Você está na página 1de 432

CONTEMPL AT ING ART

This page intentionally left blank

Contemplating Art
Essays in Aesthetics
JE RRO LD LEV INSON

CLARENDON PRESS OXFORD

Great Clarendon Street, Oxford Oxford University Press is a department of the University of Oxford. It furthers the Universitys objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With ofces in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trademark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York Jerrold Levinson 2006 The moral rights of the author have been asserted Database right Oxford University Press (maker) First published 2006 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose this same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Levinson, Jerrold. Contemplating art / Jerrold Levinson. p. cm. Includes index. ISBN-13: 9780199206179 (alk. paper) ISBN-10: 0199206171 (alk. paper) ISBN-13: 9780199206186 (alk. paper) ISBN-10: 019920618X (alk. paper) 1. Aesthetics. 2. ArtPhilosophy 3. Arts. I. Title. BH39.L492 2006 701 .17dc22 2006016276 Typeset by Laserwords Private Limited, Chennai, India Printed in Great Britain on acid-free paper by Biddles Ltd., Kings Lynn, Norfolk ISBN 019920618X 9780199206186 ISBN 0199206171 (Pbk.) 9780199206179 (Pbk.) 1 3 5 7 9 10 8 6 4 2

Contents
Acknowledgements Introduction I. ART 1. The Irreducible Historicality of the Concept of Art 2. Artworks as Artifacts 3. Emotion in Response to Art 4. Elster on Artistic Creativity II. MUSIC 5. Sound, Gesture, Space, and the Expression of Emotion in Music 6. Musical Expressiveness as Hearability-as-Expression 7. Nonexistent Artforms and the Case of Visual Music 8. Music as Narrative and Music as Drama 9. Film Music and Narrative Agency 10. Evaluating Music 11. Musical Thinking 12. Musical Chills III. PICTURES 13. Wollheim on Pictorial Representation 14. What is Erotic Art? 15. Erotic Art and Pornographic Pictures 239 252 259 77 91 109 129 143 184 209 220 13 27 38 56 vii 1

vi

Contents

IV. INTERPRETAT ION 16. Two Notions of Interpretation 17. Whos Afraid of a Paraphrase? 18. Hypothetical Intentionalism: Statement, Objections, and Replies V. AE ST HE T IC PROPE RT IE S 19. Aesthetic Properties, Evaluative Force, and Differences of Sensibility 20. What Are Aesthetic Properties? VI. HISTORY 21. Schopenhauers Aesthetics 22. Humes Standard of Taste: The Real Problem VII. OT HER MAT TERS 23. The Concept of Humor 24. Intrinsic Value and the Notion of a Life Index 389 400 419 355 366 315 336 275 288 302

Acknowledgements
The Irreducible Historicality of the Concept of Art, British Journal of Aesthetics 42 (2002): 36779. Artworks as Artifacts, in E. Margolis and S. Laurence (eds.), Creations of the Mind (Oxford: Oxford University Press, 2006). Emotion in Response to Art, (as Emotion in Response to Art: A Survey of the Terrain), in M. Hjort and S. Laver (eds.), Emotion and the Arts (Oxford: Oxford University Press, 1997), 2034. Elster on Artistic Creativity, in B. Gaut and P. Livingston (eds.) The Creation of Art (Cambridge: Cambridge University Press, 2003), 23556. Sound, Gesture, Space, and the Expression of Emotion in Music (as Sound, Gesture, Spatial Imagination, and the Expression of Emotion in Music), European Review of Philosophy 5 (2002): 13750. Musical Expressiveness as Hearability-as-Expression, in M. Kieran (ed.), Contemporary Debates in Aesthetics and the Philosophy of Art (Oxford: Blackwell, 2005), 192206. Nonexistent Artforms and the Case of Visual Music, in A. Haapala, J. Levinson, and V. Rantala (eds.), The End of Art and Beyond: Essays after Danto (Atlantic Highlands, NJ: Humanities Press, 1997), 12239. Music as Narrative and Music as Drama, Mind and Language 19 (2004): 42841. Film Music and Narrative Agency, in D. Bordwell and N. Carroll (eds.), PostTheory: Reconstructing Film Studies (Madison: University of Wisconsin Press, 1996), 25488. Evaluating Music, in P. Alperson (ed.), Musical Worlds: New Directions in the Philosophy of Music (University Park, PA: Pennsylvania State University Press, 1998), 93108. [An earlier version of this essay appeared in Revue Internationale de Philosophie 198 (1996): 593614.] Musical Thinking, Midwest Studies 27 (2003): 5968. Musical Chills. [Earlier versions of this essay were Musical Frissons, Revue Francaise dEtudes Amricaines 86 (2000): 6476; and Musical Chills and Oth e er Delights of Music, in J. Davidson (ed.), The Music Practitioner (Aldershot: Ashgate Publishing, 2004), 33551.] Wollheim on Pictorial Representation, Journal of Aesthetics and Art Criticism 56 (1998): 22733. What is Erotic Art, (as Erotic Art), in E. Craig (ed.), The Routledge Encyclopedia of Philosophy (London: Routledge, 1998), 4069.

viii

Acknowledgements

Erotic Art and Pornographic Pictures, Philosophy and Literature 29 (2005): 22840. Two Notions of Interpretation, in A. Haapala and O. Naukkarinen (eds.), Interpretation and its Boundaries (Helsinki: Helsinki University Press, 1998), 221. Whos Afraid of a Paraphrase?, Theoria 67 (2001): 723. Hypothetical Intentionalism: Statement, Objections, and Replies, in M. Krausz (ed.), On the Single Right Interpretation (University Park, PA: Pennsylvania State University Press, 2002), 30918. Aesthetic Properties, Evaluative Force, and Differences of Sensibility, in E. Brady and J. Levinson (eds.), Aesthetic Concepts: Essays after Sibley (Oxford: Oxford University Press, 2001), 6180. What Are Aesthetic Properties?, Proceedings of the Aristotelian Society. Supplement 78 (2005): 21127. Schopenhauers Aesthetics, (as Schopenhauer, Arthur), in M. Kelly (ed.), The Encyclopedia of Aesthetics (New York: Oxford University Press, 1998), 24550. Humes Standard of Taste: The Real Problem, Journal of Aesthetics and Art Criticism 60 (2002): 22738. The Concept of Humour, (as Humour), in E. Craig (ed.), The Routledge Encyclopedia of Philosophy (London: Routledge, 1998), 56267. Intrinsic Value and the Notion of a Life, Journal of Aesthetics and Art Criticism 62 (2004): 31929.

Introduction
Contemplating Art is the third of my essay collections in philosophy of art, following clearly in the line of Music, Art, and Metaphysics (1990) and The Pleasures of Aesthetics (1996). All three volumes are situated in what may be called mainstream analytic aesthetics, or aesthetics in the tradition of analytic philosophy. The present volume brings together the bulk of my work in this vein in the past ten years, and contains twenty-four essays, making it considerably larger than its predecessors. That it covers a decade of work accounts in part for its size, but also relevant is the inclusion of one essay, Film Music and Narrative Agency, that is almost a small book in itself. I have grouped the essays into seven parts, on roughly thematic grounds. Part I contains four essays on art in general, raising issues in art theory not closely tied to a particular artform. Part II, the longest in the book, contains essays dealing with philosophical problems specic to music, the art that has always been my principal occupation as an aesthetician. Part III brings together three essays that concern pictorial art, while Part IV brings together three essays that concern interpretation, and more particularly, the interpretation of literature and literary language. Part V consists of two essays on the nature of aesthetic properties, the sort of properties exhibited prominently, if not exclusively, by works of art, while Part VI consists of two essays that address issues in historical aesthetics. Finally, Part VII contains essays on two topics, humor and intrinsic value, falling somewhat outside the scope of aesthetics as usually conceived, though their relevance to central issues in aesthetics should nevertheless be apparent.
For an overview of that mainstream the reader is invited to consult J. Levinson (ed.), The Oxford Handbook of Aesthetics (Oxford: Oxford University Press, 2003), though work outside the mainstream also receives extensive coverage. It was in fact published as such in France, under the title La musique de lm: ction et narration (Pau: Presses Universitaires de Pau, 1999).

Introduction

In the opening essay, The Irreducible Historicality of the Concept of Art, I return once more to the intentional-historical theory of arthood I have championed since my rst paper on the topic in 1979. After succinctly restating the account, according to which, reduced to essentials, arthood is a matter of being projected for regard or treatment as some earlier artworks, or what are taken to be such, are or were correctly regarded or treated, I underline in particular the historical element in that account, which captures an inescapable aspect of the modern concept of art, and which dooms to inadequacy any purely formal or functional account of arthood. Most of the essay is devoted to responding to recent reservations about the intentionalhistorical theory, though since of the making of objections there is no end, I harbor no illusion of having responded to all the reservations in its regard to be found in the current literature. The second essay, Artworks as Artifacts, is concerned with that same account of arthood, but here it is the nature of the artifactuality of artworks presupposed by the account that is the focus of attention. I develop my ideas on the artifactuality of artworks in counterpoint with recent contributions on the subject by Paul Bloom and Amie Thomasson. Against Bloom, who seeks to extend the intentionalhistorical account to all artifacts, I defend the claim that artworks remain a distinctive sort of artifact in possessing, perhaps alone of all artifact kinds, only intentional-historical necessary conditions. Against Thomasson, who maintains that artifact-making necessarily involves a substantive conception of what is being made, I defend the claim that the conception of artwork necessarily involved in art-making, although not without content, is about as insubstantive as an object concept can be. Emotion in Response to Art is a survey essay on the range of philosophical problems that can be encompassed under that rubric. It details ve such problems, according most of its attention to the rst two of those, namely, the nature of the emotional responses had to art, and the puzzle of emotional responses to ctional entities known to be ctional (what is often labeled the
Dening Art Historically, British Journal of Aesthetics 19 (1979): 23250, reprinted in Music, Art, and Metaphysics (Ithaca: Cornell University Press, 1990). This was followed by two further essays expounding and defending the theory: Rening Art Historically (1989), reprinted in Music, Art, and Metaphysics, and Extending Art Historically (1993), reprinted in The Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996). Two substantial critiques that appeared after the essay was published, to which I thus do not there respond, are Nigel Warburton, The Art Question (London: Routledge, 2003), ch. 4, and Victor Yelverton Haines, Recursive Chaos in Dening Art Recursively, British Journal of Aesthetics 44 (2004): 7383.

Introduction

paradox of ction). But attention is also given to the puzzle of how people derive satisfaction from art expressive or evocative of negative emotion (what is often labeled the paradox of tragedy), and to the question of how abstract works of art manage to express or evoke emotions at all. Elster on Artistic Creativity is a study of what of general import might be said about the processes or principles of creativity in art, conducted through an examination of a thought-provoking discussion of artistic creativity by the social theorist Jon Elster. I take issue with Elsters account of creativity in art as simply a matter of optimizing choice within constraints following an earlier stage of choice of constraints, and also take issue with some of the evaluative consequences, both general and specic, that Elster draws from his account. All the essays in Part II concern principally the art of music, and most of them bear connections to earlier writings of mine. The rst two essays are both concerned with the problem of musical expressiveness, how it is to be analyzed and what it is to perceive or experience it. Musical Expressiveness as Hearability-as-Expression is continuous with an earlier essay entitled simply Musical Expressiveness, and defends the analysis of that phenomenon reected in its title, according to which music is expressive of an emotion or other mental state insofar as it induces us to hear it as the personal or personlike expression of that mental state. Along the way various competing theories of musical expressiveness, notably those of Malcolm Budd, Stephen Davies, Robert Stecker, and Roger Scruton, are submitted to critical examination. Sound, Gesture, Space, and the Expression of Emotion in Music, which in addition to drawing on Musical Expressiveness also reworks material from an even earlier essay, Authentic Performance and Performance Means, emphasizes rst the role of grasp of musical gesture in the grasp of musical expressiveness, and second, the role of spatial imagination in the grasp of musical gesture. Nonexistent Artforms and the Case of Visual Music is the oldest of the essays reprinted here, in terms of its date of composition, having been written for a conference on the future of art held in Lahti, Finland in 1990. The rst, somewhat fanciful, half of the essay is not specically concerned with music, but attempts rather to sketch a general framework for thinking about nonexistent yet possible artforms, issuing in a number of schematic formulas for generating such artforms in the abstract. The second, more concrete, half of the essay takes as a case study the relative nonexistence of visual music, despite
In The Pleasures of Aesthetics. In Music, Art, and Metaphysics.

Introduction

numerous attempts in that direction over the years, and proposes an explanation of visual musics stubborn failure to establish itself as a viable artform. The next two essays deal, from different angles, with the relationship of music to narrative. Music as Narrative and Music as Drama pointedly poses the question of whether music, especially as regards its succession of expressive properties or states, is fruitfully thought of as a narrative of some sort. The answer returned is guardedly negative, and the attractions of an alternate model, one owing to the musicologists Anthony Newcombe and Fred Maus, of expressive music as dramatized rather than narrated emotion, are touted instead. Film Music and Narrative Agency, which, as already noted, is the longest essay in this collection, is as much concerned with lm as it is with music. It seeks to illuminate, on the basis of an account of making ctional along lines laid down by Kendall Walton, and through an extensive survey of examples, the ways and means by which extrinsic lm music inects the ctional content of a lm, identifying two distinct modes in which that can occur, one in which such music is ascribed to the lms cinematic narrator, and one in which such music is ascribed, less commonly, to the lms implied director. The next essay, Evaluating Music, is an attempt to identify plausible midlevel principles by reference to which one might conceivably justify an evaluation of some music as good, where by mid-level principles I mean principles whose specicity lies between the extremes of, on the one hand, musics being good if it affords appropriate listeners worthwhile experiences, and on the other hand, musics being good if it displays this or that set of technical features held to be productive of musical worth, such as monothematic structure or coherent harmony. The perspective of Evaluating Music derives from that developed in an earlier essay, What Is Aesthetic Pleasure?, where I propose that the distinctive mark of aesthetic satisfaction in art is that it is satisfaction deriving from attention that focuses, above all, on the relation of content to form and form to content in the given work of art. The mid-level principles of musical evaluation arrived at in light of that perspective on aesthetic satisfaction are then illustrated in connection with one of Schuberts piano sonatas, the Sonata in A major, D. 959. The last two essays in Part II, Musical Thinking and Musical Chills, like Nonexistent Artforms and the Case of Visual Music, tackle questions in musical aesthetics that have not been the focus of much, if any, discussion.
In The Pleasures of Aesthetics.

Introduction

Musical Thinking, which begins with a commentary on Wittgensteins scattered remarks on musical understanding, poses the question of whether there is a distinctive, non-verbal form of thinking that music, or alternatively, the composing or performing of music, might be said to exemplify. A positive answer is returned, and three candidates for such distinctively musical thinking are sketched; these are illustrated with a number of musical examples, most notably, Beethovens Tempest Sonata and Stan Getzs rendition of The Girl from Ipanema. Musical Chills, of all the essays in this collection, is the one that has undergone the most evolution since it was rst drafted around 1998, having already been published twice, under different titles, and in truth my thinking on the subject remains in ux, despite my committing it to print once again. It is also the only essay I have written to date whose principal spur was an empirical study, one concerned with a musical phenomenon that has always fascinated me, namely, the distinctive and usually pleasurable chills, shivers, or frissons that listening to certain passages of music produces in many listeners. At any rate, after describing the phenomenon and situating it in the eld of musical pleasures as a whole, and after considering and nding wanting the explanations of the phenomenon and its value that have been offered by cognitive psychologists, I try to construct a more satisfactory explanation, one illustrated most fully in connection with a piano piece of Scriabin, his Etude in C# minor, op. 42, no. 5. Part III initiates a shift of focus to the visual arts. Wollheim on Pictorial Representation was written as a contribution to a symposium in honor of the distinguished aesthetician Richard Wollheim, and begins with a sympathetic summary of his highly inuential account of depiction in terms of the successfully realized intention that viewers have a certain sort of seeing-in experience faced with a picture depicting a given subject. While agreeing with the basic thrust of Wollheims account, which makes a certain sort of visual experience in appropriate viewers criteria of achieved depiction, I differ with Wollheim as to whether that experience is invariably one of seeingin, given the twofold attention to subject and surface that that notion, as Wollheim conceives it, necessarily involves. I sketch an alternative account, Wollheimian in spirit, but closer than most recent proposals to the classic Gombrichian view of depiction as involving something akin to illusion. What I propose specically is that a picture that depicts a subject is one fashioned so

Conducted by the neuropsychologist Jaak Panksepp.

Introduction

as to yield an experience of as-if seeing of its subject, but not an experience that engenders the false beliefs typical of illusion. As is evident, the next two essays in Part III have a common theme, namely, the erotic in art. What Is Erotic Art?, an expanded version of an encyclopedia article published in 1998, and my rst foray into this domain, straightforwardly addresses the question of its title. The answer offered is not calculated to astonish: erotic art is, rst, art, and second, erotic. In less sphinx-like terms, erotic art is art that aims to engage viewers sexually through explicit sexual content, and that succeeds at least to some extent in doing so. This answer is held up for conrmation to a range of examples of the category, some uncontroversial and some less so, and a number of useful subcategories of erotic art are identied. Erotic Art and Pornographic Pictures, which like its predecessor connes its attention to the visual, was written in response to a 2001 essay of Matthew Kieran, itself prompted in part by remarks on the distinction between the erotic and the pornographic offered at the end of What Is Erotic Art?. Whereas Kieran holds that there is no incompatibility, and even precious little tension, between somethings being pornography and somethings being erotic art, I hold, and endeavor to demonstrate, that there is indeed such tension, and that the two statuses are in fact incompatible. That said, nothing is entailed as to whether pornography, though it is not art, may or may not be, for various reasons, of value. Another shift of focus is effected in Part IV, whose three essays concern for the most part literature and literary language. The rst essay, however, is of somewhat more general scope. Two Notions of Interpretation brings into relief a distinction among semantic interpretations, or among activities of semantic interpreting, that cuts across verbal and non-verbal phenomena. The distinction is between interpretations that aim to answer the question What does such and such mean? and those that aim to answer the question What could such and such mean?, the former exemplifying the determinative mode, and the latter the exploratory mode, of interpreting. In the rest of the essay I investigate, through a range of examples literary and non-literary, the relationship between determinative and exploratory interpretation in a given inquiry, and the varying, sometimes interlocking, motivations with which determinative and exploratory interpretations are undertaken. In Whos Afraid of a Paraphrase? I turn specically to the interpretation of metaphors. My principal claim, in opposition to the well-known stance of Donald Davidson, is that metaphors, however much their force or imagery

Introduction

outstrips their semantic content, in fact usually possess relatively denite meanings, meanings which deserve the label of metaphorical, and which paraphrases can to a large extent express. The key to the stance on metaphors adopted is the conception of them as utterances in specic linguistic contexts, which acquire meanings in such contexts despite there being no rules of a semantic sort for the projection of such meanings. Examples of metaphors from both literary and non-literary contexts come in for examination. The conception of literary meaning as centrally a species of utterance meaning is the foundation stone of the view of literary interpretation dubbed hypothetical intentionalism that I have argued for in two earlier essays. In Hypothetical Intentionalism: Statement, Objections, and Replies I briey restate the view, which locates the meaning of a literary text not in what its author intended it to mean (what one can call utterers meaning), nor in what the text might be said to mean as a piece of language in the abstract (what one can call textual meaning), but roughly in what an appropriate audience would most reasonably hypothesize the contextually situated author to have meant by composing precisely the text that he or she did. I then proceed to consider a fair number of objections to the view in the literature and attempt to respond to them. But as this is currently a very active area of research I am, as with my replies to objections to the intentional-historical account of arthood, under no misapprehension that these will constitute the last words on the subject. The concerns of the two essays in Part V, which are continuous with those in my earlier Aesthetic Supervenience, are as much metaphysical as aesthetic. The central issue is the nature and objectivity of aesthetic properties, especially those belonging to works of art. In both essays I defend aesthetic realism, by which is meant the claim that aesthetic properties exist, that they are bona de properties, and that their possession constitutes the truth condition of true aesthetic attributions. In the rst part of Aesthetic Properties, Evaluative Force, and Differences of Sensibility, written for a 1997 conference in honor of the inuential British aesthetician Frank Sibley, I sketch a largely Sibleyan view of aesthetic attributions, though a more metaphysically
Intention and Interpretation in Literature and Messages in Art, both in The Pleasures of Aesthetics. An important recent set of words on the subject, for instance, of which no account is taken here, is Paisley Livingstons Art and Intention (Oxford: Oxford University Press, 2005). In Music, Art, and Metaphysics.

Introduction

committed one than Sibley was inclined to hold, underlining the extent to which a descriptive core can be located in almost all such attributions, whatever evaluative force they may carry, and however implicitly relativized they may be to perceivers of certain sorts. I then proceed to formulate and respond to a number of concerns one might have about this brand of aesthetic realism. In What Are Aesthetic Properties? I extend my defense of aesthetic realism, devoting most of my attention to the issue of how we should conceive of properties in general and of aesthetic properties in particular. What I propose is that at least paradigm cases of the latter are to be understood as higher-order perceptual ways of appearing. In the course of developing this proposal I address the vexed issue of whether or not aesthetic properties are response-dependent, or such that they cannot be conceived or analyzed except in terms of kinds of responses in relevant perceivers, and conclude by suggesting that there is a spectrum, among properties usually thought of as aesthetic, from ones that are clearly response-dependent to ones that are clearly non-response-dependent, with many gradations in-between. The two essays in Part VI take up themes from the history of aesthetics. My aim in Schopenhauers Aesthetics, written as an encyclopedia article, is largely expository. I begin with Schopenhauers relationship to Kant, and the extent to which the great pessimists aesthetic philosophy relies on Kants metaphysics even more than it does on Kants aesthetics, and then go on to highlight the breadth of Schopenhauers vision of the role of art and of the liberating aesthetic experiences it makes possible. At the end I address the puzzle of how the art of music, which according to Schopenhauer presents us with blind, ceaseless, and hateful willing in its most unvarnished form, can yet provide aesthetic experience of the highest order, justifying Schopenhauers according to music the foremost position among the arts. My aim in Humes Standard of Taste: The Real Problem, on the other hand, is more polemical than scholarly. I there formulate a persisting problem about the authority of art criticism, one that should concern anyone for whom the arts occupy an important place in life, and situate this problem in relation to Humes search for the standard of taste in his famous essay of that name. I then sketch a complex solution to this problem, somewhat provocatively labeled the real problem left us by Hume, a solution whose complexity is justied by the thorniness of the problem in question.
As seems to be the case, say, for properties like nauseatingness or disgustingness.

Introduction

The Concept of Humor, also written as an encyclopedia article, surveys the main theories of humor in the philosophical tradition, and then proposes a novel account of the essence of humorousness, often regarded as an aesthetic property. This essence is held to lie not in perceived incongruity, nor in perceived superiority, nor in the power to trigger experiences of relief, but in the disposition to produce affect of a sort tied identifyingly to laughter. This account is dubbed the affective theory of humor, and some recent objections to it are discussed and defused. So far as the causes or mechanisms through which humorousness is achieved are concerned, I discuss the pros and cons of the leading theory in that vein, the so-called incongruity theory of humor, and side in conclusion with those who hold that resolution of incongruity is perhaps closer to the heart of the matter than incongruity per se. Finally, in Intrinsic Value and the Notion of a Life, I address a problem in the general theory of value that goes beyond the concerns of aesthetics as such, though aesthetic issues at one point serve to bring into relief the nature of the thesis about intrinsic value ultimately arrived at. That thesis concerns the form that defensible judgments of intrinsic valueroughly, what is of value in itself or for its own sakemust take, or equivalently, the sort of thing that may defensibly be claimed to have intrinsic value. The thesis defended, which tries to mediate between object-based and experience-based conceptions of the intrinsically valuable, is that a richly sentient life being a certain way is the only possible subject of a defensible judgment of intrinsic value. One consequence of this thesis is a disagreement with G. E. Moore regarding the intrinsic value of a beautiful world devoid of sentience, a famous thought experiment from his Principia Ethica. But a more important consequence is the suggestion, if I am right, of an intimate connection between the notion of a richly sentient life and the very idea of intrinsic value. Thanks are owed to all the following for helpful comments on the essays collected here at various stages in their evolution: Lars-Olaf Ahlberg, Jose Bermudez, Paul Boghossian, Malcolm Budd, Nol Carroll, David Chalmers, e Ted Cohen, Jean-Pierre Cometti, Jack Copeland, Gregory Currie, David Davies, Stephen Davies, Rafael De Clercq, Sabine D ring, John Doris, o Hubert Eiholzer, John Fisher, Berys Gaut, Alessandro Giovannelli, Stan Godlovitch, Mitchell Green, Arto Haapala, Garry Hagberg, Robert Hatten, Peter Lamarque, Keith Lehrer, Paisley Livingston, Mike Martin, Derek Matravers, Fred Maus, Aaron Meskin, Daniel Nathan, Alex Neill, David Novitz,

10

Introduction

Elisabeth Pacherie, Derk Pereboom, Paul Pietroski, Diane Proudfoot, Aaron Ridley, Mark Rollins, Anthony Savile, Martin Seel, Roger Shiner, Elliott Sober, Robert Stecker, Joseph Tolliver, Saam Trivedi, Kendall Walton, Alicyn Warren, Susan Wolf, and Nick Zangwill. Finally, thanks to Peter Momtchiloff for his aid and encouragement, and to Ludmilla Kolokolova for her love and support, throughout the process of bringing this book to fruition.

PART I ART

This page intentionally left blank

1
The Irreducible Historicality of the Concept of Art
I INT RODUCT ION I claim that our present concept of art is minimally historical in the following sense: whether something is art now depends, and ineliminably, on what has been art in the past. I claim, in other words, that the concrete history of art is logically implicated in the way the concept of art operates, and that some part of that history is involved, either opaquely or transparently, in the claim to arthood made by any work of art. By contrast, the concepts square , red , pig , mountain , and so on are not obviously historical in this sense: whether something falls under them does not seem to depend in the same way on what specically fell under them in the past, and to operate those concepts correctly you do not need to invoke the concrete history of their correct application. The gist of the intentional-historical conception of art that I advocate is this: something is art iff it is or was intended or projected for overall regard as some prior art is or was correctly regarded. As is evident, such a conception attributes to art, and centrally, the property of minimal historicality sketched above. In this brief essay I will forgo defense of the sort of complete denition of art I am inclined to favor, and that I have tried to articulate in three
This chapter was rst published in British Journal of Aesthetics 42 (2002): 36779. It has even been argued recently that this may be true of all artifact concepts, artistic and nonartistic alike. See Paul Bloom, Intention, History, and Artifact Concepts, Cognition 60 (1996): 129. I comment briey on Blooms intriguing suggestion at the end of this essay, and again in the following essay. Jean-Pierre Comettis essay, Mis`reou grandeurde lhistoricisme?, in Jean-Pierre Cometti e (ed.), Denitions de lart (Brussels: La lettre vole, 2002) has helped me to see the importance of e dissociating the minimal historicism of art claimed by my theory from more robust historicisms of a Hegelian or Dantoesque sort, such as ascribe to the development of art an inherent goal, or view the development of art as governed by inherent laws of stylistic evolution. In that light, it might have been better to denominate my theory of art a retrospectivist or auto-referentialist one, rather than a historicist one, in order to avoid such unwanted associations.

14

Art

previous essays. I will also largely ignore questions regarding the sufciency of an intentional-historical condition for arthood, and questions as to the necessity of the intentional component of such a condition, in order to focus on the necessity, in some guise or other, of the historical component. My ambition in the present outing is thus modest. I aim to do only two things. One is to underline the necessity of a historical dimension in any acceptable account of arthood. Two is to sketch answers to certain objections that have been recently raised for an intentional-historical account of art, most of which offer a challenge to its insistence on an ineliminable historical element in any such account. In addition, in the course of underlining the historical character of the concept of art I hope to show that certain non-historical considerations appealed to by some theorists, for instance, institutional or functional ones, which appear to weigh importantly in some cases of arthood, in fact have an underlying or reinforcing rationale of a history-involving sort. II OBJECTIONS AND REPLIES I now consider a number of objections that have been lodged against the intentional-historical theory of art, and offer replies to them.

The Objection from the Implausibility of a Recursive Denition of Art6


Some writers have objected to the intentional-historical denition of art on the grounds that it is a recursive denition, or else entails that art can be dened
See Dening Art Historically, British Journal of Aesthetics 19 (1979): 23250, and Rening Art Historically, Journal of Aesthetics and Art Criticism 47 (1989): 2133, reprinted in Music, Art, and Metaphysics, 2nd edn. (Oxford: Oxford University Press, 2006), and Extending Art Historically, Journal of Aesthetics and Art Criticism 51 (1993): 2133, reprinted in The Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996). I have also replied to criticisms of my theory in some shorter pieces: A Reners Fire: Reply to Sartwell and Kolak, Journal of Aesthetics and Art Criticism 48 (1990): 2315; Further Fire: Reply to Haines, Journal of Aesthetics and Art Criticism 48 (1991): 767; and Art Historically Dened: Reply to Oppy, British Journal of Aesthetics 33 (1993): 3805. See also Robert Stecker, Artworks: Denition, Meaning, Value (University Park, PA: Pennsylvania State University Press, 1997), 8898, for a critical reconstruction of the intentional-historical theory of art. These were queried vigorously by Nol Carroll in his Identifying Art, in Robert Yanal (ed.), e Institutions of Art (University Park, PA: Pennsylvania State University Press, 1994), 339. See Graham Oppy, On Dening Art Historically, British Journal of Aesthetics 32 (1992): 15361. Some of Oppys criticisms are anticipated in Stephen Davies, Denitions of Art (Ithaca, NY: Cornell University Press, 1991). See, for example, Tom Leddy, The Socratic Quest in Art and Philosophy, Journal of Aesthetics and Art Criticism 51 (1993): 399410

The Irreducible Historicality of the Concept of Art

15

recursively, neither of which strikes those writers as a happy result. But strictly speaking, the charge is mistaken. My basic denition of art is a one-step affair, as is evident even in Steckers reconstruction of it. What I have proposed is that the full extension of art in a given tradition might be displayed by a recursive denition, but not that our present concept of art is to be explicated by such a recursion. Again, its true that my denition implies that the totality of art in a given tradition has a recursive structure, but that is not tantamount to my having dened art recursively. In underlining that the intentional-historical denition of art is not as such a recursive one I am thus denying that the notions of rst art and ur-art, with which such recursions can be thought to begin, are elements in our concept of art, and that what we mean by an artwork is something that either is or stands in the appropriate relation to instances of rst art or ur-art. This is, of course, all to the good, since it would be implausible to maintain that such notions are a part of the ordinary grasp of what arthood is.

The Objection from Unwanted Descendants of the Ur-Arts7


Ancestors of art activities, such as ritual cave paintings, may also turn out to be ancestors of present-day activities that are clearly non-art, such as deer hunting with high-caliber ries. But then my denition, it seems, will wrongly count these latter activities as art. My reply to this is as follows. Though that sort of misring of the denition is conceivable, it is arguable that in presumed cases of this sort the links from remote to present-day activities are not precisely of the right sort, that is, of the backward-looking-intentional-invocation-of-regard sort. In other words, the generating principle of these other sequences, ones that begin with some ur-art and issue in clearly non-art activities, is likely not precisely of the sort involved in the generation of artistic chains. One would have to examine closely a putative concrete aberrant chain, leading from unequivocal ur-art to unequivocal non-art, to assess fairly the strength of this objection. But it is unclear that any such chains survive scrutiny.

The Objection from the Obsolescence of Art-Regards


Here is a forceful statement of this objection, as put by Nol Carroll: e
Levinson supposes that something might be art now just in case it supports any type of regard, treatment, or mode of appreciation that was appropriate to at least some
See Nol Carroll, Historical Narratives and the Philosophy of Art, Journal of Aesthetics and e Art Criticism 51 (1993): 31326

16

Art

works of art in the past. The problem here is that not every mode of appreciation that was lavished on artworks in the past is eternally available. Some modes may have become historically obsolete . . .

Carroll then goes on to give a concrete counterexample to illustrate his abstract charge. This involves a benighted chicken farmer named Jones, one steeped in old-time religion, who sets out to make an artwork. Specically, he shoots a mass of chickens in record time in order to propitiate the gods. Moreover, he presents the massacre as an artwork: onlookers are invited to appreciate it, assess it, or regard it in terms of its effectiveness as a means of propitiating the gods. According to Carroll, since propitiating the gods was an aim of some past artworks, and since such artworks were correctly appreciated in terms of such an aim, my theory must acknowledge for appreciation as an instance of god-propitiation as an art-making regard, and thus Joness action as an artwork, which, given Jones cannot be understood as either a Conceptual or a Performance artist, seems wrong. [Joness] intention is simply to make something that is to be regarded . . . as a vehicle for propitiating the gods, where propitiating the gods was once an acknowledged purpose of art. But Carrolls conclusion is unwarranted. The problem with his objection is that it mistakes a single, isolated regard appropriate to some past artworks for a complete, integral ensemble of regards appropriate to some past artworks. It is only in being intentionally projected for the latter, not the former, that an object acquires the status of artwork on my theory. Although some ancient artworkssay, certain tragedies or templeswere intended, let us assume, for appreciation as instances of god-propitiation, it is certainly not the case that they were intended solely for appreciation in that respect. Surely they were also intended for other regards, involving attention to those works emotional, formal, and symbolic aspects. Hence Joness chicken-slaughtering act, being intended only for appreciation as an instance of god-propitiation, is not an act intended for regard in the way any past artworks were as a whole correctly regarded. So the charge that the denition misres when it invokes obsolete
See Carroll, Identifying Art, 334. Ibid. 345. Ibid. 35. I in fact dealt with this objection preemptively in my Rening Art Historically. I there underlined that only relatively complete or total ways of regarding are to be allowed as substitution instances of the artwork formula thing which has been seriously intended for regarding-as-awork-of-art, i.e. regard in any way preexisting or prior artworks are or were correctly regarded (p. 24). But, it may be asked, what if Jones does intend that those other sorts of regards be taken to his act as well, those for which those ancient temples and tragedies were also intended? In that

The Irreducible Historicality of the Concept of Art

17

ways of regarding art itself misres. Integral ensembles of regards appropriate to past artworks in fact never become obsolete, so far as their potential to enfranchise possible future artworks is concerned, though of course they may fall out of favor or fashion.

The Objection from the Putative Ascension of Attractive Nonartworks to the Ranks of Artworks
There are admittedly many cases of attractive purely utilitarian objects subsequently treated as artworks by some individual or individuals, counter to or in the absence of any artistic intentions on the part of their creators. But it is an error, I suggest, to think this makes such objects into artworks: audiences, appreciators, consumers cannot make things art merely by treating them as such. Of course, as far as aesthetically appealing objects of real-world human cultures are concerned, e.g. pots, knives, masks, curtains, rugs, the idea that any of them were ever conceived or projected in a purely utilitarian way is implausible. Thus, when we exhibit such objects in art museums we dont need to be thought of as transforming or altering their status, but as simply acknowledging the quasi-artistic status they already have, at least in part, as created. On the other hand, it is possible that some such objects, e.g. magic-ritual ones aimed only at invoking the spirits or manipulating natural forces, really dont belong in art museums, given their original constitutive projection by their makers, however aesthetically interesting or artistically advanced they might seem. In any event, the least that can be said is that such examples of putative artworks lacking the appropriate sort of intentional projection by their makers are too much in dispute to serve as decisive counterexamples to an intentional-historical conception of arthood.

The Objection from the Puzzling Status of First Art13


What secures for rst art its status as art such that it is art from the outset and thus capable of anchoring the chain of artworks that, according to historical theories of art, reaches from rst art to the present? Stephen Davies puts the difculty as follows: First art must be art already at the time second wave
case, it seems my theory might have to count Joness act as art; but then it would also be less counterintuitive to so count it. See Stephen Davies, First Art and Arts Denition, Southern Journal of Philosophy 35 (1997): 1934.

