, and reject the point about what imagining is, and vice versa. Because ST involves these two separable claims, we need to be careful in deciding what ST tells us about imaginative engagement with fictions. Since one claim of ST is that mind reading is done by imaginative projection, we might think it tells us this: imaginative engagement with a fiction involves putting ourselves in the shoes of the characters. I agree with Matthew Kieran that this would be a mistake. He shows how much of our response to fiction does not depend on any sort of imaginative identification with, or putting ourselves in the shoes of, any of the characters; Peter Goldie makes a similar point. But the conclusion here ought to be that we are focusing on the wrong aspect of ST. The relevant lesson from ST is that imaginative engagement with the fiction consists, not in, say, believing that Peter Pan is in danger and desiring that he escape, but in taking what I have called imaginative counterparts of these states. No one, I think, is claiming that imaginative projection into the situation of a character never takes place or is never a useful move on the part of an audience. But I agree with Kieran and others that such activity is secondary, so far as engagement with the story is concerned. What is important is that I imagine the propositions made fictional by the work. ST has a story to tell about what this involves; it is the story I have just outlined about counterpart states. It is not a story to the effect that imaginative involvement with a fiction always or usually involves the simulation of characters. But what, we may ask, is it to simulatively imagine the propositions of the story? Surely to simulate is to simulate someone. In that case, who am I simulating, if not a character? One positive answer to this question would be “you are simulating a character merely implied by the story; one who probably does not figure in its explicit content. This character is an implied or imagined reader or viewer (depending on the modality of the work): someone who thinks of the work as a source of veridical information about certain people and events. While

296 GREGORY CURRIE

we real readers and viewers merely imagine that Pip is frightened by Magwitch, our implied reader believes this”. I think this view faces a number of difficulties, but will not describe them here. A general point worth making is that, to the extent that the view is wheeled on in order to defend the idea that simulation is always simulation of someone, it is wrong headed. Simulations are always simulations of something, but there need be no interesting sense in which they are simulations of someone. When I take on the imaginative counterpart of the belief that p, I am simulating the belief that p; I need not be engaged in a project that has as its purpose that I enter imaginatively into the life of some particular person, real or imagined. But does not ST tell us that we understand the mental lives of the real people we come into contact with by simulating their mental states? If it does, surely it tells us that we understand fictional characters in the same way. And there can be no understanding of stories without understanding their characters. So ST looks, after all, to be committed to the idea that we imaginatively identify with characters. The right thing to say here is that understanding real people and understanding fictional, and especially literary characters are very different activities. Thinking about simulation as an evolved mechanism suggests that it developed in response to the problems posed by face-to-face exchanges with conspecifics in fairly small groups, where there is much shared experience and a high probability of relatedness between the parties. ST then suggests that simulation has at least a role to play in understanding the mental states of people where (i) there is little time for theorizing; (ii) there is much in common between interpreter and interpretee; (iii) there are few other reliable sources of information. Each of these conditions is likely to be violated in the case where we seek to interpret a fictional, and in particular a literary, character. Literary texts, since they dictate no particular pace of engagement, provide us with opportunities for lengthy reflection that may tip the balance in favor of theorizing; they often concern people historically and culturally distant from ourselves, or with whom we simply do not have much in common; their authors are often a source of reliable, and indeed authoritative information about the characters’ motives and other mental states. The last mentioned condition—the role of authors—introduces especially interesting and complex factors. Authors do not function merely as sources of additional evidence over and above that which might be gleaned from simulation, together with normal patterns of evidence-sifting and hypothesis formation. Their activities serve to indicate to us patterns of salience in the narrative that can strongly affect our conclusions about characters’ actions and motives. When we learn something about a fictional character because the work itself tells or suggests it, this is not at all like learning something about someone in the real world by casual observation. For what the fiction tells us comes labeled as having a special evidential relevance; we are told this for a reason. The reason might be that the author wants to suggest something about a character’s motive, and thinks that putting this piece of evidence in our way will suggest it. But this works only because we recognize that the evidence has been put there for that very

THE CAPACITIES TO PRODUCE AND CONSUME ART 297

reason. Thus the world of a fiction is not one that we engage with by applying ordinary standards of evidence and probability; what is evidence is inseparable from what is a sign of authorial intention. And so we may rightly draw conclusions about motive or behavior that would be unwarranted on the same purely evidential basis in real world situations. Elsewhere I argue that it is characteristic of a certain kind of psychosis to see the world as if it were governed by the sorts of forces that externally regulate the actions of fictional characters (Currie and Jureidini 2003). Authors also shape their narratives so as to indicate genre-membership; this can have further effects on the conclusions we draw about characters and their motives. David Lewis gives an example: In the Threepenny Opera, the principal characters are a treacherous crew. They constantly betray one another… There is also a street singer... Is he also a treacherous fellow? The explicit content does not make him so. Real people are not so very treacherous, and even in Weimar Germany it was not overtly believed that they were, so background does not make him so either. Yet there is moderately good reason to say that he is treacherous: in the Threepenny Opera, that is how people are… His treacherous nature is an intra-fictional carry-over from the treacherous nature in the story of Macheath, Polly, Tiger Brown, and the rest. (Lewis 1983:274; Goldie notes other ways in which awareness of a higherlevel, external perspective affects our engagement with the narrative.) Here, and in many other stories, there are conclusions it is reasonable to draw about characters and their motives, but which would certainly not be mandated by acts of imaginative identification; they depend on our seeing the story as having a certain narrative structure. All these are excellent reasons for being skeptical about the idea that mental simulation ought to play a role in understanding fictional characters comparable with the role it plays (or is said to play) in understanding real people. They are not any kinds of reasons for thinking that imaginative engagement with fictions is not mental simulation. All that follows is that the simulation does not have to be simulation of characters. I emphasize that these concessions should not be taken as an abandonment of the idea of character simulation. There are plenty of fictions in which the appropriate emotional effect is standardly achieved via imaginative identification with a character, though this identification need not be complete or long lasting. I suspect we step into and out of characters’ shoes all the time. For certain purposes this need not depend on any great likeness between the character and myself. As Robert Gordon has pointed out, total projections—where one makes no allowance for mental differences between oneself and the other person—work well in some situations (Gordon 1992). Even psychopaths respond very much as we do to a range of stimuli; a conversational remark made by an otherwise distinctly

298 GREGORY CURRIE

odd character can still often be made sense of by imagining myself responding in that way, and seeing what sort of motivation I would have for doing so. Few works depend on, or could reasonably mandate, our imaginatively identifying with a character throughout; what is more common is for the work to encourage a complex mosaic of simulations that are simulations of characters and simulations that are not. Imagination and emotional kinds Much effort, some of it manifest in this volume, has gone into trying to decide whether we really experience emotions like fear, hope and delight in response to fictional characters and events. No one disputes that there are similarities between our responses to fictions and our genuinely emotional responses to real life. The debate has been largely concerned with whether the dissimilarities give us reason for thinking that these responses are of different kinds. Weight has been given to the thought that responses to fictions do not have the relations to actions possessed by ordinary emotions—and this does seem to be a significant difference. One way to undermine this argument is this. Let us assume that emotions were selected for by evolution because they promoted fitness-enhancing behaviors. One way for them to do this is for them directly to drive behavior. But this is not the only way. Thoughtful creatures plan, particularly, as seems to have been the case with our recent ancestors, where they live in fast changing environments. Planning can be done in various ways, but there is evidence that emotions play an important role in the kind of planning we do. There is, for example, a group of neuropsychological patients who are apt to choose high-risk strategies even in the face of negative past experience, because contemplation of the bad outcome does not seem to generate an appropriately negative emotion (Damasio 1994). Here we can see the emotional response (or, in the case of these patients, a lack of it) as genuinely emotional even though it is not leading directly to action. Rather, my vividly imagining a predator lurking over there causes me not to go there. Thinking about how emotional responses to imagined circumstances can have an evolutionary explanation helps to make plausible the idea that fictions generate genuine emotions. If one focuses on the contrast between the fiction case and immediate response to reality, this seems implausible. Kendall Walton asks, “what is pity or anger that is never to be acted on? What is love that cannot be expressed to its object… We cannot even try to rescue Robinson Crusoe from his island, no matter how deep our concern for him” (Walton 1990:196; cf. Matravers 1997:85– 6). Think instead of cases where the imagined events are of relevance to my situation. Take my fear of heights, as triggered by my imagining climbing out onto the window ledge to rescue the cat. The result of my experiencing that episode of fear is that I very sensibly do not go out onto the windowsill. It does not seem particularly odd to say that this is genuinely an episode of fear, even though the situation to which it is a fear response is merely imagined; the connection between the fear and my real behavior is clear enough. And it is

THE CAPACITIES TO PRODUCE AND CONSUME ART 299

simply an extension from this case to that of pitying Robinson Crusoe even though I cannot go and rescue him. If we interpolate emotions-in-response-toaction-planning-via-imagining between emotions-that-drive-an-immediateresponse and (purported) emotions-in-response-to-fiction, it starts to look much more as though we have a single psychological kind than a gerrymandered union of kinds. The idea that imagination developed as a planning tool is no more than a hypothesis. There are other hypotheses, and it is plausible that imagination evolved under a number of distinct but sometimes interlocking pressures. The point I am making here about the taxonomy of the emotions does not depend on talking any particular view about this. If imagination is implicated in mind reading, then it is probable that intra-species competition for reproductive and other resources fueled its growth (Humphrey 1976). And being thus adaptive, hard-to-fake indicators of imaginative capacity may have been influential in mate choice, leading to runaway sexual selection (see Miller 1997, 1999). An imaginative capacity with mind reading as (one of) its proper function(s) might very well exhibit connections with emotions. Suppose imaginative projections into the situations of conspecifics are reasonably predictive of the actions of those conspecifics. Now emotions are among the determinants of action. So a mind reading mechanism ought to reproduce affective states as well as beliefs and desires. A mind reading system based on simulation can do that by ensuring that imaginative counterparts of beliefs and desires mirror the connections to affect that belief and desires themselves have. And sexual selection might reinforce the connection; if imaginative displays generate affect in the agent who displays them, they are likely to be more colorful and impressive. If all this is roughly correct, then the claim that fictions do not generate genuine emotions starts to look like the claim that we do not genuinely believe high-level mathematical principles, because such principles have no impact on practical life. The case for saying that believing the Generalized Continuum Hypothesis is genuinely belief is that this is merely an extended use of an adaptive capacity. While we are constrained to find an adaptive account of belief, we are not required to show that every possible belief is relevant to fitness. Our adaptationist just-so stories about imagination all make it plausible that there are circumstances in which imagination does generate real emotion; being worried about Crusoe would then be an extension no harder to swallow than belief in the Generalized Continuum Hypothesis. Morality-in-fiction Kendall Walton has pointed to an apparent asymmetry in our responses to fiction; there are no obvious limits to our capacity to imagine outlandish factual circumstances, but we seem to be limited in our capacity to imagine outlandish moralities. A work that invites us to assent, in imagination, to the proposition that slavery is morally right is likely to be seen as setting impossibly high barriers

300 GREGORY CURRIE

to entry. As Peter Goldie emphasizes, the limits to contrary-to-morality imaginings are complex and variable across context. Still, there does seem to be something to the idea that imagining alien moralities can present special difficulty. Derek Matravers adds his own to a growing number of solutions to the puzzle (see also Walton 1994a; Gendler 2000; Currie 2002). It may be that the phenomenon is somewhat more complex than has so far been recognized, and that “imaginative resistance” names a family of problems that lack a single solution. Here I will indicate two so-far unexplored directions for further work. The first suggestion is that we take a closer look at what philosophical attempts to analyze the notion of valuing might tell us about our problem. As a way into this project, I offer a suggestion that might prompt useful inquiry, though I regard the proposal as unsatisfactory, for reasons that will emerge. Valuing p is not the same as desiring p, since it makes sense to say that someone desires what he does not value, and fails to desire what he does value. Perhaps a morally ideal agent will desire all that he values and nothing that he disvalues (leaving a middle, value-neutral ground where he is free to desire as he pleases), but surely agents who are less than morally ideal can have values. Frankfurt and Lewis have argued that valuing is desiring to desire: second order desiring (Frankfurt 1982; Lewis 1989). But this position reduces the clash between value and desire to a clash between one kind of desire and another; this leaves us with no obvious way to show that the clash between value and desire ought rationally to be resolved in favor of the valuing (Smith 1994: sect. 5.7). The view that valuing is desiring of some kind has been said to have the advantage that it explains the internal, conceptual connection there is between values and desires. However, to the extent that there is such a connection, it is explicable on the view that values are beliefs. The idea is that valuing p is believing p desirable, and that what it is desirable to do is what we would do if we were fully rational (Smith 1994: sect. 5.8). In that case, my valuing p means that I rationally should desire to p. Suppose now I confront a story in which it is fictional that female infanticide is morally right. I am invited, therefore, to imagine that it is right. In being thus invited, I am to imagine female infanticide being desirable, and hence to imagine that, were I fully rational, I would desire to act so as to promote female infanticide. What might be difficult about this? What I am being asked to imagine is a conditional: that I would desire this, were I fully rational. What barriers might there be to imagining this conditional? Presumably I am to imagine evaluating the conditional, coming thereby to find it correct. To really evaluate the conditional, I am to imagine myself as fully rational and then see whether, from within the imagined position of full rationality, I desire female infanticide. But the problem is that to imagine evaluating the conditional, I need to imagine imagining myself as fully rational. And this kind of iterated imagining tends to collapse: imagining imagining p collapses into imagining p. Take the novel Five Roundabouts to Heaven by John Bingham, in which the narrator tells us of his friend’s attempt to murder

THE CAPACITIES TO PRODUCE AND CONSUME ART 301

his (the friend’s) wife. But the narrator makes it clear that there was much at stake in this for him, and confesses at the beginning to describing events he could not really know about. The story seems, primarily, to be a narrative of the narrator’s own imaginings about what happened; it is a story according to which it is fictional that various things are imagined. Yet readers do not, I think, imagine these events as imaginings; they simply imagine the occurrence of the events as imagined by the narrator. The thought, then, is this. The imagined evaluation of the conditional collapses into the actual evaluation of it. Since I evaluate it as false, I must similarly reject it in the case where I am asked merely to imagine it. While the idea of transitive collapse of imagining may be a useful one to explore here, the current proposal cannot be the solution to the problem. One reason for this is that we cannot be expected to evaluate the desirability of some outcome by the process, described above, of imagining ourselves fully rational. In the case of a conditional like “if I had more money I would be happier” it does not seem unreasonable to suppose that what we do is to take on, in imagination, the idea that we have more money and then see whether, from within this imagination-based stance, we feel happier. Here the adjustment that I make to my mental set by taking on the conditional is effected simply by adding the content of the antecedent to the stock of my beliefs, but as something imagined rather than as something believed. The adjustment I bring about corresponds to a situation in which I believe that I have more money, not to a situation in which I actually have more money. But in this case the difference does not matter, for what matters about having money is the beliefs and other psychical states that go with having it, not the mere possession; nobody thinks that having money and not knowing that you do will make you happier. But in the case where the antecedent is “I am fully rational”, the relevant “taking on” cannot be effected by imagining that I am fully rational. For even it is true that, were I to believe I was fully rational I would desire p, that is no evidence for the claim that, were I fully rational, I would desire p. The relevant adjustment that I must make to assess the conditional is therefore not an adjustment that brings me into a state that is the imaginative counterpart of believing myself fully rational; it is an adjustment that brings me into a state that is the imaginative counterpart of being fully rational. But what is that state? Plausibly, there is no difference between rationality as applied to beliefs and rationality applied to imagination. In that case the imaginative counterpart of being fully rational is, just, being fully rational. And someone who is not already fully rational cannot make themselves so by an imaginative act. Earlier the suggestion was to take the philosophical high road, and to look for a metaphysical/conceptual solution to the problem. A quite different approach would be to look for some contingent feature of the human mind on which to lay the blame. People have argued that imaginative resistance to alien morals cannot be just a species of resistance to contradictions—even assuming that moral truths are conceptually necessary truths—because we seem to be unresistant to at least certain kinds of contradictions (Gendler 2000; Currie 2002; cf. Stock chapter

302 GREGORY CURRIE

6, this volume). However, we might wonder whether the mental searchlight that looks for incoherence in our conceptual constructions shines particularly brightly on the moral cases. There is some evidence that we tend to approach questions about moral rules and obligations in rather special ways (Cosmides and Tooby 1992). Given the probable significance of social rules, their observance and violation in our recent evolutionary past, it would not be surprising to find that we are apt to dedicate special attentional and other resources when moral questions are in focus; it has even been suggested that there is a mental module dedicated to deontic reasoning (Cummins 1996; cf. Chater and Oaksford 1996). It may therefore be that it is harder to distract us from what we take to be the impossibility of a radically revisionary moral statement than it is in the case of necessities in other realms such as arithmetic. The natural history of the imagination So far, I have considered aspects of the imagination that concern us when we think about ways that art is consumed. But imagination plays a vital role in the production of art. I have already suggested that an evolutionary approach can help us individuate imaginative kinds. Recent work on the creative use of imagination, some of it by philosophers, also takes an evolutionary approach (Mithen 1996; Power 1999; Carruthers 2002). While this strikes me as a very welcome development, there are assumptions in this area that are questionable. One is that considerations of natural selection push us in the direction of thinking about imagination and other mental capacities in modular terms. Another is that the basic (and unexplained) fact of cultural prehistory is the so-called big bang at around 40,000 years before the present (BP): a sudden leap forward that combined an unprecedented symbolic flowering with new forms of tool use and social organization. Thinking in the terms provided by evolutionary psychology, we might look for some significant change in an imagination module: a special-purpose device with limited access to other systems, dedicated to solving problems in some narrowly circumscribed domain (see Barkow et al. 1992). But imagination looks a poor candidate for a modularized capacity, given its applicability to a wide range of problems from planning, through mind reading to the construction of fictional narratives. And imaginative tasks seem to draw on potentially unlimited supplies of knowledge; think how we elaborate a solution to the mystery story by bringing together forensic evidence and facts about apparently unrelated events in the characters’ pasts (see Currie and Sterelny 2000). There are complex issues I cannot go into here concerning accounts of modularity that dispense with encapsulation; just how useful these accounts are is an open question (see Samuels 2000; Currie in preparation). A different line of thought is Steven Mithen’s proposal that the cultural big bang was the outcome of a breakdown in modularity that allowed us to combine social, natural-world and technological thinking: all systems that, according to Mithen,