18

Art

pieces become art, otherwise second wave pieces could not be art as a result of standing to rst art in the art-dening relation. The intentional-historical theory of art seems to have no acceptable answer to how rst art becomes art, since rst art is, by denition, not related to any prior art by the art-making relation the theory proposes as crucial. I have elsewhere attempted to supply answers to that question, some of which were, admittedly, not acceptable. The answer I favor now is that rst art is indeed art at the time of making, but that it is art in a somewhat different sense, or for a slightly different reason, than all subsequent art. Obviously, rst art in a given tradition bears to ur-artthe ultimate non-art progenitors of artworks in that traditionsomething like the relation that second art bears to rst art, and that all subsequent art bears to art preceding it. The relation is roughly that of being projected for treatment or regard in ways that earlier objects were appropriately regarded or treated. But since ur-art objects are not art, this cannot be described univocally as projection for regard in the way earlier art was appropriately regarded. There remains an irreducible difference between ur-art and rst art apart from mere temporal precedence, which consists in the fact that, for both ur-art and rst art objects, while there are certain regards or treatments that are appropriate to them, only the latter are projected for regard or treatment in the way the former appropriately are regarded or treated. Though both rst art and ur-art objects are artifacts, whose identities are governed by intentions, the intentionality that makes rst art art ineliminably refers to earlier things and activities, but not so the intentionality that makes ur-art ur-art. Anyway, it looks like an expanded denition of art is needed if both rst art and later art are to be comprised, and that this will need to be disjunctive in form. It would be that something is art iff either (a) it satises the basic denition or (b) it is an instance of rst artthat is, one of those things from which all other art, that satisfying the basic denition, springs.
Ibid. 21. For example, the suggestion that rst art acquires its art status retroactively, after the art tradition which it stands at the beginning of gets going. (See my Art Historically Dened: Reply to Oppy.) As Davies rightly notes, an only-retroactively-conferred art status for rst art would vitiate the recursive chain of artmaking from the outset. It is only retrospectively that we can determine, given adequate archeological research, the identity of the ur-arts for a given tradition of art, but it remains true that they always were that. It is not a matter of their retroactively becoming those ur-arts. See Extending Art Historically.

The Irreducible Historicality of the Concept of Art

19

The Objection from Anthropocentricity17


It is surely the case that ancient Martians, active in the remote past, could have made art even though their artifacts would not have been intentionally related to any preceding works of human art, as the intentional-historical theory seems to require. Suppose we respond, then, by liberalizing the theory so as to allow for an objects being art in virtue of being intentionally linked in the right way to succeeding works of human art. Matters would not, however, thereby be much improved. For granted that liberalization of the theory it would still be entailed that Martians could not even have known that their artifacts were art until humans came along, thousands of years later, which seems unintuitive. The intentional-historical theory thus fails to capture the concept of art in its full generality, the objection concludes, because it takes what is contingent . . . [namely, arts concrete historical realization] . . . for something essential in art. Examples of this sort, which highlight for our attention possible artworks and art practices that predate human history altogether, do appear to call for some modication of the intentional-historical denition as originally proposed. For clearly these would be artworks that failed to possess intentional connections, even opaque-to-their-creators intentional connections, to earlier works of human art, there being by hypothesis no such earlier works. The liberalization I favor to meet that problem is not, however, that ventured above, that the works in question are art through having the right intentional relations to later works of human art. Rather, what I propose is that we might with justice consider such works as ancient Martians may have produced to be art in virtue either of (a) their having come about in a reexive practice of making-and-intending-for-regard-asearlier-such-objects-were-regarded similar to our own practice of art, or (b) their having been made and intended for particular regards acknowledged in our own contingently evolved art history, or (c) the conjunction of
See Gregory Currie, Aliens, Too, Analysis 53 (1993): 11618. For another reply to Currie, see Robert Stecker, Alien Objections to Historical Denitions of Art, British Journal of Aesthetics 36 (1996): 3058, and reprised with small changes in ArtWorks, 1078. Currie, Aliens, Too, 118. See also Oppy, On Dening Art Historically. Currie correctly notes that the intentional connections between later and earlier artifacts may very well be of this kind, whereby the intention governing later artifact B connects it to earlier artifact A even though the maker of B is oblivious of A. See Aliens, Too, 117.

20

Art

(a) and (b). Call the intentional-historical theory liberalized in some such fashion the extended intentional-historical theory. But what now about the argument from knowledge, which applies to extended intentional-historical theory as much as it applies to the theory as originally proposed? There are, it appears, two questions to address. First, could such artistic Martians have known that they were, by the theorys lights, making art, and second, to the extent that they could not, is that indeed unacceptably counterintuitive? The answer to the rst question is complicated. Here are some things that those Martians, in the course of making art, could have known: (a) that they were making things for a certain sort of regard or treatment; (b) that they were making things to be regarded or treated as certain earlier things made by them were properly regarded or treated. Here, on the other hand, are some things that they, in the course of making art, clearly could not have known: (c) that they were making things for a sort of regard or treatment featuring in some future earth art tradition; (d) that they were making things to be regarded or treated as certain earlier things made by them were properly regarded or treated, in just the same manner as that in which art on earth typically relates to preceding art on earth. But it is (c) or (d), it seems, that they would have had to have known in order to have known that they were making art on the extended intentional-historical theory of art. So, indeed, they could not have known they were making art, by our present concept of art, even if the extended intentional-historical account of that is correct. That is, they could not have known that their stuff was art in the full historically-reexive sense I claim is now ours, though they could of course have known that what they produced was *art*, where *art* is some non-historicist predecessor of our current concept of art, one that was roughly adequate to artistic production in our culture prior to the early twentieth century. Now for the second question: how counterintuitive is that? Properly viewed, not very. What we have just seen is that those Martians could not have known that their stuff was art as we now understand that, that is, art
I earlier ventured such a suggestion in Extending Art Historically, 4223. Compare a similar and later suggestion by Stecker: One can say that although Martian modern art does not stand in an appropriate relation to previous human art, it does stand in such a relation to previous Martian art . . . So Martian modern art is art in virtue of its relation to previous Martian art, and ultimately to Martian ur-art (ArtWorks, 108). As Stecker puts it, But why should [the assumed Martians] complain that they cannot apply our concept of art to their art? They would be no more able to do this than would ancient Egyptians or Greeks be able to apply our concept to their art (ArtWorks, 1078).

The Irreducible Historicality of the Concept of Art

21

in the specic actual-history-reective sense I claim is now ours. But as we also saw, they might very well, in virtue of knowing (a), have known that what they practiced was *art*. In other words, they could have known that they were making in art in a stripped-down, form-and-function-based sense not equivalent to the sense we operate with at present. Another, more recent, version of the anthropocentricity objection against historical theories of art goes as follows: their account of an art-historical relation is insufciently projectable: there could be art objects which are recognizable as such, but which stand in no art-historically signicant relation at all to any of our art. But it is not that a tradition of object-making must already stand in an art-historically signicant relation to something in our tradition of art in order to be an art-making tradition, but rather that, insofar as anything outside our art tradition is properly said to fall under our concept of art, it is because we can appropriately relate it to our tradition of art, and in particular to the normative regards that have, as a contingent matter of fact, emerged in that tradition. Consider nally the hyperbolic question: could there have been art a million years ago, on a planet of the star Betelgeuse, if human history had never occurred at all? Well, yes and no. The answer is no, I think, if one means art in the specic actual-history-conditioned sense it now has, in the early twentyrst century. The answer is yes, I think, if one just means objects made in certain ways, for certain kinds of reception, all intrinsically dened. But that, though it may once have served as our concept of art, and perhaps as recently as a century ago, no longer does.

The Objection from the Multiplicity of Art Traditions23


This objection, originally voiced by Stephen Davies, has been concisely formulated by Robert Stecker:
Davies admits that historical denitions of art explain how something is an artwork by relating it to a given tradition. However, he claims that such denitions will be incomplete until a basis is provided for distinguishing art traditions from other historically continuous cultural processes or practices. If there are different art traditions, and if something is an artwork only in relation to some of these traditions, then
Berys Gaut, Art As A Cluster Concept, in N. Carroll (ed.), Theories of Art Today (Madison: University of Wisconsin Press, 2000), 36. Davies, Non-Western Art and Arts Denition, in N. Carroll (ed.), Theories of Art Today.

22

Art

the explanation of why something is an artwork will not be complete without some account of what makes something an art tradition.

And here is the point in Daviess own words: A denition that characterizes art making as artworld-relative and that also concedes the existence of autonomous artworlds must explain how artworlds are of a single type. An account is required of what makes the various artworlds artworlds. As it turns out, Davies is sympathetic to the extension of historical denitions of art to accommodate non-Western art, provided one can say what makes a given practice of object-projection characterized by backward intentional invocation of predecessors an art practice, rather than something else. Davies suggests that this is to be done by recognizing the essentiality of aesthetic interests and regards, at least in initial stages, to any practice that might be accounted artistic, however far it may have diverged from its original aesthetic roots. This is a plausible proposal as to what, in surveying the known art traditions of the world, makes them all art traditions rather than internally historical traditions of some other sort of making. But whereas Davies views the necessary aesthetic origins of any tradition that can be recognized as artistic as a ground-oor fact, arrived at after reection on both the ubiquity of art in human culture and the universality of aesthetic interests at the beginning of art traditions, I would be inclined to historicize further, and so render more contingent, the role of the aesthetic in the characterization of artistic practices. What I want to say is, yes, possibly anything we would recognize as an artworld or an artistic practice would display aesthetic concerns, at least at its origins, but that is because aesthetic concerns emerged centrally and persisted centrally for thousands of years in the Western tradition of artmaking. In other words, it is neither a strange accident nor a conceptual truth that anything we recognize as an art practice will have involved cultivation of and attention to humanly bestowed aesthetic features of things; it is just that those are the concerns that were original to and for ages almost uncontestedly dominant in what we unreectively know as our practice of art.

The Objection from the Putative Two-Dimensional Semantics of the Term Art26
It has been claimed by Gregory Currie that if art is to be a historical concept in the relevant sense, that is, one in which the very identity of our concept
ArtWorks, 108. Davies, Non-Western Art and Arts Denition, 212. Gregory Currie, A Note on Art and Historical Concepts, British Journal of Aesthetics 40 (2000): 18690.

The Irreducible Historicality of the Concept of Art

23

of art depends on history, then it will have to be a concept (a) that is our actual concept, (b) where its being our concept depends on some contingent matter of history, (c) where some other concept or concepts might, if history had gone differently, been our concept of art, and (d) where there is some overarching concept that binds them all together, showing that they are all, indeed, concepts of art. Currie further claims that this is the conceptual structure exhibited by natural kind concepts such as water , and thus that art , if it is to be historical, must exhibit the same general feature as natural kind concepts, even though it is not thereby claimed that art is a natural kind. The general feature in question, as the Putnamian analysis of water suggests, is a two-dimensional semantics in which the meaning of a term depends on two things, a qualitative, purely observable notion like that of waterish stuff (i.e. stuff resembling water) in the case of water, and certain contingent facts about what in a given world constitutes the stuff in question, such facts being, in the case of water and the actual world, that the waterish stuff of our acquaintance is H2 O, or hydrogen oxide. Thus, as applied to art, this analysis claims that if art is to be similarly historical there must be a qualitative, purely observable covering concept like artish thing, and then certain contingent facts about what things actually fall under that concept in a given world, in order for the meaning of art to be xed in that world. But if so, then there is no avoiding the observable concept artishthe concept is presupposed in making out the putatively two-dimensional character of the concept of art and, the objection concludes, there is no reason to think the concept of art we actually use is other than that of artish thing, thus undermining the rationale for a historical analysis of arthood. At bottom, I think this critique of Curries mistakes its target, taking the intentional-historical theory of art to be a claim about the dependence of the concept of art on art histories in other possible worlds, whereas the claim is rather that the concept of art is such that what can be art at a given time in a world logically depends on what is already art at that time in that world, since artmaking minimally involves an agents intentionally relating a proposed object to the body of already existing art. In other words, what is maintained is the dependence of the possible extension of art at t on the actual extension of art prior to t, but not the dependence of our present concept of art on the
Ibid. 187. Ibid.

24

Art

actual path of art history. So far as I can see the intentional-historical theory of arthood is not committed to the claim that our present, minimalist and intentionalist, concept of art would have been different if actual art history had been different. I want to say that the concept of art with which we presently operate, if historical in the way I have in mind, is not different in different possible worlds. It is rather, in any world, the concept of an object-identifying practice where what can count as art at t depends on what, contingently, already counts as art prior to t. The variation in the possible extension of art from world to world occasioned by the contingencies of what has already fallen under the term art at a given time, to which the intentional-historical theory is committed, does not seem to me modellable on the two-dimensional meaning account apt for natural kind terms, which involves an implicitly referenced underlying nature. The concept of art, I suggest, is structurally unlike that of water. Let us look more closely at the disanalogy between water and art. In the case of water, assuming Putnams account of natural kinds, when we consider other possible worlds where persons have, let us say, the same basic idea of water as we do (i.e. watery stuff), we do not count as water everything they count as water; we only count something in that possible world as water if it has the same underlying structure or material composition as our water. In the case of art, assuming my account of arthood, it is true that in other possible worlds persons will be counting different things as art than we do, given the contingent development of their art historymost obviously, there may be objects in that world that do not exist in ours. But when we consider those possible worlds from this one, I claim, we have every reason to count or acknowledge as art what they count or acknowledge as art, that is, things intended for regard the way earlier presumed art in that world was correctly regarded. It is perhaps true that we can only identify what is their art practiceas opposed to other practices they might have that hold together through reexive-retrospective intentionsby noting that it is the one where the normative regards invoked in the reexive-retrospective intentions are those which contingently emerged as such in our art practicee.g. aesthetic, formal, expressive, communicative ones. This is the key way in which art as we now deploy it is anchored in the contingency of what is our own art history. Still, it seems true that we will count as art in that world what they count as art in that world.
In particular, what I above labeled extended intentional-historical theory.

The Irreducible Historicality of the Concept of Art

25

Thus the contrast with how water works remains. Given what we mean by water, where underlying constitution is essential, what persons on TwinEarthwhere XYZ, not H2 O, lls the water rolethink of as water, nevertheless isnt water. But given what I maintain we now mean by art, and have meant by art for at least half a century, then if persons in another possible world think of something as artthat is, they label things that way within a practice we can recognize as an art practice because of its structural (backward-looking reexive intentions) and substantive (involving the sort of normative art regards, ones it is correct to bring to such objects) parallels with our contingently developed art practicethen those things are art, albeit art in that world, even by our lights. Thus art, even if historical in the ` way I propose, is not plausibly regarded as a two-dimensional concept a la water. Shorn of technicalities, the gist of Curries critique is that a concept operative in a given world and dependent on the contingent history of that world regarding what falls under it cannot be shown to be a concept of art unless a qualitative notion like artish is assumed to operate in conjunction with the rst concept. More concretely, the implication seems to be that the backward-looking act of reference posited by the theory, which by linking present objects to past artworks makes of the former artworks as well, must involve a content on the order of preexisting artish things, thus undermining the theorys pretension to outline a purely historical, non-qualitative concept of art. But that is not so. Intentional reference of the art-making sort need not be secured by appeal to an independent, qualitative notion of the artish. Rather, it can be secured purely demonstratively, as by a speech act or thought of form as those things are properly regarded, where the things demonstrated are in fact artworks, or else via paradigms, as by a speech act or thought of form as Beethovens 5th Symphony and the like are properly regarded. In other words, on the theory of arthood I espouse the facticity of arthood, including not only what things are art but what ways of regarding them are normative for them, goes all the way down. Contra Currie, then, the intentionalhistorical theory of art need not fall back self-defeatingly on a quasi-observable notion such as that of the artish. On my theory, a qualitative notion like artish plays no role in how, in any given world, the future extension of art necessarily relates to its past extension.
Or at the least, all the way to the ur-arts, whatever those turn out to be.

26

Art

III ARE ALL ARTIFACT CONCEPTS ESSENTIALLY HISTORICAL? The psychologist Paul Bloom has proposed extending the intentionalhistorical theory of artworks to cover artifacts of all kinds. Bloom proposes, roughly, that for any artifact kind K, to be a K is to be successfully intended to be a K, where what it is to be a K in the latter context is given, inevitably, by past instances of K. Bloom is persuasive in pointing up the superiority of his proposal to existing competitors, those which analyze artifact concepts in terms of necessary and sufcient conditions, family resemblances, characteristic functions, or prototypes. But if Bloom is right, then what remains of the special historicality of the concept of art as opposed, say, to those of chair, pencil, or house? Two things, it seems. First, it should be observed that on Blooms analysis something is a K in virtue of being intentionally related in the right way to preceding Ks generally. But something can be an artwork, on my analysis, through being intentionally connected in the right fashion to some particular past artwork or artworksthe history of art thus enters more concretely into what art is and can be at any point than does the history of a given artifact kind into what is or can be an instance of that kind at any point. To know that something was art thus might require tracing relations to a particular episode or domain of arts history, but nothing comparable would seem to be required to establish that a candidate chair, pencil, or house was indeed a bona de instance thereof. Second, it is arguable that artifact concepts, in contrast to that of artwork, usually retain at least rough necessary conditions as regards either form or function; for example, a chair must exhibit shape within a given broadly circumscribed rangecertain shapes, e.g. that of a javelin, would seem to be excludedand must answer to or have been designed to answer to a certain purpose or useto wit, that of being sat upon. But that is not the case with the current concept of artwork, which, if I am right, and in contrast to perhaps every other artifact concept, retains only certain purely historical and intentional necessary conditions.
See P. Bloom, Intention, History, and Artifact Concepts, Cognition 60 (1996): 129.

2
Artworks as Artifacts
I T HE INTENTIONAL-HISTORICAL CONCEPTION OF ART What kind of artifact is an artwork? The answer to that question depends, clearly enough, on the conception of art that one is inclined to adopt. Past conceptions of art, according to which art was essentially a mode of representation, or a vehicle of emotional expression, or a display of skill in fashioning, or an exploration of form as such, or the pursuit of the beautiful, no longer seem remotely adequate to the nature and range of what have been accounted artworks in the past hundred years or so. The abstract canvases of Kupka and Kandinsky are almost a century old; John Cages aleatoric music of the 1960s seems devoid of emotional expression; Robert Rauschenbergs Erased De Kooning Drawing displayed no notable skill; Tolstoys The Death of Ivan Ilych does not strike one for its exploration of form; and Francis Bacons tortured portraits are anything but beautiful. Accommodating arts development since the nineteenth century seems to call for a more circumspect approach, one that is noncommittal as regards medium, style, form, content, and artistic objectives. I have defended a conception of art of this sort, one along intentionalhistorical lines, according to which something is art in virtue of being governed by certain intentions with an essential historical, or backward-looking, content. More specically, what I claim is that an artwork, in the current understanding of the term, is something that has been intended by someone for regard or treatment in some overall way that some earlier or preexisting
First published in E. Margolis and S. Laurence (eds.), Creations of the Mind (Oxford: Oxford University Press, 2006).

28

Art

artwork or artworks are or were correctly regarded or treated. The artmaking intention involved may be either of an opaque sort, having roughly the content just expressed, one that simply references prior art as such, or of a transparent sort, invoking specic ways of regarding or treating objects that, as a matter of fact, and whether known to the agent or not, gure in the set of correct ways of regard or treatment for earlier or preexisting artworks. In either mode of artmaking, the concrete history of artmaking up to a given time is thus ineliminably implicated, in whole or in part, in any artmaking undertaken at that time. This conception of arthood has obvious points in common with the art-theoretical and social-institutional conceptions of arthood elaborated earlier by the philosophers Arthur Danto and George Dickie. Like those conceptions, it looks for a relational, situational, or contextual dening feature of art, rather than a formal, intrinsic, or perceivable one. All three conceptions have their roots in the enforced revision of traditional ideas about art that was effected by certain revolutionary ventures in artmaking in the early and middle twentieth century, notably those of the Dadaists, Marcel Duchamp, Andy Warhol, Jasper Johns, Robert Rauschenberg, and others. By appropriating, reframing, reconguring, and reprojecting as art any number of things theretofore assumed to lie outside of the ambit of art, artists such as the preceding managed to establishsince their ventures must be regarded, at least from our present vantage point, as undeniably successfulthat more or less any object could be made into or could become a work of art, if suitably repositioned, reconceived, or, in Dantos famous term, transgured. Among the objects that were thus transgured into art in those years, with little or no physical alteration or manipulation, were the following: a urinal, a snow shovel, a bottle rack, a beer can, a coffee cup, a disordered bed, and a postcard reproduction of the Mona Lisa. It no longer seemed necessary, then, that an artwork be fashioned by its maker with technical skill, that it make use of traditional materials in its construction, that it display form of any notable complexity, that it have any obvious aesthetic appeal, or that it inevitably reect in its handling the individual personality of the artist. It is difcult to deny that the concept of art that emerged in the wake of those developments, now almost a century old, was an altered and notably broadened one, covering
See Dening Art Historically, British Journal of Aesthetics 19 (1979): 23250, and Rening Art Historically, Journal of Aesthetics and Art Criticism 47 (1989): 2133, reprinted in Music, Art, and Metaphysics (Ithaca: Cornell University Press, 1990).

Artworks as Artifacts

29

all that had previously been recognized as art, to be sure, but much that would not have been recognized as art under the traditional concept that had held sway, with only minor modications, since at least the Renaissance. The intentional-historical conception of art differs from the art-theoretical and social-institutional ones, though, in positing as the crucial contextual condition of arthood not a relation to some prevailing artistic theory, nor a relation to a surrounding social institution, but a relation to the concrete history of artmaking and art-projection into which the candidate object hopes to enter. The intentional-historical conception differs also from its contextualist predecessors in taking its most direct inspiration not so much from the readymade and appropriational modes of artmaking that had been established by Duchamp and others, but from the subsequent, more radical activities of Conceptual artistssuch as Robert Barry, Robert Morris, John Baldessari, Joseph Kosuth, Sol LeWitt, Vito Acconciwhich seemed to establish that art per se had no need even of any concrete object, whether appropriated, readymade, or fashioned from scratch, but could apparently consist merely in concepts, words, statements, gestures, thoughts, and the like, with the apparent consequence that anything, or at least anything thinkable, demonstrable, or designatable, of whatever metaphysical or logical sort, could be, or at least could become, a work of art. Not surprisingly, the intentional-historical conception of art has elicited a certain number of critiques, turning on such issues as the apparent circularity of such a conception, the status of rst art on such a conception, the extendability of the conception to cultures or histories other than our own, and the problematic recursiveness of the procedure for identifying objects as art that the conception appears to entail. I have addressed those critiques elsewhere, if perhaps not to the satisfaction of all, so will not address them again here. I will instead simply assume that the intentional-historical conception is more
The ontology of Conceptual Art is not as simple as it seems. In particular, the identity of a work of Conceptual Art cannot be equated with that of the concepts it invokes or deploys. There is arguably always something concrete involved in the making of a work of Conceptual Art, and in which its identity as that artwork, of that artist, is anchored. This applies even to one of the most emblematic of Conceptual artworks, Robert Barrys All the things I know but of which I am not at the moment thinking: 1:36 p.m., 15 June 1969, New York. For this work was created by a particular individual at a particular time and place, through a concrete act of articulation, one embodied in some particular physical inscription, to which its identity as an artwork would seem to be tied. See Extending Art Historically, Journal of Aesthetics and Art Criticism 51 (1988): 41123, reprinted in The Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996), and The Irreducible Historicality of the Concept of Art, British Journal of Aesthetics 42 (2002): 36779 (reprinted in this volume).

30

Art

or less adequate to what it now is to be an artwork, in the most comprehensive sense, in order to ask what that implies for the status of artworks as artifacts, and for the extent to which the artifactuality of artworks differs, if at all, from the artifactuality of artifacts in general. II ART WORKS VERSUS OT HER ARTEFACTS The psychologist Paul Bloom has attempted to extend the intentionalhistorical theory of artworks so as to cover artifacts of all kinds. Bloom proposes that for any artifact kind X, to be an X is to be an object successfully created with the intent that it be an X, where what it is to be an X at a given time is informed inescapably by past instances of X. Blooms insight is thus that all artifact concepts, and not just that of artwork, have an essential historical component, so that the past deployment of such concepts ineluctably enters into their present and future deployment, through the backward-directed intentions that the makers of such artifacts must of necessity possess. Blooms explicit statement of his proposal is as follows: We construe the extension of an artifact kind X to be those entities that have been successfully created with the intention that they belong to the same kind as current and previous Xs. Bloom is persuasive in pointing up the superiority of his proposal to existing competitors, those which analyze artifact concepts in terms of necessaryand-sufcient conditions, family resemblances, characteristic functions, or prototypes. But it remains to be seen whether Blooms own, original and sweeping, proposal is entirely acceptable. Suppose for the moment that Bloom is right, and that an analysis of the sort that captures what it is to be an artwork also captures what it is to be an artifact of any sort. What, if anything, would remain of the special historicality of the concept of artwork, as opposed to those of chair, pencil, house, or other standard artifacts? Two things, it seems. First, on Blooms analysis something is an X in virtue of being intentionally related in the right way to preceding Xs generally. But on the intentional-historical analysis of arthood, something can be an artwork through being intentionally connected in the right manner to a particular past artwork or artworks, whether or not intentionally connected to past art invoked generally. For example, someone could make an artwork of a sculptural sort by assembling pieces of wood
See Bloom, Intention, History, and Artifact Concepts, Cognition 60 (1996): 129. Ibid. 10.

Artworks as Artifacts

31

and plastic with the intention of the assemblage being regarded in an overall manner appropriate to Henry Moores Reclining Nude, but without any intent explicitly invoking the category of art or even the subcategory of sculpture. The history of art, it appears, enters more concretely into what can be art at a given point in time than the history of a given artifact kind enters into what can be an instance of that kind at that time. So to establish that something was an artwork might require tracing intentional relations to a particular item or episode in the history of art, but nothing comparable would seem to be required to establish that a candidate chair, say, was an instance of that kind. Second, it is arguable that standard artifact concepts, in contrast to that of artwork, retain at least some necessary conditions as regards form or function, whatever the historical dimension of their correct deployment. For instance, a chair must exhibit shape within a given broadly circumscribed range, with certain shapes being excluded in advance. And a chair must answer to a certain purpose in the case of chairs, that of being sat upon with some degree of comfort, or at the very least, be aimed at answering to such purpose. But I maintain that that is not the case with artworks as such, which in contrast to perhaps every other sort of artifact, retain only certain purely intentional-historical necessary conditions. In other words, nothing can be declared a failed artwork, in the sense of not succeeding in being an artwork at all, through failing to display a certain broadly specied form or a particular sort of functionality. But something can be declared a failed chair, in the sense of not even being a chair, if shaped like a javelin or if incapable of being sat upon at all. Thus even were an intentional-historical account of artifacts in general to be accepted, artwork would remain distinctively historical, in contrast with other artifact concepts, in respect of the creation involved requiring only the satisfaction of certain intentional-historical conditions. It is difcult to say whether these differences between artworks and other artifacts, which amount to the latter being less purely intentionallyhistorically determined, would be contested by Bloom. And that is because of the specic way he formulates his intentional-historical theory of artifacts, in which crucial appeal is made to the action of successfully creating an X. What is it, though, to successfully create an X? Does successfully creating an X differ from simply creating an X? If so, it should be possible to create an X, but unsuccessfully, which is not, I think, what Bloom is after. So successfully creating an X, it would appear, is just creating an X. What the adverb serves

32

Art

to call attention to, however, is a minimal success condition that Bloom apparently regards as constraining intentional-historical artifact making in general. This is evident in his illustration of how chair-making, for example, might fail, even though the intentional-historical condition of such making was in place: If someone intends to create a chair, but it falls to pieces as soon as it is nished, the person would not view this creation as successfully fullling his or her intent, and thus has not created a chair. Thus for Bloom, something intended to be a chair but that was merely a heap of materials incapable of being sat upon would not be a chair regardless of how rm the intention involved that it belong to the category of chairs. That seems right, but Blooms justication of this judgment is rather peculiar. He implies that such an object would fail to be a chair not because it could not fulll the basic function of affording single seating, but rather because its creator would not recognize it as the successful product of an intention to create a chair, that is, something effectively affording single seating. This is peculiar because it seems that whether or not something counts as a chair, though it may depend crucially on the intentions of its maker, should not ultimately depend on whether from the point of view of its maker those intentions are fullled, but rather on whether, from some objective point of view, those intentions really are fullled. For after all, a would-be chair maker may be deluded or confused, thinking that a pile of nails or a coiled length of rope for which he is responsible conforms well enough to past chairs to count as a successfully created new one. It seems that what is relevant to satisfaction of the minimal success condition is not the makers conception of a chair based on past acquaintance with them, but rather the conception of a chair endorsed by competent users of the term chair in general, one that imports at least some minimal features of form or function. Given satisfaction of that condition, the identity of a candidate thing as a chair may indeed be entirely determined, as Blooms account would have it, by an appropriate chair-history-invoking intention in its making. But the insistence on a minimal success condition, which rules out piles of nails, lengths of rope, decks of cards, javelin-shaped rods, and so on, as chairs, shows that some non-purely-historical conception of chairhood is in play in circumscribing the boundaries of the category. As regards artworks, however, it is far from clear that any such non-purely-historical conception of
See Bloom, Intention, History, and Artifact Concepts, Cognition 60 (1996): 129.

Artworks as Artifacts

33

arthood is in play, or that there is any minimal success condition of a substantive sort on the making involved. Why is that so? It helps to recall what, in the Post-Duchampian era, are two salient features of artmaking in contrast to standard artifact-making. First, one can more or less simply declare something a work of art, and it becomes such. Or at least one can, in certain contexts, or with a certain standing, do so. Second, anything, whatever its material constitution, cultural category, or ontological status, can become, or can be incorporated into, a work of art. These features are arguably enough to distinguish the concept of artwork from that of other artifacts, even if such artifacts, if Bloom is right, share with artworks the primary determining of their categorial status by a historical, or past-invoking, intention in their making. What is special about the artifact concept artwork, one might say, is that it is a wholly relational one; it is more like those of observed thing or beloved object or prize winner than it is like those of standard artifacts, such as chair or cup or cabin, for which there are at least minimal conditions of form as regards nished shape, of constitution as regards material, of making as regards the activity of the maker, or of functional success as regards usability of the nal product. III ARTMAKING AND SUBSTANTIVE CONCEPTIONS If the above is correct, then however sound the inspiration of Blooms intentional-historical theory of artifacts, he errs in blurring the difference between artworks and other artifacts, failing to appreciate that though minimal success conditions, rooted in some not-purely-relational conception of the kind of artifact in question, are ineliminably involved in the making of the latter, that is not the case with the former. Amie Thomasson, in a careful essay highlighting the insufciently acknowledged role that background conceptions of artifacts play in their creation, holds Blooms analysis at fault precisely for not sufciently acknowledging that role, and for thinking that artifact creation can proceed in a more conceptually thin or purely historical way than it in fact can. According to Thomasson, even the intention to create something of a given artifact kind K cannot consist merely in intending the object to belong to the same
See Thomasson, Artifacts and Other Creations, in E. Margolis and S. Laurence (eds.), Creations of the Mind (Oxford: Oxford University Press, 2006).

34

Art

kind as existing examples of K to which one can point or refer. Thomasson instead maintains, as a completely general principle, that a substantive conception of an artifact kind K must be involved in the intentional production of an artifact of that kind: the relevant sort of intention to make a thing of artifactual kind K must thus involve a substantive, and substantially correct, concept of what a K is, including an understanding of what sorts of properties are Krelevant and an intention to realize many of them in the object created. She adds that, for an artifact of that kind to be created, the intention in question must be largely successfully realized, which is the minimal success condition implicit also in Blooms account of artifactuality. Assuming that Thomassons principle is true for the making of gardenvariety artifacts, is it also true for the making of artworks, that most elusive species of artifact? Can one create an artwork without a substantive conception of what artworks are? More specically, does artmaking on an intentional-historical account of it require that an artmaker have a substantive conception of what he or she is making? The answer depends, in part, on how substantive is substantive. Can one make an artwork merely by intending something for the sort of regard or treatment appropriate to artworks, but without knowing what artworks are, in any qualitative sense, but only that there are such things, and that they are some sort of artifact, and without knowing what sorts of regards or treatments are appropriate to them, but only that there are such? I claim one can, and if so, one neednt have a substantive concept of what an artwork is, one implicating characteristic properties or functions. Does one need to possess in any ` measure a theory of art, a la Danto, in order to make art, or need one only know that there are such things as artworks and that there are ways it is correct to approach them? I claim not, and if so, once again an artmaker need not possess a substantive concept of what an artwork isthough of course virtually all artmakers will possess such, which concept will vary from artist to artist, and from artform to artform. Elsewhere in her discussion Thomasson offers an argument that could be seen as directly aimed at undermining the possibility just afrmed, one that according to the intentional-historical theory is sometimes realized in the making of art, that an artifact of kind K might be made merely by intending an object to stand in certain relations to existing instances of K:
See Thomasson, Artifacts and Other Creations, in E. Margolis and S. Laurence (eds.), Creations of the Mind (Oxford: Oxford University Press, 2006).

Artworks as Artifacts

35

If a later artisan does know of the existence of extant Ks, then he may have the intention to create one of those, with an implicit reference back to earlier K. Yet even so . . . his intention cannot be a mere transparent intention to create one of those . . . without any substantive concept of what those are, of what features are relevant to being of artifactual kind K. In order for his making to be controlled and directed, his intention to make a K must be lled out with intentions regarding what features are to be imposed on the object of his creation in order to succeed at realizing his intention to make a K.

But I think it clear that the stricture Thomasson here invokes, of a making lled out with feature-conscious or feature-directed subsidiary intentions, whose satisfaction is necessary for the making to succeed, though applicable to the making of standard non-art artifacts, such as chairs, as well as traditional art artifacts in established media, such as paintings, is inapplicable to artmaking in an appropriational or conceptual mode. Arguably nothing more is needed for successful artmaking in that mode than the belief that there is a practice of art, that various things are exemplars of it, and that there are correct ways of regarding, treating, or interacting with those things. So far as I can see, this necessarily involves the maker in some conceptions about art, to be sure, but not in substantive conceptions, in the sense Thomasson seems to have in mind, about the nature of artworks and their characteristic properties. IV TRADITIONAL ARTMAKING It would be remiss to end this short essay on the nature of artworks as artifacts without some remarks on the special character of artmaking of a traditional sort, that is, all artmaking before Duchamp, Warhol, and Conceptual Art, and most artmaking after them as well. In artmaking as traditionally conceivedand for simplicity I conne my attention to artmaking in the visual artsthere are distinctive raw materials, e.g. paint, clay, charcoal; there are distinctive techniques, e.g. carving, etching, impasto; and there are distinctive aims, such as visual beauty, representational verisimilitude, and emotional expression. But the making of chairs and pencils also involves distinctive materials, distinctive techniques, and distinctive aims, albeit utilitarian ones. So even if artmaking in the comprehensive, postDuchampian sense distinguishes itself from other sorts of artifact-making
Ibid.

36

Art

by its presumed purely intentional-historical character, is artmaking of the traditional sort, though issuing in a physical object whose interest is primarily aesthetic rather than utilitarian, fundamentally different from the making of physical artifacts generally, including the making of craft objects such as rugs or pots? To a degree. Dewey and Collingwood were two philosophers of art who had insightful and consonant things to say on the distinctive character of the making involved in traditional art, especially as in contrast with the making involved in the overtly similar activity of craft. What both thinkers stressed is that the making of an artwork is an open-ended, indenitely extended creative-critical process, with alternating phases of making and assessing, or doing and undergoing, but one not governed by any xed goal or preconceived idea of what the artwork must be, or how it must turn out. An artist making a sculpture, for example, in contrast with a craftsman making a rug or a pot, need not envisage what its dimensions will be, what it will look like, or what form it will have. This is unsurprising if one recalls that making a traditional artwork is, as much as anything, an expressive activity, but one in which, as Collingwood underlined, the artist does not know precisely what he has expressed until the process is completed. The maker of a craft object, though, must rst and foremost assure the creation of a usable object of the craft in question, some of whose features, such as atness or water-holding capacity, are accordingly non-negotiable, thus enjoining a preconception of some specicity on the craftsmans part of the object to be created. Granted the above, the upshot for our discussion is this. If, as Thomasson urges, the making of standard artifacts is always governed by a substantive conception of the artifact in question, one that sets clear terms for success and failure in such makings, and if traditional artworks are accounted standard enough in that respect, then what is most noteworthy about the making of such artifacts is that the substantive conceptions involved in their creation are relatively insubstantial, that is, not such as to constrain notably them in formal, material, or functional ways. A sculpture, say, needs to be physical, perceivable, and perhaps smaller than the planet, but apart from that, it can be of any size, any composition, any shape, any color, and any subject. The relative insubstantiality of the conceptions governing the making of traditional
See John Dewey, Art as Experience (New York: G. P. Putnam, 1934), and R. G. Collingwood, The Principles of Art (Oxford: Oxford University Press, 1938).