THE CAPACITIES TO PRODUCE AND CONSUME ART 303

had to that point developed as modules. The result was a huge and sudden increase in mental flexibility that enabled us to design tools for specific purposes, see the natural world as suffused with intentionality and—a less attractive development—see our conspecifics as means to technological ends (Mithen 1996). Mithen’s picture of the pre-big bang mind as modular depends on the idea that modularity is the best explanation of the (assumed) uncreativity of the mind at this time. Why should we believe this? A non-modular but not very powerful mind might very well exhibit the lack of creativity Mithen claims to see in the record. Fodor famously argued that, while perceptual systems are modular, belief fixation is not; it is of the nature of belief fixation in humans to be sensitive to evidence from all over the place (Fodor 1983). But Fodor surely did not mean by this that, when we form a belief that p, we consider how all of our other beliefs might bear on p; a mind that worked like that would never get anything done. A non-modular Fodorian central processor is a system in which there are no architectural barriers between beliefs from one domain (e.g. technology) and beliefs from another one (e.g. folk psychology). But how effectively beliefs from one area are brought to bear on another, perhaps only distantly related, area will depend on features that might vary considerably from one non-modular mind to another. The less computationally powerful the mind is, the less we will expect it to exhibit the bringing together of bits of information that are not obviously related. And creative problem solving depends on more than merely the capacity to bring disparate beliefs together, however novel the bringing together may be; a creature may have the capacity to bring disparate thoughts together but never do anything creative with them. If our ancestors were pretty uncreative folk, that tells us little about the extent to which their minds were modular. Thinking about the evolution of creativity has been dominated in recent years by a view about what the paleological record tells us. According to this view, the shift to a creative mind was both dramatic and sudden. While stone-tool use developed very slowly over more than two million years, between 38,000 and 35, 000 BP we see emerging beads and pendants of stone, animal teeth, ivory and marine fossils. Carved animals, what may be animal-human hybrids, and the socalled Venus figurines follow, with the marvels of Chauvet cave at 32,000–29,000 BP. Rock art in the Northern Territory of Australia may date from 20,000 BP. Chauvet is particularly staggering because the creatures depicted there are done with a vibrant and sophisticated realism not found again until the Renaissance. It is as if human kind somehow lost a language underpinned by innate capacities and which everyone had picked up with ease, and then had to go through a slow and painful process of language reacquisition via general learning. The standard timetable, placing the big bang at 40,000 BP, is problematic, partly because it is generally accepted that anatomically modern humans with brains no smaller than our own were in existence at least by 100,000 years ago, and partly because the signs of cultural change are evident in highly dispersed populations, inviting implausible stories of parallel evolution. How strong is the

304 GREGORY CURRIE

evidence for a big bang at 40,000 BP? There is certainly evidence of proto-artistic activity well before this. McBrearty and Brooks argue for a radical reassessment of the record (McBrearty and Brooks 2000). They claim that the elements of culture visible in Europe and elsewhere at 40,000 emerged slowly in Africa from about 250,000, with sophisticated stone-tool use and the processing of pigments, probably for bodily adornment, in place by this time. (Others have suggested that pigment use indicates a symbolic cultural tradition going back at least 100,000 years; see Watts 1999.) Bone tools, including some with barbs, and the incision of surfaces for notational purposes appear around 100,000 BP. Other aspects of the slow dawning of modernity include evidence of the survival into adulthood of a handicapped infant at 125,000 BP; capacity to bring down large and dangerous game; exchange of goods between distant groups. Indeed, it is difficult not to see aesthetic forces driving the construction of much older artifacts such as the wildly nonfunctional hand axe from Hoxne, dated at 400,000, in the British Museum: a bravura of twisted symmetry. Yet even on the skeptical account of McBrearty and Brooks, pictorial representations do not appear until close to 40,000 BP, and their own date for depiction in Africa is much earlier than those of some other authorities (McBrearty and Brooks 2000:526). This leaves us with the puzzle of the apparently simultaneous emergence of this crucial development in dispersed populations. If we take the archaeological record at face value, it looks as if we still have at least a depictive big bang to explain. And this narrowing of focus may constitute progress. Whatever the historical facts, we do well to distinguish the forms of cultural expression, avoiding the use of a generalized category of the symbolic, which has resulted in some distortion of our explanatory ambitions. It is not useful, for example, to describe pretense as “symbolic play”, since pretend acts do not stand for or symbolize the acts they are pretended versions of (Currie 1995b). Depiction, pretend play, bodily adornment, belief in the supernatural and the marking out of social rank depend on distinct and to some extent independent capacities—or so we should assume, unless there are very clear indications to the contrary. Figuration depends on the capacity to see one thing in another, a capacity no doubt first exercised in response to natural objects such as frost and clouds, wherein one can often see a face (Wollheim 1987: chapter 1). It is natural to say that the first intentional depictions were attempts to reproduce the experiences of seeing-in. One puzzle here is that the capacity to see one thing in another must surely be rooted in a contingent fact about the tuning of our perceptual recognition capacities. For good reasons, the exercise of these capacities is apt to produce false positives; under the influence of a vaguely lion-like stimulus, your visual recognition system can tell you “lion ahead” when there is none (and, indeed, when you do not believe there is one; this is the modularity of perception again). This means that it is possible to create artifacts that will have this effect. Now face recognition seems to be on a hair-trigger; overwhelmingly, it is faces that we see in the moon, the clouds and on stained walls. However, faces are just what we do not find in the earliest depictions so far discovered.

THE CAPACITIES TO PRODUCE AND CONSUME ART 305

We cannot account for the firmly embedded practice of ice-age depiction in terms of seeing-in alone. For one thing the capacity to see an object in a marked surface is phylogenetically ancient; chimps, with whom we share a common ancestor at around 6 million BP, respond to pictures in ways that indicate that they see what is there represented. For another, depiction is a manual activity (see Sheets-Johnstone 1990, chapter 9), and the transition from seeing objects in surfaces to intentionally marking surfaces in such a way as to achieve the same effect is nontrivial. The transition is especially problematic since the means chosen to achieve the effect was the use of lines, a use not suggested by natural displays such as clouds and frost. That lines will deliver the experience of seeing-in depends on a second contingent fact about our perceptual recognition capacities: that perceptual recognition involves several stages, one of which is edge extraction. Artificial visual systems have been developed which perform this function, and it is found that the result of edge detection operations applied to a visual scene corresponds closely to a line drawing of it. This suggests that line drawings exploit the modularity of vision by being pre-packaged for perception; they are effortlessly recognizable because they have done for the scene represented something which the visual system would do anyway if exposed to the scene itself (Hayes and Ross 1995). The crucial step may have been the accidental creation of lined surfaces that have the effect of enabling seeing-in to take place. Here uniquely human capacities are relevant. The capacity of humans to point, not displayed by other primates in their natural surroundings, is important here since pointing is easily translated into drawing the index finger across a surface. Once a depictive effect is achieved by accident, the agent—or another agent—has to be able to reproduce the performance. And cultural embedding of the practice requires spreading and elaborating it through imitation. Nonhuman primates engage in behaviors that enable them to learn from one another and which can seem to be imitative. But the evidence suggests that genuinely imitative learning, which depends on the capacity to understand the goals of the agent and not merely their physical movements, is either uniquely human or is so unusual or limited in other ape communities that it cannot sustain an accumulation of knowledge and skills (Boyd and Richerson 1996; Tomasello 2000: chapter 2). We still have an explanatory gap, since pointing and imitation are much older than depiction seems to be; an imitative capacity was required for the establishment of all the cultural phenomena we find much earlier than 40,000 BP. What cannot be ruled out is a late enrichment of that capacity, carrying human capacities for cultural learning to a new level. Acknowledgment My thanks are due to Jill Cook of the British Museum and to Peter Carruthers for discussion.

BIBLIOGRAPHY

Allen, R. (1995) Projecting Illusion: Film Spectatorship and the Impression of Reality, New York: Cambridge University Press. ——(1997) “Looking at Motion Pictures,” in R.Allen and M.Smith (eds) Film Theory and Philosophy, Oxford: Oxford University Press. Aristotle (1958) On Poetry and Style, trans. G.M.A.Grube, Indianapolis: Hackett. Aristotle (1987) The Poetics, ed. and trans. S.Halliwell, London: Duckworth. Austen, J. (1992) Sense and Sensibility, New York: Everymans Library. Bal, M. (1997) Narratology: Introduction to the Theory of Narrative, 2nd edn, Toronto: University of Toronto Press. Barkow, J., L.Cosmides, and J.Tooby (eds) (1992) The Adapted Mind, New York: Oxford University Press. Barnes, H. (1959) The Literature of Possibility, Lincoln: University of Nebraska Press. Batson, C. (1991) The Altruism Question: Toward a Social Psychological Answer, Hillsdale NJ: Lawrence Erlbaum. ——and L.Shaw (1991) “Evidence for Altruism: Towards a Pluralism of Prosocial Motives,” Psychological Inquiry 2:107–22. Beardsmore, R.W. (1984) “Literary Examples and Philosophical Confusion,” in A.P.Griffiths (ed.) Philosophy and Literature, Cambridge: Cambridge University Press. Bechara, A., A.Damasio, H.Damasio, and S.Anderson (1994) “Insensitivity to Future Consequences Following Damage to Human Prefrontal Cortex,” Cognition 50:7–15. Berbaum, K. and C.C.Chung (1981) “Müller-Lyer Illusion Induced by Imagination,” Journal of Mental Imagery 5:125–8. Blum, L. (1994) Moral Perception and Particularity, Cambridge: Cambridge University Press. Booth, W. (1961) The Rhetoric of Fiction, Chicago: Chicago University Press. Bordwell, D. (1989) Making Meaning, Cambridge: Harvard University Press. ——and K.Thompson (1990) Film Art: An Introduction, New York: McGraw-Hill. Bouldin, P. and C.Pratt (2001) “The Ability of Children with Imaginary Companions to Differentiate Between Fantasy and Reality,” British Journal of Developmental Psychology 19:99–114. Bourchier, A. and A.Davis (2000a) “Individual and Developmental Differences in Children’s Understanding of the Fantasy-Reality Distinction,” British Journal of Developmental Psychology 18:353–68. ——(2000b) “The Influence of Availability and Affect on Children’s Pretence,” British Journal of Developmental Psychology 18:137–56. Bouwsma, A.K. (1954) “The Expression Theory of Art,” in W.Elton (ed.) Aesthetics and Language, Oxford: Blackwell. Boyd, R. and P.Richerson (1996) “Why Culture Is Common but Cultural Evolution Is Rare,” Proceedings of the British Academy 88:77–93. Bradley, A.C. (1974) Shakespearean Tragedy, London: Macmillan.

BIBLIOGRAPHY 307

Brook, D. (1969) “Perception and the Appraisal of Sculpture,” Journal of Aesthetics and Art Criticism 27:323–30. Bruell, M.J. and J.D.Woolley (1998) “Young Children’s Understanding of Diversity in Pretence,” Cognitive Development 13:257–77. Budd, M. (1989) “Music and the Communication of Emotion,” Journal of Aesthetics and Art Criticism 47:129–38. ——(1995) Values of Art: Pictures, Poetry, and Music, London: Penguin. Bullough, E. (1957) “‘Psychical Distance’ as a Factor in Art and an Aesthetic Principle,” in E.Wilkinson (ed.) Aesthetics: Lectures and Essays, Stanford: Stanford University Press. Camus, A. (1948) The Plague, trans. S.Gilbert, New York: Modern Library. Carpenter, P.A. and P.Eisenberg (1978) “Mental Rotation and the Frame of Reference in Blind and Sighted Individuals,” Perception andPsychophysics 23:117–24. Carpenter, R. (1960) Greek Sculpture, Chicago: University of Chicago Press. Carroll, L. (2000) The Annotated Alice: The Definitive Edition, ed. Martin Gardner, London: Penguin. Carroll, N. (1990) The Philosophy of Horror, Or Paradoxes of the Heart, London: Routledge. ——(1991) “Art, Narrative, and Emotion,” in M.Hjort and S.Laver (eds) Emotion and the Arts, New York: Oxford University Press. ——(1996a) “Prospects for Film Theory,” in D.Bordwell and N.Carroll (eds) Post-Theory: Reconstructing Film Studies, Wisconsin: University of Wisconsin Press. ——(1996b) Theorizing the Moving Image, Cambridge: Cambridge University Press. ——(1997) “Simulation, Emotions and Morality,” in G.Hoffmann and A.Hornung (eds) Emotion in Postmodernism, Heidelberg: Universitätsverlag C.Winter. ——(1998a) A Philosophy of Mass Art, New York: Oxford University Press. ——(1998b) “Art, Narrative, and Moral Understanding,” in J.Levinson (ed.) Aesthetics and Ethics, Cambridge: Cambridge University Press. ——(1999) Philosophy of Art: A Contemporary Introduction, London: Routledge. ——(2001a) “Moderate Moralism,” in Beyond Aesthetics, Cambridge: Cambridge University Press. ——(2001b) “The Wheel of Virtue: Art, Literature, and Moral Knowledge,” Journal of Aesthetics and Art Criticism 60:3–26. Carruthers, P. (2002) “Human Creativity: Its Cognitive Basis, Its Evolution, and Its Connection with Childhood Pretence,” British Journal for the Philosophy of Science 53: 225–49. ——and P.Smith (eds) (1996) Theories of Theories of Mind, Cambridge: Cambridge University Press. Ceci, S.J. and M.Bruck, (1993) “Suggestibility of the Child Witness,” Psychological Bulletin 113:403–39. ——and R.Friedman (2000) “The Suggestibility of Children,” Cornell Law Review 86: 33– 108. Chambers, D. and D.Reisberg (1985) “Can Mental Images Be Ambiguous?,” Journal of Experimental Psychology: Human Perception and Performance 11:317–28. Chater, N. and M.Oaksford (1996) “Deontic Reasoning, Modules and Innateness: A Second Look,” Mind and Language 11:191–202. Chute, C. (1985) The Beans of Egypt, Maine, New York: Warner. Cohen, J. (1999) “Holism: Some Reasons for Buyer’s Remorse,” Analysis 59:63–77.

308 BIBLIOGRAPHY

Cosmides, L. and J.Tooby (1992) “Cognitive Adaptations for Social Exchange,” in J.Barkow, L.Cosmides, and J.Tooby (eds) The Adapted Mind, New York: Oxford University Press. Cook, J. (1969) “Human Beings,” in P.Winch (ed.) Studies in the Philosophy of Wittgenstein, London: Routledge. Crane, T. (1988) “The Waterfall Illusion,” Analysis 48:142–7. Craver-Lemley, C. and A.Reeves (1992) “How Visual Imagery Interferes with Vision,” Psychological Review 99:633–49. Cummins, D.D. (1996) “Evidence for the Innateness of Deontic Reasoning,” Mind and Language 11:160–90. Currie, G. (1990) The Nature of Fiction, Cambridge: Cambridge University Press. ——(1991) “Visual Fictions,” The Philosophical Quarterly 41:129–43. ——(1995a) Image and Mind: Film, Philosophy and Cognitive Science, Cambridge: Cambridge University Press. ——(1995b) “Imagination and Simulation: Aesthetics Meets Cognitive Science,” in M.Davies and T.Stone (eds) Mental Simulation, Oxford: Blackwell. ——(1997) “The Paradox of Caring: Fiction and the Philosophy of Mind,” in M.Hjort and S.Laver (eds) Emotion and the Arts, New York: Oxford University Press. ——(1998) “Reply to my Critics,” Philosophical Studies 89:355–66. ——(2002) “Desire in Imagination,” in T.Gendler and J.Hawthorn (eds) Imagination, Conceivability and Possibility, New York: Oxford University Press. ——(forthcoming) Meeting of Minds: Thought, Imagination and Perception, Oxford: Oxford University Press. ——(in preparation) “Art and the Evolution of Mind,” Journal of Aesthetics and Art Criticism. ——and J.Jureidini (2003) “Making the World Fictional,” Monist. ——and I.Ravenscroft (2002) Recreative Minds: Image and Imagination in Philosophy and Psychology, Oxford: Oxford University Press. ——and K.Sterelny (2000) “How to Think about the Modularity of Mindreading,” Philosophical Quarterly 50:145–60. Dadlez, E. (1996) “Fiction, Emotion and Rationality,” British Journal of Aesthetics 36: 290– 304. Damasio, A. (1994) Descartes’ Error: Emotion, Reason and the Human Brain, New York: Harper Collins. ——(1999) The Feeling of What Happens: Body and Emotion in the Making of Consciousness, New York: Harcourt. ——, D.Tranel, and H.Damasio (1991) “Somatic Markers and the Guidance of Behavior: Theory and Preliminary Testing,” in H.Levin, H.M.Eisenberg, and A.L.Benton (eds) Frontal Lobe Function and Dysfunction, New York: Oxford University Press. Darwall, S. (1998) “Empathy, Sympathy, Care,” Philosophical Studies 89:261–82. Davidson, D. (1974) “Belief and the Basis of Meaning,” Synthese 27:309–23. ——(1984) “What Metaphors Mean,” in Inquiries into Truth and Interpretation, Oxford: Oxford University Press. Davies, D. (1999) “Living in the ‘Space of Reasons’: The ‘Rationality Debate’ Revisited,” International Studies in the Philosophy of Science 13:231–44. Davies, D. (2000) “Fiction,” in B.Gaut and D.M.M.Lopes (eds) Routledge Companion to Aesthetics, London: Routledge.

BIBLIOGRAPHY 309

Davies, M. and T.Stone (eds) (1995a) Folk Psychology: The Theory of Mind Debate, Oxford: Blackwell. ——(1995b) Mental Simulation: Evaluations and Applications, Oxford: Blackwell. Davies, S. (1994) Musical Meaning and Expression, Ithaca: Cornell University Press. Debellis, M. (1995) Music and Conceptualization, Cambridge: Cambridge University Press. De Renzi, E. (1982) Disorders of Spatial Exploration and Cognition, New York: Wiley. de Sousa, R. (1987) The Rationality of Emotion, Cambridge: MIT Press. ——(1997) “Fetishism and Objectivity in Aesthetic Emotion,” in M.Hjort and S.Laver (eds) Emotion and the Arts, New York: Oxford University Press. Dennett, D. (1984) Elbow Room, Cambridge: MIT Press. Devereaux, M. (1998) “Beauty and Evil: the Case of Leni Riefenstahl’s Triumph of the Will,” in J.Levinson (ed.) Aesthetics and Ethics, Cambridge: Cambridge University Press. Dickie, G. (1964) “The Myth of the Aesthetic Attitude,” American Philosophical Quarterly 1: 56–66. Dodd, J. (1995) “McDowell and Identity Theories of Truth,” Analysis 55:160–5. Epstein, E.J. (1992) “Epitaph for Jim Garrison: Romancing the Assassination,” New Yorker, November 30. Esrock, E. (1994) The Reader’s Eye, Baltimore: Johns Hopkins University Press. Estes, D., H.Wellman, and J.Wooley (1989) “Children’s Understanding of Mental Phenomena,” in H.W.Reese (ed.) Advances in Child Development and Behavior, San Diego: Academic Press. Farr, R. and S.Moscovici (eds) (1984) Social Representations, Cambridge: Cambridge University Press. Fauconnier, G. and T.Mark (1998) “Conceptual Integration Networks,” Cognitive Science 22:133–87. Feagin, S. (1996) Reading with Feeling: The Aesthetics of Appreciation, Ithaca: Cornell University Press. Finke, R.A. (1990) Creative Imagery: Discoveries and Inventions in Visualization, Hillsdale NJ: Lawrence Erlbaum. Fiske, S.T. and S.E.Taylor (1984) Social Cognition, New York: Random House. Flew, A. (ed.) (1979) A Dictionary of Philosophy, London: Pan. Fodor, J. (1983) The Modularity of Mind. An Essay on Faculty Psychology, Cambridge: MIT Press. ——(1998) Concepts, Oxford: Oxford University Press. Frankfurt, H. (1982) “Freedom of the Will and the Concept of a Person,” in G.Watson (ed.) Free Will, Oxford: Oxford University Press. Fraser, A. (2001) Marie Antoinette: The Journey, London: Weidenfeld and Nicolson. Freud, S. (1959) The Standard Edition of the Complete Psychological Works of Sigmund Freud, ed. and trans. James Strachey, London: Hogarth. Friedman, M. (2000) A Parting of the Ways: Carnap, Cassirer and Heidegger, Chicago: Open Court Press. Friend, S. (2000a) “Real People in Unreal Contexts,” in A.Everett and T.Hofweber (eds) Empty Names, Fiction and the Puzzles of Non-Existence, Stanford: CSLI Publications. Friend, S. (2000b) Review of Fiction and Metaphysics by Amie L.Thomasson, Mind 109: 997–1000. Fuchser, L.W. (1982) Neville Chamberlain and Appeasement, New York: Norton.