Artworks as Artifacts

37

artworks ts well, of course, with the innovative and exploratory aim often ascribed to art, both traditional and non-traditional. V CONCLUSION So what sort of artifact is an artwork? In the past, and thinking primarily of the visual arts, one might have answered: a physical object, fashioned with skill, involving a recognized medium, designed to be of aesthetic interest, and whose making is governed by a fairly substantive conception of the genre of artwork in question. And such an answer would still be largely adequate to at least traditional artmaking today. But at present, and just conning ourselves to the activities of visual artists, such an answer is no longer even remotely adequate. That is because of alternate modes of artmaking that have become entrenched in the past hundred years, whereby artworks need not be fashioned by their creators, need not involve recognized artistic media, need not be aimed at satisfying aesthetic interests, and whose making need not be governed by any very substantial conception of a genre in which the artist is working. Those modes of artmaking have revolutionized the concept of art, making it the case that the concept-of-art-2005 is something fundamentally, and not just marginally, different from the concept-of-art-1905. Artworks are necessarily artifacts, since they are things intentionally brought into being through human agency. That much remains true. But if I am right, to be an artwork today is simply to be something governed by an intention relating it in a certain way to what have been accounted artworks in the past. By contrast, more is required to be an artifact of a standard sort, such as a chair, even if the intentional-historical connection sufcient for being an artwork plays a crucial role there as well, in the manner that Bloom has underlined.

3
Emotion in Response to Art
Responding emotionally to artworks is a familiar enough occurrence, and hardly seems puzzling, recalled at that level of generality. Why should not works of art, in company with people, animals, natural objects and political events, produce emotions in us? Philosophers have, however, raised questions about emotional responses to art in particular contexts, or when viewed from certain angles. These questions suggest that there is indeed something puzzling about such responses. One such context is that of response to ctions, whether literary, dramatic, or cinematic ones, where emotions appear to be had, not only for the work or representation itself, but for the ctional characters or situations represented therein, even though these are perfectly well understood not to exist. A second such context is that of abstract or non-representational art, with music the example par excellence, where it is unclear both what could elicit such a response and what its object could be. A third context is that in which artworks expressive of negative emotion, for example tragedies, requiems, and tales of horror, engender parallel responses in perceivers without evoking avoidance or disapproval. And a fourth context in which emotional response to art has struck philosophers as problematic is where the proper appreciation of art is at issue, since such appreciation may be thought to be incompatible with experiencing life emotions of a familiar sort that art seems capable of raising in us. We might formulate the main philosophical questions concerning emotion in response to art as follows. (1) What kind of emotions are had in response to works of art? (2) How can we coherently have emotions for ctional persons or situations, given that we do not believe in their existence? (This query relates to what is known as the paradox of ction.) (3) How
First published as Emotion in Response to Art: A Survey of the Terrain, in M. Hjort and S. Laver (eds.), Emotion and the Arts (Oxford: Oxford University Press, 1997), 2034.

Emotion in Response to Art

39

do abstract works of art, especially musical ones, generate emotions in audiences, and toward what do audiences then have these emotions? (4) How can we make sense of the interest that many appreciators have in experiencing empathetically art that is expressive of negative emotions? (A particular form of this query is the paradox of tragedy.) (5) Is there a conict between responding emotionally to art and what the aesthetic appreciation of art demands? Answers to these questions depend, to some extent, on the conception of emotion adopted. I thus begin this essay by sketching a conception of the emotions that mediates between the sensationalist model of early twentiethcentury psychology and the cognitivist model widely favored in current philosophy. With that as background, various answers to the above questions will be critically reviewed. I T HE NATURE OF EMOTIONS In order to fruitfully assess the varieties of emotional response to art it is obviously of use to have some account of what exactly emotionsin the occurrent, as opposed to dispositional, senseare. Philosophical debate on the nature of emotions, informed to greater or lesser degree by available work in psychology, has in the past thirty years or so revolved around an opposition between feeling (or sensation) based, and thought (or cognition) based, approaches. The former holds that at the core of an emotion is an internal feeling or set of sensations, while the latter holds that at its core an emotion is a particular kind of thought, judgment, or evaluation. While the feeling approach has trouble accommodating the intentionality (or objectdirectedness) and amenability to reason of many emotions, the thought approach has trouble with the experiential aspect of emotions, that is, with what it is to feel them, as opposed to merely having the beliefs or entertaining the thoughts that may be associated with them, with the evident inertia and passivity of many emotional conditions, as well as with states of desire, whose
See, for example, among major studies in a cognitivist vein, Robert Solomon, The Passions (Garden City, NY: Anchor Books, 1978); William Lyons, Emotion (Cambridge: Cambridge University Press, 1980); Robert Gordon, The Structure of Emotions (Cambridge: Cambridge University Press, 1987); Ronald De Sousa, The Rationality of Emotion (Cambridge, MA: MIT Press, 1987); and Patricia Greenspan, Emotions and Reasons (London: Routledge, 1989). A useful review of this literature, to which I am indebted, is John Deigh, Cognitivism in the Theory of Emotions, Ethics 104 (1994): 82454. An even more extreme cognitivism about emotions is presented in Martha Nussbaums recent Upheavals of Thought (Cambridge: Cambridge University Press, 2001).

40

Art

connection with many emotions seems more than contingent. Still, while the feeling approach can be faulted for too mindless a picture of emotions, it is right to insist on bodily response and inner affect of some sort as a sine qua non of emotion, and while the thought approach can be faulted for too mindful a picture of emotions, it is right to emphasize that many emotions include cognitive elements essentially, e.g. thoughts with specic contents, which contents are, in many cases, socially shaped. At present there appears to be some consensus that, in perhaps the majority of cases, an emotion is best thought of as a bodily response with a distinctive physiological, phenomenological, and expressive prole, one that serves to focus attention in a given direction, and which involves cognition to varying degrees and at various levels. The level of cognitive involvement runs from mere registering of presence, to ways of seeing or regarding that which is registered, to propositional conceptions of the object responded to, to articulate beliefs about or attitudes toward the object of response. Alternatively put, an experienced emotion can be said to have as its core a bodily reactioncomprising physiological sensations, feelings of comfort and discomfort, and orientings of attentionwhich reaction is often caused or modied by, and is sometimes necessarily bound up with, cognitions of various sorts and strengths, depending on the type of emotion involved. Note that on such a view of emotion, in which cognitive representations on the order of beliefs (or for that matter, desires) are seen as characteristic of, but not essential to, experienced emotion, the intentionality or directedness of emotions (as opposed, say, to moods) is preserved by the root feature of orientation of attention to or focusing of concern on that which the subject registers as signicant. On the other hand, there is also a growing acknowledgement that the pretheoretically recognized emotions constitute an irreducibly heterogeneous class, i.e. that they do not form a natural kind. It seems reasonable to recognize a spectrum of emotional states experienced by humans, from the startle reaction, involving minimal cognition, at one end, to pride, envy, shame, jealousy, grief, remorse, embarrassment and the like, involving complex and often morally conditioned cognitions, at the other end, with hunger, surprise, lust, fear, anger, joy, sorrow, and so on lling in the vast
See, most relevantly, Jenefer Robinson, Startle, Journal of Philosophy 92 (1995): 5374. See on this, in addition to Robinson, ibid., Paul Grifths, What Emotions Really Are: From Evolution to Social Construction (Chicago: University of Chicago Press, 1996).

Emotion in Response to Art

41

middle ground. The emotional responses typical of engagement with art, though, tend to be of the moderately, or highly, cognitively involved sorta fact relevant to some recent attempts to dissolve too quickly the paradox of ction by appeal to what need not be true of all cases of emotion. Although it is convenient to speak of emotions having elements or components of various sorts, e.g. thoughts, sensations, desires, feelings, pleasures, pains, shifts of attention, these should not be thought of as merely bundled together, and the emotion as a mere conglomeration. The truth is rather that an emotion is an ordered complex or structure of the elements it is taken to comprehend, with causal relations prominent among those in which this order consists. My anger at my daughter for having carelessly misplaced my keys, for example, is a bodily response, rooted in physiology and reected in countenance, involving a focusing of attention on her, and feelings of agitation and displeasure, which feelings result jointly from my thought of her action and my desire that she not have so acted, while fueling, perhaps, my desire that she in some way pay for having so acted.

II EMOTIONAL RESPONSE TO REPRESENTAT IONAL ART: THE PARADOX OF FICTION The much-discussed paradox of ction can be formulated as a set of three propositions, to each of which we seem to have strong allegiance, but which are jointly inconsistent, and thus impossible to maintain coherently as a set. Solutions to the paradox, then, typically take the form of rejecting one or more of the propositions, with a reasoned justication for doing so. The propositions are these. (a) We often have emotions for ctional characters and situations known to be purely ctional; (b) Emotions for objects logically presuppose beliefs in the existence and features of the objects in question; (c) We do not harbor beliefs in the existence and features of objects known to be ctional. In the extensive discussion of this conundrum in the literature, almost every possible solution to it has been essayed. The following comprise most of the solutions that have found adherents. (1) The Non-Intentionalist solution: emotional responses to ctions are not, despite appearances, instances of emotions as such, but rather of less complex states, such as moods (e.g. cheerfulness) or reex reactions (e.g. shock), which lack the full intentionality and cognitivity of emotions per se. As is evident, this solution involves the denial of (a). But the diagnosis it

42

Art

offers seems to comfortably apply to only a small portion of the full range of developed responses to ctions. (2) The Suspension-of-Disbelief solution: while caught up in ctions, consumers thereof temporarily allow themselves to believe in the nonexistent characters and situations of the ction, and thus to have bona de emotions for them, reverting to standing beliefs in their nonexistence once the ction no longer actively engages them. Such a solution, turning on a denial of (c), though popular in the nineteenth century, unacceptably depicts consumers of ction as having both a rather tenuous grip on reality and an amazing ability to manipulate their beliefs at will. (3) The Surrogate-Object solution: emotional responses to ctions take as their real objects not known-to-be-nonexistent persons and events in ctions, but other, existent and believed-to-be-existent, objects. This solution, in one way or another, thus calls (a) into question. On one version of this solution, the object of response is simply the ctional work or artistic representation itself, or parts thereof. On another version, the objects of response are rather the descriptions, images, propositions, or thought contents afforded by the ction or representation. And on a third version, different enough from the preceding two to deserve a separate labelthe Shadow-Object proposalthe objects of response are real individuals or phenomena from the subjects life experience, ones resembling the persons or events of the ction, and of which the ction puts the subject covertly or indirectly in mind. The Surrogate-Object solution in its rst two guises distorts the logic and phenomenology of emotional response to ctions. Whatever the nature or status of our response to ctional characters or situations, it is an emotional response to them, not to something else. Our responses, however ultimately analyzed, have those characters and situations as their evident objects, and not the vehicles that bring them to us or the thoughts through which they are delineated. Much the same complaint can be brought against the Shadow-Object proposal, though here it is clear that the sort of response to which the proposal draws attention does indeed often accompany and
For discussion, see Eva Schaper, Fiction and the Suspension of Disbelief , British Journal of Aesthetics 18 (1978): 3144, and Nol Carroll, The Philosophy of Horror (London: Routledge, 1990). e See, for example, Peter Lamarque, How Can We Fear and Pity Fictions?, British Journal of Aesthetics 21 (1981): 291304, and Carroll, The Philosophy of Horror, ch. 2. See, for example, William Charlton, Feeling for the Fictitious, British Journal of Aesthetics 24 (1984): 20616.

Emotion in Response to Art

43

underlie the emotional response to ctional matters per se. Still, despising a ctional character, say, is not simply reducible to despising people of that sort generally, or to despising some actual similar individual of ones acquaintance. (4) The Anti-Judgmentalist solution: emotional responses to objects do not logically require beliefs concerning the existence or features of such objects, but only weaker sorts of cognitions, e.g. seeing a certain way, or conceiving in a certain manner, or regarding as if such and such; thus, there is no good reason not to categorize as standard emotions the emotional responses had toward ctions, since they satisfy the demands of a more relaxed cognitivism about emotions. This approach to the paradox, which directly challenges (b), has a growing number of proponents, and merits extended discussion. The instances of emotional response that challenge judgmentalismthe view that the cognitive element involved in all emotions is a judgment or beliefare mostly of two types. The rst type is where there is insufcient time for substantial cognition, so that no real representation of the object responded to is formed, there being only a virtually instantaneous reaction, instinctive or reexive in nature, unmediated by conscious thought (examples: apprehension at a suddenly looming shape, disgust at an accidentally felt slug). A second type is where, though cognition is involved in generating the response, the representation thus formed is either not propositional in nature, or else does not have the status of a judgment, or both (examples: phobic fear of garter snakes, unfounded resentment of female superiors). As noted earlier, the emotions involved in responding to ctions, ones such as pity, sorrow, love, admiration, anger, hate, hope, lie in the main in the middle and upper ranges of cognitive complexity for emotions. It thus seems undeniable that, whether or not they involve beliefs, such emotions are centrally mediated by representations of various sorts, such as views, conceptions, or evaluations, which serve to characterize the object of response. But now even if emotions at this cognitive level do not necessarily involve beliefs of a characterizing sort about their objects, such emotions, it seems, must still involve existential beliefs in regard to those objects, or something very close to that, i.e. attitudes or stances on the order of taking to exist or regarding as existent. Otherwise, the state attributed becomes unintelligible,
See, for example, John Morreall, Fear Without Belief , Journal of Philosophy 90 (1983): 35966; for a response, see Alex Neill, Fear and Belief , Philosophy and Literature 19 (1995): 94101.

44

Art

whether as an emotion or anything else. How can one be said to pity, fear, admire, or hate something that one does not, concurrently with ones emotion, at least take or regard as existing, now or at some other time? If indeed that cannot be said, then the problem resurfaces, despite what is right in the critique of judgmentalism: since sane consumers of ction do not take, regard, or view ctional characters as existing, even when fully engaged with them appreciatively, they cannot really be in the full-edged emotional states they are casually said to inhabit. The paradox of ction is proof against antijudgmentalist dissolution, even if we grant that emotions can occur without characterizing beliefs. The sticking point of the paradox of ction is the dimension of existence and nonexistence, as this connects to the cognitive characterization that emotions of the sort in question minimally require. When we view or conceive an object as having such and such properties, whether or not we strictly believe that it does, we must, on pain of incoherence, be taking said object to exist or regarding it as existent. For nothing can coherently be viewed or conceived as having properties without at the same time being treated as existent. A case of genuine emotion of a cognitively mediated sort, unlike a corresponding emotional response to a ctional character, involves at least viewing or conceiving an object as having such and such features, which thus in turn presupposes regarding it as existent or taking it to exist. But I do not, when reading Dostoyevskys The Brothers Karamazov, take Smerdyakov to exist, and so cannot strictly be viewing or conceiving him as having properties, such as being base or being a murderer. And though my evaluative Smerdyakov-thoughts, generated as I read, may largely be what causes my hateful response, directed ostensibly at him, for that response strictly to have him as its object, and so count clearly as an instance of hatred of Smerdyakov, requires, once more, that I take him to existwhich I clearly do not.
It is a mistake, in particular, to try to assimilate cases of ctional fear to cases of phobic fear, even apart from their evident divergence in behavioral consequences. With phobic fear we can say that, though the subject doesnt believe the animal in question is dangerous, the subject at least conceives or views the animal as dangerous, all the while clearly believing that the animal exists, that there is something the subject is so conceiving or viewing. With ctional fear, however, it is not open to us to say that the subject even conceives or views some ctional individual as dangerous or threatening, since the subject does not, if in his right mind, believe that such an individual exists; he does not believe there is any such thing to be so viewed or conceived. Full-blown fear of X has an irreducible cognitive component, one part of which is a viewing or perceiving X as dangerous, but another part of which is a taking or regarding X to exist.

Emotion in Response to Art

45

It may, however, be the case that I imagine, or make-believe, that someone of that sort exists, and that as a result I imaginarily, or make-believedly, experience for him an emotion of hate. Elaborating, my response to Dostoyevskys character can be interpreted, not as truly one of hatred for Smerdyakov, but as various sensations, feelings, and focusings of attention caused in me by my Smerdyakov-thoughts in the course of making-believe that he and his world exist, which may as a result amount to my make-believedly, or imaginarily, hating Smerdyakov. (For more on this, see section (7) below.) (5) The Surrogate-Belief solution: certain emotional responses to ctions, e.g. that of pity, require only beliefs that, in the ction, the character exists and is or does such and such, and those beliefs are indeed widely held by rational consumers of ction. This solution thus rejects (c), though not in the manner of Suspension-of-Disbelief theorists. However, the beliefs this proposed solution highlights, ones about what is ctionally the case, can only ground the truth of ones ctionally, or imaginarily, pitying a character, not of ones literally doing so. Furthermore, that such beliefs play a role in generating emotional responses to ctions does not touch the heart of the paradox, which is that intelligible emotions for objects of the sort typical of engagement with ction conceptually require beliefs in the existence of such objects, or at a minimum, an existential stance toward them. Beliefs about how things are ctionally can cause emotional reactions of some sort, to be sure, but they cannot logically ground intelligible emotions for entities whose existence is denied. Even where the emotion in question is such as to constitutively require beliefs, they are the wrong sort of beliefs to partly constitute such emotions. The beliefs I have in connection with Anna Karenina, say, cannot coherently make her the proper object of any pitiful reaction I might have. Pity involves concern for the welfare of and distress at the suffering of some creature. If one doesnt believe that such welfare or suffering is actual, what can one be concerned for or distressed about? Pity may likewise involve wishes or desires, with respect to the thing pitied, but absent a belief in the thing, or more loosely, an existential stance toward it, there cannot coherently be any such wishes or desires. (6) The Irrationalist solution: while caught up in ctions, consumers of ction become irrational, responding emotionally to objects that they know do
See, for example, Alex Neill, Fiction and the Emotions, American Philosophical Quarterly 30 (1993): 113; and Robert Yanal, The Paradox of Emotion and Fiction, Pacic Philosophical Quarterly 75 (1994): 5475.

46

Art

not exist and thus do not have the features they are represented as having. Irrationalists either implicitly deny (c), proposing that we in some manner do endorse the existence of ctional characters and eventswhile apparently at the same time disavowing themwhich qualies as irrational, in the sense of inconsistent, or else implicitly deny (b), holding that we can have emotions for such as ctional characters and events, toward which we lack the usual beliefs, but qualifying such emotions consequently as irrational, in the sense of unwarranted. On the rst construal the Irrationalist solution approaches closely that of Suspension-of-Disbelief, with the difference, perhaps, that no attempt is made to mitigate the clash of existential stances involved by suggesting that they are not simultaneously in full force. On the second construal the Irrationalist solution holds appreciators of ction at fault, not for believing what they already believe the negation of, namely, that ctional characters and events exist, but for emotionally responding to such characters and events in ways contraindicated by their beliefs. It might seem that the Irrationalist solution, on this second construal, is saved from being a non-starter by the rejection of judgmentalism, since otherwise it could be held to be simply impossible, rather than just possibly irrational, to experience full-edged emotions in the absence of certain beliefs. But as suggested earlier, if the critique of judgmentalism, applied to emotions of the sort that ction typically elicits, shows only that characterizing beliefs, as opposed to existential ones, may be absent in such cases, the logical space this construal hopes to occupy may not be available. In any event, in the judgment of most commentators, portraying the normal consumer of ction as fundamentally enmeshed in irrationality, however this be understood, is too high a price to pay for this to be an acceptable solution to the paradox. Before completing our survey of responses to the paradox of ction, it is worth revisiting an issue we may have too quickly settled in setting up the terms of the problem. The issue is this: Is what underlies the paradoxicality of emotional responses to ction a matter of the conceptual impossibility of a response being an emotion if it does not include, or is not premised on, certain beliefsas embodied in (b) above in our formulation of the
See Colin Radford, How Can We Be Moved By The Fate of Anna Karenina?, Proceedings of the Aristotelian Society Suppl. 49 (1975): 6793, and Harley Slater, The Incoherence of the Aesthetic Response, British Journal of Aesthetics 33 (1993): 16872.

Emotion in Response to Art

47

paradoxor is it, rather, a matter of the irrationality of responding with emotion to things believed not to exist or not to be as they are described? We are now in a position to see that this may depend, to some extent, on whether one focuses on existence-beliefs or characterizing-beliefs regarding creatures of ction. If the formerfor instance, that Anna Karenina exists or existedthe problem is perhaps best understood as one of rationality: it is simply irrational to have an emotional response to something whose existence you dont credit in any degree. If the latterfor instance, that Anna Karenina suffers or sufferedthen although the issue might again be construed as one of rationality, since pitying someone you dont believe suffers seems to qualify as irrational, it is perhaps better understood as concerned with the logic of the emotional concept in question. For the emotion pity for X might reasonably be held to strictly presuppose the belief X suffers, so that whatever one feels toward an item X, it cant be pity unless one believes X to be, as it were, a logically t object of pity. (7) The Make-Believe, or Imaginary, solution: emotional responses to ctions cannot, despite appearances, be instances of the ordinary emotions with whose names we tend to label them, but are instead instances of imaginary, or make-believe, emotions. For rst, the standard emotions of life arguably have belief or belief-like presuppositions, notably existential ones, that are not fullled in normal engagement with ctions, and second, such emotions have motivational or behavioral consequences that are not in evidence in the course of such engagement. The proposal is that in our interactions with works of ction we experience make-believe emotions, or make-believedly experience emotions, for ctional characters and situations; it is thus (a) that is rejected on this solution. Make-believedly experiencing fear, say, is enough like really experiencing fear, especially internally, that it is easily confused with it, and yet make-believedly experiencing fear can be reconciled, while really experiencing fear cannot, with the absence of existential endorsement and motivational upshot vis-`-vis a the ctions that are feared. In this way is the paradox nally resolved. In considering this solution, it is important to distinguish the claim that what we feel for ctional characters is some kind of emotion, or constitutes emotional response in the broad sense, from the claim, here disputed, that what we feel for ctional characters and describe with some ordinary emotion
See Kendall Walton, Mimesis as Make-Believe (Cambridge, MA: Harvard University Press, 1990) and Gregory Currie, The Nature of Fiction (Cambridge: Cambridge University Press, 1990).

48

Art

word is literally an example of such emotion. We are indeed moved, this solution afrms, but not strictly to the standard emotions whose names come to our lips. The issue is not whether making-believe can cause various emotional reactions, but whether those reactions, given certain cognitive conditions are not satised, qualify as full-edged emotions of the ordinary sort. Note also that to classify our emotions for ctions as imaginary is to say that they are ones we imagine ourselves to be having, on the basis of experiences, contributory to emotion, that we are actually having, but does not imply that such emotions are illusory or unreal. What makes some reluctant to accept that our emotional relations to ctional objects might be of a different stripe from our emotional relations to objects we take as existent, as the Make-Believe theory insists, is the sense that, on the inside, they seem very much the samethey feel the same, we might say. But as has been observed, there is more to emotional conditions than feelings. Cognitive and conative commitments play a role in the identity of many, though not all, emotions; thus, if those commitments vary, so may the emotion which is present. But suppose it is replied that consumers of ction do take or regard the characters encountered in ctions as existing, precisely insofar as they imagine them to exist as they engage with them? Very well. This can only mean that, imaginarily, they take them to exist, or that, in the ction, they take them to existnot that they take them to exist, period. A dilemma presents itself, in short, for those who would resist the MakeBelieve solution to the paradox of ction. Either such taking to exist of their objects as these emotions must be understood to involve amounts to belief, in which case the subject, who denies that ctional entities exist, is mired in inconsistency, or else such taking to exist amounts to making-believe to exist, in which case any emotion both based on that stance and directed on its object will be make-believe emotion, of the appropriate sort. Though the Make-Believe proposal thus probably provides the best resolution to the paradox of ction as such, a full account of our emotional responses when engaged with ctionsas opposed to our emotions for ctional characters per se will want to acknowledge what is called to our attention by the Non-Intentionalist and Surrogate-Object proposals as well. And even the Irrationalist proposal, on the rst construal, may contain a grain of
For discussion, see Jerrold Levinson, Making Believe, in The Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996).

Emotion in Response to Art

49

truth, for perhaps we are, at least at moments of maximum involvement, in the incoherent states of mind it postulates as ours throughout. III EMOTIONAL RESPONSE TO ABSTRACT ART: MUSIC AND FEELING Emotional response to abstract art is puzzling, principally, because the strategies that provide obvious explanations of both why we respond emotionally, and what we are responding to, in the case of representational art, here seem not to be available. A novel, lm, or Impressionist landscape gives me the image of a human world, elements of which I can empathize or identify with, react to sympathetically or antipathetically, or even mirror unthinkingly, by a sort of natural contagion. But with a symphony, sonata, minimalist sculpture, or Abstract Expressionist painting such explanations appear to have no purchase. Human beings and their predicaments are notably absent, at least as far as representation is concerned. So why, or how, does perception of such artworks raise emotion in us, and on what is such emotion directed? Concentrating for brevitys sake on the art of music, rough answers are as follows. Insofar as music is capable of eliciting emotions in listeners, this appears to work through two different routes or mechanisms, typically operating in tandem. The rst we may label the sensory, or cognitively unmediated route, and the second the perceptual-imaginative, or cognitively mediated route. It seems undeniable that music has a certain power to induce sensations, feelings, and even moods in virtue of its basic musical properties, virtually without any interpretation or construal on the listeners part. Particular timbres, rhythms, intervals, dynamics, and tempi exemplify this power most
See Malcolm Budd, Music and the Emotions (London: Routledge, 1985) and Values of Art (London: Penguin, 1995); Peter Kivy, Sound Sentiment (Philadelphia: Temple University Press, 1989) and Music Alone (Ithaca: Cornell University Press, 1990); Colin Radford, Emotions and Music: A Reply to the Cognitivists, Journal of Aesthetics and Art Criticism 47 (1989): 6976; Jerrold Levinson, Music and Negative Emotion and Hope in The Hebrides , in Music, Art, and Metaphysics (Ithaca: Cornell University Press, 1990), and Musical Expressiveness, in The Pleasures of Aesthetics; Stephen Davies, Musical Meaning and Expression (Ithaca: Cornell University Press, 1994); Jenefer Robinson, The Expression and Arousal of Emotion in Music, Journal of Aesthetics and Art Criticism 52 (1994): 1322; Francis Sparshott, Music and Feeling, Journal of Aesthetics and Art Criticism 52 (1994): 2335; Alan Goldman, Emotions in Music (A Postscript), Journal of Aesthetics and Art Criticism 53 (1995): 5969; Aaron Ridley, Music, Value, and the Passions (Ithaca: Cornell University Press, 1995); and Roger Scruton, The Aesthetics of Music (Oxford: Oxford University Press, 1997).

50

Art

clearly. Such properties need only be registered, as it were, to have their effect, at least for one acclimatized to a given musical culture. The rise in heartbeat caused by rapid tempo, the discomfort occasioned by dissonant intervals, the kinetic impulses induced by dancing rhythms, the excitement produced by quick alternations of soft and loud, the relaxation engendered by a certain tone color or manner of articulation, are all familiar phenomena. But if the capacity of music to elicit emotion were exhausted by the direct effects of sensing basic musical features, it would be a poor thing, falling far short of the evocation of emotions proper, or even the semblance of such. The gap is lled by the second, or cognitively mediated, route to such evocation. In addition to presenting an array of sonic features, simultaneously and successively, much music offers the appearance of human emotion, or of persons outwardly manifesting emotional states; arguably, that is what the expressiveness of music largely consists in. In other words, music is often heard as, or heard as if, or imagined to be, the expression of emotion by an unspecied individual, whom we may call the musics persona. The degree of resemblance between the shape of music and the behaviors through which emotions are commonly expressed in life will have something, though not everything, to do with our being disposed to hear music in such ways. In any event, once this occurs, the mechanisms mentioned above and familiar from appreciation of representational artmirroring, identication, empathy, sympathy, antipathycan come into play, resulting in the arousal in the auditor of those same emotions, or else the feelings characteristic of them, or else those emotions on an imaginary plane. The sensory aspect of music alone indeed seems capable of inducing in us at least a number of simple states of arousal typically identied as constituent elements of one or another emotion. But it is the perceptual-imaginative aspect, manifested in our disposition to hear emotion or emotional expression in music, that is surely primarily responsible for the complex, more robustly emotional responses to music, whether mirroring or reactive, that so many listeners report. It remains to add that these mechanisms do not operate in total isolation from each other. The emotion I hear a passage as expressing may soften or accentuate the particular psychological effect some basic musical feature produces on me, while the effect induced in me, largely unthinkingly, by some basic musical feature may inuence and constrain the emotion I am disposed to hear an image of in the music. But if emotions are often produced in listeners in virtue of auditing emotionally expressive music, toward what are such emotions directed? Music

Emotion in Response to Art

51

neither supplies any objects, nor appears itself to be an appropriate object, for at least the vast majority of such emotions as are putatively aroused. In addition, music does not seem to provide anything that would justify the beliefs or attitudes toward objects that many objects can be held to require. Among the ways of responding to this difculty are the following. It can be held that music produces in listeners only moods, which intrinsically lack intentionality, e.g. anxiety or elation, or else objectless emotions, ones characteristically taking objects but somehow lacking them when aroused by music, e.g. sadness or joy directed on nothing, or nothing in particular. Alternatively, it can be held that music produces in listeners just the feeling component of an emotion, together with the sense of focus or directedness inherent in the bodily response at the emotions core, but not the cognitions which characteristically accompany or even partly constitute the emotion. Finally, it could be maintained that what music occasions in many listeners are states of imaginary emotion. The idea is that listeners readily erect, upon a basis of feelings produced in them by music whose expressiveness they empathetically grasp, imagined emotions of a corresponding sort, and that they do this through imagining, usually tacitly, objects and thoughts suitable to the emotions in question. The object of musical emotion, then, is not missing, but merely posited indenitely in imagination, or perhaps appropriated, as it were, from the emotion imaginarily ascribed to the musics persona.

IV EMOTIONAL RESPONSE TO NEGATIVELY EMOTIONAL ART: THE PARADOX OF T RAGEDY The paradox of negative emotion in artof which the paradox of tragedy is a classical illustrationis this. Art that is negatively emotional, i.e. art that represents, expresses, or otherwise deals with emotions such as shame, grief, horror, sorrow, anger, remorse, despair and the like, seems to have a propensity to elicit parallel responses in appreciators. But if that is so, one would expect appreciators to avoid, or at any rate judge as inferior, art of this nature. Yet not only do they not do so, but often they hold such art to be the highest or most rewarding art of all. A number of possible explanations have been given for how it is that persons rationally desire or value the empathic experience of negatively emotional

52

Art

art, given the ostensibly negative character of that experience. Here is a rough categorization of such explanations. (i) Compensatory explanations: negative emotion aroused by negatively emotional art is, as such, unpleasant, but undergoing it offers other rewards that compensate for this. (ii) Conversionary explanations: negative emotion, which is initially or ordinarily a disagreeable response, is transformed, in the context of artistic appreciation, into something that is in fact agreeable, or at any rate, capable of being enjoyed. (iii) Organicist explanations: negative emotion aroused by negatively emotional art is an essential element in a total experience, an organic whole that is desired or valued. (iv) Revisionary explanations: neither negative emotions, nor the feelings they include, are intrinsically unpleasant or undesirable, and thus there is nothing odd about appreciating art that induces such emotions or feelings. (v) Deationary explanations: despite appearances, neither negative emotions, nor the feelings they include, are really aroused in us by negatively emotional art. Compensatory explanations include Aristotles doctrine of catharsis, understood as a purging or purication of excess or unruly emotions of pity and fear through engagement with tragic drama, which justies the raising of such emotions in the course of that engagement. Another such explanation appeals to the value of knowledge of important truths of human existence that emotional engagement with negative art is said to afford. A third explanation endorses such engagement, not for the knowledge of life it may afford, but rather for the knowledge of the artwork it facilitates, emotional engagement with a work being seen as a necessary cost, in many cases, of fully understanding it. A fourth such explanation invokes the moral exercise that is provided, or the moral deepening that results, as a benet of engagement with negatively emotional art. And a fth explanation appeals to purely aesthetic pleasures in the beauty, lifelikeness, virtuosity, or cognitive interest of the representation or expression itself, positing these as enough to outweigh whatever negative emotion is undergone in their appreciation. Conversionary explanations include Humes explanation of the appreciation of tragedy; like that just noted, Humes explanation highlights the pleasure in artistic representation and expression as such, but premises that this
See, for instance, Mark Packer, Dissolving the Paradox of Tragedy, Journal of Aesthetics and Art Criticism 47 (1989): 21219. See, notably, Nelson Goodman, Languages of Art (Indianapolis: Bobbs-Merrill, 1968), ch. 6. Nol Carrolls resolution of the paradox of horror, a rst cousin of the paradox of tragedy, is e largely of this sort. See his The Philosophy of Horror.

Emotion in Response to Art

53

pleasure, being greater than the pain of the negative emotions concomitantly raised, does not simply offset that pain, but rather overwhelms and absorbs it, leaving an experience of uniformly positive character. A rather different conversionary explanation proposes that since the negative emotions raised by a work of art have no life implications for spectators, calling for no actions and betokening no real harms, such emotions must evidently be so altered by the artistic conditions under which they issue that, though still recognizable as this or that negative emotion, disagreeable affects intact, they are yet capable of being relished or enjoyed for experiences sake. An example of an organicist explanation would be one invoking a satisfaction in some negative emotions being raised in one by a work of art, perhaps because the emotion strikes one as appropriately raised in such circumstances, and oneself as admirably human for being thus susceptible. Such a satisfaction would obviously be inseparable from the negative emotion raised, in the fact of which satisfaction is taken. Another such explanation would appeal to the value of working through negative emotions in connection with a work of art, via immersion in its formal, narrative, or dramatic structure, the emotions raised thus being an essential element in the experience valued as a whole. Revisionary explanations go something like this. The experience of negative emotions is not intrinsically unpleasant; the affects, that is, sensations and feelings, involved are not in themselves disagreeable, and can be unproblematically savored as such, in appropriate contexts. What is negative about negative emotions is only the evaluation of their objects that is central to such emotions. Thus, there is no special difculty about people seeking these emotions from art. Deationary explanations come in at least three varieties. One hypothesizes artistic analogues of the life emotions, distinct from them in hedonic tone, conative connectedness, and behavioral implication, and proposes that only these are raised in us by engagement with emotional art, and not the life emotions themselves. Another deationary explanation simply atly denies that
See Marcia Eaton, A Strange Kind of Sadness, Journal of Aesthetics and Art Criticism 41 (1982): 5163, and Levinson, Music and Negative Emotion. See in this vein Susan Feagin, The Pleasures of Tragedy, American Philosophical Quarterly 20 (1983): 95104. See John Morreall, Enjoying Negative Emotions in Fiction, Philosophy and Literature 9 (1985): 95103; Jenefer Robinson, Leducation sentimentale, Australasian Journal of Philosophy 73 (1995): 21226; and Levinson, Music and Negative Emotion. A suggestion of this sort can be found in Kendall Walton, Mimesis as Make-Believe (Cambridge, MA: Harvard University Press, 1990), ch. 7, and also in Berys Gaut, The Paradox of Horror, British Journal of Aesthetics 33 (1993): 33345.

54

Art

anything like the garden-variety emotions are evoked in subjects in the course of engaging with emotional art, and suggests that the subjects response, insofar as it is emotional, is exhausted by properly appreciative reactions, such as being moved by a works beauty of expression. A third deationary explanation maintains that spectators are always only make-believedly in states of negative emotion in virtue of engaging with a work of art, and that on the assumption that make-believe emotions of the negative sort are not inherently displeasing, there is no special problem about people tolerating, or even actively pursuing, such experiences. Detailed assessment of these proposals awaits another occasion, but in my view there is more merit in compensatory and organicist explanations, and in the second of the conversionary explanations sketched above, than in revisionary or deationary ones. V E MOTION AND THE APPRECIATION OF ART Are there emotions unique to the appreciation of art, or aesthetic emotions per se, had when and only when a work is apprehended aesthetically? Past theorists, notably Clive Bell, have posited something of this sort, but such a posit has not lately found favor, nor does it appear to answer to any pressing theoretical problem about art. There may, on the other hand, be an interesting category of positive emotions that, if not had uniquely for art, are both distinctive of the appreciation of art and not of the sort that typically gure in the content of art. Candidates for membership in this category would be emotions such as admiration for a works skill, fascination with a works form, delight in a works beauty, or awe at a works depth of expression. What might also gure here are experiences, remarked by many, of momentary will-lessness or self-transcendence occasioned by intense absorption in a work of art. The question may also be raised as to the appropriateness of emotional responses to art of the ordinary sort. One form of this question concerns an apparent tension between the familiar picture of an emotion as a disturbing temporary derangement of the psyche and the image of aesthetic appreciation as a state of calm and unclouded attention to a work of art.
A suggestion of this sort, applied to music, can be found in Kivy, Music Alone, ch. 8. Suggestions of this sort are also to be found in Walton, Mimesis as Make-Believe. But see the debates on this, pro and con, in Monroe Beardsley, The Aesthetic Point of View (Ithaca: Cornell University Press, 1983), and George Dickie, Art and the Aesthetic (Ithaca: Cornell University Press, 1974).