310 BIBLIOGRAPHY

Fuller, G. (1995) “Simulation and Psychological Concepts,” in M.Davies and T.Stone (eds) Mental Simulation: Evaluations and Applications, Oxford: Blackwell. Gaut, B. (1993) “The Paradox of Horror,” British Journal of Aesthetics 33:333–5. ——(1998) “The Ethical Criticism of Art,” in J.Levinson (ed.) Aesthetics and Ethics, Cambridge: Cambridge University Press. ——(forthcoming) “Philosophy of the Movies: Cinematic Narration,” in P.Kivy (ed.) The Blackwell Guide to Aesthetics, Oxford: Blackwell. Gelman, S.A. and H.M.Wellman (1999) “Insides and Essences: Early Understandings of the Non-Obvious,” in E.Margolis and S.Laurence (eds) Concepts, Cambridge: MIT Press. Gendler, T.S. (2000) “The Puzzle of Imaginative Resistance,” Journal of Philosophy 97: 55– 81. ——(forthcoming) “Fictional Emotions,” in M.Kieran (ed.) Contemporary Debates in Aesthetics and the Philosophy of Art, Oxford: Blackwell. Genette, G. (1980) Narrative Discourse, trans. J.Lewin, Ithaca: Cornell University Press. German, T.P. and A.M.Leslie (2001) “Children’s Inferences from ‘Knowing’ to ‘Pretending’ and ‘Believing,’” British Journal of Developmental Psychology 19:59–83. Gerrig, R. (1993) Experiencing Narrative Worlds, New Haven: Westview. Gibson, J.J. (1950) The Perception of the Visual World, Cambridge: Riverside Press. ——(1986) The Ecological Approach to Visual Perception, Hillsdale NJ: Lawrence Erlbaum. Goldie, P. (1999) “How We Think of Others’ Emotions,” Mind and Language 14:394–423. ——(2000) The Emotions: A Philosophical Exploration, Oxford: Oxford University Press. ——(2002a) “Compassion: A Natural, Moral Emotion,” in S.Doering and V.Mayer (eds) Die Moralitaet der Gefuehle, Special Issue of Deutsche Zeitschrift fur Philosophie 4, Berlin: Akademie. ——(2002b) “Emotion, Personality and Simulation,” in P.Goldie (ed.) Understanding Emotions: Mind and Morals, Aldershot: Ashgate. ——(forthcoming) “Narrative and Perspective; Values and Appropriate Emotions,” in A.Hatzimoysis (ed.) Philosophy and the Emotions, Cambridge: Cambridge University Press. Goldman, A. (1993) “The Psychology of Folk Psychology,” Behavioral and Brain Sciences 16: 101–8. ——(1995a) “Interpretation Psychologised,” in M.Davies and T.Stone (eds) Folk Psychology: The Theory of Mind Debate, Oxford: Blackwell. ——(1995b) “Empathy, Mind and Morals,” in M.Davies and T.Stone (eds) Mental Simulation: Evaluations and Applications, Oxford: Blackwell. Golomb, C. and R.Kuersten (1996) “On the Transition from Pretence Play to Reality: What Are the Rules of the Game?,” British Journal of Developmental Psychology 14: 203– 17. Goodman, N. (1976) Languages of Art, 2nd edn, Indianapolis: Hackett Gordon, R. (1992) “The Simulation Theory: Objections and Misconceptions,” Mind and Language 7:11–34. ——(1995a) “Folk Psychology as Simulation,” in M.Davies and T.Stone (eds) Folk Psychology: The Theory of Mind Debate, Oxford: Blackwell. Gordon, R. (1995b) “Simulation Without Introspection or Inference from Me to You,” in M.Davies and T.Stone (eds) Mental Simulation: Evaluations and Applications, Oxford: Blackwell.

BIBLIOGRAPHY 311

——(1996a) “Sympathy, Simulation, and the Impartial Spectator,” in L.May, M.Friedman, and A.Clark (eds) Mind and Morals, Cambridge: MIT Press. ——(1996b) “Radical Simulation,” in P.Carruthers and P.Smith (eds) Theories of Theories of Mind, Cambridge: Cambridge University Press. Greenspan, P. (1988) Emotions and Reasons: An Inquiry into Emotional Justification, London: Routledge. Hagberg, G.L. (1994) Meaning and Interpretation, Ithaca: Cornell University Press. Hamlyn, D.W. (1994) “Imagination,” in S.Guttenplan (ed.) Companion to the Philosophy of Mind, Oxford: Blackwell. Hampson, P.J. and C.Duffy (1984) “Verbal and Spatial Interference Effects in Congenitally Blind and Sighted Subjects,” Canadian Journal of Psychology 38:411–20. Harris, P.L. (1994) “Understanding Pretence,” in C.Lewis and P.Mitchell (eds) Children s Early Understanding of Mind, Hillsdale NJ: Lawrence Erlbaum. ——(1995) “From Simulation to Folk Psychology: The Case for Development,” in M.Davies and T.Stone (eds) Folk Psychology: The Theory of Mind Debate, Oxford: Blackwell. ——(2000) The Work of the Imagination, Oxford: Blackwell. —— and R.D.Kavanaugh (1993) “Young Children’s Understanding of Pretense,” Society for Research in Child Development Monographs, 231. ——, R.D.Kavanaugh, and M.Meredith (1994a) “Young Children’s Comprehension of Pretend Episodes: The Integration of Successive Actions,” Child Development 65: 16– 30. ——, A.Lillard, and J.Perner (1994b) “Commentary: Triangulating Pretence and Belief,” in C.Lewis and P.Mitchell (eds) Children s Early Understanding of Mind, Hillsdale NJ: Lawrence Erlbaum. Harvey, N. (1992) “Wishful Thinking Impairs Belief-Desire Reasoning: A Case of Decoupling Failure in Adults?” Cognition 45:141–62. Hayes, A. and J.Ross (1995) “Lines of Sight,” in R.Gregory (ed.) The Artful Eye, Oxford: Oxford University Press. Heal, J. (1998) “Co-Cognition and Off-Line Simulation: Two Ways of Understanding the Simulation Approach,” Mind and Language 13:477–98. Hegel, G.W.F. (1974) Aesthetics: Lectures on Fine Art, trans. T.M.Knox, Oxford: Oxford University Press. Herman, B. (1993) The Practice of Moral Judgment, Cambridge: Harvard University Press. Hildebrand, A. (1932) The Problem of Form in Painting and Sculpture, New York: Stechert. Hjort, M. and S.Laver (eds) (1997) Emotion and the Arts, New York: Oxford University Press. Hopkins, R. (1994) “Resemblance and Misrepresentation,” Mind 103:421–38. ——(1998) Picture, Image and Experience, Cambridge: Cambridge University Press. ——(2002) “Sculpture,” in J.Levinson (ed.) Oxford Companion to Aesthetics, Oxford: Oxford University Press. Horowitz, T. and G.Massey (1991) Thought Experiments in Science and Philosophy, Savage MD: Rowman and Littlefield. Hume, D. (1978) A Treatise on Human Nature, ed. L.A.Selby-Bigge and P.H.Nidditch, Oxford: Oxford University Press. Hume, D. (1985) “Of the Standard of Taste,” in E.F.Miller (ed.) David Hume: Essays Moral, Political, and Literary, Indianapolis: Liberty Classics. —— (1993) An Enquiry Concerning Human Understanding, Indianapolis: Hackett