Emotion in Response to Art

55

The traditional notion of the aesthetic attitude, whose roots are in Kant, Schopenhauer, and the eighteenth-century theorists of taste, depicts a frame of mind characterized by disinterestedness, detachment, and disengagement from the practical. Charitably construed, such a notion demands only that ones personal situation or condition not be what primarily drives or directs ones response to a work, but instead, the humanly signicant material that the work presents. In other words, such a notion need not call for suppression of emotional receptivity generally. So long as ones emotional response is a way of connecting to a work, of tracing its expressive outline, or grasping its dramatic import, rather than a means of being distracted from it, a springboard to simply wallowing in ones private concerns, then there is no conict between responding to a work with a range of ordinary emotions, on the basis of ones life experience and individual sensibility, and appreciating a work in an aesthetically appropriate manner, as the specic embodiment of human content it is. By contrast, disinterestedness or detachment understood not as a principle for maintaining focus on a work rather than ones own circumstances, but as a desired end-state of impassivity or imperturbability, is nothing that an account of artistic appreciation need embrace. Finally, accounts of the value of emotional response to art can be divided roughly into those that exploit the value of emotional experience generally, for example in contributing to a full life, and those that seek instead to identify a particular value of emotional experience in the context of art and its appreciation, for example as a mode of understanding a work more fully or a means of reaping more efciently the benets a work has to offer.
For further discussion, see Jerrold Levinson, What Is Aesthetic Pleasure?, in The Pleasures of Aesthetics. See, for example, the discussion as applied to music in Davies, Musical Meaning and Expression, ch. 6.

4
Elster on Artistic Creativity

1. The French novelist Georges Perec wrote a novel, La Disparition, in which the
letter e was nowhere used . . . This is an extreme example of the more general idea that artists may impose constraints on themselves in order to create better works of art. In Perecs case the constraints were entirely idiosyncratic. In other and more frequent cases the constraints take the form of conventions that dene a particular genre. Although freely chosen, in the sense that it is up to the artist whether to submit to the laws of the genre, they are not invented by the artist. In still other cases the constraints are imposed from outside . . . (1756)

So begins Jon Elsters discussion of creativity and constraints in art in the third part of his recent study, Ulysses Unbound: Studies in Rationality, PreCommitment and Constraint. I return later to La Disparition, indicating my differences with Elster on the virtues of that singular work, but I begin by exploring how far we might take the tripartite division of artistic constraints that Elster here suggests, into chosen, invented, and imposed ones. Chosen constraints are ones adopted by the artist from a pool of preexisting styles, genres, or forms. Invented constraints are ones devised by the artist, and to which the artist subsequently adheres. And imposed constraints are ones rooted in external conditions that cannot be altered or evaded by the artist. Some examples of chosen constraints are: sonata form; sonnet; haiku; iambic pentameter; still-life painting; charcoal drawing; two-person play, comedy of manners; Greek temple; Roman arena. Some examples of invented constraints are: twelve-tone composition; musique concr`te; prepared piano e ` ` pieces a la John Cage; collage painting; kinetic sculpture; readymades a la
First published in B. Gaut and P. Livingston (eds.), The Creation of Art (Cambridge: Cambridge University Press, 2003), 23556. Jon Elster, Ulysses Unbound: Studies in Rationality, Pre-Commitment and Constraint (Cambridge: Cambridge University Press, 2000). Page references to this book in the text are given in parentheses.

Elster on Artistic Creativity

57

` ` Duchamp; rayograms a la Man Ray; and absurdist theatre a la Beckett or Ionesco. Some examples of imposed constraints, nally, are these: the maximum length of breath possible in singing; the maximum loudness attainable in solo piano music; the maximum height of jump in dance; the rectangularity of frames or shots in lm; the irreversibility of shaping in carved sculpture; the prohibition of live rearms in installation art; the prohibition of certain subject matters in PG-rated lm. With both chosen and invented constraints there is evidently an element of voluntary participation by the artist, whereas with imposed constraints this appears to be lacking. But it is worth observing that even with imposed constraints, the artists agency is not entirely annulled. In the rst place, the artist is still faced with the choice between various attitudes that can be taken to the imposed constraint. For example, cheerful acceptance, stoic resignation, angry deance, or blithe obliviousness. In the second place, the artist is always free to decline the gambit entirely, electing not to create subject to the constraint in question, and simply passing on or forgoing the artform or endeavor in which that constraint is unavoidable. If an artist chooses a constraint, says Elster, he presumably believes he will benet artistically from having a smaller sphere of choice. This is a case of standard pre-commitment, and the constraint is thus an essential one, i.e. consciously chosen for the expected benets of it. In other cases a constraint is adopted or accepted, but not with an eye to reaping a benet; it is then an incidental constraint. An incidental constraint may end up being of benet, but such benet is not the reason for its adoption or acceptance. And an incidental constraint may also evolve into an essential constraint, if endorsed by the artist for its expected benets after initial adoption. 2. One of Elsters main themes is the contrast between choice of constraints, at the outset of the creative process, and choice within constraints, occurring throughout the creative process:
A nice example of a fruitful attitude one might adopt toward imposed constraints is provided by the following anecdote, from liner notes by Harry Halbreich to Arthur Honeggers Le Roi David (Erato 2292-45800-2). Honegger, having been commissioned to write music for the theatre piece in question, but chang under the rather peculiar musical conditions the commission involved, asked his fellow composer Stravinsky for advice. This was the advice offered: Its quite simple . . . Act as if you had wanted exactly that ensemble, and compose for a hundred singers and seventeen instrumentalists. Honegger comments on this advice that it constituted for him an excellent composition lesson: never consider the givens of a commission as something imposed, but on the contrary as a personally chosen stricture, or an interior necessity.

58

Art

The creation of a work of art can in fact be envisaged as a two-step process: choice of constraints followed by choice within constraints. The interplay and back-and-forth between these two choices is a central feature of artistic creation, in the sense that choices made within the constraints may induce an artist to go back and revise the constraints themselves. (176)

Elsters description has the ring of truth: there are undoubtedly phases in the process of creating works of art, and the phases may typically be related in the way that Elster proposes. But still we must ask, if the artist is free to go back at will, at any point, and revise the constraints under which he had been operating, is it really useful to think of them as constraints, rather than as, say, provisional guidelines, or working assumptions? Elsters idea of constraints as a constant and dening feature of the creative process seems somewhat too . . . constraining. To be fair, Elster acknowledges that there is an interaction between choice of constraints and choice within constraints in the making of art. An example he gives is this:A writer may initially plan to develop an idea in a full-length novel, and then, nding that it will not sustain that format, turn it into a short story of thirty pages. (201). Even so, Elster arguably underestimates the purely heuristic value for artistic creativity of mere opening moves or posits, ones not thought of as binding, and thus not reasonably construed as constraints. The self-binding of an artist, in other words, is rather unlike the self-binding which served as a paradigm in Elsters earlier studies of the rationality of precommitment, namely that undertaken by Ulysses in having himself bound to the mast of his ship as a prophylactic against the sirens. For Ulyssess selfbinding was entered into (a) deliberately and explicitly, (b) categorically and unrevisably, and (c) with a denite objective in view. But the self-binding of an artist, if we consent to call it that, is typically (a) inexplicit or only partly explicit, (b) eminently revisable, and (c) often entered into with no very denite objective in view. In short, Elsters model of precommitment is unnecessarily inexible, and sorts ill with the way artists actually go about creating works. 3. Elster is explicitly committed to an account of artistic creation that sees it as a project of knowingly engaged-in maximization:
See, in particular, his Ulysses and the Sirens: Studies in Rationality and Irrationality (Cambridge: Cambridge University Press, 1990). Michael Bratman, by contrast, offers a more exible model of pre-commitment, one that is perhaps a better t for the artistic case. Bratman stresses the need for rational agents to frame future-directed intentions that have a certain inertia or weight, yet not be bound to pour good money after bad. See his Faces of Intention: Selected Essays (Cambridge: Cambridge University Press, 1999).

Elster on Artistic Creativity

59

. . . both choice of constraints and choice within constraints can be represented as a form of maximization. Specically, artists try to maximize artistic value. (178) The process of artistic creation is guided by the aim of maximizing aesthetic value under constraints . . . Creativity is the ability to succeed in this endeavor. (200)

The idea that artists seek to maximize aesthetic or artistic value certainly has some plausibility. If agents are generally, to a rough approximation, utility-maximizers according to their own, often idiosyncratic, utility functions, then it is not too big a stretch to suggest that artists are, to some extent and in some contexts, maximizers of the utility that corresponds to aesthetic or artistic value. But is not Elster guilty of making the creation of art out to be a too predominantly rational affair? First, it is not clear that the maximizing aim stated above can plausibly be attributed to most successful artists, even at a less than fully conscious level. Whether on the basis of express avowals, observed behavior, or the character of completed works, the evidence does not make such an attribution inescapable. Second, even those most committed to rational explanation of human behavior usually allow that there is some share of the irrational, or at least nonrational, in creative activity, as in related activities of a sexual or religious kind, with their not wholly transparent bases. Now Elster might try to accommodate that by suggesting that acceding to a limited irrationality in the service of creative ends can be seen as a strategy that is overall rational, and perhaps even maximizing of aesthetic value, but surely this is to ascribe too calculating an attitude to artists who open themselves to the irrational in the name of art. In other words, though it is true that irrational methods are sometimes rationally chosen by artiststhink of the automatic writing experiments of the Surrealistssurely not all irrationality issuing in something of artistic value has been deliberately opted for by the artist on rational grounds. Third, and most importantly, there are positive aims we have good grounds to ascribe to artists as motivating them that do not reduce to that of maximizing aesthetic or artistic value. Some artists are driven to discover new relationships between elements of their chosen media, or new ways of expressing states of mind, or new techniques for achieving realistic representation, where their commitment to those goals as such overrides any direct concern they might have to maximize aesthetic value within given constraints. In even simpler terms, artists are most often driven to make works the works they want
Elster is not too careful of this distinction, a matter to which I return below.

60

Art

them to be, whether or not they are seen to maximize aesthetic value, even by the artists own lights. 4. Elster, like some other recent writers, takes a fairly rm stand against originality: I distinguish creativity (working within constraints) from originality (changing the constraints), arguing that the latter has no intrinsic relation to aesthetic value (180). Elster sustains that judgment even in light of this later, ostensibly approving gloss of what originality involves: Originality . . . can be dened as a durable break with existing conventions rather than as a momentary departure from them. The emergence of free verse, nongurative art, and atonal music are obvious examples from the past century (224). Elster recognizes that originality in a work of art can coexist with greatness, and that the search for originality may even enhance the pursuit of creativity, but he remains opposed to the idea that originality can be part of the value of a work of art. Elsters judgment against originality per se, however, fails to recognize artmaking as an activity that, like most activities, involves the possibility of accomplishment or achievement, and the artwork that results as the embodiment of such accomplishment or achievement, which makes the work logically and appreciatively inseparable from the activity that generates it. Part of the achievement of some works of art is precisely the striking originality, whether of means, of ends, or of ends-in-relation-to-means, that they manifest. Once the achievement in question is solidied and secured, usually through the production of a series of related works, then later works, however similar, do not partake of that achievement and do not exhibit that
This observation, which I owe to Richard Wollheim, abstracts from purely extrinsic or strategic considerations, such as might induce an artist to make a work a certain way to please a patron, to secure a commission, or to annoy a rival. In a sense, an artist in such cases is making a work the way he wants it to be, but not for its own sake; in other words, he is not making quite the work he wants, qua artist, to make. See, for example, Frank Sibley, Originality and Value, British Journal of Aesthetics 25 (1985): 16984; and Bruce Vermazen, The Aesthetic Value of Originality, Midwest Studies in Philosophy 16 (1991): 26679; and more evenhandedly, Paul Crowther, Creativity and Originality in Art, British Journal of Aesthetics 31 (1991): 3019. Sibleys argument against originality as such goes like this: original is very commonly used of style, manner, technique, medium; Cubism, Abstract Expressionism, Pointillisme, the twelve-tone system exemplify large innovations . . . In themselves, these were extremely original but aesthetically quite neutral . . . The value of such innovations is instrumental; it lies in what new aesthetic characters and values they render possible . . . So the enthusiastic praise often lavished on works employing innovatory techniques [or] . . . novelties of form or medium is misplaced unless these bring, or have the potential to bring, some new aesthetic character of value (1745). There is some irony, though, in Elsters opposition to artistic originality, given the reputed creed of Elsters musical hero, Lester Young: Youve got to be original, man.

Elster on Artistic Creativity

61

originality, and hence are not as admirable artistically. The locus of originality in art is the historically rooted artwork itself, not the works maker as an individual, and not the works abstract form or detachable content. Arguments that originality is not an artistic value, moreover, tend to equate artistic value, or value attaching to art as art, with narrowly aesthetic value, that is, value derived from sensuous-perceptual-imaginative experience of an objects formal and aesthetic qualities. But the two are not equivalent. Artistic value is a broader notion than aesthetic value, though it includes aesthetic value as its core. For clearly the experiential rewards afforded an audience by an artwork, and by design, are a large part of its value as art. But some of the value of an artwork is not anchored in its character as an object for perceptual or imaginative experience, which it may share with portions of nature, but is instead rooted in its being a human artifact, in virtue of which it, say, displays originality, expresses singular attitudes, communicates moral insight, or inuences in a positive manner artworks to come. These contribute to its artistic value, though not to its aesthetic value in the narrow sense. 5. Aristotle famously offered a criterion of organic unity, the essence of which is that a work possessed of such unity can only be altered for the worse. Elster, it appears, is inclined to accept Aristotles dictum: One piece of evidence for the view that artists are engaged in maximizing is the widespread belief that in a good work of art nothing can be added and nothing be subtracted (Aristotle, Poetics.) . . . Another piece of evidence is the widespread practice of artists of experimenting with small variations until they get it right (202). But Aristotles dictum should not in fact be so uncritically accepted. From the fact that in a good artwork every element plays some role in the functioning of the work or the emergence of its artistic content, one cannot validly infer that only those elements could have served that function or have resulted in that artistic content. More generally, one cannot conclude that only the specic ensemble of elements that constitutes the nished work can be thought to optimize the value such a work might possess. This is because there may be several, equally effective ways of getting it right, that is, bringing a work to successful completion once it has reached a certain point. The practice of trying out small variations need not be seen as a search for a unique
For more on artistic value versus aesthetic value see my Art, Value, and Philosophy, Mind 105 (1996): 66782. On the related though not equivalent distinction that can be made between the artistic and the aesthetic properties of works, see my Artworks and the Future, in Music, Art, and Metaphysics (Ithaca: Cornell University Press, 1990).

62

Art

local maximum. In fact, it need not even be viewed as a search for any local maximum, but might instead be seen as governed by a principle of good enough, that is, as conformable to a satiscing, rather than optimizing, model of human behavior. Even if getting it right generally motivates the trying out of small variations of a work in progress, that may not require making it best, but only, say, really quite good. The error of unqualied adherence to the doctrine of organic unity, which maintains that successful works of art are so nely tuned that no alteration in them can fail to have an adverse effect, is related to that involved in unrestricted allegiance to the principle of aesthetic uniqueness, which holds that any two perceptually different works of art belonging to the same genre necessarily differ in some aesthetic respect or other. But though close to true, that principle falls short of universal validity. Though perceptually different artworks in a given genre necessarily afford different experiences and perhaps even experiences that differ aesthetically, if the perceptual basis of an aesthetic property always comes within the scope of its appreciationthat does not yet imply that the works themselves necessarily differ aesthetically. Think of two abstract, calligraphic paintingssay in the vein of Mark Tobeywhere one is a mirror image, or else a slight internal rearrangement, of the other. Is it clear that these will differ at all in aesthetic impact, despite their being easily perceptually discriminable? Both such paintings might work, so to speak, and in exactly the same way, and without inclining one to think that no other arrangement of similar elements could have worked as well. 6. Elsters discussion of the dimensions of value in temporal artforms, and their particular manifestation in the sphere of classic jazz, is to my mind the most satisfying and stimulating part of his inquiry into artistic creativity in Ulysses Unbound. I thus take the liberty of quoting from it at some length.
I distinguish between two sources of emotional satisfaction through the arts. On the one hand, many works of art can generate non-aesthetic emotionsjoy, grief, and the like. On the other hand, all works of art, if they have any artistic value at all, induce specically aesthetic emotions by means of rhythm, echoes, symmetries, contrasts, repetitions, proportion, and similar devices . . . The specically aesthetic emotions include wonder, amazement, surprise, humor, relief, and release. (206) In literature and music, sublime artistic effects are created when these two emotional effects go together and reinforce each other. As if by magic, the pleasure of a
See my Aesthetic Uniqueness, in Music, Art, and Metaphysics. See my What Is Aesthetic Pleasure?, in The Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996).

Elster on Artistic Creativity

63

rhyme that falls into place adds to the poignancy of the words. Although the components are analytically separable, they are not experienced separately. (206) In the case of jazz, the non-aesthetic emotions in the listener are produced by music that has what I shall call emotional depth. The aesthetic emotions can arise at two levels. At the simplest level, they are produced by music that possesses what I shall call taste. At a more advanced level, the aesthetic emotions are produced by the story told by the music . . . (247) Tastethe sense of order, balance, proportion, timingis an essential prerequisite for the production of the specically aesthetic emotions . . . Emotional depth refers to the capacity to generate strong non-aesthetic emotions in the listener. (248) Taste and emotional depth are only two of the relevant dimensions of quality in jazz. The thirdand the only one in which jazz differs radically from other musical performancesis inventiveness in storytelling. (252) The ability to tell a story through melodic innovation is related to taste, but goes well beyond it. (253)

There is much with which to agree in the above extracts, and the discussion from which they are drawn is unquestionably an insightful contribution to the aesthetics of jazz. But I have, nevertheless, certain qualms. My rst qualm concerns Elsters claim that the storytelling dimension of a jazz solo, its ability to suggest, through its melodic, harmonic, and dynamic evolution, an abstract narrative of some sort, enters into the generation in the listener of aesthetic emotions, but not non-aesthetic ones. This seems unwarranted. For the narratives that music can suggest are often themselves emotional ones, and it stands to reason that our response to an emotional sequence might be a further non-aesthetic or life emotion that would be appropriate to the sympathetic contemplation of such a sequence. For example, a solo conjuring up a narrative of trouble, followed by hope for that troubles disappearance, leading only to the dashing of that hope, would naturally conduce to a feeling of sorrow or distress in a listener, apart from any purely aesthetic reaction on the listeners part to the rightness of the emotional narrative or its musical underpinning. My second qualm relates to the degree to which taste, characterized as a formal or congurational matter, and emotional depth, characterized
What precise distinction Elster has in mind by aesthetic vs. non-aesthetic emotions is not entirely clear. Some emotions which he denominates aesthetic, for instance wonder, can certainly be had in connection with objects that are not works of art. However, I will accede in the distinction as drawn, taking it to be roughly this: aesthetic emotions are ones characteristically had for or directed on the aesthetic aspects, including formal ones, of objects, most notably works of art. For an emotional scenario of that sort, and its role in the generation of musical frissons, see the discussion of a Scriabin etude in my Musical Chills (Ch. 12, this volume).

64

Art

as a material or substantive one, are entirely independent and summable dimensions of musical value, as is suggested by Elsters amusing diagram (251) locating various jazz artists in a two-dimensional space whose y-axis is taste and whose x-axis is emotional depth. I think these dimensions are rather interlocking ones, with a musicians formal sense of timing, proportion, and ow impacting on the non-aesthetic emotions his discourse will be able to elicit. For example, the concision, clarity, and drive of John Coltranes Giant Steps, presumably the result of the great saxophonists taste, have quite a lot to do with the bold, life-afrming condence the music expresses. And the relative structural freedom and relaxation of Coltranes A Love Supreme is surely not separable from the serenely rapturous sense of the goodness of things that that music projects, especially in its rst two sections. In addition, we may note that the formal and emotional dimensions of musical value in jazz or any other music are arguably subordinate to another, overarching and non-summative, dimension of value we can label emotionin-relation-to-form, or content-in-relation-to-conguration. That is, our ultimate judgment of the value of a piece of music arguably does not rest only on our separate estimations of the quality of its content and the excellence of its form, but also, and most particularly, on the specic relation or ttingness of the former to the latter. 7. Elster has interesting things to say on the topic of pre-commitment to chance or randomness in the making of art:
. . . each choice made in the creation of a work of art serves as a constraint on later choices. An artist may decide, however, to generate the constraints randomly rather than intentionally. (242)
Although I agree with most of the critical judgments embodied in Elsters diagram, and concur in particular with the placement of Lester Young, Billie Holiday, Johnny Hodges, and Django Reinhardt in its favored upper right corner, I would explain the relative undervaluation of Art Tatum, Sarah Vaughan, and Frank Sinatra, who though accorded high y-values are accorded only middling x-values, in terms of Elsters simply undervaluing certain realms of expression, namely the lithe/blithe/carefree, relative to others, namely the heavy/doleful/careworn. Only if emotional depth is narrowly equated with the latter does Elsters downgrading of Tatum, Vaughan, and Sinatra seem justied. I should also note that Elster in fact recognizes that his two dimensions of musical value are not entirely independent: Taste and emotional depth do not . . . vary entirely independently of each other. Total lack of taste is incompatible with great emotional force (250). But I submit that he still overestimates the degree of their independence, as I try to show in the text. See my Evaluating Music, in Philip Alperson (ed.), Musical Worlds (University Park, PA: Pennsylvania State University Press, 1998), and Ch. 10 in this volume. Understood here to include also its degree of inventiveness in storytelling, which is Elsters third proposed dimension of musical value.

Elster on Artistic Creativity

65

Thus Francis Bacon began his pictures by throwing paint at the canvas, so that the resulting blots would serve as constraints on the further work, limiting his freedom and presumably enhancing his creativity. Jackson Pollocks poured paintings may be seen as a variant of this idea. (243)

But as regards the rst of these cases, Bacons accidentally produced blots need not be taken as constraints on, rather than prompters of, further creative decisions, since it is in the nature of the medium of painting that any such blots can be covered over or scraped off, if need be. And to a lesser extent that would be true in the case of Pollocks paintings as well. Elster next turns his attention to that enfant terrible of modern music, John Cage:
To the extent that it relies on randomizing or on producing periods of silence, Cages work is entirely unserious. The most charitable interpretation is that it is a gigantic and successful put-on. (245) . . . the use of objective or epistemic randomness to select within constraints has no aesthetic justication . . . by removing choice rather than restricting it one destroys creativity rather than enhances it. (246)

This strikes me as a fairly uncomprehending reaction to Cages work as an artist, however one rates that work in the last analysis. Cages creativity manifests itself, in part, in the astonishing variety of evocative means he devised to circumvent the operation of human choice at the level of sound selection, in the name of an individual quasi-Buddhist philosophy of music and life. The irony of Cages work, though, is that his very distinctive personality shows itself, at a second level, through the specic range of meansdice, star charts, paper imperfections, the I Ching he devises for the elimination of personality at the rst level, and even through the very undertaking of the project of eliminating subjectivity or preference from music to begin with. Richard Wollheim, in a famous essay on art of the kind that subsequently became known as minimal, sketched a conception of negative work in artmaking, that is, the sort of work appropriate to the production of objects of excessive simplicity and relative absence of articulation. Such negative work, suggests Wollheim, is largely a matter of decisions about what a piece will be like, ones involving negative acts such as renouncing, eschewing,
Personality is a imsy thing on which to build an art, Cage once wrote. Yet he himself seems to make his most direct and appealing impact through his own personality (John Rockwell, All American Music (New York: Vintage Books, 1984), 58). Minimal Art, Arts Magazine, January 1965; reprinted in Gregory Battcock (ed.), Minimal Art: A Critical Anthology (New York: Dutton, 1968).

66

Art

withholding, or restraining, in relation to media, techniques, styles, or aims available in the artists milieu. If it is possible for an artwork whose genesis predominantly involves negative work to exhibit creativity, then surely the creativity involved in some of Cages more radical musical self-removals is of that sort. As it happens, Elsters skepticism about the value of avant-garde modes of art is not restricted to music: [literary] modernism fails because it rests on an erroneous criterion of adequacy of a text to its object. Just as a awed world is not better represented by a awed text, a uid or chaotic world is not better represented by a uid or chaotic text (190). But this blanket condemnation ignores, among other things, the exemplifying and expressing functions of a literary text, as opposed to its representing function per se. As Goodman and others have stressed, artworks mean in ways other than by representing. Thus, a work of art may at least need to approach, if not wholly embrace, chaos and uidity in order to exemplify or express, rather than just represent, those properties. 8. I return now to Elsters discussion of the experimental novel La Disparition, invoked at the beginning of this essay:
Preexisting or preset constraints enhance and stimulate the creative process . . . Yet not all constraints will do equally well. The constraint of writing a novel without the letter e may not make the task of the writer more difcult than writing in the demanding form of terza rima. Yet the latter constraint, unlike the former, can contribute directly to aesthetic value, over and above the indirect contribution that follows from the focus-enhancing effect. Because rhythm and meter generate an organized form, they have intrinsic aesthetic potential; the absence of a given letter in the alphabet does not. (209)

But one can argue that the systematic withholding of a certain letter, especially one as central to the French language as the letter e, gives by its absence a kind of form to Perecs text, though obviously not form in the sense of meter or rhythmic scheme. And that form can even be expressed in a positive manner, as the requirement to craft sentences using only the vowels a, i, o, and u.
For most writers, having to write around the letter e would be an irritating distraction rather than a stimulation, an obstacle to be overcome rather than a challenge to be met. (210)
See his Languages of Art (Indianapolis: Hackett, 1976) and How Buildings Mean, in Reconceptions in Philosophy and Other Arts and Sciences (Indianapolis: Hackett, 1988).

Elster on Artistic Creativity

67

That is no doubt true. But even so, it only means that willingly avoiding the letter e in writing would not, for most writers, be a spur to creativity. This says nothing, however, as to whether it was in fact a spur to creativity in a given case, namely that of Perec. Recalling Elsters useful trichotomy of chosen, invented, and imposed constraints, we can ask whether the express avoidance of e in La Disparition counts as a chosen or an invented one. The question, in effect, is whether that sort of restriction preexisted Perecs writing of his novel, or whether in so writing it, and in writing it so, he instituted that sort of restriction on literary production. Though Elster seems inclined to treat La Disparition as a case of invention, there is actually reason to see it as a choice to operate in an existing category, that of the lipogram, but in a more thoroughgoing way than had been previously attempted. A lipogram is precisely a text produced under the constraint of eschewing a given letter or letters of the alphabet, and was a form much favored by the group of avant-garde French writers to which Perec belonged, the Oulipo or Workshop of Potential Literature. 9. Constraints like that adhered to by Perec in writing La Disparition are sometimes derided as purely arbitrary and unmotivated, and thus as issuing in nothing of any artistic worth. But if artistic creativity is, as Elster proposes, centrally a matter of operating innovatively within constraints, then how does one know which constraints are likely to give rise to artistic value and which not? Can there be a reasoned prospective criterion of goodness of constraints, rather than the obvious retrospective one of judging them by their fruits? These are difcult questions, which I am unsure how to answer. I want instead just to consider whether the success of Perec in crafting La Disparition should indeed be regarded as without artistic point, a pure exercise in self-binding for its own sake. I think not. For one, there is no reason to regard the skill, cleverness, and ingenuity of Perec in meeting the constraint he set himself as somehow beyond the pale of artistic appreciation. Surely we are practicing artistic appreciation when we admire the interest that Barnett Newman gave his paintings of very peculiar vertical format, roughly six feet high and four inches wide, or the elaborately constructed montage in the
The name is a contraction of Ouvroir de Littrature Potentielle. e This is a good point to mention that there has been, astonishingly enough, a translationor really, recreationof La Disparition in English. It is by the novelist Gilbert Adair, and is cleverly titled A Void (San Francisco: Harper, 1995). Adairs work presumably has some of the virtues of Perecs, but not all, and arguably has some that Perecs does not.

68

Art

Odessa Steps sequence from Eisensteins Battleship Potemkin, or the resourcefulness with which Mozart manages the voices in the sextet from The Marriage of Figaro, or the magnicence of Racines alexandrines, unfailingly ten syllables to a line and a rhyme every time. Virtuosity, which broadly construed covers those cases and that of Perec, is unquestionably an artistic virtue, if not the highest of them, and is something appreciable for its own sake. The interest of La Disparition, in other words, is surely not merely that which Samuel Johnson infamously allowed that the phenomenon of female sermonizing might be said to have. But the interest of La Disparition may in fact go beyond the attention its virtuosic surmounting of challenges properly merits. This is due to special features of both the writer and the context of writing. Born in 1936, Perec was the child of Polish-Jewish parents who died during the Second World War, one in the Resistance and the other in Auschwitz; he was subsequently raised in a Catholic boarding school in the Is`re during the Occupation. Is it too e much of a stretch, then, to see La Disparition, published in 1969, as referring not only to the victim in the murder mystery it ostensibly relates, and not only to the letter e mysteriously banned from its pages, but to the disappearance from Europe not long before of millions of individuals as arbitrarily declared persona non grata as that unlucky vowel? Might not La Disparition be seen as a meditation on how easy it was to get rid of, or do without, six million Jews, as much as an exercise in getting rid of, or doing without, tens of thousands of es? We know also that every one of Perecs published works is an exercise in a different style. Does this not t all too well with one who had to dissimulate and reinvent himself from a very early age? In any event, if such further dimensions are attributable to what might seem the purely formal stylistic exercises of a Perec, they give those exercises added value, in virtue of the ttingness between content and form invoked earlier as a primary locus of aesthetic value.
See Thomas Mark, On Works of Virtuosity, Journal of Philosophy 77 (1980): 2845. Sir, a woman preaching is like a dog walking on hind legs. It is not done well, but you are surprised to nd it done at all. The roman nouveau of Alain Robbe-Grillet, Nathalie Sarraute, and others might be mentioned here as another example of an artform subject to self-imposed constraintsones as Draconian in their way as those under which Perec workedwhere the resulting stylistic spareness and sobriety of detail has expressive value. Robbe-Grillets novel La Jalousie, for example, arguably expresses a world of individuals locked into their separate consciousnesses, their phenomenological identication with the objects of their experience suggesting a kind of ultimate isolation and alienation. But such expression is clearly a product of the severely detached and methodically descriptive mode of narration adopted by the author.

Elster on Artistic Creativity

69

10. Shifting to cinema for another case of chosen constraints of an ostensibly limiting sort, consider those with which Sidney Lumet operated in making his powerful lm of jury deliberation, 12 Angry Men: a single, sparsely furnished room; a xed set of characters; a continuous segment of time; and black and white cinematography. Yet 12 Angry Men is as transxing and rewarding a lm as cinema has to offer. Clearly in this case Lumets chosen constraintshis form of provisional self-bindingwere a spur to creativity rather than a shackle on it. Some of Lumets specic techniques for sustaining cinematic and dramatic interest given those constraints deserve remark: (a) focusing on different characters, and different character pairs, at different points; (b) temporarily subtracting characters through the device of exits to the restroom; (c) rhythmic but not mechanical varying of long shots, medium shots, and close-ups; (d) alternating of evidence-sorting and reason-weighing discussions, on the one hand, with character-revealing and tension-provoking personal encounters, on the other hand, so far as the pattern of scenes is concerned. So Lumets lm ts nicely enough Elsters picture of artistic creativity as successful maneuvering within constraints. But it is worth observing that even if all artworks involve choices made within constraints of some sort or otherwhether self-imposed or imposed from without, whether revisably or unrevisably adopted, whether conventional or idiosyncraticthere is a useful distinction to be made between artworks which foreground the fact of their operating under constraints, and those which do not do so, in some cases even managing to make us unaware of the constraints that governed their creation. In the rst instance works are partly about their constraints, i.e. about virtuosity or the surmounting of challenges, whereas in the second instance they are not. In the rst category belong, obviously, La Disparition; Hitchcocks Rope, shot virtually in one continuous take; and Machauts rondeau Ma n est mon commencement, a musical palindrome. In the second category, more likely, belong 12 Angry Men, Coltranes Giant Steps, and the sonnets of Keats or Millayworks produced within formal connes but which do not especially draw attention to them. Twelve-tone music, the basic principle of which is that a given pitch cannot be repeated until all eleven other pitches of the chromatic scale have been soundeda method which effectively destroys the feeling of tonality or keyis sometimes offered as a mode of musical composition that is too conning, one involving constraints that fetter rather than abet

70

Art

creativity. But it seems as if there are simply different cases here, which make any such generalization suspect. For Schoenberg the method was a way to inject new local interest into old global forms associated with Brahms and his predecessors in the Viennese Classical tradition, an effort in which Schoenberg succeeded notably in the Third and Fourth String Quartets, the String Trio, the Piano Concerto, and the Serenade Op. 24. For Berg the method was from the outset something to be intuitively modied and relaxed in the interests of dramatic expression, not a code requiring strict adherence, and yet it remains essential to the rigor and power of his twelve-tone masterwork, the opera Wozzeck. For Stravinsky, who came to twelve-tone music late and after much resistance, the method offered itself as something to selectively adapt and fuse with his preexisting neo-classical style, with perhaps greatest success in the ballet Agon; in other cases, e.g. the Movements for Piano and Orchestra, the results were crabbed and less appealing. Finally, for some composers, like the American George Rochberg, twelve-tone musicor rather, its subsequent generalization, serial musicbecame an indirect prod to creativity by serving as an orthodoxy against which to rebel, leading to that musics total abandonment in Rochbergs largely tonal and pointedly neoromantic Third String Quartet. 11. One of Elsters examples of an externally imposed, yet ultimately fruitful, artmaking constraint concerns the nonfeasibility, prior to 1940, of recording popular music over three minutes in length, due to the limits of the old 78 rpm technology. In relation to this he comments that jazz improvisation at a high level of quality is so hard to sustain that the 78 record with three minutes playing time was just about optimal (194). To my ears, this sounds like a too convenient retrospective Leibnizian explanation. For my impression of even the best solos of the great jazz players in the 1930s big bands, such as Coleman Hawkins, Ben Webster, Johnny Hodges, or Lester Young, is that those solos would in fact usually have beneted from being somewhat longer, that they sometimes hardly get going before they are over. Certainly, there are many cases of mesmerizingly good improvised jazz solos extending over large stretches of time, for example, those by Coltrane and McCoy Tyner in My Favorite Things, which last upwards of four minutes each. Is it not rather plausible to think we might have had even more sublime solos from Hawkins and Young had they been able to dilate, on occasion, for two minutes or even three? 12. Let us turn to two other examples, ones which, at least at rst blush, comport well with Elsters thesis. The rst is Michelangelos celebrated

Elster on Artistic Creativity

71

Laurentian Library vestibule. Michelangelo, who had no formal training as an architect, accepted as a constraint to employ the vocabulary of High Renaissance architecturecolumns, corbels, pilasters, balusters, pediments, niches, and architravesbut not to employ them in the traditional ways. Rather, he set out to employ them, it seems, in pointedly illogical, that is, non-structurally justied ways, with the evident aim of more effectively stimulating the free play of the imagination in their regard. Michelangelo, untrammeled by habits he might have acquired had he been of the architects guild, was instead freed to create expressive form for its own sake: this anteroom, with its unusual somber white and grey color scheme, its high and narrow spatiality, its owing and spreading staircase, and its constellation of familiar yet oddly deployed constructional elements, is one of the most imposing architectural ensembles I know. Its success indeed seems to be a function of adhering to certain constraints derived from the prevailing culture, while simultaneously explicitly outing others. As Vasari apparently said of Michelangelos architecture, it broke the bonds and chains of common usage. A second example is Picassos creation of the Portrait of Kahnweiler as a response to both a highly general, tradition-rooted Charge and a more specic, circumstances-involving Brief, in the terms of the convincing analysis of Michael Baxandall. Picassos Charge, suggests Baxandall, was the standard painterly one of producing an object with intentional visual interest, while Picassos Brief, largely self-assumed, was the three-pronged one of reconciling the claims of color and form, reconciling the claims of twodimensional patterning and three-dimensional representation, and being true to painting as both cumulative process and completed image. According to Baxandall this posited Brief, together with the assumed background Charge, explain much about how Picassos portrait actually turned out. Still, with both Michelangelos vestibule and Picassos portrait, there is a real worry as to whether our hypothetical reconstructions of the constraints under which and in response to which those works emerged are not merely just-so stories. Certain constraints are perhaps undeniable and documentable, particularly material and physical ones. But hypotheses about constraints at the discretion of the artist, however, seem more liable to be products of
Quoted in H. H. Janson, History of Art (New York: Harry Abrams, 1962). See Patterns of Intention: On the Historical Explanation of Pictures (New Haven: Yale University Press, 1985).

72

Art

unfettered backward thinking in light of what actually results. The fact that hypotheses are usually underdetermined by evidence urges caution here, at least, as to the reliability of such speculative reconstructions. Consider now a case of maximizing artistic value by violating a freely chosen constraint. This is illustrated by an anecdote told of Arthur Rubinstein during one of his rst concerts in Paris in the 1920s. The rst piece on his program, the then recently composed Valses Nobles et Sentimentales of Ravel, had not been well received, so following a second half consisting of pieces of Chopin that was rather better received, Rubinstein answered the clamor for an encore by playing the Valses Nobles et Sentimentales once again. Now the art in question here is, of course, a performing one, that of recital-giving, and the constraint in question, which comes with the territory, is that encores be both shorter than, and other than, the pieces making up the main program. But arguably Rubinstein gave a ner recital on that occasion precisely by outing one of the dening norms of the performing art he was engaged in. 13. Is there a necessary connection between creativity and problemsolving? Assume for the sake of argument that there is a necessary connection between creativity and constraints, that creativity is roughly, as Elster suggests, a matter of value-maximizing choice within constraints. The question then becomes that of the relation between problem-solving and constraints. Are all cases of constraints in operation cases of problems calling for solution? One reason to deny this is that a problem, unlike a mere set of constraints, has something like a principle of unity, being often a matter of resolving some conict, or meeting some need, or nding the answer to a denite question. Another reason is that a problem has at least a prima facie claim on ones attentionthat it is something that, at least on the face of it, merits addressing. Thus it is a problem how to compose comprehensible and engaging music outside the bounds of tonality, or how to maintain the integrity of the picture plane without giving up blatant imageryproblems that were solved respectively by Schoenberg and Berg, on the one hand, and Johns and Dubuffet, on the other. But writing a poem of fteen lines that employs the word dog in every line, or making a lm that last fteen minutes and costs less than $150,000, would seem to be cases of operating under constraints, but not really cases of solving problems. The two notions, then, should not be thought interchangeable. 14. By way of conclusion, let me move from the sphere of ne art to that of applied art, and briey contemplate constraints on the design of chairs,

Elster on Artistic Creativity

73

those varied receptacles for our backs and bottoms. The case is well chosen, I think, to illustrate the reciprocity between constraints and creative activity conducted under such constraintsa reciprocity to which Elsters analysis of creativity, however illuminating, seems insufciently sensitive. For highly innovative chairs challenge, question, and refashion the very idea of a chair, subjecting its presumed necessary featurese.g. possessing legs, being sturdy, having a seat, being articulatedto intense conceptual pressure. Consider, in this connection, the intentional-historical theory of artifact concepts proposed by the psychologist Paul Bloom. Blooms theory takes off from and generalizes upon my own intentional-historical theory of the concept of artwork, whose basic idea is that an artwork is something that is intentionally connected by its maker to a preceding practice or tradition of artmaking, possibly identied only through exemplars rather than a covering description, and whose art status is taken for granted. On Blooms related theory a chair, say, is not something of a certain circumscribed form or something fullling an unambiguously understood function, but roughly something intended to belong to the category of chairs as antecedently exemplied, that is, something intended for regard or treatment as preexisting chairs were regarded or treated. The immense profusion of chairs displayed in the astounding book 1000 Chairs, which presents the most comprehensive survey of chair design from 1808 to the present, not only makes Blooms theory attractive, but makes a theory of that sort almost unavoidable, rendering doubtful any conception of chairs in terms of essential shape or structure or purpose. The assault on chairhood represented by the many outr examples of chairs in 1000 e Chairs might protably be compared with the challenge to what a building is provided by the most envelope-pushing architectural offerings of postmodern architects such as Frank Gehry, Peter Eisenmann, and Rem Koolhaus. Now Elster does, it is true, recognize a distinction between what he calls rebellion and revolution in art, which he glosses as follows: Rebellions violate existing conventions, whereas revolutions abolish them and create new ones (p. 223). What I am gesturing at in the case of chairs and buildings, however, is something in-between, neither the simple abrogation of existing rules
Intention, History, and Artifact Concepts, Cognition 60 (1996): 129. See Dening Art Historically and Rening Art Historically, in Music, Art, and Metaphysics, and Extending Art Historically, in The Pleasures of Aesthetics. Charlotte and Peter Fiell, 1000 Chairs (Cologne: Taschen, 1997). The quote is from the back cover.

74

Art

for such things nor the institution of new rules in their place, but the forced reevaulation, by an ostensible chair or building, of what the rules for such things really were or are. Creativity, in short, is sometimes a matter or reconceiving or reinterpreting or reconstruing given constraints, and not always a matter of either remaining inventively within them or entirely abandoning them.

PART I I MUSIC

This page intentionally left blank

5
Sound, Gesture, Space, and the Expression of Emotion in Music

I INT RODUCT ION The main concern of this essay is musical expressiveness, and the role that the imaginative construal of sound in gestural ways plays in grounding that expressiveness. But I will come to that main concern obliquely, beginning instead with some reections on the nature of sounds and the sense of hearing. What I will be arguing, in the central sections of this essay, is that to hear the expressiveness of music is to hear it as personal expression; that to hear it as personal expression is to hear a sort of gesture in the music, or the music as gesturing in a certain manner; that to hear such musical gesture is to deploy a capacity to imagine in spatial terms, most obviously because that is required for apprehension of the behavioral gestures and performing gestures that underlie musical gesture. Thus, grasp of expressiveness is music is not detachable from a comprehension of the range of expressive human gesture and a comprehension of the means by which music is actually produced, each of which implicates possession by the hearer of a robust spatial imagination.
First published as Sound, Gesture, Spatial Imagination, and the Expression of Emotion in Music, European Review of Philosophy 5 (2002): 13750. This essay, originally presented at a colloquium entitled Objets et espaces sonores held at the Ecole Polytechnique in Paris in March 2001, draws signicantly on two earlier essays of mine, Authentic Performance and Performance Means, in Music, Art, and Metaphysics (Ithaca: Cornell University Press, 1990), 393408, and Musical Expressiveness, in The Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996), 90125. For a broader perspective on problems of emotion in relation to art see my Emotion in Response to Art: A Survey of the Terrain, in Mette Hjort and Sue Laver (eds.), Emotion and the Arts (Oxford: Oxford University Press), 2034 (and reprinted as Ch. 3 in this volume).

78

Music

II THE NATURE OF SOUNDS At the heart of their book, La Philosophie du son, Roberto Casati and Jerome e Dokic propose what they call a thorie evnementielle of sounds. Accorde ing to Casati and Dokic, sounds are neither auditory sensations occurring in our heads, nor sound waves arriving at our ears when we hear, nor secondary qualities of any sort, but rather something happening in the resonant object itself. They summarize their theory in the following formulation: sounds are vibratory events involving objects. I believe Casati and Dokic are right in that contention. Sounds are indeed events, and they are located just where the objects that produce them or in which they occur are located. But it seems to me that the proposal they advance is in need of some qualication, since not every vibratory event that concerns an object can reasonably be accounted a sound. For example, the atomic and subatomic vibrations that characterize every parcel of matter, the slow oscillation of an aspic de legumes, the gentle trembling of the hand of a sufferer from tremor, are all vibratory events in objects, but not thereby sounds involving those objects. On the other hand, as Casati and Dokic rightly insist, we should not require of sounds that they be humanly audible, otherwise we would unfairly exclude ultrasounds and infrasounds, which we have reason to consider sounds from a philosophic point of view even though they do not lend themselves to human hearing. We might thus add to the formulation offered that sounds are vibratory events in objects that are either humanly audible or of a physical order comparable to those that are humanly audible, and so at least a possible object of audition to other creatures we might credit with a sense of hearing. One might also remark that, without modication, Casatis and Dokics proposal, identifying sounds with vibratory events in objects, has the consequence that the music of the spheres, of which the ancients spoke, is not after all a metaphor, but something to be taken literally. At least that is so if one supposes, as seems reasonable, that the movements of revolution and rotation of the planets are a sort of vibration involving them. However, I do
Roberto Casati and Jerome Dokic, La Philosophie du son (N mes: Editions Jacqueline Chambon, 1994). e Les sons sont des evnements vibratoires intressant un objet (La Philosophie du son, 49). e The same thesis, one may note, is defended in Robert Pasnau, What is Sound?, Philosophical Quarterly 49 (1999): 30923, though with less precision.

The Expression of Emotion in Music

79

not regard it as a plus for a theory of sounds to count among sounds the supposed music of the spheres.

III THE APPRECIATION OF MUSIC AND THE SENSE OF HEARING One of the main claims in La Philosophie du son is that the sense of hearing, properly speaking, involves the formation of beliefs regarding the spatiality of sounds understood as vibratory events in objects, and thus presupposes an image of the space in which such events are located. Casati and Dokic go on to relate this conception of hearing to the phenomenon of music, in the following striking observation:
If what counts in the appreciation of music is being able to have auditory sensations, rather than being able to form beliefs regarding the location of sounds in the environment, it follows that to appreciate music there is no need, strictly speaking, of a sense of hearing. One could in effect conceive of a perfectly solipsistic musical subject, one without any beliefs concerning external sonic events, but who, in prey to hallucinations, would hear whole symphonies.

In my opinion this observation is partly just and partly not. I agree with Casati and Dokic that the sense of hearing, robustly understood, goes beyond the capacity to have auditory sensations, and involves the ability to form beliefs about the sources of sounds heard. But these beliefs, we may note, are of two kinds: one concerns the spatial location of the source of the sound relative to the hearer, while the other concerns the spatial nature of the source of the sound, apprehended in terms of size, shape, movement, and orientation. There are thus at least two separable abilities involved. Let us label the rst of these abilities specic spatial imagination and the second generic spatial imagination.
Si ce qui compte, pour apprcier de la musique, cest davoir des sensations sonoresplut t e o ` que dtre capable de former certaines croyances quant a la localisation des sources sonores prsentes e e dans lenvironnementalors pour apprcier de la musique, point nest besoin de lou On peut en e e. e effet concevoir un sujet musical solipsiste, qui naurait aucune croyance sur des evnements sonores, ` mais qui, en proie a des hallucinations, entendrait des symphonies (La Philosophie du son, 28). An illustration of the distinction is the following. In perceiving a particular sound I might have the thought that it is situated down the street, and moving in my direction, and also the thought that it is the sound of a car engine, understood as an object of a certain shape consisting of parts moving relative to one another in a certain way. The former thought is the product of specic spatial imagination, the latter of generic spatial imagination.

80

Music

Now it is true that to appreciate music of the standard sort we may perhaps dispense with beliefs of the rst kind, and so with the rst kind of spatial imagination. But it is not at all clear that we can dispense with beliefs, and so spatial imagination, of the second kind, and thus at least a part of the sense of hearing robustly understood. The reason is a fact about the appreciation of music that I will be foregrounding in this essay, namely that to appreciate traditional instrumental music adequately we have need at least of generic spatial imagination, enabling us to imagine possible sources of heard sounds, and possible agents and actions that could be responsible for those sounds. Thus, although Casati and Dokic may be right that appreciation of such music does not necessarily call upon the full sense of hearing, they are wrong to think that the mere capacity to experience auditory sensations could be adequate to such appreciation. A satisfactory account of a listeners grasp of musical expressiveness entails that the listener has the capacity to imagine in spatial terms the possible sources of sounds heard. More broadly, we can take the question to be whether the appreciation of music is detachable from the normal operation of the senses with which human beings have been endowed. Assuming a capacity to imagine actions in a real space as possible sources of sounds heard is appreciatively relevant to most music, and that that capacity could only be rmly in place if at least one sensehearing, vision, or touchcapable of providing spatial information and of priming the spatial imagination was at the disposal of the subject, the answer must be no. Contra the solipsistic supposition entertained by Casati and Dokic, then, the appreciation of music is not entirely separable from the senses in their normal functioning. A key premise in the argument just sketched is that the imagination of the sources of sounds, in terms of generating actions and objects, is relevant to the appreciation of music of the ordinary sort. I now devote attention to establishing that.

IV PE RFORMING GESTURE, MUSICAL GESTURE, AND SPATIAL IMAGINATION What I hope to show in this section is that a part of the expressive character of a passage of traditional instrumental music as heard derives from the impression one has of the manner in which it has been produced and the correlation of that impression with the sound of the passage narrowly construed. The

The Expression of Emotion in Music

81

sound of the sound, so to speak, does not sufce in itself to x the expressive character of the passage, but only that sound in conjunction with a presumed manner of production. The same sounds present different appearances, and affect us differently, according to the notions we entertain at the same time regarding the actions or processes that have engendered them. Mozarts Serenade in E-at, K. 375, begins with a unison statement of the full cohort of windsoboes, clarinets, horns, and bassoonsa ve-note dotted phrase with marked sforzandi. That beginning has an assertive quality, and suggests a call to attention. But from what, exactly, does that assertive, attention-getting character derive? It derives, in part, from the fact that that stretch of music comes across as an instance of honking, as passages involving double reed instruments played forte often do. Given the manner in which the sounds have actually been produced, that is, by actions of wind players similar to those of waterfowl engaged in honking, that passage could in fact be considered to be a honking, broadly understood. Hearing the sounds of which that stretch of music is composed as having been produced by the rapid passage of air through narrow openings in wooden tubes, that is, as a honking, rather than by some entirely other process, makes for an experiential difference. Were we convinced that the sounds we were hearing did not result from the playing of wind instruments in the normal manner, but issued instead from a perfect synthesizer, that stretch of music would make a somewhat different impression on us. The reason is that a synthesizer, however powerful or accurate, cannot truly honk. That is to say, it cannot do what oboes, klaxons, and geese all naturally can. And this modies, in a subtle manner, how the sounds the synthesizer emits are received, since those sounds are gauged against the background of a presumed manner of production. I have been arguing that knowledge of the actual sources of sounds affects how those sounds strike us. I have also been suggesting that such knowledge properly enters into aesthetic assessment of passages of music, which are presumably to be taken as what they are, say the product of windplaying, and not what they are not but only appear to be, as when a synthesizer mimics such windplaying. But in fact that is a stronger thesis than is needed for present purposes, so let me also articulate a weaker, less controversial one. All that is needed to show the indispensability of generic spatial imagination to a grasp of the expressive character of music is that a sense of the possible sources of sounds, not necessarily knowledge of their actual sources, is required if the expressiveness of sequences of sound is to be registered.

82

Music

Clearly the assertiveness of the opening of Mozarts serenade only emerges if that opening is heard as a woodwind proclamation, leaving aside whether on a given sounding occasion it actually is, and whether it is known to be. This point is easily generalized. Descriptions of musical passages in terms of actions are extremely widespread. Passages can be described as sighing, whispering, chirping, purring, squawking, roaring, sawing, hammering, pounding, slashing, caressing, swooping, and so on. Now obviously one cannot recognize the applicability of such descriptions without possessing an image of the action literally denoted. For example, one cant hear a passage as sighing unless one has a conception of what sighing is. But the applicability of such descriptions is also predicated on a sense of related performing actions from which the sounds in question are assumed to result. Absent a sense of those background generative actions such passages would not seem to merit those descriptions, or at any rate, would not do so unequivocally. Ones sense of a passage as a slashing one is not completely detachable from the impression that it has been produced by a violin played staccato, and ones impression of a passage as a purring one is similarly not entirely independent of the impression that it has been produced by an oboe played legato. A slashing or purring sound not thought of as the product of those sorts of performing actions does not slash, does not purr, in quite the way it does when believed to issue from violin or oboe. Now just as generic spatial imagination is required to form a conception of the actions literally denoted in the action-based descriptions recalled above, so is it required to envisage the background performing actions behind musical passages which contribute to the applicability of those descriptions. For both sorts of actions, e.g. the literal roaring we may hear in a passage of music, and the rising declarations in the brass that underlie it, necessarily manifest themselves in space. And these action-based descriptions, in turn, have an obvious bearing on the expressive character musical passages are perceived to have, given the aptness of many such actions to the expression of emotion. Let me approach the conclusion that spatial imagination is necessarily involved in a grasp of musical expression in another way. The expressiveness of music is grounded on the factwhich I will shortly explore in more detailthat the actions or gestures that one hears in a passage of music recall the actions or gestures that serve as behavioral expressions of emotions, which allows us to hear the former as the latter, and so the passage as expressive of

The Expression of Emotion in Music

83

those emotions. But the gestures rightly heard in music are only heard in all their specicity if the apparent performing gestures behind the sequences of sounds per se are taken into account. For the gestures we are right to hear in musical sequences are those we hear in them when we are cognizant of the instrumental actions understood to generate such sequences. It is important to be clear on a crucial step in the preceding argument. It is not that the gestures heard in a passage of musicwhat in effect we hear the music as doing are identical with the performing gestures that we imagine as responsible for the sounds we hear. It is rather that those gestures, which for clarity I label musical gestures, are in part a function of the performing gestures we are right to imagine in auditing a given passage. In other words, the gestures we are correct to hear in a passage of music, and on which depends our estimation of its emotional expressiveness, are partly determined by what we take performers of the passage to literally be doing in producing it. Some additional examples will serve to support this thesis of the intimate relationship between performing and musical gestures, and thus ultimately, of the necessary involvement of spatial imagination in the grasp of musical expression. Consider the rapid upward keyboard glissando. This familiar musical formula conveys most often an impression of gaiety or insouciance. The best-known instances of it are probably Chico Marxs one-nger antics in various Marx Brothers movies, but a number of examples enliven as well the scores of Ravel, Prokoev, and Gershwin. I claim that the characteristic impression produced by the upward keyboard glissando has its source, in part, in an image we form of the icking or sweeping gesture behind the rising tones that we hear, a gesture that occupies a certain place for us in the eld of expressive behavior as a whole. The imagined executive action melds with the rising tonal movement so as to create a musical gesture in which we readily hear a gay insouciance. One of the most vivid portraits in all music is that found in the sixth section of Job: A Masque for Dancing by Ralph Vaughan Williams. There the composer
A word about the gestures of conductors during performance is in order. Such gestures, probably the ones most on display in the musical arena, do not count as examples of what I mean by performing gestures. That is, such gestures do not gure among those a listener must have an image of in order to grasp the expressiveness of the music being heard, that is, in order to grasp its musical gesture. However, a conductors gestures might very well symbolize the musical gestures contained in the music being conducted, and reect the conductors understanding of those musical gestures, but it is only rarely that there is anything approaching identity between musical and conductorial gesture at a given point.

84

Music

depicts the hypocrites gathered around the suffering Job. The slimy nature of these false comforters and the avor of their viscous consolations could not be more clearly conveyed. The musical phrases charged with achieving thisslow, descending seconds and thirds, alternately major and minor, in a mellow timbreare perfectly apropos, but they only make their optimal effect if they are taken precisely as the saxophone gestures they are. It is only in the light of that image that the music presents itself unequivocally as a stylization of the whining, honey-tongued vocal behavior typical of hypocrites. Take nally the standard expressive contribution of percussion in orchestral music. Notes on the timpani carry a powerful association of pounding or batteringits a matter of mallet strokes, after alland in passages where timpani are prominent this effect can be overwhelming. However, it is not timpani sound per se that is efcacious, but the conjunction of such sound and the action of which that sound is the sign. That a phrase sounds a certain way, in the narrow sense, is not the only thing that counts so far as the musical gesture we will be disposed to hear in the phrase is concerned; equally important is our sense that the phrase has been sounded in a certain way. In the scherzo of Beethovens Ninth Symphony the timpanis statements are aggressive and interruptive, while in the rst movement of Nielsens Fifth Symphony those of the snare drum are warlike and menacing, in virtue of their maddening repetitiveness. But in both cases, these effects are precisely what they are only if the utterances in question are heard as strikings and hammerings, and not as pure disembodied sequences of sound. This is a good place to offer a further observation on musical gesture, which as we have seen depends on performing gesture though without reducing to it. In many cases in which musical gesture is perceived one is able to characterize the musical gesture in familiar terms, metaphorically employed; for example, one might say of a passage of music that it was dancing, or striding, or thrusting, or hectoring, or meditating, or recoiling, or questioning, etc. But this will not be true in all cases in which one nonetheless registers gesture in music; in such cases one might perceive music to be gesturing in a certain manner, and as a result hear the music as, say, heroic, but without being inclined to identify, or even being able to identify, the manner of gesturing in terms of some prior category of action.
Of course it will always be possible to characterize retrospectively the manner of gesturing as heroic, but that would not be to invoke an antecedently familiar action concept so as to characterize the manner of gesturing heard in question.

The Expression of Emotion in Music

85

V M USICAL EXPRESSIVENESS AND HEARABILIT Y-AS-PERSONAL-EXPRESSION At several points in the preceding discussion I have leaned on a theory of musical expressiveness, but without elaborating it. It is now time to lay out that theory more clearly. I cannot here undertake a full defense of the theory, which I have argued for at length elsewhere, but I will try to make its attractions evident. The central idea is that musical expression of mental statesand for reasons of simplicity I stick to the most important case, that of emotionsmust model itself on the primary expression of emotions by human beings, that is, expression of emotions through behavior or other outward manifestation, including countenance, posture, bearing, demeanor, actions, gestures, and modications of voice. That is not to say that music is expressive in precisely the same sense as is behavior, for several reasons. One, music doesnt actually exhibit behavior as such; two, musical gesturing is not literal but rather metaphorical; three, the character of expressive music is neither the immediate result of emotion experienced by the music, music not being sentient, nor always the upshot of emotion experienced by the composer; and so on. But despite these manifest differences between musical expression of emotion and human expression of emotion, it seems we cannot consider a piece of music to be strictly expressive of an emotional state Srather than, more loosely, simply possessing a correlative emotional quality Qunless we regard it as analogous to a being endowed with sentiments capable of announcing themselves in an external manner. In short, music expresses an emotion only to the extent that we are disposed to hear it as the expression of an emotion, but through different means, by a person or personlike entity. More formally, what I have proposed as an analysis of expression in music goes like this: a passage of music P is expressive of an emotion E iff P, in full context, is readily heard, by a listener appropriately backgrounded in the musical genre in question, as the expression of E in a sui generis, purely musical manner, by an indenite agent, what we can call the musics persona. The concept of musical expressiveness is thus a complex one, in which the ideas of
See Hope in The Hebrides , in Music, Art, and Metaphysics (Ithaca: Cornell University Press, 1990), 33675, and Musical Expressiveness, in The Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996).

86

Music

personal expression, human gesture, musical gesture, imaginary agency, and hearing-as each plays a role. What is it to hear a stretch of music as something else? For present purposes it sufces to locate hearing-as among perceptual acts that partake freely of or substantially enlist the imagination. The agents one hears in music when one hears it as an expression of emotion are thus inescapably imaginary ones, ones displaying the indeniteness characteristic of all imaginary objects. To hear music as such and such is, perhaps, to imagine that the music is such and such, and, more specically to imagine of the music, as you are hearing it, that it is such and such. In earlier discussions of the problem of musical expressiveness I allowed myself to alternate the locution hear X as Y with the more counterfactuallyavored locution hear X as if it were Y, suggesting that in connection with music these come to more or less the same thing. I am now less sure of that, so here conne myself to the rst locution. The notion of hearing X as if it were Y may not be an entirely happy one. I suspect that is because the modier as if it were is more appropriate to acts of treating, regarding, or behaving towards than it is to acts of perceiving as such. For example, though one clearly understands what it would be for X to treat Y as if she were the only person in the world, it is less clear that one understands what it would be for X to see Y as if she were the only person in the world. What about the evocation of emotion by music? Is not music that is expressive of an emotion precisely music that evokes in us this same emotion, or that at any rate has a disposition or tendency to evoke in us that emotion? This is not the place to demonstrate that this popular theory of musical expressiveness is erroneous. However, neither is it necessary. Because we can willingly admit that the emotion, and also the feelings and sensations, that a passage of music has a tendency to evoke in a listener inuences what emotion the listener will be disposed to imagine the music to be the expression of, without it being necessary to identify the emotion expressed with that emotion, or those
For further ref lections on the concept of musical expressiveness, and doubts whether the expression that music is heard as when its expressiveness is being perceived should be held to be a sui generis mode of expression, rather than just expression tout court, see my Musical Expressiveness as Hearability-as-Expression (Ch. 6, this volume). Those cited in n. 9. But see Stephen Davies, Musical Meaning and Expression (Ithaca: Cornell University Press, 1994), and Peter Kivy, New Essays on Musical Understanding (Oxford: Oxford University Press, 2001). Still, the evocation theory of musical expression continues to have advocates, of whom Derek Matravers, Art and Emotion, Oxford: Oxford University Press, 1998), is the most sophisticated.

The Expression of Emotion in Music

87

feelings and sensations. What one has a tendency to imagine, and more particularly to hear-as, in relation to a passage of music one is auditing, remains the key to its expressiveness, and not whatever power it may possess to provoke in us a corresponding emotion. Such is the theory of musical expressiveness to which I subscribe, and which gured in the background of my earlier remarks. What follows, as already intimated, is that generic spatial imagination is indispensable to the appreciation of traditional instrumental music. The argument to that conclusion, in summary form, is this: appreciation of traditional instrumental music requires grasp of musical expression; musical expression of emotion presupposes the notion of personal expression of emotion and rests on the emergence of musical gestures; personal expression of emotion, on the one hand, is fundamentally behavioral and so necessarily manifested in space, while musical gesture, on the other hand, is a function of both performing gestures understood as the sources of musical sequences heard and perceived resemblances between musical sequences and behavioral expressions; hence grasp of performing gestures and of behavioral expressions, and thus ultimately of musical gestures, requires generic spatial imagination; hence grasp of musical expression requires generic spatial imagination; hence spatial imagination is necessary to the appreciation of traditional instrumental music. VI MUSICAL EXPRESSIVENESS AND MENTAL SIMUL ATION A question that we can now pose is this: does the imagining of indeterminate personlike agents in music, who seem to be expressing their own emotionswhich I take to be the core of an experience of musical expressivenessinvolve us in acts of mental simulation, and if so, of what are these acts the simulation? Though I am unsure how to answer this question, I will permit myself a measure of speculation in connection with it. Faced with a passage of music that strikes us as behaving or gesturing in such and such fashion in virtue of its musical movement, its underlying performing actions, and other aspects of its sonic appearance, perhaps we frame for ourselves, without being fully
For further illumination on this point, see my Musical Expressiveness and Jenefer Robinson, The Expression and Arousal of Emotion in Music, in Philip Alperson (ed.), Musical Worlds (University Park, PA: Pennsylvania State University Press, 1998), 1322.

88

Music

conscious of it, the hypothetical query, If someone was comporting himself in that way, what emotion is it likely he was feeling? However, perhaps it is the case that faced with such a passage we do not theorize in this manner, even subconsciously, in the aim of inferring what the musics expressiveness must be, but that we rather try to imagine ourselves in the place of the music, that we assume as our own the musical gestures we hear the passage to be suffused with, as a consequence of which we nd ourselves feeling, in imagination, such and such emotion, and so in that way come to know what the music expresses. It remains to add only that such postulated acts as putting oneself in the place of the music and assuming as ones own the musics perceived gestures might well be realized by some sort of simulation procedure. More precisely, in order to effect such a simulation one would have to enter into the presumed mental simulator in off-line mode thoughts or beliefs such ], [I am making the as [I am the music], [I am behaving in the manner musical gesture ], the content of the blanks being supplied by acts of mental ostension or demonstration, with the expectation of nding oneself consequently in such and such mental state, a state one would then be in a position to say was the state expressed by the music. I am not prepared to conjecture that we in fact do something of this sort, but such a scenario seems to me at least possible. At any rate, the theory of musical expressiveness here defended is certainly amenable to interpretation in a simulationist manner, if not such as to require that. VII EXPRESSION IN ELECTROACOUSTIC MUSIC The necessity of a robust sense of hearing, including the ability to localize sounds in space, to the overall appreciation of electroacoustic musicmusic written not for instruments, but composed on computer, tape, or synthesizer for direct transmission to loudspeakerscan be taken for granted, given that such music uses space as a primary structural material and derives probably its most striking effects from spatial manipulations. Appreciating such music draws on not only generic but specic spatial imagination in the most obvious
The process of imaginative projection by which we endeavor to seize the expressiveness of music is convincingly described in Roger Scruton, The Aesthetics of Music (Oxford: Oxford University Press, 1997), though in terms of Verstehen rather than of simulation. I should make clear that I am here thinking of a music of multiple channels, typically six or eight, capable of generating an ample acoustic space. Another term for electroacoustic music, especially in France, is acousmatic music, though the latter has a slightly narrower meaning.

The Expression of Emotion in Music

89

way. What is at issue at present, rather, is whether generic spatial imagination is necessary to a grasp specically of the expressiveness of such music. With instrumental music the need for generic spatial imagination in gauging expressiveness is a consequence of the need to imagine the performing gestures behind the sounds, and behind those performing gestures, the gestures and actions that serve as the behavioral expressions of emotions. It is because those gestures and actions take place in space that a capacity to imagine them goes hand in hand with a capacity for at least generic spatial imagination. How, then, do matters stand with electroacoustic music? Since such music is not performed, grasp of its expressiveness cannot require us to imagine performing gestures from which it is presumed to issue. Still, though the expressiveness of electroacoustic music thus necessarily differs somewhat in its basis from that of traditional music, an argument can be made in its regard connecting a grasp of expressiveness to a spatial imagination of sources. As we have seen, the capacity to imagine the presumed sources of sounds, in terms of instruments, agents, and actions, plays a crucial role in estimating the expressiveness of traditional music. With electroacoustic music, by contrast, we generally have only a vague idea as to the sources of the sounds heard, because the sounds in question are usually unfamiliar and difcult to categorize, and information about their provenance is rarely indicated in ancillary material. Nevertheless, it seems that our habits of hearing-as are likely to be transferred, to some degree, from traditional music to electroacoustic music, without our being conscious of any categorization of sounds according to their likely sources. For instance, a sound resembling that of a chainsaw but not consciously classied by us as such would still probably be subconsciously assimilated by us in those terms, inducing us to take it as emanating from something like a chainsaw, with all that that implies expressively. Thus, even with electroacoustic music, where it is not incumbent upon listeners to hear stretches of sound in terms of their presumed sources, me may conclude that the capacity to imagine in a spatial manner possible sources of sounds heard will play a role in the expressiveness such music wears for us. On the other hand, insofar as the expressiveness of music rests in part on the sources one imagines sounds to have had, one can also see there an explanation of why the expressiveness of electroacoustic music, which would be hard to deny, remains yet more elusive than that of traditional instrumental music. It would be because in this music, as just noted, we nd it difcult to classify the majority of sounds encountered as to their ostensible sources, and

90

Music

probably even our subconscious classiers of sounds, such as they are, often rest undecided before a good portion of the sounds that escape conscious categorization. If so, then by consequence we will also nd it difcult to grasp with any deniteness the musical gestures embodied in a piece of electroacoustic music, in contrast with those embodied in a piece of music composed for and sounded on familiar instruments. However, and there is perhaps appropriate compensation in this, to the degree that the expressiveness of electroacoustic music remains elusive to us we are in a certain sense liberated, and so free to deploy in an even more creative manner than usual the auditory imagination that is ours.

6
Musical Expressiveness as Hearability-as-Expression
I E XPRESSION AND EXPRESSIVENESS As a number of philosophers have rightly underlined, expression is essentially a matter of something outward giving evidence of something inward. In other words, expression is essentially the manifesting or externalizing of mind or psychology. The scope of expression is thus not, pace Goodman and others, properties in general, or properties metaphorically possessed, but rather psychological properties, those pertaining to the mental states of sentient creatures. For only those can be intelligibly expressed, whoever or whatever is doing the expressing. And this holds as well for expressiveness, which we can initially understand as the sort of expression that some objects, and perhaps most notably musical works, manage to achieve, despite their not literally having inner lives. Exactly what sort of expression expressiveness amounts to will emerge as the argument proceeds. My discussion focuses on the case of musical expressiveness, but at a later point I offer some reections on artistic expressiveness in general. II MUSICAL EXPRESSIVENESS I here defend an account of musical expressiveness elaborated in earlier papers, replying to certain objections it has elicited, and underlining its
First published in M. Kieran (ed.), Contemporary Debates in Aesthetics and the Philosophy of Art (Oxford: Blackwell, 2006), 192206. For instance, Alan Tormey, The Concept of Expression (Princeton: Princeton University Press, 1971); Bruce Vermazen, Expression as Expression, Pacic Philosophical Quarterly 67 (1986): 196224; and Aaron Ridley, Expression in Art, in Jerrold Levinson (ed.), Oxford Handbook of Aesthetics (Oxford: Oxford University Press, 2003), 21127. Hope in The Hebrides , in Music, Art, and Metaphysics (Ithaca: Cornell University Press, 1990) and Musical Expressiveness in The Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996).

92

Music

superiority to the most plausible competing accounts on offer. Crucial to my account is the idea that the expressiveness of music resides in the invitation that music extends to the listener to hear it as expression in the primary sensethat is, expression, by persons, of inner states through outer signs. Thus, to hear music as expressive is to hear it as an instance of personal expression. Expressive music is music we are disposed to hear as expressing; and that is what its expressiveness fundamentally consists in. Since the primary vehicle of the expressing of states of mindas opposed to articulate thoughtsis gesture, broadly understood, and since music is naturally assimilated to that kind of expressing, that means that we hear expressive music as gesturing of some sort, that we hear a sort of gesture in expressive music. Call this musical gesture. It is important to stress that though musical gesture is related to both the ordinary behavioral gestures connected to expression of states of mind and the specic performing gestures involved in the sounding of music, it is not equivalent to either of those. It is a matter, at base, of what we hear the music to be doing, in virtue, most importantly, of the movement we hear in music. Summon up in your mind the opening of Brahmss First Symphony. You are inescapably presented, in listening, with the image of someone in the throes of emotion, which emotion is being manifested to you, through what one might call musical gestures. You may perhaps not know, or be able to articulate, what emotion the agent heard in the music is in the grip of, but in the grip of it he is, and this is something you directly hear. It is as if an emotion is being expressed, in the most literal sense, though it is somehow happening through music. The musical expression of states of mindincluding emotions, feelings, attitudes, desires, and beliefsmust thus be modeled on the primary expression of such states by persons or human beings, whereby such states are revealed or evinced through behavior or other outward manifestation. In the case of emotions, the usual focus of musical expression, such manifestations include countenance, posture, bearing, demeanor, actions, gestures, and modications of voice. Naturally there are differences between human expression and musical expression. For one thing, music is not literally behavior, and musical gesture
See Stephen Davies, Musical Meaning and Expression (Ithaca: Cornell University Press, 1994) and Roger Scruton, The Aesthetics of Music (Oxford: Oxford University Press, 1997) on musical movement, and Scruton, ibid., and Jerrold Levinson Sound, Gesture, Space and the Expression of Emotion in Music (Ch. 5, this volume) on musical gesture.