312 BIBLIOGRAPHY

Humphrey, N.K. (1976) “The Social Function of Intellect,” in P.P.G.Bateson and R.A.Hinde (eds) Growing Points in Ethology, Cambridge: Cambridge University Press. Iseminger, G. (ed.) (1992) Intention and Interpretation, Philadelphia: Temple University Press. Ishiguro, H. (1966) “Imagination,” in B.Williams and A.Montefiore (eds) British Analytical Philosophy, London: Routledge. James, H. (1944) “The Author of Beltraffio,” in P.Rahv (ed.) The Great Short Novels of Henry James, New York: Carroll and Graf. John, E. (forthcoming) “Fiction and the Critique of Evaluative Concepts,” in D.A.Masolo (ed.) Philosophy and Pluralism: Recent Perspectives and Debates, Bloomington: Indiana University Press. Joyce, R. (2000) “Rational Fear of Monsters,” British Journal of Aesthetics 40:209–24. Kahneman, D., P.Slovic, and A.Tversky (eds) (1982) Judgment Under Uncertainty: Heuristics and Biases, New York: Cambridge University Press. Kant, I. (1997) The Critique of Pure Reason, trans. Paul Guyer and Allen Wood, Cambridge: Cambridge University Press. ——(2000) Critique of the Power of Judgment, trans. Paul Guyer and Eric Matthews, Cambridge: Cambridge University Press. Kerr, N.H. (1983) “The Role of Vision in ‘Visual Imagery’ Experiments: Evidence from the Congenitally Blind,” Journal of Experimental Psychology: General 112:265–77. Kieran, M. (2001) “Value of Art,” in B.Gaut and D.M.M.Lopes (eds) The Routledge Companion to Aesthetics, London: Routledge. ——(forthcoming) “Forbidden Knowledge: The Challenge of Immoralism,” in J.Bermúdez and S.Gardner (eds) Art and Morality, London: Routledge. Kivy, P. (1989) Sound Sentiment, Philadelphia: Temple University Press. ——(1997) Philosophies of Arts: An Essay in Differences, Cambridge: Cambridge University Press. Knight, D. and G.McKnight (1999) “Suspense and its Master,” in R.Allen and S.I.Gonzales (eds) Alfred Hitchcock: Centenary Essays, London: British Film Institute. Kripke, S. (1980) Naming and Necessity, Oxford: Blackwell. Lamarque, P. (1981) “How Can We Fear and Pity Fictions?” British Journal of Aesthetics 21: 291–304. ——(1991) “Essay Review of Mimesis as Make-Believe” Journal of Aesthetics and Art Criticism 49:161–6. ——and S.H.Olsen (1994) Truth, Fiction, and Literature, Oxford: Oxford University Press. Langer, S. (1953) Feeling and Form, New York: Scribner. Laurence, S. and E.Margolis (1999) “Concepts and Cognitive Science,” in E.Margolis and S.Laurence (eds) Concepts, Cambridge and London: MIT Press. LeDoux, J. (1996) The Emotional Brain: The Mysterious Underpinnings of Emotional Life, New York: Simon and Schuster. Leslie, A. (1987) “Pretense and Representation: The Origins of ‘Theory of Mind,’” Psychological Review 94:412–26. ——(1994) “Pretending and Believing: Issues in the Theory of ToMM,” Cognition 50: 211–38. Lessing, G.E. (1962) Laocoon, New York: Dover. Levinson, J. (1996a) “Musical Expressiveness,” in The Pleasures of Aesthetics, Ithaca: Cornell University Press.

BIBLIOGRAPHY 313

——(1996b) “Pleasure and the Value of Works of Art,” in The Pleasure of Aesthetics, Ithaca: Cornell University Press. ——(1997) “Emotion in Response to Art: A Survey of the Terrain,” in M.Hjort and S.Laver (eds) Emotion and the Arts, New York: Oxford University Press. ——(2001) “Who’s Afraid of a Paraphrase?,” Theoria 67:7–23. Lewis, C. and P.Mitchell (eds) (1994) Children’s Early Understanding of Mind, Hillsdale, NJ: Lawrence Erlbaum. Lewis, D. (1983) “Truth in Fiction,” in Philosophical Papers, Oxford: Oxford University Press. ——(1989) “Dispositional Theories of Value,” Aristotelian Society Supplementary Volume 63: 113–37. Lillard, A. (1993) “Young Children’s Conceptualization of Pretend: Action or Mental Representational State?,” Child Development 64:372–86. ——(1994) “Making Sense of Pretence,” in C.Lewis and P.Mitchell (eds) Children’s Early Understanding of Mind, Hillsdale NJ: Lawrence Erlbaum. Locke, J. (1975) An Essay Concerning Human Understanding, ed. P.Nidditch, Oxford: Oxford University Press. Loftus, E. (1996) Eyewitness Testimony, Cambridge: Harvard University Press. Lopes, D.M.M. (1998) “Imagination, Illusion, and Experience in Film,” Philosophical Studies 89:343–53. ——(2000) “From Languages of Art to Art in Mind,” Journal of Aesthetics and Art Criticism 58:227–31. Lyons, W. (1980) Emotion, Cambridge: Cambridge University Press. McBrearty, S. and A.Brooks (2000) “The Revolution that Wasn’t,” Journal of Human Evolution 39:453–63. McDowell, J. (1981) “Non-Cognitivism and Rule-Following,” in S.H.Holtzman and C.M.Leich (eds) Wittgenstein: To Follow a Rule, London: Routledge. ——(1994) Mind and World, Cambridge: Harvard University Press. McGinn, C. (1997) Ethics, Evil, and Fiction, Oxford: Oxford University Press. McGinn, M. (1997) Wittgenstein and the Philosophical Investigations, London: Routledge. Marks, J. (ed.) (1986) The Ways of Desire, Chicago: Precedent. Marmor, G.S. and L.A.Zaback (1976) “Mental Rotation by the Blind: Does Mental Rotation Depend on Mental Imagery?,” Journal of Experimental Psychology: Human Perception and Performance 2:515–21. Martin, D.F. (1976) “The Autonomy of Sculpture,” Journal of Aesthetics and Art Criticism 34: 273–86. Martin, M. (1995) “Imagination,” in T.Honderich (ed.) The Oxford Companion to Philosophy, Oxford: Oxford University Press. Martin, M.G.F. (1998) “An Eye Directed Outward,” in C.Wright, B.C.Smith, and C.MacDonald (eds) Knowing Our Own Minds, Oxford: Oxford University Press. Matravers, D. (1991) “Who’s Afraid of Virginia Woolf?” Ratio (n.s.) 4:25–37. ——(1997) “The Paradox of Fiction: The Report Versus the Perceptual Model,” in M.Hjort and S.Laver (eds) Emotion and the Arts, New York: Oxford University Press. ——(1998) Art and Emotion, Oxford: Oxford University Press. Medin, D.L. (1989) “Concepts and Conceptual Structure,” American Psychologist 44: 1469– 81. Mele, A. (1987) Irrationality: An Essay onAkrasia, Self-Deception, and Self-Control, New York: Oxford University Press.

314 BIBLIOGRAPHY

Merleau-Ponty, M. (1962) The Phenomenology of Perception, trans. C.Smith, London: Routledge. Milgram, S. (1963) “Behavioural Study of Obedience,” Journal of Abnormal and Social Psychology 67:371–8. ——(1974) Obedience to Authority: An Experimental View, New York: Harper and Row. Miller, G. (1997) “Protean Primates,” in A.Whiten and R.Byrne (eds) Machiavellian Intelligence II: Extensions and Evaluations, Cambridge: Cambridge University Press. ——(1974) Obedience to Authority: An Experimental View, New York: Harper and Row. ——(1999) “Sexual Selection for Cultural Displays,” in R.Dunbar, C.Knight, and C.Power (eds) The Evolution of Culture, Edinburgh: Edinburgh University Press. Mithen, S. (1996) The Prehistory of the Mind, London: Thames and Hudson. Moran, R. (1989) “Seeing and Believing: Metaphor, Image and Force,” Critical Inquiry 16: 87–112. ——(1994) “The Expression of Feeling in Imagination,” Philosophical Review 103: 75– 106. Morreall, J. (1993) “Fear Without Belief,” Journal of Philosophy 90:359–66. Morrison, T. (1970) The Bluest Eye, New York: Washington Square. Murdoch, I. (1992) Metaphysics as a Guide to Morals, New York: Allen Lane. ——(1998) Existentialists and Mystics, New York: Allen Lane. Nabokov, V (1953) Pnin, New York: Vintage. Neill, A. (1991) “Fiction and Make-Believe,” Journal of Aesthetics and Art Criticism 49: 47– 56. ——(1992) “Who’s ‘Afraid’ of Virginia Woolf?,” Ratio (n.s.) 5:94–7. ——(1993) “Fiction and the Emotions,” American Philosophical Quarterly 30:1–13. New, C. (1999) Philosophy of Literature, London: Routledge. Nichols, S. and S.Stich (2000) “A Cognitive Theory of Pretense,” Cognition 74:115–47. Nietzsche, F. (1967) The Birth of Tragedy, trans. W.Kaufmann, New York: Random House. Novitz, D. (1987) Knowledge, Fiction and Imagination, Philadelphia: Temple University Press. Nussbaum, M. (1990) Love’s Knowledge, New York: Oxford University Press. O’Neill, O. (1986) “The Power of Example,” Philosophy 61:5–29. Oatley, K. and M.Gholamain (1997) “Emotions and Identification: Connections Between Readers and Fictions,” in M.Hjort and S.Laver (eds) Emotion and the Arts, Oxford: Oxford University Press. Packer, M. (1989) “Dissolving the Paradox of Tragedy,” Journal of Aesthetics and Art Criticism 47:211–19. Peacocke, C. (1985) “Imagination, Experience and Possibility,” in J.Foster and H.Robinson (eds) Essays on Berkeley, Oxford: Oxford University Press. Perner, J. (1991) Understanding the Representational Mind, Cambridge: MIT Press. ——S.Baker, and D.Hutton (1994) “Prelief: The Conceptual Origins of Belief and Pretence,” in C.Lewis and P.Mitchell (eds) Children’s Early Understanding of Mind, Hillsdale NJ: Lawrence Erlbaum. Peterson, M.A., J.F.Kihlstrom, P.M.Rose, and M.L.Glisky (1992) “Mental Images Can Be Ambiguous: Reconstruals and Reference-Frame Reversals,” Memory and Cognition 20: 102–23. Piaget, J. (1945/1962) Plays, Dreams and Imitation in Childhood, trans. C.Gattegno and F.M.Hodgson, New York: Norton.