Musical Expressiveness as Hearability-as-Expression

93

is not literal gesture. For another, expression in music is not the result of emotion experienced by the music, since music is not sentient, nor is it invariably the upshot of emotion experienced by the composer, since lived emotion and musically imagined emotion can readily diverge. But despite those evident differences between musical expression of emotion and human expression of emotion, we should not consider a piece of music to be strictly expressive of an emotionrather than standing in some other, weaker, relation to it, such as possessing a perceptual quality associated with the emotionunless we regard it as analogous to a being endowed with sentiments capable of announcing themselves in an external manner. In short, music expresses an emotion only to the extent that we are disposed to hear it as the expression of an emotion, although in a non-standard manner, by a person or personlike entity. More formally, what I have proposed as an analysis of expression in music goes something like this: a passage of music P is expressive of an emotion E if and only if P, in context, is readily heard, by a listener experienced in the genre in question, as an expression of E. Since expressing requires an expresser, this means that in so hearing the music the listener is in effect committed to hearing an agent in the musicwhat we can call the musics personaor to at least imagining such an agent in a backgrounded manner. But this agent or persona, it must be stressed, is almost entirely indenite, a sort of minimal person, characterized only by the emotion we hear it to be expressing and the musical gesture through which it does so. It is important to keep that in mind when entertaining skepticism as to whether understanding listeners normally hear or imagine personae when they apprehend expressiveness in music. My basic analysis of musical expressiveness, again, is that music expressive of E is music heard as, or as if, someone expressing E. Of course one could fairly expand this as music heard as, or as if, someone experiencing, and as a result expressing, E, since A expresses E presupposes A experiences E. But it is presumably the expressing part alone that enters into what the music can intelligibly be heard as, since we can, it seems, have no idea what it would be to hear music or musical process, as if it, the music or musical process, were an experiencing of something. It has been suggested, incidentally, that not all expressions are such that we think of their possessors or bearers as engaged in acts of expressing, in the sense of intentionally trying to communicate a state of mind. Since I
Vermazen, Expression as Expression, 1989.

94

Music

analyze expressiveness in terms of as-if expression, not as-if-expressing, then even if that suggestion is valid, it is immaterial for my account. The suggestion can, in any case, be challenged. For it can be argued, pace Vermazen, that all expressions ascribable to an agent in fact are cases of expressings by the agent. It is just that not all such expressions are cases of intentional or self-conscious expressing. Earlier formulations of my proposal appealed to the idea of a sui generis mode of expression of emotion, suggesting that music, when heard as expressing, was such a mode of expression. But that idea may be an unfortunate one, and has attracted its share of criticism. Some commentators, for instance, have charged that there is something incoherent in the suggestion that one hears a passage of music as a sui generis expression of some emotion, on the grounds that we can form no conception of a mode of expression declared to be sui generis; others have charged that the appeal to a sui generis mode of expression in connection with music implies that we experience music as a novel corporeal means of sound production or else as an odd creature that somehow behaves musically. In light of these charges, it is probably a mistake to insist that the expression that music is heard as when its expressiveness is being perceived is sui generis expression, rather than just expression simpliciter. Little appears to be gained, and only confusion sowed, by such insistence. And yet the notion of a sui generis mode of expression, which the singularity of our experience of music as expressive suggests to us, may still have a role to play in the full explanation of musical expressiveness. What must be avoided, it seems, is making the idea of a sui generis mode of expression part of the content of the hearing-as experience involved in registering musics expressiveness. (After all, thinking in Latin can hardly be a prerequisite for the grasping of expression in music!) The sui generis aspect of musical expression, such as it is, will have to be reected in the analysis in a different manner.

Derek Matravers, Art and Emotion (Oxford: Oxford University Press, 1998), 131. Kendall Walton, Listening with Imagination: Is Music Representational?, Journal of Aesthetics and Art Criticism 52 (1994): 4761 (p. 56); Malcolm Budd, Values of Art (London: Penguin, 1995) 132. Kendall Walton, Projectivism, Empathy, and Musical Tension, Philosophical Topics 26 (1999): 40740 (p. 435).

Musical Expressiveness as Hearability-as-Expression

95

Compare hearing the sound of a trains wheels as a baby whining (but in an unusual, because regular and rhythmic, way), and hearing a stretch of music as an expression of sadness (but of an unusual, because non-corporeal and extraordinarily uid, sort). Now, do the parentheticals in those cases give part of the content of the hearing-as experiences in question? Perhaps yes, if we view such contents as highly backgrounded ones. But we might with more justice answer no, viewing those parentheticals not as giving part of the contents of those experiences, but rather part of the contents of subsequent reections on those experiences. Thus perhaps the qualication sui generis that attaches to the expressing that we hear music as doing when we register its expressiveness does not enter into the hearing-as experience itself, but only into characteristic further thoughts about that experience. The modier, in sui generis fashion, plausibly belongs to the content of a subsidiary thought on the expressing we hear in expressive music, not part of the content of the core experience of hearing the music as, or as if it were, an expression of some sort. The notion of hearing-as, it will have been noted, has been relied on rather heavily in the preceding discussion. So what is it to hear a stretch of music as something elseor alternatively, to hear that something else in that stretch of music? This remains a difcult matter, but for present purposes it sufces to locate hearing-as and hearing-in among perceptual acts that partake freely of, or that substantially enlist, the imagination. The agents one hears in music when one hears it as an expression of emotion, sui generis or not, are thus inescapably imaginary ones, ones displaying the indeniteness characteristic of all imaginary objects. To hear music as such and such is, perhaps, to imagine that the music is such and such, and more specically, to imagine of the music, as you are hearing it, that it is such and such. The worry is sometimes voiced, concerning the appeal to imagination in the analysis of musical expressiveness, that imagination is too unconstrained to secure the degree of objectivity that musical expressiveness appears to enjoy. But this worry can be put to rest by recalling that the appeal is not to what a passage might perhaps be imagined to be the expression of, but rather, to what a passage is most readily and spontaneously imagined to be the expression of, in which case the exercise often has a fairly unequivocal outcome.
On the effective equivalence of hearing-in and hearing-as in regard to music, see my Musical Expressiveness, in Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996).

96

Music

III RESEMBL ANCE-BASED VIEWS OF MUSICAL EXPRESSIVENESS Malcolm Budd, in a trenchant discussion of our topic, distinguishes a minimal, or basic, concept of musical expressiveness, and then identies three accretions to that minimal concept, resulting in three more elaborate conceptions of musical expressiveness, the third of which is more or less the conception defended here. The minimal conception goes like this: a stretch of music is expressive of E if one hears the music as sounding like the way E feels, or perceives a likeness between the music and the experience of E, and it is correct to do so. My main objection to Budds minimal conception, which is a resemblancebased one, is that it is simply too minimal. What it denes isnt yet musical expressiveness, but at most a rough precondition or typical upshot of musical expressiveness. Perceiving a likeness between two things A and Band especially where, as in the present instance of musical passages and emotional states, it is a matter of cross-categorial perceptionis not sufcient for hearing A as B. The latter is a distinct occurrence, which neither entails nor presupposes the former. Resemblance in various respects between the sound and shape of a passage of music and the inner experience or outer expression of an emotion is undoubtedly one of the chief grounds of musical expressiveness, but neither the resemblance as such, nor the capacity to make listeners aware of that resemblance, constitutes the expressiveness in question. I can perceive that or acquire the perceptual belief that a leafy tree resembles a bushy beard, for instance, and not have the experience of seeing the tree as a beard. I can notice the likeness between the two and yet not see the one in the other. But surely we cannot speak of a musical passage being expressive of an emotion unless listeners are induced to hear the emotion, or more precisely, an expressing of the emotion, in the passage, whatever degree of resemblance they might note, in whatever respects, between the passage and the emotion. Note further that even if a resemblance-based account of musical expressiveness could deliver the right verdicts in individual casesthat is, even if the degree of resemblance that an ensemble of musical features needs to have to some emotion for a passage possessing such an ensemble to be expressive of the emotion in question could be specied in some
Values of Art, ch. 4.

Musical Expressiveness as Hearability-as-Expression

97

general mannerthat would not constitute an acceptable analysis of musical expressiveness. For it would not elucidate what musical expressiveness was, but only what ensembles of musical features were coextensive with and underlay such expressiveness. Stephen Davies, the most prominent defender of a resemblance-based view of musical expressiveness, maintains that emotion words used to describe appearances, whether in persons, natural objects, or works of art, are parasitic on the use of such words to refer to felt emotions; they thus represent a secondary, though literal, use of such words. In this secondary use, says Davies, emotion words describe emotion-characteristics-in-appearance, and it is in these that musical expressiveness lies. According to Davies, the expressiveness of music consists in its presenting emotion characteristics in its appearance. . . . These expressive appearances . . . are not occurrent emotions at all. They are emergent properties of the things to which they are attributed. Davies goes on to explain that such musical expressiveness depends mainly on a resemblance we perceive between the dynamic character of music and human movement, gait, bearing, or carriage. In conclusion, Davies afrms that emotions are heard in music as belonging to it, just as appearances of emotion are present in the bearing, gait, or deportment of our fellow humans and other creatures. Daviess view might, I think, be stated as follows: P is expressive of E iff P exhibits an emotion-characteristic-in-sound associated with E, that is, exhibits a sound-appearance analogous to the human emotion-characteristic-inappearance of E. Though I am not unsympathetic to the basic thrust of this account of musical expressiveness, I have major qualms about the central notion in terms of which it is framed, namely that of musical emotion-characteristicsin-appearance. The problem is that the appearance of a passage of music is not precisely that of a person, or a persons face or body in any condition, or a persons behavior at any moment. It is instead a matter, when a passage displays an emotion-characteristic in its medium of sound sequences, of an appearance similar to that presented by a person in some state. But since everything is similar to everything else to some degree, the issue then becomes one of how similar such an appearance must be to one presented by human behavior in order to constitute an emotion-characteristic-in-sound of the
See his Musical Meaning and Expression. Ibid. 239. Ibid. 228. Ibid. 229.

98

Music

emotion in question, or else, as Davies sometimes puts it, of how similar the experience of musical movement and of expressive behavior must be, in order for the appearance generated by such movement to constitute an emotioncharacteristic-in-sound of the emotion in question. I think it is plain that there is no answer to this question except by appeal to our disposition to hear that emotionrather than another, or none at allin the music, that is, by appeal to our disposition to aurally construe the music as an instance of personal expression, perceiving the human appearances in the musical ones, in effect animating the sounds in a certain manner, to use a phrase given currency by Peter Kivy. Only if this occurs does the music have the expressiveness in question, regardless of the degree of similarity between the musics appearances and the human appearances by relation to which it ends up being expressive, or alternatively, the degree of similarity between the experiences of those appearances. There is simply no independent conception of and no access to what Davies calls musical emotion-characteristics-in-appearance apart from satisfaction of the hearability-in-the-music-of-an-expressing-of-emotion condition vis-`a vis attuned listeners; the latter is what gives content, ultimately, to the former, however familiar the appearances in question might be. What the musical emotion-characteristic-in-appearance of sadness is, in general, cannot be derived from the behavioral emotion-characteristic-in-appearance of sadness in persons. There is no translation rule from behavioral appearance-characteristics to musical appearance-characteristics; only the act of perceiving in music the outlines of the former gives rise, so to speak, to the latter. Though Davies does not want to be committed to the view that musical expressiveness consists in analogy or resemblance to literally expressive behavior, his invocation of emotion-characteristics-in-sound as something founded in and emerging out of such analogy or resemblance in any event suggests that such characteristics are speciable independently of experiences of hearing emotional expression in music, like the behavioral emotioncharacteristics-in-appearance associated directly with felt emotion. But that is to overlook the real differences between human emotion-characteristics-inappearance and the supposed musical emotion-characteristics-in-appearance being appealed to; the former can to some extent be catalogued independently

See Kivy, Sound Sentiment (Philadelphia: Temple University Press, 1989).

Musical Expressiveness as Hearability-as-Expression

99

of individual judgments of expressive import on the part of perceivers, whereas the latter cannot. Let me elaborate on this. We can give content to sad human appearance by glossing it as the appearance or kind of appearance sad humans typically display. But we cant analogously give content to sad musical appearance. There is no such thing as the appearance or kind of appearance that sad music typically displays. There is no extractable prole of sad musical appearance, as there is of sad human face or sad human posture; sad musical appearance, unlike sad human face and sad human posture, is not, as it were, paraphrasable. The only way to anchor sad musical appearance, I submit, is in terms of our disposition to hear such music as sad. The analysis of musics expressiveness must thus foreground that perceptual-imaginative experience, and not the resemblances that, no doubt, underlie it. Musical emotion-characteristics-in-appearance, if they are supposed to be something identiable apart from experiences of hearing such and such emotion or expression of emotion in music, and as parallel to human emotion-characteristics-in-appearance, are an illusion. Take sadness. Sadness is an emotion, that is, a mental condition with various cognitive, conative, affective, evaluative, behavioral, and possibly physiological aspects. Next we have sad look, which is more or less the lookin face and bodythat sad people typically wear, or wear when they are not trying to conceal or suppress their sadness. And then we have sad sound, which is more or less the soundvocal, for the most partthat sad people typically make, at least when not trying to conceal or suppress their sadness. So characterized, sad look and sad sound are human emotioncharacteristics-in-appearance, as is their conjunction, which we might label sad appearance. But now we come to the alleged corresponding musical emotioncharacteristic-in appearance, or sad musical sound. How is that to be cashed out, in light of the characterizations of sad look and sad sound just given? One possibility would be, as the musical sound that sad people typically make. But that cant be right, since sad people dont typically make musical sounds of any sort, and the sounds they do makeweeping, sighingare clearly distinguishable from music of an ordinary sort. A second possibility would be, as the sound that sad pieces of music have in common. But that is an even worse proposal, since it presupposes the prior identication of pieces of music as sad. A third possibility would be to appeal to purely technical or structural features of music, such as those of melody, harmony, tempo and texture. But that would be of no use, since even if there is a complex disjunction of technical

100

Music

or structural features coextensive with sadness in music, such a disjunction would not serve to explicate the concept of sad musical sound, nor would such a disjunction appear to play a role in our identifying passages as sad. The only proposal with any chance of success, then, must be that sad musical sound is sound resembling sad sound (the standard aural appearance of sadness) or, cross-modally, sad look (the standard visual appearance of sadness), or both. But to what degree? Theres the rub, for as we all know, everything resembles everything else, yet the degree of resemblance to an emotion required to make a musical appearance a musical emotion characteristic in appearance of that emotion cannot be specied in terms of some xable degree of resemblance between the two. It can only be specied, it seems, as whatever resemblance is sufcient to induce appropriately backgrounded listeners to hear the music as sad, or as expressing sadness. But that, surely, is to give up the idea that there is a recognizable musical emotioncharacteristic-in-appearance of sadness, somehow analogous to the human emotion-characteristic-in-appearance of sadness, on which the analysis of musical expressiveness can rest. In sum, if there really are musical emotion-characteristics-in-appearance, to which the explication of musical expressiveness must advert, we should be able to identify them other than simply as being appearances in which the corresponding emotion can be heard. Thus, for sad musical sound, there should be some possible specication or prole, however schematic, of what sort of sound that is, other than sound that invites hearing-as-sad. But there isnt.

IV INFERENCE-BASED VIEWS OF MUSICAL EXPRESSIVENESS Bruce Vermazen, Jenefer Robinson, and Robert Stecker are philosophers who subscribe to inference-based views of musical expressiveness. In the view of each of them, the expressiveness a passage of music possesses is something like the conclusion of an inference to the best explanation. In Vermazens case it is an ascription of a state of mind to the imagined utterer of the passage that best explains its distinctive features; in Robinsons case it is an ascription of a
See Vermazen, Expression as Expression; Jenefer Robinson, The Expression and Arousal of Emotion in Music, in Philip Alperson (ed.), Musical Worlds (University Park, PA: Pennsylvania State University Press, 1998), 1322; and Robert Stecker, Expressiveness and Expression in Music and Poetry, Journal of Aesthetics and Art Criticism 59 (2001): 8596.

Musical Expressiveness as Hearability-as-Expression

101

state of mind to the imaginary protagonist of the passage that gures in the best interpretation of the musical work taken as a whole; in Steckers case it is a best hypothesis on the part of an ideal listener as to what state of mind the composer of the music intended such a listener to hear in the passage. Though I share with Vermazen and Robinson the commitment to personae in the analysis of musical expressiveness, and with Stecker the notion that certain aspects of artistic meaning are amenable to analysis in hypotheticalist terms, I disagree with all three that basic musical expressivenessthat is, the expressiveness of individual passagescan be constitutively tied to the sustaining of inferences about the music. That is due to my belief that basic musical expressivenessthough perhaps not all sorts of expressiveness, such as that more typical of literature, involving articulate states of mind, nor that perhaps attaching to works of music as wholesis something directly heard, not inferred, by attuned or properly backgrounded listeners. Otherwise put, inferentialist views of expressiveness fail to capture the immediacy with which we register basic musical expressiveness. I thus continue to think, pace Stecker, that immediacy is a proper desideratum for an account of musical expressiveness, and that my ready-hearability-as-expression account acknowledges that better than the hypotheticalist account that Stecker proposes instead. It is true, as Stecker points out, that hearing emotion in music and judging the expressiveness of music are not the same things. But there is a third thing, perceiving the musics expressiveness, and that is not the same as judging its expressiveness to be such and such. The expressiveness of music, I claim, and not just the emotion in music, is standardly something directly registered, not just conjectured about. My account can accommodate this fact, as can also a resemblancebased account, but an inferentialist account cannot. Now Stecker, seeking to hoist me with my own petard, argues that my ready-hearability-as-expression proposal is ultimately as inferentialist as his own, since in order to determine that a work was expressive of E on the basis of his hearing an expression of E in it, even a properly backgrounded listener would have to engage in inference, taking the fact of his so responding as a premise together with the premise that he was in fact a properly backgrounded listener listening properly. But I anticipated and replied to that line of criticism in my previous essay on this topic. I there emphasized that
Stecker, Expressiveness and Expression in Music and Poetry, 92. Musical Expressiveness, 11819.

102

Music

a qualied listener who hears the expression of some emotion in music normally acquires, without reection, the conviction that the music is expressive of that emotion, that is, would be readily hearable as such by other qualied listeners. And that, I suggest, counts as perceiving the musics expressiveness, or as much as anything could. Whether the listener in addition thereby knows that the music is thus so expressive is another matter, one that may indeed involve further reection or investigation. Here is a variant of that reply. Qualied listeners arguably tacitly assume, while listening to music, that they are qualied listeners, and are listening appropriately. Thus, on the view I favor, hearing an expression of E in a stretch of music becomes, for such listeners, tantamount to hearing the musics expressiveness of E. All qualied listeners need do to hear the expressiveness of the music is to readily hear expression of emotion in it. Since they are, and unreectingly assume they are, qualied listeners, listening appropriately, their readily hearing such and such an expression in the music directly manifests its being readily so hearable by such listeners! Note, however, that this line would not work to secure the immediacy of musical expressiveness on a hypotheticalintentionalist view of it, even were we to grant a parallel tacit assumption on the listeners part, to the effect that he was an ideal listener. For arriving at a best hypothesis of what the historically rooted composer intended for one to hear in a given passage remains an ineliminably inferential affair. There is still another reason to resist a hypothetical-intentionalist account of musical expressiveness. Expressive content in music and expressive content in poetry certainly seem to be quite different sorts of things, hence it would not be surprising if they lent themselves to different sorts of analyses. The latter is largely propositional, and so reasonably assimilated to the kind of meaningbasic literary meaningthat a view like hypothetical intentionalism is designed to account for, whereas the former is largely non-propositional, hence not reasonably assimilable to basic literary meaning. I am inclined to think that a perceivability-as-if-expression account is apt not only for music, but for non-representational art generally. However, where we have to do with representational art, and perhaps especially, literary artincluding narrative painting, epic poetry, theatre, cinema, and the novelemphasis on the immediacy of expressiveness seems less apt, and the merit of a more inferentialist account of expressiveness seems correspondingly greater. The reason may simply be that immediacy of expression is an
See chs. 17 and 18 of the present volume.

Musical Expressiveness as Hearability-as-Expression

103

appropriate demand for affective states or attitudes, the core expressive content of non-literary arts such as music or abstract painting, but not for propositional beliefs or thoughts, the core expressive content of literary arts such as cinema or the novel. We can now offer an observation about expressiveness in art applicable across the arts, whether representational or non-representational, literary or non-literary. Remember that expression in general can be characterized roughly as the evidencing of a state of mind through some sort of external manifestation. In the case of art, the external manifestation is not behavior, but rather the artwork itself, in all its perceivable particularity. But just as we grasp literally expressed states in and through the behavior, verbal and otherwise, that expresses them, so with expressive art we grasp the asif expressed states in and through the concrete vehicle of the work, its metaphorical body. With expressive art, expressiveness is grasped through perceiving the work in its specic detailwhether of word, paint, sound, or stoneand the states expressed are ones that perceivers are thus aided to enter into in imagination precisely in virtue of perceiving the work through which, either immediately or inferentially, they grasp what states those are. V SOME OBJECTIONS TO THE HEARABILIT Y-AS-EXPRESSION VIEW OF MUSICAL EXPRESSIVENESS The rst and most common objection is this. A number of writers charge that competent listeners do not in factor at least not all of them, all of the timehear or imagine personae in music whose expressiveness they are registering, and thus that such imaginative hearing cannot be constitutive of hearing music as expressive and the disposition to induce such hearing as constitutive of the musics being so expressive. Well, that may be how it sometimes seems, or seems on the surface, but if expressive music is, as I maintain, music readily heard as, or as if, expression, and if, in addition, expression requires an expresser, then personae or agents, however minimal, just are presupposed in the standard experience of such music. But, of course, one may not always notice or acknowledge what
Stephen Davies, Contra the Hypothetical Persona in Music, in Mette Hjort and Sue Laver (eds.), Emotion and the Arts (Oxford: Oxford University Press, 1997), 95109; Walton, Projectivism, Empathy, and Musical Tension; and Stecker, Expressiveness and Expression in Music and Poetry.

104

Music

is presupposed in ones imaginative hearing of music. The claim is not that listeners are always explicitly aware that personae are involved in their hearing music as expressive. For people are often not entirely aware of what is implicated in a perception or experience they are having. A point I have made before is worth recalling here. It may indeed be true that listeners who recognize the expressiveness of a passage of music do not invariably hear it, then and there, as the expression of an emotion. But all that the theory requires is that they recognize the music as readily hearable as the expression of an emotion, even if, for one reason or another, they do not themselves give in to that inducement on a given occasion. Moreover, quite possibly all such cases are ones in which listeners are recognizing passages they have previously heard as the expression of given emotions, or else as ones highly similar to such passages, thus presupposing occasions on which the expressiveness in question was in fact grasped through an experience of hearing-as-the-expression-of. Furthermore, some of the discrepancy with the sincere avowals of listeners on this subject, I suggest, is that a passage of music may more loosely have an emotional quality, in virtue of suggesting an emotion through its appearance, without being strictly speaking emotionally expressive, understood as being such as to induce hearing-as-expression of that emotion. The nale of Beethovens Fifth Symphony, to take a stock example, is expressive of something like triumphant joy, and I think it is hard not to hear that nale as if there is someone, or some agent, there who is expressing his, her, or its triumphant joy in those familiar musical gestures, the character of which is rendered especially vivid in virtue of the movement in which they occur being the successor and culmination of the three that precede it. By contrast, the opening Prelude of Bachs Well-Tempered Clavier has an emotional quality one might describe as contentment or equanimity, and yet perhaps one is not induced to hear it as, or as if, the expression of that state of mind. But then it is probably also right to deny that it is expressive of contentment or equanimity, in addition to just possessing the corresponding emotional quality. The second objection is this. It has been suggested that the appeal to apt hearability-as-expression as a benchmark of real, as opposed to merely apparent, expressiveness, a feature of my original formula, will either not do the work that is required of it or else is called upon to do too much work.

Stecker, Expressiveness and Expression in Music and Poetry, 914.

Musical Expressiveness as Hearability-as-Expression

105

I am at this point inclined to agree. The qualication of apt in my original formula was an evasion. What I now think is that the burden of securing the objectivity of expressiveness, or equivalently, the normativity of judgments of expressiveness, must simply rest on the properly backgrounded listener and his or her hearing of a passage in its proper intrawork and extrawork context. Where there is ready hearability-as-expression under such conditionsevidenced most clearly by convergence in experiences among such listenersthen objectivity and normativity are present. Where not, then not. Exactly how large is the domain of objective expressiveness in music, which depends on such convergence, thus remains an open question. The third objection is centered on a related worry about my original formula voiced by Roger Scruton, to the effect that appeal to a reference class of qualied listeners whose ready hearing of a passage as expression of E serves as the mark of the passages being truly expressive of E is doomed to vacuity, because the reference class can only be characterized as the class of listeners who in fact perceive the passages expressiveness. But that is not so. It is like suggesting that the only way to characterize the class of appropriate, objectivity-anchoring perceivers for the color of a given patch of greenish paint is as perceivers who correctly perceive that the patch in question is greenish. The reference class of listeners anchoring the objectivity of expressiveness in a given musical genrewhat I mean by properly backgrounded listenersis roughly that of listeners demonstrably competent at understanding such music, such competence being manifested through various recognitional, continuational, and descriptive abilities, and whose other judgments of expressiveness are in line with established ones in uncontroversial cases. There is perhaps a certain amount of bootstrapping involved in this picture of the qualied listener for a given musical genre vis-`-vis the expressa ive and other qualities of works in that genre of whose objectivity the qualied listener is to serve as a benchmark, but there is nothing, I think, fatally circular in it. At any rate, it seems the sort of difculty that affects all attempts to analyze perceptual properties in terms of appearances or dispositions to appear relative to a class of perceivers of a certain sort. Paul Boghossian has recently seconded Scrutons worry in a more general form. Boghossian charges that the appeal to qualied listeners in the analysis
Scruton, The Aesthetics of Music, 353. Boghossian, Musical Experience and Musical Meaning, in Kathleen Stock (ed.), Philosophers on Music (Oxford: Oxford University Press, 2006).

106

Music

is vacuous because there is no way to characterize what makes for a qualied listener without presupposing an understanding of musical expressiveness that does not invoke such listeners or the ways such listeners are disposed to hear or otherwise respond to music. Again, qualied listeners for a given piece of music are naturally not to be characterized, unhelpfully, as ones who correctly hear the expressiveness of the given piece. How, then, are they to be characterized? As suggested earlier, perhaps as musically competent listeners whose judgments of expressiveness in the tradition in question accord with accepted ones in paradigm cases. But what are paradigm cases of musical expressiveness? Well, one conception of them would be as pieces in a given tradition on whose expressiveness almost all at least minimally musically competent listeners agree. Boghossian retorts that even on this suggestion, paradigm cases having the expressiveness they do remains unanalyzed, thus presupposing some ultimately non-experiential, non-response-dependent notion of such expressiveness. But I am not so sure. For the expressiveness of paradigm cases, we may suggest, comes to precisely the same thing as it does for non-paradigm cases, namely, the musics being most readily hearable as the expression of such and such emotion by qualied listeners, the only difference being that, since these are paradigm cases of expressiveness, they will be so heard by virtually all listeners who are at least minimally musically competent. The fourth objection is another difculty Scruton has articulated for my proposal that is worth addressing here. It is a particular elaboration of the skepticism acknowledged earlier as to whether we can form an idea of the singular way of expressing emotions that expressive music on my proposal is made out to be, at least in imagination, and whether or not in so doing we characterize it to ourselves as sui generis. I quote the objection in full:
When we hear expression in music, Levinson suggests, this is like hearing another person express his feelings. But in what way like? We have no prior conception of what it would be to express feelings in music: if we can think of someone doing this, it is because we have an idea of the expressive character of music, and therefore can imagine someone choosing just this piece of music, to convey just this state of mind. Our ability to imagine a subject expressing his feelings in just this way is predicated upon our ability to recognize the expressive content of music. Only if we can independently recognize the emotional content of music, therefore, can we embark on the thought-experiment required by Levinsons denition.
The Aesthetics of Music, 352.

Musical Expressiveness as Hearability-as-Expression

107

Scruton claims that my account is committed to listeners being able to conceive what it would be to express an emotion, in the literal sense, in or through music. He then suggests that their only way of doing so would be by imagining someone choosing suitable music to convey the emotion in question, which obviously presupposes an antecedent grasp of the musics expressiveness, thus rendering the putative account of such expressiveness otiose. But both points in this objection are misplaced. First, my account does not imply that listeners who register expressiveness in music possess a concrete conception of what literally expressing emotions through music instead of behavior would amount to, or how such expressing would work. The account requires only that listeners are able to imagine music to be such a literal expression. Second, listeners who do that are not constrained to think of the musics persona as somehow choosing from available musical items ones suitable to convey his changing moods, like a sound editor selecting tracks to go with the successive scenes of a lm. Scruton has simply misunderstood the nature of the thought-experiment that, if my account of musical expressiveness is correct, one is effectively called upon to perform in order to grasp the expressiveness of music. It is not imagine someone choosing from among preexisting music to convey a given state of mind; now, what state of mind would that be if he chose the music you are hearing? It is, rather, imagine the music you are hearing to be the literal expressing of a state of mind; now, what state of mind does that appear to be?, or perhaps, imagine that the musical gesture you hear in the music was your own; now, what state of mind do you appear to be in? That thoughtexperiment, which my proposal is committed to, grounds identication of the musics expressiveness, but does not, like the one Scruton saddles me with, presuppose such identication. If I am right, one grasps what a musical passage expresses precisely in virtue of imagining, or being disposed to imagine, a mental state it is as if the literal expression of. That there is no algorithm or procedure for this thought-experiment, unlike the one Scruton would substitute for it, does not entail that it cannot be carried out. We carry it out, in fact, every time we attend to some musics expressive dimension and attempt to articulate for ourselves what it is. VI PE RSONAE IN MUSIC Many writers, including those sympathetic to imagined expressions and their personae, have voiced concerns regarding the indeterminacy of

108

Music

musical personae and the indeniteness of the principles governing their postulation. When is the persona of one passage the same as the persona of another passage? When is there continuity and when discontinuity of persona, as a work progresses from beginning to end? Might there be multiple personae present in a single passage? Might personae heard in different passages be related to one another through recognition, solidarity, or opposition? I cannot address these questions here, which go well beyond the scope of this essay. What I wish to underline, in closing, is just that the sustainability of the thesis of a minimal persona we are induced to hear in expressive music, and typically do hear in it when listening attentively, is not affected by worries of this sort, even if they are ultimately unresolvable. The persona implicated in the ready-hearability-as-expression account of musical expressiveness is merely the agent of the expression we hear in expressive music, or the owner of the musical gesture that is the vehicle of that expression. Whether that persona persists as the music proceeds, whether a given persona is accompanied by others, whether personae enter into relation with one another, and so on, are matters on which the account of basic musical expressiveness here defended can remain agnostic.
See Walton, Listening with Imagination; Davies, Contra the Hypothetical Persona in Music; Fred Maus, Music as Drama, Music Theory Spectrum 10 (1998): 5673; and Gregory Karl and Jenefer Robinson, Shostakovichs Tenth Symphony and the Musical Expression of Cognitively Complex Emotions, in J. Robinson (ed.), Music and Meaning (Ithaca: Cornell University Press, 1997), 15478.

7
Nonexistent Artforms and the Case of Visual Music
There is a difference between the realm of art as a whole, however we characterize that, and the class of all artforms that exist, or have existed, at a given point in time. That difference, seen from one angle, is the class of nonexistent but possible artforms. This chapter will be devoted at the outset to exploring this latter class. My tools, however, will be merely those of a philosopher, not a soothsayer. The problem concerning nonexistent artforms of perhaps greatest interest is why certain theoretically possible ones have not been essayed, or else have not proven successful even if essayed to some extent. Examples of nonexistent artforms, by at least the second of these tests, are kinetic painting, wordless song, danced poetry, olfactory art, and visual music. One reason this problem is interesting is that insofar as art as a whole has often evolved explicitly through the emergence of new artforms, the problem bears directly on the future of art; if there are fewer untried artforms out there, or if those which remain unpracticed even though tried are indeed inherently problematic, then the future of art is, in this respect, somewhat more dim than it would otherwise be. This essay has two parts. In the rst part I propose some ways of thinking about the eld of nonexistent arts as wholein effect making some advance inroads into itand in the second part I try to cast light on why certain artforms that seem eminently possible in fact fail to exist, through a case study of one such would-be art, that of visual music, i.e. a structured organization of colored presentations in time, such as might be provided through the medium of color lm. Why doesnt visual music, despite experiments in that direction,
First published in A. Haapala, J. Levinson, and V. Rantala (eds.), The End of Art and Beyond: Essays after Danto (Atlantic Highlands, NJ: Humanities Press, 1997), 12239.

110

Music

exist as an entrenched artistic activity, as a medium with promise and a loyal audience? One may be forgiven for suspecting, in a pessimistic vein, that the potential for a signicant artistic visual analog to music is just not there. I will investigate, rst, why this should be so, and second, how such an art might best proceed anywayagainst the odds, as it were.

I T HE FIELD OF NONEXISTENT ARTS A central question is how the domain of nonexistent but possible artforms is to be circumscribed or delineated. How can we get a handle on, or sketch some of the features of, the domain we are interested in? One way to construe this question is as a request for acceptable principles of generating potential artforms. Taken that way, it seems there are two approaches we could adopt, which we might characterize, respectively, as the combinatorial and the extrapolational. The rst, combinatorial, approach would be to set up an abstract matrix of artistic properties (materials, modalities)e.g. spatial, temporal, representational, sonic, monochromatic, gurative, narrative, one-stage, stoneusing, wood-using, human-body-employing, etc.in which every existing and also nonexisting artform would be represented as a particular conjunction of such properties (or their negations). Thus, classical Greek sculpture would be approximated by the conjunction: spatial/static/gurative/stone-using/ manually-worked, while a presently nonexistent artform would be represented by the conjunction: stone-using/temporal/dramatic/multi-colored. The other, extrapolational, approach would be to go beyond existing artforms by the use of various intelligible rules of projection, such as juxtaposition, reduction, fusion, opposition, or mutual accommodation. Thus, we might have an artform which involved the simple addition of dance to painting, or which resulted from the fusion of poetry, music, and calligraphy. On either approach, of course, the issue will remain of whether what is formally generated by combination of analytic elements or extrapolation from actual artforms is conceptually coherent, practically feasible, and artistically promising. That issue will be engaged, at least in one instance, in my consideration of visual music in the second part of this essay. One problem with the combinatorial approach to identifying nonexistent potential artforms is that we have not said where the primitive termsthe properties whose combination is to represent any given artformare to come

Nonexistent Artforms and the Case of Visual Music

111

from. What justies the particular set of elements out of which the matrix would be constructed? Why are stone-using, temporal, and narrative reasonably included in this set but not, say, water-soluble, high-risk, expensive, or innate? The answer must be that we only arrive at a plausible set of artformconstituting properties by working from the artforms we are acquainted with, and by relying on our background understanding of the nature of art in general, its typical objectives and achievements, and its overall place in human affairs. That is to say, we start with familiar, understood arts, then decompose them into their salient features, ones that both differentiate them from other arts and are important in criticism and appreciation. Having arrived at such elements, we can then mentally experiment with recomposing them in different waysgrouping them in different packagesand reecting on the outcome. It must be observed, though, that we will not, by this method, really have constructed all possible arts from the ground upthat is, from a purely abstract eld of possibilities. For the terms of our matrix will inevitably come from the arts we know and love, as they are, or have been, constituted. The combinatorial method, despite initial appearances, could not be deployed ahistorically, at least not with any hope of interesting results. There is something unappealingly inefcient, as well, in the combinatorial method of generating possible artforms as I have sketched it. For if we are going to have to think in terms of existing artforms anyway, why not skip the step of decomposition into elements and just consider combinations of, or confrontations between, existing artforms themselves? The combinatorial method seems inefcient, further, in that given a certain constellation of dening features, for a given position on the matrix, the actual shape of the artform or artforms so conforming to that position or set of properties will usually be excessively indeterminate, hard to form an image of. Third, the pure combinatorial approach seems quite likely to yield a numberand perhaps a majorityof combinations which are either conceptually problematic or practically infeasible. All told, then, I suspect we are perhaps better off pursuing the extrapolational methodcombining or modifying existing artforms directlyas a way of thinking our way toward the eld of nonexistent artforms. Such a procedure has also the advantage of paralleling, to a much greater degree, the
For example, we know that the arts are activities, that they typically aim at pleasure or satisfaction of some sort, that they usually involve skills, that they are modes of meaning and articulating the world, and so on.