BIBLIOGRAPHY 315

Pietroski, P. (1996) “Critical Notice of McDowell’s Mind and World,” Canadian Journal of Philosophy 26:613–36. Pitino, R. and B.Reynolds (1997) Success is a Choice: Ten Steps to Overachieving in Business and Life, New York: Broadway. Plato (1961) Ion, in E.Hamilton and H.Cairns (eds) The Collected Dialogues, Princeton: Princeton University Press. Power, C. (1999) “‘Beauty Magic’: The Origins of Art,” in R.Dunbar, C.Knight, and C.Power (eds) The Evolution of Culture, Edinburgh: Edinburgh University Press. Powers, L.H. (1970) Henry James: An Introduction and Interpretation, New York: Holt, Rinehart, and Winston. Pressy, A.W. and A.E.Wilson (1974) “The Poggendorff Illusion in Imagination,” Bulletin of the Psychonomic Society 3:447–9. Price, H.H. (1960/1969) Belief, London: George Allen and Unwin. Putnam, H. (1973) “Meaning and Reference,” Journal of Philosophy 70:699–711. Radford, C. (1975) “How Can We Be Moved by the Fate of Anna Karenina?” Proceedings of the Aristotelian Society Supplementary Volume 49:67–80. Read, H. (1961) The Art of Sculpture, Princeton: Bollingen. Reese, H.W. (ed.) (1989) Advances in Child Development and Behavior, San Diego: Academic Press. Ricoeur, P. (1984–) Time and Narrative, 3 vols, trans. K.McLaughlin and D.Pellauer, Chicago: University of Chicago Press. Ridley, A. (1995) Music, Value and the Passions, Ithaca: Cornell University Press. Roberts, R.C. (1988) “What An Emotion Is: A Sketch,” Philosophical Review 97:183–209. Robertson, R. (1999) “Introduction to Freud’s ‘The Interpretation of Dreams,’” in S.Freud (ed.) The Interpretation of Dreams, Oxford: Oxford University Press. Rorty, A.O. (ed.) (1980) Explaining Emotions, Berkeley: University of California Press. Rozin, P. and C.J.Nemeroff (1991) “The Laws of Sympathetic Magic: A Psychological Analysis of Similarity and Contagion,” in J.Stigler, G.Herdt, and R.A.Shweder (eds) Cultural Psychology: Essays on Comparative Human Development, Cambridge: Cambridge University Press. Ryle, G. (1949) The Concept of Mind, London: Hutchinson. ——(1971) “Jane Austen and the Moralists,” in Collected Papers, vol. 1, New York: Barnes and Noble. Samuels, R. (2000) “Massively Modular Minds: Evolutionary Psychology and Cognitive Architecture,” in P.Carruthers and A.Chamberlain (eds) Evolution and the Human Mind, Cambridge: Cambridge University Press. Santayana, G. (1988) The Sense of Beauty: Being the Outlines of Aesthetic Theory, ed. W Holzberger and H.Saatkamp, Jr., Cambridge: MIT Press. Sartre, J.-P. (1948) The Psychology of Imagination, New York: Philosophical Library. Savile, A. (1998) “Imagination and the Content of Fiction,” British Journal of Aesthetics 38: 136–49. Schier, F. (1986) Deeper Into Pictures, Cambridge: Cambridge University Press. Schopenhauer, A. (1969) The World as Will and Representation, New York: Dover. Scruton, R. (1992) “Imagination,” in D.Cooper (ed.) Companion to Aesthetics, Oxford: Blackwell. ——(1997) The Aesthetics of Music, Oxford: Oxford University Press. Segal, S.J. and V.Fusella (1970) “Influences of Imaged Pictures and Sounds on Detection of Visual and Auditory Studies,” Journal of Experimental Psychology 83:458–64.

316 BIBLIOGRAPHY

Sinaiko, H.L. (1998) “The Ancient Quarrel: Socrates’s Critique of Poetry in Republic, Book X,” in Reclaiming the Canon, New Haven: Yale University Press. Smith, Michael (1994) The Moral Problem, Oxford: Blackwell. Smith, Murray (1997) “Imagining from the Inside,” in R.Allen and M.Smith (eds) Film Theory and Philosophy, Oxford: Oxford University Press. Sorensen, R. (1992) Thought Experiments, New York: Oxford University Press. Stecker, R. (2001) “Expression and Expressiveness in Music and Poetry,” Journal of Aesthetics and Art Criticism, 59:85–96. Stokes, A. (1934) “The Stones of Rimini,” in The Critical Writings of Adrian Stokes, London: Thames and Hudson. Strawson, P.F. (1966) The Bounds of Sense, London: Methuen. ——(1974) “Imagination and Perception,” in Freedom and Resentment and Other Essays, London: Methuen. Tanner, M. (1994) “Morals in Fiction and Fictional Morality,” Aristotelian Society Supplementary Volume 68:51–66. Tarkovsky, A. (2000) Sculpting in Time, Austin: University of Texas Press. Thomas, N.J.T. (2000) “Imagination,” in C.Eliasmith (ed.) Dictionary of Philosophy of Mind, . Thomson, J.J. (1986) “A Defense of Abortion,” in P.Singer (ed.) Applied Ethics, Oxford: Oxford University Press. Tomasello, M. (2000) The Cultural Basis of Human Cognition, Cambridge: Harvard University Press. Travers, P.L. (1962) Mary Poppins, New York: Harcourt, Brace and World. Trivedi, S. (2000) “Review of Music, Value and the Passions by Aaron Ridley,” Mind 109: 434. ——(2001) “Expressiveness as a Property of the Music Itself,” Journal of Aesthetics and Art Criticism 59:411–20. Truffaut, F. (1984) Hitchcock, New York: Simon and Schuster. Tversky, A. and D.Kahneman (1982) “Availability: A Heuristic for Judging Frequency and Probability,” in D.Kahneman, P.Slovic, and A.Tversky (eds) Judgment Under Uncertainty: Heuristics and Biases, New York: Cambridge University Press. Tye, M. (1995) 10 Problems of Consciousness, Cambridge: MIT Press. Vance, R.D. (1995) “Sculpture,” British Journal of Aesthetics 35:217–25. Velleman, J.D. (2000) “The Aim of Belief,” in The possibility of Practical Reason, Oxford: Oxford University Press. Wallace, B. (1984) “Apparent Equivalence between Perception and Imagery in the Production of Various Visual Illusions,” Memory and Cognition 12:156–62. Walker-Andrews, A.S. and R.Kahana-Kelman (1997) “The Understanding of Pretence Across the Second Year of Life,” British Journal of Developmental Psychology 17: 523– 36. Walton, K. (1973) “Pictures and Make-Believe,” Philosophical Review 82:283–319. ——(1978) “Fearing Fictions,” Journal of Philosophy 75:5–27. ——(1988) “What is Abstract About the Art of Music?,” Journal of Aesthetics and Art Criticism 46:351–64. Walton, K. (1990) Mimesis as Make-Believe, Cambridge: Harvard University Press. ——(1994a) “Morals in Fiction and Fictional Morality,” Aristotelian Society Supplementary Volume 68:27–50.

BIBLIOGRAPHY 317

——(1994b) “Listening with Imagination: Is Music Representational?,” Journal of Aesthetics and Art Criticism: 52:47–62. ——(1997) “Spelunking, Simulation, and Slime: On Being Moved by Fiction,” in M.Hjort and S.Laver (eds) Emotion and the Arts, New York: Oxford University Press. Watts, E. (1999) “The Origin of Symbolic Culture,” in R.Dunbar, C.Knight, and C.Power (eds) The Evolution of Culture, Edinburgh: Edinburgh University Press. Wellman, H. and D.Estes (1986) “Early Understanding of Mental Entities: A Reexamination of Childhood Realism,” Child Development 57:910–23. Wharton, E. (1920) The Age of Innocence, New York: Collier. Wicker, T. (1991) “Does ‘J. F.K.’ Conspire Against Reason?” New York Times, December 15. Wiggins, D. (1987) “A Sensible Subjectivism,” in Needs, Values, Truth, Oxford: Blackwell. Wilkes, K. (1988) Real People, Oxford: Oxford University Press. Williams, B. (1972) “The Truth in Relativism,” Reprinted in Moral Luck, Cambridge: Cambridge University Press, 1981, pp. 132–43. ——(1973a) “Critique of Utilitarianism,” in J.J.C.Smart and B.Williams (eds) Utilitarianism: For and Against, Cambridge: Cambridge University Press. ——(1973b) “Imagination and the Self,” in Problems of the Self, Cambridge: Cambridge University Press. ——(1981) “The Truth in Relativism,” in Moral Luck, Cambridge: Cambridge University Press. ——(1985) Ethics and the Limits of Philosophy, Cambridge: Harvard University Press. Wilson, C. (1983) “Literature and Knowledge,” Philosophy 58:489–96. Wollheim, R. (1974) On Art and the Mind, Cambridge: Harvard University Press. ——(1980) Art and its Objects, 2nd edn, Cambridge: Cambridge University Press. ——(1984) The Thread of Life, Cambridge: Harvard University Press. ——(1987) Painting as an Art, London: Thames and Hudson. Woolf, V. (1955) To the Lighthouse, New York: Harcourt Brace Jovanovich. Wright, C. (1998) “Self-Knowledge,” in C.Wright, B.C.Smith, and C.MacDonald (eds) Knowing Our Own Minds, Oxford: Oxford University Press. Wynn, T. (2000) “Symmetry and the Evolution of the Modular Mind,” in P.Carruthers and A.Chamberlain (eds) Evolution and the Human Mind, Cambridge: Cambridge University Press. Yablo, S. (1993) “Is Conceivability a Guide to Possibility?,” Philosophy and Phenomenological Research 53:1–42. Yovel, Y. (1980) Kant and the Philosophy of History, Princeton: Princeton University Press. Zimler, J. and J.M.Keenan (1983) “Imagery in the Congenitally Blind: How Visual Are Visual Images?,” Journal of Experimental Psychology: Learning, Memory and Cognition 9: 269–82.