112

Music

evolutionary and synthetic process by which new artforms (e.g. lm, kinetic sculpture, video art) have actually been generated in the past. Advances in art on the level of artforms have regularly come about through the impact of disparate arts on one another, or through the modication of existing artforms in the direction of others, but not through positing artforms de novo to correspond to abstract congeries of artistic properties not previously bundled together. In an earlier essay, Hybrid Artforms, I explored the issue of existing artforms that have arisen through the combination of two (or more) preceding oneslabelling these hybrid artformsand suggested a threefold categorization of such hybrids, into juxtapositional (additive), synthetic (fusional), and transformational ones. The distinction between them is roughly as follows. In a juxtapositional hybrid the arts are combined in a largely discrete, non-interpenetrating manner, so that the contributions of each artform entering into the result are independently identiable. In an object of the hybrid art, objects of the contributing arts are still discernible as simply parts, spatial or temporal, of the hybrid art object. In a synthetic hybrid, by contrast, the contributing artforms are more or less fused in the result, so that an object of the hybrid fails to belong clearly to either of the contributing artforms, in their standard conceptions, and fails also to exhibit parts which can be simply assigned to one artform or the other. Finally, a transformational hybrid is intermediate between the previous two cases, and occurs when there is some interpenetration, rather than mere addition, of two artforms, as evidenced in the nature of the resulting art object, but where one artform clearly holds sway in the end and can still claim the product as one of its own, if singularly modied in the direction of the other art. Examples offered of these three categories were the following: (a) symphony with light show, dance with calligraphic drawings [juxtapositional]; (b) music drama, concrete poetry, shaped canvas [synthetic]; (c) kinetic sculpture [transformational]. Visual music, in the form of abstract color lm, could probably be considered a transformational caseas color lm pushed in the direction of pure music, adopting both its means and the sort of experience and expressiveness characteristically aimed at. This would be transformational, rather than
See Jerrold Levinson, Hybrid Artforms, in Music, Art, and Metaphysics (Ithaca: Cornell University Press, 1990). This essay was originally published in 1984. I also count as transformational hybrids cases of new artforms in which paradigm objects of the art are altered relative to those of , the parent art, in a manner producing tension with the implicit structural norms of , even if not in the direction of some identiable second art, . (These are in effect cases of self-induced hybridization, or differentiation.)

Nonexistent Artforms and the Case of Visual Music

113

synthetic, because the result would remain unequivocally a lm, though one whose obvious self-restriction (e.g. non-photographicness, non-representationality) would appear bizarre except for the intended parallel with music. More obviously, such visual music would be transformational, rather than juxtapositional, because music there serves as a model or inspiration, and not as a literal component in the result. I start, then, with these three categories of hybrid, in terms of the sorts of extrapolation they suggest for artforms which do not yet exist, and then proceed beyond them to consider further, or ner-grained, ways of projecting from existing artforms. So, rst, we can imagine all those artforms that would involve merely juxtaposing objects or activities of preexisting arts that have not till now been placed in conjunction. Thus, we might have musidrawing, the art of drawings exhibited for prearranged lengths of time simultaneously with selected or newly composed stretches of music, or pulpture, the art of pairs of paintings and sculptures made to be presented and experienced together, in denite spatial relation to one another. An important variable even for artforms such as these, arrived at merely through unprecedented juxtapositions, would be the degree of interrelatedness envisaged between the components brought into conjunction. This could range anywhere from selfcontained, oblivious standing-alongside-of, to delicately nuanced mutual sensitivity and adjustment, of the components vis-`-vis one another. a The synthetic paradigm of hybridization, in which the identities of contributing arts are partly dissolved, is a more complicated one, and we can derive from it at least two distinct rules or procedures for generating notyet-existent arts. One is this. Given two artforms, and , posit an artform whose objects are structurally or formally halfway between those of and those of . The genres of shaped canvases (e.g. those of Frank Stella) and canvaseswith-attachments (e.g. those of Jasper Johns), are ones that have come about through that sort of cross-fertilization. Both cases mediate between painting and sculpture, in that objects of these genres are neither as at and rectangular as the paradigms of the former, nor as three-dimensionally rounded and conceived as paradigms of the latter. But consider now the art of danceotry, intermediate between dance and poetry. A typical work of danceotry would involve a performer who goes through a series of movements during her recitation, but ones of more restricted scope and ambit than are generally found
Sound lm in the familiar sense, though, could be considered as roughly a juxtapositional hybrid of silent lm, music, and dialogue.

114

Music

in dance, and whose recitation, in turn, would be less elaborate and selfcontained, more sparse, than that normal for poetry offered on its own. Or consider archisculpture, whose objects are house-like constructions which both stand erect and enclose an interior space, but which otherwise out many of the exigencies of the architecture of dwellingse.g. intact roofs, usable entrances, sound construction materialsin favor of the formal freedom of sculpture. Or, third, consider sonopoetry, which involves the production of measured sequences not of words but only non-verbal vocal sounds, e.g. hiccups, snorts, gasps, whistles, whispers. Danceotry, archisculpture, and sonopoetry would all be applications of the rule of projecting an artform that is structurally intermediate between two given ones. A close cousin of this rule, though still distinct from it, would be one which projected artforms which in some way joined or compounded salient structural features of the objects of two arts, rather than, so to speak, splitting the difference between them. Of course in some cases mediationthe procedure involved in the previous ruleis the only manner of combination that makes sense; for example, if it is a question of a two-dimensional art (painting) and a three-dimensional one (sculpture), a synthetic hybrid can be more readily conceived whose objects are somewhere between two- and three-dimensional, loosely speakinge.g. Johns paintings with protuberancesthan one whose objects would be clearly both two-dimensional and three-dimensional. But in other cases, combination of arts could conceivably involve amalgamation, or compounding, of their distinctive traits or potentials. Consider aquarelloetry, a given specimen of which displays both the ordered sequence of carefully chosen words denitive of poetry and, superimposed upon it, an abstract pattern of water colors designed to complementto enhance and be enhanced bythe sounds and sense of the words on the page. Or novellasong, which would involve short stories composed explicitly to be chanted, or intoned, by an unaccompanied performer; one can imagine this being done in such a way that the result seemed roughly equally a musical object and a literary narrative.
An object that may be an advance member of this artform was on display in the National Gallery of Art East, in Washington, D.C., during the summer of 1990. It consisted roughly of a cube about 10 feet on each side, whose walls were mirrors inside and out, and whose rather cramped interior one could enter through a single opening. Perhaps fun houses and chambers of horror in certain amusement parks would also come close to qualifying. This genre was actually devised and practiced by Dadaist pioneer Hugo Ball early in the twentieth century. Some entries under the category of Performance Art perhaps already exemplify this description. Note also that the description is intended to exclude lieder, obviously already an existing,

Nonexistent Artforms and the Case of Visual Music

115

I turn now to the transformational model of hybridization for additional suggested rules for generating new artforms. The distinction between transformational hybrids and synthetic ones, recall, is that in the former case the new artform is most naturally seen as a variant of just one of its parents, the other, if any can be identied, serving as more of a godfather. One principle of generation with wide application might be this. Given an existing artform, whose objects exhibit certain standard (or dening) features, and certain variable (or elective) features, and for which certain features are contra-standard (or disqualifying), posit another artform that is just like the rst except that some standard feature of the objects of the rst artform is now allowed to become a variable, or contra-standard, feature, or else some variable or contra-standard feature allowed to become a standard feature. Thus, if we consider traditional sculpture and imagine that one of its standard (or category-dening) features, immobility and xed spatial relationship of its parts, is allowed to become a contra-standard (or category-disqualifying) one, everything else remaining the same, then we end up roughly with kinetic sculpturesculpture turned in a danceward direction. Or take traditional oil painting, for which a variable feature is the use of reds, and convert this to a standard feature, thus generating a monochrome subgenre of painting in which reds are de rigueur, and the use of other colors a agrant violation of an implicit norm. Or take plays, for which number of characters is variable, and standardize that at two, thus generating a two-person subgenre of theatre, in which the appearance of a third party, or a lengthy absence of the rst or second parties, is strongly contraindicated. The general idea behind this principle is one of modifying existing artforms by either imposing a restriction on a dimension where freedom had previously reigned, or else opening up a dimension which serves as a dening constraint, either positively or negatively, on the artform as normally constituted. Another transformational rule, perhaps a special case of the above, is familiar because it has often been invoked in this century. Take an existing artform and envisage a reductive or minimalist version of it, where this involves signicantly limiting the structures, subjects, or contents of the objects produced, relative to normal procedure in the art.
well-established art, whose allegiance to the category of music clearly dominates its relationship to literary storytelling. The notions of standard, variable, and contra-standard properties relative to a category of art (e.g. artform, genre, medium, style) are taken from Kendall Walton, Categories of Art, Philosophical Review 79 (1970): 33467.

116

Music

The minimal sculpture of artists such as Donald Judd or Carl Andr can e be seen to have developed by taking constructivist or assemblage (as opposed to carved or cast) sculpture as it then existed and restricting it to a small diet of geometrical forms, a unity of material and color, and a severe set of possible spatial interrelationships. The early minimalist music of Steve Reich or Philip Glass dened itself as a genre by explicitly eschewing the melodic and harmonic complexity, the reliance on key movement, and the driving, developmental, and goal-oriented aspect of traditional tonal music, leaving manipulation of pulse, rhythm, and phase as almost its sole sphere of operation. The eld of abstract painting generally, obviously enough, can be thought of as having arisen as an application of this rule, by deliberate exclusion of representational matter from the possible content of a painting. An important variety of minimalist cinema (as exemplied by some lms of Michael Snow) can be understood to have come about through taking the lmmakers usual freedom of shot and camera movement and jettisoning those in favor of a self-imposed, spartan limitation to very long takes and xed camera positions. The appropriational art of the photographer Sherrie Levine, nally, can be construed as an instance of the rule, where photography in the robust sense is transformed into a peculiar minimalist subgenre governed by the restriction that the only possible subject for a photograph will be a photograph of some earlier, usually famous, photographer. Imagining further applications of the rule of minimalist or reductive versions of existing arts is not difcult. Take kinetic sculpture, already a clear product of redrawing the lines around existing arts, and posit a subgenre which allows motion only in a single plane, where that plane is perpendicular to the line of sight of the viewer; the objects of such a restricted kinetic sculpture would tend to establish a kinship with painting, equal to any they retained with traditional sculpture in the round. Take the genre of the short, descriptive poem (e.g. haiku) and impose on it the restriction that all words except nouns be banned, thus conning the poet to getting his effects, creating images, conveying ideas, etc. merely through carefully judged and paced lists of things. Take modern dance and imagine a form of it in which all movements were to be carried out extremely slowly, almost at the limit of detection of movement. Now consider another principle for projecting a
The writer Georges Perec initiated a genre of this sort, by producing a viable instance himself: a novel written without using the letter e, entitled La Disparition. For additional discussion, see Elster on Artistic Creativity, Ch. 4 in this volume.

Nonexistent Artforms and the Case of Visual Music

117

transformation of a given art. Instead of imagining a reshufing of dening and optional features of objects in an artform, or imagining a reductive version of a given artform, we might try to imagine how some artforms could remain themselves and yet be transformed so as to achieve something of the distinctive effects or characteristic experience of another artform. Kinetic sculpture has already been offered as an art viewable as the upshot of the rule of transforming a standard feature of an artformimmobilityinto a variable or contra-standard one, but we could alternatively see it as originating in the notion of making sculptures that would achieve some of the distinctive effects of dance, and not initially in an envisaged structural reconception of the art. Now reversing this last thought we could ask, in accord with our latest principle, how dance, while in some sense remaining dance, might yet answer to some of the aims of sculpture. Possibly through the medium of choreotableaux, or static dance. A work of choreotableaux would consist of an unmoving arrangement of individuals on a stage, with all aspects of posture and gesture xed, which would be assumed by dancers and held, for contemplation, during a period of a few minutes. Visual music, in the form of abstract color lm, could be laid down to this same projective principle: take some visual art and modify it in such a way that it achieves some of the characteristic effects or experiences provided by pure music. Obviously this kind of thought projection will be successful in some cases but not in others; certain artforms are structurally and thematically so different that we get no purchase on the idea of the one realizing somehow, in its own medium, the qualities of the other. Thus, if we ask whether the novel might be modied so as to achieve the characteristic objectives and results of sculpture, we would have, so far as I can see, no foothold at all. I would suggest just one further transformational principle for generating, in the abstract way we have been pursuing, artforms that do not currently existone that, once more, yields visual music as an output. It is this: posit an artform, as far as possible like an existent artform, in which the sense modality centrally appealed to in the latter is replaced by some other sense modality. The idea is to see whether unprecedented arts, modeled on established ones, can be built around senses other than those already in play. This idea is perhaps particularly apt in connection with those unexploited, socalled lower senses, e.g. touch, taste, smell, proprioception, which are not the primary basis of any established art at present. In the rst case, if we start with ordinary sculpturewhich is, by and large, for seeing, i.e. ambulatory visual

118

Music

beholdingand replace seeing by touching, we arrive at tactile sculpture. In tactile sculpture objects are fashioned for palpation and caressing with no concern for, or even explicit disregard of, how they look. What is only just tolerated, or even proscribed, for ordinary sculpture, i.e. tactile exploration, would be here the invited and exclusively prescribed mode of address. Nor would such tactile sculpture be hard to arrange, in practical terms: blindfolds, or enclosures with small openings for the hands, could be made good use of. I leave it to your imaginations what olfactory music, gustatory painting, and the like, might involve. I have not mentioned, in the foregoing, principles for projecting nonexistent artforms that could someday exist by appeal to radically new means and media that technological advances will make available. Most of the new artforms of the future, it is fair to say, will have such advances as a partial cause. The only problem, from present point of view, is that there is no way to anticipate what these might beas noted earlier, my toolkit as essayist contains no crystal balland so nothing contentful on this wavelength to add to my sketch of the domain of possible but unrealized arts. It is hard not to appreciate the essential technological contributions to, or at least preconditions of, the emergence of photography, etching, cinema, computer music, musique concr`te, video art, holography, earthworks, and jazz; these did not e come about merely because some overlooked combination or transformation of existing arts was nally hit upon, some underutilization of an existing sense capacity nally noticed, or some approach of one art to the aims of another nally worked out. Still, in many of these cases, one might locate the spur to develop technological means enabling these arts to exist, or at least a rationale for their emergence after the fact, in projections of the kind I have been pursuing. On this way of thinking, photography arises in part because a visual artform capable of the detail and precision of painting or engraving, but executable with the spontaneity and speed of, say, a charcoal sketch, was really asking to be brought into existence. Holography comes to ll the space opened up logically by positing a cross, in the biological sense, between photography and sculpture; a hologram is a synthetic hybrid that no longer comfortably resides
I am aware that artforms along these lines have already been pursued among and for the blind, and that the great sculptor Constantin Brancusi created some pieces in this vein, but that does not alter the fact that tactile sculpture does not exist as a ourishing artform in mainstream artistic culture. I am thinking here of the crucial role of the saxophone, a late nineteenth-century Belgian invention, in the evolution of jazz sound.

Nonexistent Artforms and the Case of Visual Music

119

under either banner, photography or sculpture, but exhibits features of both: fashioning by light, and three-dimensional form. Musique concr`te, or cole lage music, while obviously unrealizable without the magnetophone or tape recorder, just as obviously gets off the ground conceptually by attempting to transpose the collage and constructivist procedures of modern painting and assemblage to the realm of sounds in time. In another vein, it could also be objected to the domain-exploring procedure I have adopted that it mainly looks for new artforms to generate themselves out of formal impulses, rather than expressive, moral, social, or political ones. That is largely so, but I think the form-based procedure adopted provides the only effective means to sketch any signicant portion of the uncharted terrain. In any event, certainly the motivation toward new artforms can be either formal or material. In searching for new formal combinations, we may discover that there are things we can and want to express that werent possible before, or ways of embodying moral attitudes or advancing social claims that hadnt been open to us. On the other hand, in searching for outlets for new expressive impulses, or in striving to put forward novel moral, social, or political perspectives, we may equally well induce new formal combinations in response. There is, in short, a healthy dialectic between formal/structural and material/expressive impulses toward, and justications of, previously nonexistent artforms. New vessels may call forth new contents, and new contents may call forth new vessels capable of holding them. In summary, we have seen that there is a rich domain of nonexistent artforms that can be at least abstractly mapped by reference to the eld of existing artforms. If and are existing arts, then there are those whose objects involve mere juxtaposition of s and s, ones whose objects are structurally intermediary between s and of s, ones whose objects are structural compounds of s and s, ones derived from or by revaluation of what properties are standard, variable, or contra-standard for objects of the art, ones that represent minimal or reductive versions of a given artform, ones that aim at the characteristic effects and experience of a given artform through other means, and ones that derive from a given artform by substitution of the sense modality involved. II THE CASE OF VISUAL MUSIC I turn now to the question of why nonexistent yet possible artforms remain nonexistent, or more exactly, why any number of them that seem appealing

120

Music

in prospect, conceptually unproblematic, and technologically feasible are yet not pursued with any vigor. I think it is clear that there is no informative general answer to be given. Naturally we can always say that either the potential for fruitful development is simply inadequate, or that for historical, cultural, or even political reasons such potential as is present is simply not exploited, but such blanket, almost dormatively virtuous, explanations are singularly unilluminating. Case-by-case investigations thus seem to be what are called for instead. So what I do next is examine the case of visual musicto me the most intriguing of the envisagable yet stubbornly absent artsto see what insight we can get into at least its nonexistence. We may, in the course of this investigation, glean some hints as to what distinguishes those arts, conceivable in the abstract but remaining stubbornly unrealized and unpursued, from those that have a chance of emerging from the shadows of the merely possible. Why, then, are there no successful analogs to pure music in other sensory realms, no enthralling temporal patterns of colors, smells, feels, tastes? Why, in particular, isnt the art of abstract color lmat least in any manifestation known to mecomparable to music in interest and value? Why dont pure patterns of colors in timetemporal successions of huesgrab us as do temporal patterns of pitched sound? Why dont they transgure us the way great music can, or at least captivate us, as music that is merely good does? A short answer would be that there is surely some insufciency in the structures of which abstract color lm is capable, or some insufciency in the human visual system relative to the auditory, or else some mismatch between those structures and the capacities of that visual system. But let us look to more specic explanatory factors, in the hope of giving more body to the schematic diagnosis just tendered. First, there is the fact that musictraditional Western music, at any rate makes use of tones at xed pitch levels, not of sounds drawn at will from just anywhere in the pitch continuum. For abstract color lm to mirror this feature would require at least a prior systematic decision to employ a certain subset of hues, and no others, at precise points along the color spectrum.

The hero of J. K. Huysmanss novel A Rebours, Des Esseintes, has a mouth organ constructed that squirts a sequence of liqueurs on his palate in analogy with the peal of a carillon. That ctional example is about as close as anything known to me gets to what one might call taste music. For the purposes of this essay when I speak of music I have predominantly Western tonal music in mind.

Nonexistent Artforms and the Case of Visual Music

121

Second, there is the crucial role in tonal music of scales, chords, octaves, or more generally, preferred or privileged relations between tones, many of which are grounded in the harmonic series of a vibrating string. These relations, which are the precondition of so many of the fundamental qualities and effects of such music, e.g. tension and release, consonance and dissonance, cadence and closure, tendency and resolution, do not seem to have any obvious parallel in the color realm. The only possible candidate, perhaps, would be relations of complementarity among hues, but that is not nearly as rich as the web of relationships governing how the different notes of a musical scale stand to one another. Third, there is the notion of key (and key feeling), and the consequent possibility of key change (or modulation), which is the basis for so much of the large-scale dialectic of a piece of music, and for a good deal of its emotional power. It is hard to see what might do duty for this in the color realm, unless it would be something like the contrast between the warm colors (red, yellow, orange) and the cool ones (blue, green, violet), and the potential force of shifts from one set to the other. Fourth, phenomenologically speaking the series of pitches employed in tonal music form a space-like dimension, in which pitches seem to exist at denite distances from one another, and in which an experience of motionrather than just one of successionis thus enabled in connection with change of pitch over time. We hear movement in the course of a melody, and not just one note replacing another; we hear the music, as it were, go up and down, and do so with various predicates of pace, strength, intensity, and effort. The series of hues in the spectrum, by contrast, does not appear to form a space-like dimension in the same way; there is no clear impression of directionality in the succession of hues from red to violet, nor is there the sense of an underlying variable property which binds together the different hues, and in respect of which alone they differ from one another. By contrast,
Several of the factors I will cite as militating against the potential of the color realm for musicality are discussed in chapters VII and VIII of Edmund Gurneys magnum opus, The Power of Sound ([1880], New York: Basic Books, 1966). (See my Gurney, Edmund, in Edward Craig (ed.), Routledge Encyclopedia of Philosophy (London: Routledge, 1998).) One development of this idea is Gurneys notion of Ideal Motion, the distinctive characteristic of music as heard. An essay that deals with the essential spatiality and motion of music as experienced is Roger Scruton, Understanding Music, in The Aesthetic Understanding (London: Methuen, 1983), 77100. See also his monumental Aesthetics of Music (Oxford: Oxford University Press, 1997). To spell this out a bit more, I am suggesting that there is no inherent necessity in the order of the colors in the spectrum, though of course that order is physically xed and grounded, and that different orders, within limits, could be allowed to be as acceptable progressionsnot violating

122

Music

when one proceeds from C to C there is an unavoidable impression of rising, and the strong sense of a binding propertyhigh/low-ness of pitchthat is being progressively modulated. For these reasons, it seems, succession of hues is not inevitably, and certainly not readily, perceived as a kind of motion or movement, but only as change simpliciter. If this is so, then there are more weighty consequences as well. Without a sense of dimensionality and directionality in the realm of pitches, i.e. without there being a space of pitches, there could be no experiences of movement through music, or even of melodic shape or contour. But without the possibility of hearing shape and movement in music, we would be barred, a fortiori, from hearing such shape and movement as gesture, and in turn, such gesture as the imagined expression of emotion. Thus, music without its spatial/kinetic aspect would be incapable of embodying expressiveness in the chief way that it seems to do. So if color sequences are unable, through measured succession of colors, to provide the experience of shape and motion in the color realmand not just of rhythmic patternthen their potential for emotional expression even faintly approaching that of music seems dim. In general, the prospects of color sequences ending up meaning anything much, even if they managed to be perceptually absorbing, seem less than bright if we end up being unable to perceive such sequences in a spatially dynamic manner, which further resonance with worldly and human matters would seem to require. If in light of the above we accept that there can be nothing like the spatial and motional possibilities we are familiar with in the auditory realm available to us in the chromatic one, the artistic burden carried by rhythm per sethe temporal patterning of color presentationswill be correspondingly greater. Unfortunately, though, there is reason to believe that our sensitivity to visual temporal rhythm is considerably poorer than our sensitivity to rhythm in the auditory sphere. Small alterations in auditory rhythm are much more readily perceivable by us, it appears, than comparable alterations in visual rhythms.
any phenomenological impression of underlying directionas that which nature has ordained. For instance, the order of hues might have been given as GBIVROY, rather than ROYGBIV, without affronting any indelible sensory intuition. Also adding to the hearability of gesture in musical progression itself is the fact that in traditional performed music a listener has a sense of the instruments and physical motions involved in the rendering of the sounds. It is clear that this, as well as the spatiality of the pitch dimension, funds the hearing of gesture in music on the abstract plane we are here concerned with. For additional discussion, see Sound, Gesture, Space, and the Expression of Emotion in Music, Ch. 5 in this volume. For claritys sake I acknowledge that there is a use of visual rhythm in connection with static visual presentations, for instance, the progression of colors across a canvas, or the repetition

Nonexistent Artforms and the Case of Visual Music

123

Our nely hewn sensitivity to speech, as a patterning of sound over time, in all its inections and nuances, undoubtedly carries over to hearing and discriminating temporal relationships in music. There is nothing, however, in the natural repertoire of the developing human that calls upon ne sensitivity to timed sequences of light. This bears directly on the critical issue of recognizable basic units in our proposed visual music. Can the realm of visual music sustain something analogous to the reidentiable melodic-rhythmic unitsmotives, phrases, gureswithout which musical experience in the ordinary sense would be unthinkable? To what degree could such visual motives or phrases be remembered as such and recognized, if only subconsciously, upon reoccurrence? How ne may our discriminations for such temporal visual motives be? How easily would we tell apart two color motives which are the same except for a small difference in internal rhythm, e.g. a ratio of two-to-one versus three-to-onea difference, however, which is salient to the ear? I suspect that our discriminations of such motives would be much coarser, and our retention of them less reliable, than of motives in music. Curiously enough, this might even have something to do with the fact that musical phrases are something we have the capacity to give back or reproduce by singing, whistling, hummingwhereas we obviously have no natural capacity to generate or mimic measured sequences of color or light; the capacity to reproduce may facilitate, or even be part and parcel of, our grasp and retention of temporal sonic form. There is, next, the question of transposition of such visual motives or phrases to other hue levels. Motives and phrases in music can be identied when transposed to another pitch level or harmonic setting, and this possibility is clearly crucial for most kinds of musical development (most obviously, for melodic sequencing). It is not clear, however, whether a temporal visual motivea sequence of colors set to a rhythmcould have that kind of reidentiability. To compound matters, the color spectrum is not extensive enough to allow for much moving around of a color phrase in hopes of reidentication at a lower (redder) or higher (violeter) position; if we think of it, charitably, as having two octaves, red-to-green and green-to-violet,
of ornamental gures on an architectural facade, but when I speak of visual rhythms here I mean temporal visual rhythms, that is, patterns of durations of color presentations in time. The fact that we are well adapted to picking out the motions of persisting objects, such as hands or birds, through space does not automatically argue for a capacity to grasp the temporal patterns of changing visual presentations at xed positions in space.

124

Music

then this allows at a maximum twenty or so hue levels at which such a phrase might be perceived, as opposed to around a hundred for the standard compass of a piano, string quartet, or symphony orchestra. But it is quixotic to think that such a phrase would, in virtue of the relational invariants it embodied, be reidentied at any other hue level; we just dont seem to be sensitive to, and abstractive of, rhythmic color sequences in the right way. The limited extent of the color spectrum, in terms of what seem signicant differences in hue across its span, would also, obviously, impact negatively on the number of theoretically possible color phrases or motives. I mention just three more factors which seem to militate further against the possibility of a successful visual music. First, the experience of duration, or sustainedness, of merely a held tone, is more vivid than that we receive from a held visual presentationthe former is perceived as a happening or event, even while in some sense unchanging, while the latter strikes us as a non-happeningless of an event than a fact or condition. Thus, even where there is no progression in a straightforward sense, a steady musical state or situation will always seem more alive to us than a corresponding visual one. Second, the ease and naturalness of combining separate lines individual progressions of tonesin the auditory realm, yielding polyphony or counterpoint, is in strong contrast with the articiality of an analogous procedure in (say) an abstract color lm. Two or more tones or progressions of tones can be heard simultaneously and distinctly in the same region of pitch space (or register), whereas two or more colors cant occupy the same region in two-dimensional space, or in ones visual eld, at the same time. Thus, to realize a multiplicity of chromatic lines in an abstract color lm would necessarily require their presentation side-by-side, or one atop the other, or in some other spatial conguration, it being arbitrary which was elected, and it being far from clear that such independent and spatially separate streams could be integrated by the eye into anything like the blending and interpenetration of lines we experience in the sonic domain. Third and nally, counting once more against the possibility of a comparably gripping visual music, is the close connection of real, audible music with primitive emotional utterance. The prominent place of sound, especially vocal sound (yells, roars, laughs, grunts, squeals), in the expression of basic emotions and the communication of simple feelings and attitudes, is something that music, though a highly elaborate and rareed use of sound,
Differences of timbre obviously have a lot to do with enabling this.

Nonexistent Artforms and the Case of Visual Music

125

undoubtedly taps into. There seems nothing comparably visceral in human behavior for temporal patterns of color to resonate with. This might partly explain why color sequences, however diverting, seem relatively inert, without life impact, by comparison with sound sequences. The strikes against visual music just reviewed, uncovered in the course of our abstract examination of the question, seem a plausible explanation of why visual music is not a thriving enterprise. But it would be premature to foreclose on visual music completely, simply because of the possible handicaps we can expect it to labor under. For only empirical experimentation, after all, carried out with artistic imagination, will really show conclusively what can, or cannot, be achieved in this direction. Since I am not equipped to conduct such experimentation, I will just conclude with a few suggestions guesses reallyas to how ventures in abstract color lm might best proceed, with the aim of achieving something of musical effect. First, it seems clear that a limited set of positions along the spectrum of hues should be selected and adhered to, as a background matrix, analogous to the domain of xed pitches out of which tonal music is made. Second, the kinds of rhythms employed should be fairly simple, and the differentiation between different rhythmic gures fairly blatant. Third, moderate tempos are probably advisable, if there is to be much hope of xing on isolatable temporal color patterns as distinctive and memorable, in partial analogy with motives and melodies. Fourth, efforts should be made to exploit the relations of complementarity among colors, as the only likely analog of those special relations of kinship and remoteness, attraction and repulsion, among pitchesthe fth, the third, the leading tone, the octavewhich have so great a role in giving tonal music its teleological aspect, its tendency and directedness. Fifth, the potential of split-screen, in the attempt to generate something like polyphonic and interactive effects between two color streams, or even just in emulation of the effect of tune-and-accompaniment, should not be ignored, though the problem of the inherent articiality of such a device should be somehow faced head-on. Sixth, incremental use could certainly be made of variation in parameters which I have assumed held xed until now, namely those of brightness and saturation, as opposed to exact hue, of each color presentation or visual note. If variation in hue serves, in color music, as the only possible analog of movement in pitch, then variation in brightness and variation in saturation might, suitably managed, come to seem analogous to dynamic (loudness) and timbral (tone color) variation in the musical realm.

126

Music

Finally, if all choices in the dimensions just reviewed were to prove in vain, so far as generating a sequence of color presentations that would have the interest, appeal, and expressiveness of even the simplest music was concerned, one might then open the door to manipulations of size, shape, and two-dimensional patterning of color presentations as wellexpanding and contracting color elds, swirling and intermingling chromatic constellations, many-hued patches dancing from left to right, and corner to corner, etc. But note that such visual music would no longer be a strict analog of monophonic or simple polyphonic instrumental music, whose basic material is just sequences of pitches-with-durations. The sort of visual music we have just envisaged, which is in effect a kind of abstract painting in time, might well be expected to have more power and impact than the strict, spartan music of colors whose prospects have been my proper target in this section; for one, it would partake of all sorts of representational resonances which forms in space, and especially changing ones, possess. Thus, whatever greater success such expanded visual music may achieve does not thereby vindicate visual music in our original conception. Moreover, it is far from clear, from extant examples of the genre, whether spatially unrestricted sorts of abstract color lm, even if aesthetically more engaging than what can be achieved by pure chromatic sequences per se, are more engaging in virtue of affording an experience closer to that which musicliteral musicaffords than is open to us through visual music narrowly conceived. III AFTERTHOUGHTS Not long after completing this essay I discovered some reections of Ernst Gombrich on the subject of visual or color music. He notes that attempts to build a color piano may go back to the sixteenth-century painter Arcimboldo, that the idea was revived in the eighteenth century in the form of
The lms of the Canadian lmmaker Norman McLaren, which might be thought to be an actual counterexample to my pessimistic diagnosis of the possibility of pure color music in this section, are, so far as I can tell, of this impure type. That is to say, they make use of resources, particularly spatial ones, which go far beyond the idea of a rhythmic succession of simple hue presentations, or even two or three such successions simultaneously. In addition, a number of them even have soundtracks, which work in counterpoint with the complex color images constituting the purely visual component of these lms. See Gombrich, The Sense of Order. (Ithaca: Cornell University Press, 1979), epilogue.

Nonexistent Artforms and the Case of Visual Music

127

a keyboard instrument that produced colored ribbons as well as notes with every touch, and that Sir David Brewster, the inventor of the kaleidoscope, predicted in 1819 a great future for an art wherein combinations of forms and colors may be made to succeed each other in such a manner as to excite sentiments and ideas with as much vivacity as those which are excited by musical composition. Gombrich wonders, as I have done, whether color music can be viable in the absence of something which, in chromatic terms, would be a functional equivalent of the tonal system:
Could a visual music be developed which would not only offer a succession of moods of pleasing kaleidoscopic patterns, but set up a eld of force which could lead to that interplay of expectation and fullment, of tension and resolution, which is the stuff of music? He would be a bold man who would dare to predict that such an experiment could never succeed.

I am not that bold man, but on the other hand, neither would I venture the opposite prediction, that such experiments must inevitably succeed. Another writer who has treated this topic, I belatedly discovered, is Peter Kivy. Speaking of his sense of wonder at the very existence of music alone, Kivy suggests that it begins with what seems to me to be the genuine, if insoluble, mystery of why we have pure music at all, and why, since we do, we dont have music for our other sense modalities. Nevertheless, Kivy proposes an evolutionary explanation for why humans have not developed visual music, in particular. It is, according to Kivy, because we have evolved hardwired to see defensively, and so to unstoppably place a representational (or realistic) interpretation on visual perceptions, thus foreclosing on a temporally extended art of purely visual phenomena. This explanation, even on its own terms, seems to me to fail. For in fact we do not, faced with abstract color sequences, have a noticeable inclination to perceive them representationally or realistically. They remain for us what they at rst seem to be, namely, temporally evolving chromatic patterns. The explanation for their relative non-involvingness has to be sought elsewhere than in their putatively putting us frustratingly at cross-purposes with what evolution has disposed us to do with visual data.
Ibid. 287. Ibid. 305. See Kivy, Music Alone. (Ithaca: Cornell University Press, 1990), ch. 1. Ibid. 12. Ibid. 4.

128

Music

More importantly, even were this explanation on the mark, it would remain too external. Surely we need to appeal to something about the intrinsic nature of sounds as opposed to colors, or at least to salient features of familiar musical systems that have been constituted on the basis of the inherent potentialities of the sound medium. It is that sort of structural, and thus largely internal, explanation that has been offered here for the persisting absence of visual music.

8
Music as Narrative and Music as Drama
I INT RODUCT ION To what extent can instrumental music be viewed as narrative in character, or understood as involving some sort of narrative? That is the central question of this essay, but a number of subsidiary questions will also guide my reections. I will be interested in the varying potential for narrative construal of different forms of music, some of which may invite, and some of which may resist, such construal. I will be interested in what musical narrative, when present, might be a narrative of. I will be interested in whether musical narratives can possess certain of the features of standard narratives, such as literary or cinematic ones. But I will nally also explore, at some length, the appeal of an alternative construal of music: as dramatic, rather than narrative, in nature. It is salutary to ask ourselves, at the outset of an inquiry like this, exactly how often pure instrumental music impresses itself upon us as needing to be construed narratively in order to be understood. The answer, it seems to me, is not very often. This is of course not to deny that in appreciating such music we are made to focus on sequence and progression, from note to note, phrase to phrase, and section to section. For that is what following music by ear largely consists in. But it is a large step from that to the claim that music is, in its sequence and progression, narrating a story of some kind as it unfolds. Still, the idea that instrumental music, and especially the extended musical essays of composers of the eighteenth, nineteenth, and twentieth centuries, might be understood as narratives is a staple of humanist music criticism, as exemplied by Donald Tovey, George Bernard Shaw, Leonard Bernstein, Antony Hopkins, Charles Rosen, Andrew Porter, Alex Ross, and others. It is, after all, not surprising that music, as an intentionally arranged, temporally
First published in Mind & Language 19 (2004): 42841.

130

Music

extended sequence of sounds, one that often displays a character of utterance, is readily thought of as recounting something or other, and likely something that is itself temporally extended, such as a sequence of actions, events, or mental states.

II THE CONCEPT OF NARRATIVE How should narrative be characterized? For our purposes, the more simply the better. One recent writer suggests that a genuine narrative requires the representation of a minimum of two events and some indication of the ordering in time of the events depicted. If so, then there are three crucial elements to narrative: representation, events, and temporal relations. Another recent writer has proposed that a narrative must, in addition, indicate causal relations obtaining among represented events: the basis of the narrative connection is that earlier events and/or states of affairs are at least causally necessary conditions, or contributions thereto, for the occurrence of later events in the relevant stories. Applying this to music, then, if music is to be narrative (a) it must represent; (b) it must represent events or states of affairs; and (c) it must represent temporal and/or causal relations among those events or states of affairs. Accepting these as the minimal features of narrative, the task would then be to assess whether any pure instrumental musicthat is, music without program or textin fact displays them. The prospects do not seem bright. The third condition, in particular, seems especially hard to meet, as it looks as if it would require the sort of temporal and singular referential devices that language, but not music, possesses. Yet that is where the distinctive feature of narration, as opposed to non-narrative representation, would seem to lie.
George Wilson, Narrative, in Jerrold Levinson (ed.), Oxford Handbook of Aesthetics (Oxford: Oxford University Press, 2003), 392407, at p. 393. Nol Carroll, On the Narrative Connection, in Beyond Aesthetics (Cambridge: Cambridge e University Press, 2001), 11833 at p. 133. An alternative tack would be to locate the specicity of narrative representation in there being a discernible narrator internal to the representation, who tells or recounts the events and relations in question. But since a number of philosophers have argued for the possibility of narratorless narratives (see Kendall Walton, Mimesis as Make-Believe, (Cambridge, MA: Harvard University Press, 1990); Gregory Currie, The Nature of Fiction (Cambridge: Cambridge University Press, 1990); and George Wilson, Le Grand Imagier Steps Out: The Primitive Basis of Film Narration, Philosophical Topics 25 (1998): 295318), it is not a tack to be taken lightly.