NAME INDEX

Allen, R. 224, 229, 236, 241, 244 Aristotle 31, 65, 93, 150, 168, 176, 184, 197 Austen, J. 77, 106, 141, 172

Carroll, N. 7, 14, 18, 35, 37, 45, 64, 69, 148, 172, 254, 258, 259, 271 Carruthers, P. 302, 305 Cave, P. 105 Ceci, S.J. 135 Chambers, D. 208 Chater, N. 302 Chung, C.C. 11 Chute, C. 153 Clingon, W. 43 Cohen, J. 105 Collingwood, R.G. 193 Conant, J. 205 Cook, J. 305 Cosmides, L. 302 Crane, T. 122 Craver-Lemley, C. 217 Cummins, D.D. 302 Currie, G. 4, 10, 14, 21, 43, 56, 58, 67, 90, 95, 105, 124, 139, 218, 224, 231, 243, 258, 294, 297, 300, 304

Bal, P. 54 Barber, A. 105 Bardsley, K. 244 Barkow, J. 302 Barnes, H. 141 Batson, C.D. 81, 83 Beardsley, M. 167 Beardsmore, R.W. 142 Bechara, A. 135 Berbaum, K. 11 Blum, L. 151 Bonzon, R. 159 Booth, W. 54 Bordwell, D. 251 Bouldin, P. 128 Bourchier, A. 131 Boyd, R. 305 Brooks, A. 303 Brown, J. 272 Bruck, M. 135 Budd, M. 167, 176, 178, 182, 266, 270 Bullough, E. 199 Burge, T. 139

Damasio, A. 135, 298 Darwall, S. 80 Davidson, D. 237, 260 Davies, D. 99, 105, 224, 268 Davies, M. 117, 243, 294 Davies, S. 259, 263, 267, 270 Davis, A. 131 DeBellis, M. 11 Dennett, D. 149 De Renzi, E. 212 de Sousa, R. 14, 105, 146 Devereaux, M. 65, 67, 192 Dickens, C. 46, 173

Calhoun, C. 159 Camus, A. 141 Carnap, R. 259 Carpenter, P.A. 216 Carpenter, R. 278

318

NAME INDEX 319

Dickie, G. 167, 204 Dodd, J. 243 Dove, G. 158 Duffy, C. 217 Dummett, M. 214 Eisenberg, P. 216 Epstein, E.J. 42 Esrock, E. 222 Estes, D. 128 Farr, R. 71 Fauconnier, G. 138 Feagin, S. 69, 80, 84 Finke, R.A. 208 Fiske, S.T. 71 Flaubert, G. 190 Flew, A. 268 Fodor, J. 111, 113, 303 Frances, B. 88 Frankfurt, H. 300 Freud, S. 178 Friedman, M. 135, 259 Friend, S. 46, 105 Fusella, V. 217 Gaut, B. 14, 17, 38, 43, 46, 49, 61, 64, 172, 205, 244 Gelman, S.A. 112 Gendler Szabó, Z. 139 Gendler, T.S. 90, 104, 105, 123, 133, 135, 138, 188, 300 Genette, G. 54 Gholamain, M. 69 Gibson, J.J. 286 Gill, R.R.R. 88 Goldie, P. 54, 59, 82, 85, 294, 300 Goldman, A. 124 Goodman, N. 260, 274 Gordon, R. 124, 297 Greenspan, P. 14, 18, 33 Hagberg, G.L. 141 Hamlyn, D.W. 266, 269 Hampson, P.J. 217 Harris, P.L. 124, 127, 131, 137 Hawthorne, J. 139

Hayes, A. 305 Hegel, G.W.F. 279 Heidegger, M. 259 Herman, B. 151 Hills, D. 53 Hitchcock, A. 245, 247, 251, 254, 259 Hopkins, R. 105, 215, 223, 244, 273, 276, 279, 289 Howard, R. 35 Hume, D. 63, 90, 130, 167, 176, 188, 193, 197, 204, 247, 286 Humphrey, N.K. 298 Hursthouse, R. 292 Iseminger, G. 54 Ishiguro, H. 114 James, H. 141, 164, 166, 170, 172 John, E. 141, 223 Joyce, R. 14, 31 Jureidini, J. 297 Kahana-Kelman, R. 127 Kahneman, D. 134 Kant, I. 152, 198, 205, 218, 245, 252, 255, 286 Kavanaugh, R.D. 127 Keenan, J.M. 216 Kerr, N.H. 216 Kieran, M. 53, 60, 66, 69, 167, 223, 272, 294 Kierland, B. 176 Kivy, P. 167, 190, 221, 259, 267, 270 Knight, D. 255 Kolb, D. 272 Kripke, S. 118 Lamarque, P. 2, 7, 19, 59, 114 Langer, S. 281 Laurence, S. 110 LeDoux, J. 135 Leslie, A. 123, 128 Lessing, G.E. 273, 277 Levinson, J. 14, 19, 33, 167, 260, 266, 270, 272 Lewis, D. 2, 297, 300 Lillard, A. 124, 131

320 NAME INDEX

Locke, J. 142, 173, 189, 192 Loftus, E. 135 Lopes, D.M.M. 11, 53, 69, 88, 139, 168, 176, 205, 224, 233, 236, 241, 267, 272 Lyons, W. 14 McBreatry, S. 303 McDowell, J. 168, 243 McGinn, C. 162, 171, 175 McKnight, G. 255 Magee, B. 173 Margolis, E. 110 Marmor, G.S. 216 Martin, D.F. 111, 279, 281 Martin, M. 266 Martin, M.G.F. 215 Matravers, D. 14, 20, 31, 32, 38, 43, 51, 61, 90, 95, 105, 259, 263, 267, 269, 272, 298, 300 Matthen, M. 223 Medin, D.L. 112 Mele, A. 182 Mellor, H. 98 Merleau-Ponty, M. 283, 286 Milgram, S. 65 Miller, G. 298 Mithen, S. 302 Moore, H. 286 Moran, R. 9, 14, 18, 56, 62, 64, 69, 90, 188, 219, 260 Morreall, J. 14 Morrison, T. 141, 155 Moscovici, S. 71 Mukherji, N. 272 Mulvey, L. 240 Murdoch, I. 151, 160, 165, 173 Nabokov, V 147 Nathan, D. 272 Neill, A. 21, 31, 50 Nemeroff, C.J. 131 New, C. 265 Nichols, S. 124, 139 Nietzsche, F. 56, 178 Nolan, D. 139 Novitz, D. 142

Nussbaum, M. 142, 155, 159, 161, 164, 171, 175 Oaksford, M. 302 Oatley, K. 69 Olsen, S.H. 59 O’Neill, O. 144, 155, 157 Packer, M. 177 Papineau, D. 69 Peacocke, C. 211 Perner, J. 123, 128 Peterson, M.A. 208 Piaget, J. 127 Pietroski, P. 243 Pitino, R. 126 Plato 9, 23, 175 Power, C. 302 Powers, L.H. 170 Pratt, C. 128 Pressy, A.W. 11 Price, H.H. 130 Proust, M. 190 Putnam, H. 142 Radford, C. 6, 14, 18, 24 Ravenscroft, I. 294 Raymond, D. 272 Read, H. 278 Reeves, A. 217 Reisberg, D. 208 Reynolds, B. 126 Richards, I.A. 222 Richerson, P. 305 Ricoeur, P. 54 Ridley, A. 263, 266, 270 Robertson, R. 179 Roberts, R.C. 14 Rorty, A.O. 14 Ross, J. 305 Rozin, P. 131 Ryle, G. 130, 141, 187, 265 Rysiew, P. 223 Samuels, R. 302 Santayana, G. 177 Sartre, J.-P. 207

NAME INDEX 321

Savile, A. 59 Schier, F. 274 Schopenhauer, A. 178 Scruton, R. 259, 265, 267 Segal, S.J. 217 Shaw, L. 81 Sheets-Johnstone, M. 305 Shelley, J. 176 Sibley, F. 167 Sider, T. 139 Sinaiko, H.L. 176 Smith, M. 245, 300 Socrates 23, 175, 189 Sorensen, R. 157 Stairs, A. 272 Stecker, R. 266, 270 Sterelny, K. 302 Stich, S. 124 Stock, K. 105, 300 Stokes, A. 278 Stone, T. 117, 294 Strawson, P.P. 198, 243, 286 Suppe, F. 272

128, 136, 184, 195, 199, 218, 266, 270, 276, 284, 294, 298 Weil, S. 165 Wellman, H.M. 112, 128 Wharton, E. 141 Wicker, T. 42, 47 Wiggins, D. 168 Wilkes, K. 149 Williams, B. 93, 100, 105, 159, 165, 168, 173, 175, 210, 269 Williams, C. 187 Wilson, A.E. 11 Wilson, C. 141 Wittgenstein, L. 225, 268 Wollen, P. 240 Wollheim, R. 56, 85, 122, 240, 268, 274, 287, 304 Wooley, J. 128 Woolf, V 145, 149, 152, 156 Wright, C. 106

Tanner, M. 100 Tarkovsky, A. 245, 258 Taylor, S.E. 71 Thomas, N.J.T. 265 Thompson, K. 251 Thomson, J. 159, 165, 173 Tolstoy, L. 5, 172, 190, 196 Tomasello, M. 305 Tooby, J. 302 Trivedi, S. 259, 263, 266, 268, 270 Turner, M. 138 Tversky, A. 134 Twain, M. 190 Tye, M. 118

Zaback, L.A. 216 Zimler, J. 216 Zinkin, M. 245

Vance, R.D. 276 Velleman, J.D. 133, 139 Viner, L. 158 Walker-Andrews, A.S. 127 Wallace, B. 11 Walton, K. viii, 6, 14, 16, 19, 21, 34, 49, 53, 55, 64, 90, 95, 99, 105, 119, 124,

Yablo, S. 118 Yovel, Y. 257