Music as Narrative and Music as Drama

131

III OBJECTS AND CONTENTS OF MUSICAL NARRATIVE If musical narratives narrate events, a question arises immediately as to whether those events are musical or non-musical ones. In other words, one possibility is that music somehow tells a tale of musical events, such as the inversion of a motive, or the arrival of a cadence, or a modulation from B-at to E-at. Another is that music somehow tells a tale of non-musical events. Presumably the narratives of interest in music are of the second kind, for those of the rst kind, to the extent they could be made out, would seem redundant. Since the sequence of musical events is directly present and immediately heard, what would be the point of its being narrated as well? In order to locate more plausible objects of musical narration, it will be helpful to offer a brief sketch of the musical apprehension of the ordinary, though musically sensitive and practiced, listener, at least as other than a strict formalist would conceive it. The experience of apprehension can be seen as comprising different stages or levels, though it is a mistake to suppose that these are either entirely separable or clearly sequential. In any case, rst there is the level at which the most elemental properties of tones, such as pitches, durations, and timbres, are perceived. Then, following immediately upon that, is the level at which rhythms, motives, phrases, melodies, and harmonies, the fundamental building blocks of music, are perceived; note that that level involves the hearing of motion in music, rather than mere succession or alteration. Those two, perhaps not fully distinguishable, stages might be labeled the congurational level of musical apprehension. Next there is the level of gesture or action, that at which one hears the music doing something, something it is not literally doing. Note that the sense of the real gestures of the performers who are performing the music enters into the gesturecall that musical gesture that one hears the music to be engaged in, but the musical gesture and the conjectured or imagined performing gesture are not the same. We can label that the gestural (or actional) level of musical apprehension. Next there is the level of states of mind heard behind the gestures or actions perceived at the preceding level of which those gestures or actions are the expression; that is the expressive level of musical apprehension. Assuming this rough picture of musical experience, what are then possible narrative objects of music? They would seem, at a minimum, to be these: (a) gestures, (b) actions, (c) expressions, (d) mental states. As for the content

132

Music

of a musical narrative, it would presumably be some sort of sequence of the preceding. Suppose for the sake of argument we focus on option (d), and construe the narrative content of music to be a sequence of mental states. A typical specimen would then go something like this: First S1, then S2, then S3 . . . , where the Sns are mental states, plus whatever relations among the Sns might be implied, such as S1 evolves into S2, S2 results from S1, S2 is a reaction to S1, and the like. So, does expressively varied music indeed relate such a narrative? Do we, at any rate, hear such a narrative in such music? Agnosticism about that seems highly warranted, to say the least. We should observe, in addition, that certain features central to standard narratives of a literary or cinematic sort, even if not accounted essential to narrative, seem virtually impossible to locate in instrumental music. These include the capacity to predicate of a subject, the capacity for reective selfcommentary, and the capacity to clearly signal pastness or futurity. To those three features difcult to imagine pure music exhibiting, we can add others, such as narrative voice, narrative point of view, and narrative true-to-lifeness.

IV MUSICAL REPRESENTAT ION We need to examine more closely the idea of representation presupposed in the foregoing discussion. Taking representation to be a sine qua non of narration, if instrumental music is to count as narrative we will have to establish, rst, that such music does indeed represent, second, that it represents nonmusical events, and third, that such music represents in a narrative mode, that is, by somehow conveying or recounting temporal or causal relations among those non-musical events.
See Fred Maus, Narratology, Narrativity, in Stanley Sadie (ed.), New Grove Dictionary of Music (Oxford: Oxford University Press, 2003). Indication of pastness may not, for all that, be beyond the powers of music to achieve. The most promising case is probably the emergence of nostalgia in music, whereby a passage of music appears to relate regretfully to an earlier one, or to the sentiment or action with which the earlier passage was associated. This is not to deny that point of view, one attributable to the implied composer, plays a role in music involving humor, parody, allusion, and the like. Examples include Mozarts Musical Joke, Bart ks Sixth Quartet, Haydns Joke Quartet, and parts of Ivess symphonies. Such cases o typically involve implicit commentary on something musically referenced then and there. Despite such examples, it is hard to see how there could be scope for the operation in music of full-edged narrative point of view. As for true-to-lifeness in music, for an attempt to theorize that in regard to facts concerning the realm of emotions, see my Truth in Music (1982), in Music, Art, and Metaphysics (Ithaca: Cornell University Press, 1990). Though I do not entirely repudiate it, I am now less sanguine about the approach taken there.

Music as Narrative and Music as Drama

133

But with the very rst task we confront an obstacle. Even if expressive music, according to the sketch of musical experience offered above, readily induces appropriately backgrounded listeners to hear gesture and expression in it, is that quite enough to say that the music represents those things? Unfortunately not. For artistic representation, as exemplied most clearly by pictorial representation, is a strongly intentional notion. Thus, for a picture to represent an oak tree an oak tree must be seeable in it, to be sure, but the maker of the picture must also actually have intended such seeing-in. Artistic expressiveness, on the other hand, is not a strongly intentional notion, nor is that of gestural content; a passage of music might be expressive of an emotion or might embody musical gestures, without its composer having intended the hearing of those gestures or that emotion in it, and in some cases, without the composer even having foreseen or anticipated such hearing. In light of that, it may be that the relation between music and the gestures and expressions it induces us to hear in it is more properly one of suggestion, rather than representation. Be that as it may, I will put aside this reservation for present purposes, and continue to speak of expressive music as representing the gestures, actions, and expressions that it perhaps strictly speaking only suggests. That will allow the question of musics narrativity to remain open.

V M USIC AS EXTERNALLY NARRATIVE VS. MUSIC AS INTERNALLY NARRATIVE A distinction regarding musics narrativity that will prove to be of some use is that between music as externally narrativeas something that is being told, by the composerand music as internally narrativeas something that is telling of something else. In the rst case, the sequence of musical events is the story; in the second case, the story, if any, is what the musical events are about. Of course, where music can be made out to be both externally and internally narrative, the responsibility for the internal narrative ultimately rests with the composer as well, the composer being the teller, in the rst instance, of the sequence of musical events, and in the second instance, though sometimes unknowingly, of the sequence of non-musical events the musical sequence represents. In connection with musics external narrativitythat is, the idea that the events of which the music consists are themselves being narratedthere is also the role of the performer to consider. The performer might be thought

134

Music

of as occurrently narrating the events that the composer has only narrated in a standing fashion, or alternatively, as the only proper narrator of musical events on any occasion of performance, the composer then not being a narrator of any sort, but only the designator in the abstract of events for narration. Having for the moment gone as far as I can with the idea of music as narrative, I want now to consider an alternative idea, that of music as drama. VI MAUS AND NEWCOMB ON DRAMA IN MUSIC In this section I summarize two analyses of musical compositions by philosophically informed musicologists that can be seen as recommending a dramatic, as opposed to narrative, model of the events musical compositions appear to image forth. The rst is due to Fred Maus, and concerns the opening movement of Beethovens Quartet in F minor, Op. 95; the second is due to Anthony Newcomb, and concerns the scherzo of Mahlers Ninth Symphony. Mauss analysis, which focuses rather minutely on the arresting rst eighteen measures of Beethovens quartet, identifying in it individual gestures standing in certain relations to one another, issues in the following general conclusion:
It would be natural to call the quartet a conspicuously dramatic composition. The analysis makes the sense of drama concrete by narrating a succession of dramatic actions: an abrupt, inconclusive outburst; a second outburst in response, abrupt and coarse in its attempt to compensate for the rst; then a response to the rst two actions, calmer and more careful, in many ways more satisfactory.

Maus notes that his analytical description of the passage explains events by regarding them as actions, and by venturing motivations for those actions. But to whom, Maus asks, are those actions and motivations being ascribed? Neither the composer nor the performers, it seems, for though they are the authors of certain music-related actions, namely, composing and performing, they are not comfortably thought of as the authors of the actions heard in the music, actions such as asserting, objecting, responding. Rather, such actions, and associated motivations, are to be ascribed to a persona, or personlike
Maus, Music as Drama, in Jenefer Robinson (ed.), Music and Meaning (Ithaca: Cornell University Press, 1997), 10530, at pp. 11819 (originally in Music Theory Spectrum (1988)). Note that although Maus invokes narrating in this quotation, the narrating is being done by the analysis, not by the music being analyzed.

Music as Narrative and Music as Drama

135

agent, who is imagined as present, performing those actions in real time. In listening to a piece, Maus observes, it is as though one follows a series of actions that are performed now, before ones ears, not as though one merely learns of what someone did years ago. Maus proposes that a piece of instrumental music such as the quartet movement he analyzes be assimilated to a play, and thus seen to have an essentially dramatic character, rather than assimilated to a narrative form like the novel. Plays have a number of salient features which music can share. They present a series of actions, which are the actions of ctional characters or personae, which actions are experienced as occurring as they are perceived, and which form a plot, or at least, make some kind of sense as a whole. Of course the actions heard in music are not as concrete or detailed as the actions of a stage play, which more nearly approximate those of life. The agents, objects, and motivations of musically embodied actions remain much more indeterminate. Music is, as Maus suggests, a kind of drama that lacks determinate characters. This difference in degree of indeterminacy of the characters or personae involved in music and in theater should not be thought to undermine the validity of the analogy. For, after all, playwrights such as Strindberg and Beckett have created stage plays, such as Endgame and The Ghost Sonata, whose personae are almost as indeterminate as those to be heard in the expressive instrumental music with which we are concerned. Anthony Newcomb, for his part, maintains that music is heard is a reenactment of a complex pattern of intentional human action, and his analysis of Mahlers movement is designed to illustrate that. According to Newcomb, the imagination of agency in music, in schematic form, goes as follows:
[rst] the selection of [or focusing on] musical attributes . . . the interpretation of these musical attributes as attributes of human character or behavior . . . the combination of these human attributes in various congurations as possible or plausible human agencies . . . [and nally] the understanding of . . . these ctional agencies as relevant in the unfolding of a plausible chain of human actions and events.
Maus, Music as Drama, 121. Ibid. 128. Anthony Newcomb, Action and Agency in Mahlers Ninth Symphony, Second Movement, in Jenefer Robinson (ed.), Music and Meaning (Ithaca: Cornell University Press, 1997), 13153, p. 131. Newcomb hastens to add this qualication: It is important to realize that in music, as in other arts . . . aspects of agency are not continuously displayed . . . Even the most expressive music . . . at times simply swirls or dreams or chugs along in its decorative function (ibid. 133).

136

Music

Newcomb locates in Mahlers scherzo three dances of different character, which are introduced individually, and then interwoven as the movement proceeds. The rst is a medium Lndler (A), the second a fast Waltz (B), and a the third a slow Lndler (C). When the second dance rst appears, it suggests a an agency distinct from that of the rst dance, but there are two reasonable possibilities for what that agency is: either a second, entirely distinct persona, or else an element within the personality of the rst persona. The culturalhistorical context of early twentieth-century Vienna, awash in the ideas of Schopenhauer and Freud, and sympathetic to the notion of hidden sides of the human psyche, favors the latter way of reading the contrast between the agency embodied in dance A and that embodied in dance B. Overall, Newcomb suggests, the movement offers a picture of a clumsy and coarse rustic personality swept away by a sophisticated and condent urban one, followed in due course by a sober and reective personality, associated with dance C, who serves to rein in the second and perhaps restore, in some measure, the honor of the rst. In the remainder of this essay I attempt to bring into relief further aspects of the dramatic and narrative models of musical content, and to weigh the respective merits of those models in regard to music of different kinds. VII MUSIC AS DRAMATIC VS. MUSIC AS NARRATIVE Music is expressive, I maintain, when it prompts us to hear the music as animated by agency of a certain sort, more specically, when it induces us to hear the music as the expressing of a mental state, or perhaps equivalently, when it induces us to imagine a persona expressing a mental state through the vehicle of music. I call this the hearability-as-expression view of musical expressiveness. But is a sequence of passages that are expressive of a sequence of states of mind thereby an emotional narrative? Only, it seems, if we have the sense that the rst sequence involves acts of relating or telling, ones attributable to an agent who stands apart from the imagined agent or agents who are the subjects of the mental states and acts of expression that constitute the musics expressive substance. Otherwise, as the analyses of Maus and Newcomb
See Musical Expressiveness as Hearability-as-Expression, in Matthew Kieran (ed.), Contemporary Debates in Aesthetics and Philosophy of Art (Oxford: Blackwell, 2005), and Ch. 6 in this volume. That is, in the terms introduced earlier, if the music appears internally, and not just externally, narrative.

Music as Narrative and Music as Drama

137

suggest, we would seem more justied in construing the music according to a dramatic model, on which personae and their expressive actions appear directly on the musical stage, than according to a narrative model, on which those personae and expressive actions are instead being recounted to us by a narrative agent. The locus classicus of the distinction between the narrative and the dramatic is Aristotles Poetics, the rst serious attempt in Western thought to offer a taxonomy of the arts. Aristotle invokes the distinction to explain how epic and tragedy are distinct. He notes that though they have the same objects of imitation, namely, noble human beings in serious situations, and roughly the same means of imitation, namely, persons uttering words, they differ nonetheless in the manner of imitation: whereas epic speaks of the events with which the epic is concerned, tragedy presents those events directly. Alternatively, the epic reciter describes the events of the story, while tragic actors enact those very events. Again, on the view of musical expressiveness I hold, we hear expressive music precisely as if it were an expression of mental states. More generally, we hear such music, and much not specically expressive music, as imbued with action. That is, we standardly hear music as acting in ways it is not literally acting, or doing things it is not literally doing, or gesturing in ways it is not literally gesturing. In the case of expressive music, the actions are, naturally, expressive actions, but in other cases they are actions of other sorts. From such a perspective, clearly, the most natural way to view instrumental music is as dramatic, that is, as offering a sequence of actions to be directly imagined, rather than as narrative, that is, as offering a sequential relating of such actions. VIII CONDITIONS FAVORING NARRATIVE CONSTRUAL OF M USIC When are we inclined to regard the actions that we hear in music as directly present to us, enacted by personae as we listen, and when as matters that are not directly present, but instead represented in a narrative conveyed to us by a narrating agency, whether the composer, the performer, or a narrator internal to the music? A most difcult question. Let me, then, pose a simpler version of it. When do we have the sense that unadorned instrumental music is relating a story to us, that such music is, in terms invoked earlier, narrating, and not just being narrated?

138

Music

I suggest, rst, that the music must have a marked character of utterance, of seeming to speak, if that sense is to emerge. And not all music displays that. But second, there must also be a character more specically of storytelling. What I mean by that is a measured, deliberate, reective character, such as is conjured by phrases like once upon a time or long ago and far away. Music of storytelling character recruits features traditionally associated with the storytelling mode of discourse, one we think of perhaps above all as unhurried. Thus slowish tempo, relaxed rhythm, and restrained dynamic are among the musical features conducive to the emergence of a storytelling character. An example of music with something like this storytelling character would be the opening of Smetanas Ma Vlast. This is, admittedly, an explicitly programmatic work, but non-programmatic examples may also be found. The openings of Bruckners Seventh Symphony, Francks Symphony in D minor, Mendelssohns Reformation Symphony, or Dvoraks New World Symphony will perhaps serve. Another generalization we might try on for size is this: The more music strikes us as a direct utterance from the composer, the more we are likely to construe it as narrative, as in effect the composer testifying to or recounting something in sound; and the more music strikes us as constituting a world of its own, in which events occur independently of a guiding force, the more we are likely to construe it as dramatic, that is, as involving agencies that appear, interact, and depart before our ears. Or again: the more the agents imagined in connection with expressive music seem autonomous or self-directed, the more apt is a dramatic construal of the music; and conversely, the more the agents imagined seem framed or subject to outside control, the more a narrative construal of the music recommends itself. In any event, plainly not all instrumental music displays narrative character or lends itself to narrative interpretation. The extent to which it does seems to depend to a fair degree on the genre of music involved. With minuets, scherzos, toccatas, etudes, canons, variations, and the like, the form of such pieces dictates certain structural repetitions or certain kinds of musical ller which tend to block narrative suggestion or undercut narrative momentum, whereby such pieces come close to tting Peter Kivys characterization of instrumental music as sonic wallpaper. With sonata movements, by contrast, internal narrativity has more purchase, sonata form
Peter Kivy, The Fine Art of Repetition, in The Fine Art of Repetition (Cambridge: Cambridge University Press, 1993), 359.

Music as Narrative and Music as Drama

139

being inherently progressive and developmental, and so more susceptible to narrative construal. Still, the most natural construal of a sonata movement is as dramatic rather than narrative. Sonata movements generally come across like a stage on which actors appear, directly express their emotions, and move on. (Alternatively, a single actor may be thought present, who passes through a sequence of states somewhat in the manner of a soliloquy.) In the opening movement of CharlesValentin Alkans little-heard Symphony for Piano, for example, one readily hears a protagonist, embodied most clearly in a particular ve-note motive, express its peculiar mixture of yearning and suspicion. But one does not seem to hear a voice that tells of a third party. One does not have the sense of a mediator between oneself and the expressive gestures that one perceives in the music. Perhaps it would be best to admit that music can generally be heard as either strictly narrative or strictly dramatic, even if most music lends itself more readily to the latter. When music is regarded as the utterance or voice of the composer, then construing it as a narrative, one whose content consists of the gestures, actions, or emotions of the composers alter ego, is natural enough. On that model, the composer is analogous to a lyric poet, and the performer to a reciter or rhapsode. When music is regarded rather as the organized product of creative activity offered for our engagement, then construing it as a drama, one whose content consists in the gestures, actions, or emotions of various shifting personae, is arguably more natural. On that model, the composer is analogous to a playwright, and the performer to a director or producer. In the last analysis, construing music either narratively or dramatically, and not simply expressively, might best be regarded as an appreciative option, not something correct appreciation absolutely enjoins. Even when the option is exercised it is often irresolvable whether the music in question is better construed one way or the other. One likely source of this irresolvability as between narrative and dramatic construals of music is the fact that there are so many different candidates in the domain of articulate actions for what a given passage of music might be heard as ctively engaged in: monologue,
A similar claim could be made for fugue, which also avoids strict repetition, and which one could hear as relating a story consisting in the vicissitudes of its theme, and for jazz improvising on standard tunes, given that such improvising is often understood as at least in part a kind of commentary on the tune or its associated chord changes. (For interesting reections on the narrative dimension of jazz improvisation, see the third part of Jon Elster, Ulysses Unbound (Cambridge: Cambridge University Press, 2000).)

140

Music

dialogue, recitation, aside, soliloquy, meditation, diatribe, ballad, chronicle, report, confession, exhortation, peroration, elegy, admonition, and so on, some of which are dramatic actions, and some of which are narrative actions. IX NARROW AND BROAD SENSES OF NARRATIVE IN MUSIC Experienced from within, so to speak, a sonata movement strikes us as a drama, with actors corresponding to themes or motives. This is the way sonatas are usually experienced, as illustrated by the Alkan example above. Yet experienced from without, as the deliberate product of a creator, a sonata movement may take on the aspect of a narrative being related by the composer through the musical drama more immediately apprehended in listening. Thus, though not narrative in the narrow sense in which it contrasts with dramatic, the music of a sonata can be considered narrative in a broad sense, one in which narration is effected by means of drama. Consider an army commander narrating a past battle, using actors or puppets with recognizable identities or powers. This looks to be analogous to the stage-managing a composer engages in regarding his musical material and the personal agents such material gives rise to for imaginative hearing. What this amounts to, then, is the conveying of a narrative by the creating of a drama. Of course in this broad sense of narrative playwrights, who directly fashion dramas for enactment, are also engaged in narrative, even more clearly than are composers. Macbeth, for instance, certainly conveys a story, and there seems no objection to thinking of Shakespeare as, in some sense, the conveyer of that story. But neither in theatre nor in music does this amount to narrating in the stricter sense that applies to, say, literary ction, in which a narrator can be identied in the ction, in which there is a discernible narrative voice, in which narrative point of view is robustly present, and in which there is a clear distinction between narrative discourse and narrated story. X M USICAL PE RSONAE What disposes us to think of a musical persona heard in one stretch of music or portion of the musical fabric as the same as the musical persona heard in another such stretch or portion? It is very hard to generalize here, but perhaps at least a contiguous series of musical gestures of similar character will be interpreted, ceteris paribus, as the gestures of a single, continuing persona.

Music as Narrative and Music as Drama

141

Yet there is very likely ineliminable indeterminacy in such matters. Speaking of the dramatic personae hearable in the opening of Beethovens Op. 95, Fred Maus observes that:
. . . nowhere is it clear whether the response to the rst outburst is made by the same agent or agents. If the continuity in the performing forces suggests continuity of the musical agency, the registral discontinuity and utterly different treatment of the ensemble may suggest that a distinct agent or agents have entered to respond . . . . The actions that a listener follows in listening to the Beethoven passage do not belong to determinately distinct agents. More precisely, as the listener discerns actions and explains them by psychological states, various discriminations of agents will seem appropriate, but never with a determinacy that rules out other interpretations.

Nevertheless, music for a single performer is more often than not heard as communicating a slice of the psychological life of a persisting persona. Even where there is a clear melody and accompaniment structure, or a contrapuntal texture with several distinct lines, one generally has the impression of a single expressive agency, at the service of which are the various distinguishable musical components. By contrast, in certain genres, such as that of the concerto, with its manifest oppositional structure, playing off soloist against orchestra, the presumption of single expressive agency is just as clearly overridden. And duo sonatas and trios may also often invite the kind of hearing normative for concertos, in which different instruments or performing forces are heard as the vehicles of different personae. XI A SONATA OF SCHUBERT I end with a musical example, the rst movement of Schuberts Piano Sonata in a minor, D. 845. It is instructive to trace the experience of gesture, expression, and action in this stretch of music, one that, like most movements in sonata form, sustains hearing in dramatic mode rather better than it does hearing in narrative mode. A pragmatic difculty looms, though, in the effort to convince an audience of this through description of what appears to be going on in the music. For in doing so one is inevitably involved in narrating the succession of agents and actions that one hears. Still, the fact that in making a brief for the dramatic content of the music one perforce narrates ones experience of that content does not turn the music itself into a narrative rather than a kind of drama.
Maus, Music as Drama, 123. See n. 6 for the same caution.

142

Music

Three contrasting themes or theme groupscall them A, B, and Ccan be discerned in this movement, with which three personae or personality aspects can be readily associated. The opening motive, A, is of a grim, inward character, while the next, B, is of restless character overall, consisting of three sub-motives of respectively rocking, owing, and declamatory nature. The next motive, C, has a martial and strutting character, which is then succeeded by a fusion of C and B, which, though still vigorous, assumes a more lyrical character than anything that has come before, and which suggests in its fusion of motives a kind of rapprochement between them. Eventually C drops out and the owing, arpeggiated sub-motive of B predominates, until the return of A in a key a third above the tonic. From there until the end of the exposition, another twenty or so measures, there is an alternation between A and C which has the air of a dialogue, one concluded gently by C. The beginning of the development section sees an alternation of A with itself, distinct contrasts of register and dynamic demarcating the two guises of A, giving the passage the sense of a soliloquy in which different sides of a question are being weighed. This is followed, after two held chords serving as a transition, by a section of more pronounced fantasy character: here the persona of A, though shadowed by the rocking, syncopated sub-motive of B in the bass, takes wing, ventures into unknown regions, and beginning around measure 120, wrestles with doubts, the upshot of which is by no means clear. Dark thoughts continue to accumulate, and the sense of crisis is accentuated by the fragmentation of A, reduced to its last four descending notes, which plunge somewhat desperately, again and again. This eventually subsides, allowing A to reassert itself once more in full, though now uncertain, spent of energy, and winding in on itself in a series of remote and inhospitable keys. There I shall leave our motives and their associated personae, with the return to the initial order, in the recapitulation, still some sixty measures off. In the foregoing narrative of a listeners experienceor at any rate, this listeners experienceof this music, I have mixed technical, expressive, and agential vocabulary. But the technical bits in the narrative serve only to pinpoint what is going on in the music on the other two levels, those of expression and agency. And what is going on, I would hope your experience conrms, is not so much doings that occurred in another place and time, of which one is receiving a report, but a drama of events happening here and now, with indenite personae which are the shifting loci of the emotions and actions encountered throughout.

9
Film Music and Narrative Agency
I In this essay I address certain issues about paradigmatic lm music, that is, the music that is often heard in the course of a ction lm but that does not originate in or issue from the ctional world revealed on screen. What most interests me is the question, which confronts every lmgoer at some level, and to which he or she must, explicitly or implicitly, accord an answer, of who or what is responsible for such music. That is to say, to what agency is lm music assigned by a comprehending viewer, and what is this music understood to be doing, in relation either to the lms internal narrative, the viewers experience of that narrative, or the lm as an aesthetic whole? Furthermore, by what principle does a viewer assign, however tacitly, responsibility for the music he or she hears? It will turn out that different answers to this question of agency are in order from one lm to another, and even from one cue to another within a given lm. The upshot is a basic division within the realm of lm music, one I have not seen marked elsewhere, but which is probably more fundamental than others that are regularly noted. II I begin with some preliminaries. First, the music I am concerned with is usually designated nondiegetic lm music, that is, music whose source is not the story (or diegesis) being conveyed by the lms sequence of images. It is sometimes also designated soundtrack as opposed to source music, and sometimes

First published in D. Bordwell and N. Carroll (eds.), Post-Theory: Reconstructing Film Studies (Madison: University of Wisconsin Press, 1996), 25488.

144

Music

extrinsic, as opposed to intrinsic, music. Second, the lms I am concerned with are all narrative ction lms, both of the classical (or Hollywood) sort, and the modernist (or art lm) sort, though not any of the more extreme examples of the latter, in which the bounds of ctionality or narrative coherence are stretched to their limits. Third, I consider lm music here only as an integral component of a complete lm, and not as a genre of music which, in the form of suites or soundtracks, might be enjoyed and evaluated on its own. Certain kinds of answers to our opening queries can be put aside immediately as not to the point. For instance, the source of nondiegetic lm music might in one sense be said to be the composer who composes it, or the producer who commissions it, or the sound editor who integrates it into the nished lm, but this does not address the question of where, in relation to the ctional world projected, the music is situated or positioned in comprehending the lm. Similarly, the function of nondiegetic lm music might be said to be, somewhat vaguely, the aesthetic enhancement of the lm, or more specically, the emotional manipulation of the lm viewer, or more crassly, the augmentation of the lms marketability and secondary prots, but none of these answers addresses the question of how such music is understood to function in relation to the central narrative of sight and sound, and thus to contribute ultimately to a lms meaning. It should be noted straight off that there are two basic sorts of musical score regularly encountered in the domain of the sound lm: the rst, more traditional, sort consists of music composed specically for the lm in question, and generally tailored by the composer to the rough cut, scene by scene; the second sort consists of pre-existent music chosen by the lmmaker, often in conjunction with a musical consultant, and applied or afxed to scenes or parts thereof. Call the former sort a composed score, and the latter an appropriated score. Directly we can make at least two observations about these two types of score. First, with appropriated scores the issue of specic imported associations, deriving from the original context of composition or performance or distribution, rather than just general associations carried by musical style
For more on these categories, see David Bordwell and Kristin Thompson, Fundamental Aesthetics of Sound in Cinema, in Film Art: An Introduction, 4th edn. (New York: Knopf, 1993). For instance, Michael Snows Wavelength or Alain Resnaiss LAnne Derni`re a Marienbad. e e `

Film Music and Narrative Agency

145

or conventions, is likely to arise. Second, with appropriated as opposed to composed scores, ironically there will generally be more attention drawn to the music, both because it is often recognized as such and located by the viewer in cultural space, and because the impression it gives of chosenness, on the part of the implied lmmaker, is greater. To these two observations I add a third, more contentious one, that later discussion will support: music composed for a lm (e.g. the soundtracks of Vertigo or The Heiress or On the Waterfront or La Strada), is more likely to be purely narrative in function than pre-existing music appropriated by a lmmaker (e.g. the soundtracks of A Clockwork Orange or Barry Lyndon or Love and Death or Death in Venice). III There are some theoretical claims prevalent in the recent literature on lm with which I will be disagreeing, and it is best I signal what they are at the outset. One is that nondiegetic lm music is standardly inaudible, i.e. is not, and is not meant to be, consciously heard, attended to, or noticed. This seems to be clearly false, or at any rate, false for a wide range of lms in which soundtrack music calls attention to itself unmistakably, or requires the viewer to attend to it explicitly if he or she is not to miss something of narrative importance. The inaudibility claim seems most true for what is called underscoring, music at a low volume that serves as a sort of aural cushion for dialogue that remains the main order of business, or for melodically and rhythmically unmarked music helping to effect transitions between scenes of notably different character. Even here, when the music hovers in the penumbra of consciousness, it is rarely very far from being consciously focused, as is perhaps reected in the fact of being immediately noticed if stopped. If nondiegetic lm music were generally unheard, or not consciously noted by the viewer, then there would not be much of an interpretive issue for the viewer of how to construe such music in relation to the rest of what is going on in the lm. But there manifestly is an issue of some signicance, with respect to many lms. Finally, even if it were the case that casual viewing of lms with signicant music tracks often goes on without a viewers explicit awareness of
This is a central thesis in Claudia Gorbmans Unheard Melodies: Narrative Film Music (Bloomington: Indiana University Press, 1987), and is echoed by other recent psychoanalytically oriented writers on lm.

146

Music

that music, it hardly follows that an aesthetically justied or optimal viewing of such lms remains similarly oblivious. Another idea with some currency is the disavowal of what might be called narration properthe conveying of a story by an intelligent agentas actually characterizing the standard ction lm. One variant of this has it that such lms are not really narrated by anyone or anything within the lm world, but instead narrate themselves. A second variant insists that such lms are constituted as narratives only by the viewer, and contain no narration apart from that. A third variant maintains that such lms are not only constituted as narratives by viewers, but are in fact narrated by viewers to themselves as well, in the course of viewing. I reject the rst sort of disavowal on grounds of incoherence; if narration means anything, it is the conveying or imparting of a story by means that are distinct both from the story being conveyed and from that which is doing the conveying; if the lm, or its processes, are the means of narration, then it, or they, cannot also be conceived to be the agent or source of narration. I reject the second and third sorts of disavowal because they seem based on conating the viewers actual task of comprehending a lms story and signicance by actively reconstructing or piecing together the narrative on offer, with the viewers literal creation of that narrative, which would thus not exist apart from the viewer. But this is unnecessarily fanciful; our responsibility as lmgoers is to grasp what the narrative is, so as to reect further on what it might signify, rather than to create that narrative for ourselves. Furthermore, were we really to create the narrative for ourselves, its signicance would not, at any rate, be that of the lm we were putatively attempting to understand. So I am going to assume, following Seymour Chatman, that if there is narration in a ction lm, if a comprehensible story is being conveyed to us, then there is an agency or intelligence we are entitled, and in fact need, to imagine
Though he argues vigorously against the third form of disavowal in his attack on nonciation e theorists, there remains something of the rst and second in the constructivism about lm meaning defended by David Bordwell; see his Narration in the Fiction Film (Madison: University of Wisconsin Press, 1985) and Making Meaning (Cambridge, MA: Harvard University Press, 1989). For a critique of this aspect of Bordwells otherwise eminently sane approach to lm, see Berys Gaut, Making Sense of Films: Neoformalism and Its Limits, Forum for Modern Language Studies 31 (1995): 823. Bordwells rejection of narrative agents in lm as such is also criticized by Seymour Chatman, Coming to Terms (Ithaca: Cornell University Press, 1990), ch. 8. See his Coming to Terms, especially chs. 5, 7, and 8. Another writer who seems to accept the necessity of positing narrative agency in narrative lm, though he verges on abstracting this to the point of abandonment, is Edward Branigan, Point of View in the Cinema (The Hague: Mouton, 1984).

Film Music and Narrative Agency

147

is responsible for this, i.e. doing the narrating. That is, a narrator, though not necessarily an ordinary human being. There are, of course, alternatives to this assumption. As noted above, there are those who propose that lms or lmic processes are themselves the performers or executants of narration, there not being of necessity any narrator within the lms world, on the same plane as the events being displayed. But in addition to the fundamental incoherence remarked above, this proposal, to the extent it can be made out, is simply less interpretively useful than that of a narrator, however minimally characterized, for every successful narration. My response to yet another alternative, that in many cases of lmic narration, we imagine we are presented directly with the events of the story, without imagining there is any agent presenting them to us, is much the same: the postulate of narrative agency in cinema does a better job of accounting for how we, admittedly largely implicitly, make sense of lms as conveyors of stories. For those who yet balk at this postulate, I would offer this. What I want to say about assigning nondiegetic music to narrative agents as opposed to implied lmmakers can, I believe, be translated so as to require instead only the assumption of narrative processes or mere appearances-of-being-narrated. So even if one does not regard the positing of internal narrators or presenters in lm as inevitable, the issue will still remain of whether soundtrack music is to be thought of as an element in the narrative process or appearance of narrative presentation, as opposed to an element in the construction of the lm by a lmmaker, standing outside both the story and its narration. It is that issue I hope to illuminate here.

IV I now review Chatmans brief for both cinematic narrators and cinematic implied authors (that is, implied lmmakers). Chatman begins with an appeal to ordinary language, one it is hard to gainsay:
It stands to reason that if shown stories are to be considered narratives, they must be narrated . . . I would argue that every narrative is by denition narratedthat is,
A position taken, for example, by Gregory Currie, in Visual Fictions, Philosophical Quarterly 41 (1991): 12943. I respond to Currie in Seeing, Imaginarily, at the Movies, Philosophical Quarterly 43 (1993): 708. All text references in this section are to Coming to Terms.

148

Music

narratively presentedand that narration . . . entails an agent even when the agent bears no signs of human personality. (115)

If narrative lms are then necessarily narrated by a narrator, what kind of narrator is this? Not, of course, the essentially linguistic narrator of a standard literary ction:
Film often has nothing like a narrative voice, no tell-er. Even the cinematic voiceover narrator is usually at the service of a larger narrative agent, the cinematic shower. But that shower can reasonably be called a presenter . . . (113) Films, in my view, are always presentedmostly and often exclusively shown, but sometimes partially toldby a narrator or narrators. The overall agent that does the showing I would call the cinematic narrator . . . The cinematic narrator is not be identied with the voice-over narrator . . . . (1334)

Chatman also proposes that a cinematic narrator, operating mainly through the affording of sights and sounds, is closely analogous to the mute presenter of dialogue in a purely dialogic short story. So though in lm a teller, whose standard format is that of the voice-over, is usually absent or secondary, a shower or better, because the term covers more comfortably aural information, a presenter can be taken to be invariably in place, and the primary agent of narration. The presenter in a lm presents, or gives perceptual access to, the storys sights and sounds; the presenter in a lm is thus, in part, a sort of perceptual enabler. Such perceptual enabling is what we must implicitly posit to explain how it is we are, even imaginarily, perceiving what we are perceiving of the story, in the manner and order in which we are perceiving it. The notion of a presenter, whose main charge is the providing of perceptual access on the ctional world, is simply the best default assumption available for how we make sense of narrative ction lm.
This formulation of Chatmans is actually somewhat off the mark: its not the lm that is presented by the narrator, but various perceptual contents, various sights and sounds, i.e. what one is enabled to see and hear, courtesy of the presumed powers of such a narrator. The lm as such is rather presented by the lmmaker or, interpretively, the implied lmmaker. This is not to deny that it is sometimes in the purview of the cinematic narrator to present the mental contents of some character, e.g. memories, fantasies, dreams, visualizations. But two points about this should be noted. One, it may be unclear in such cases whether it is the cinematic narrator, acting on the characters behalf, who shows the characters mentation, or rather the character, acting as his own narrator, who is doing so. Two, the possibility of this sort of presenting requires a background of presentings of perceptual reality at a more basic story level. Problems of terminology loom here which a preemptive strike of clarication might dispel. Of the three ideas, cinematic narrator, lmic presenter, and perceptual enabler, the rst is perhaps the broadest and the third the narrowest. Certainly there are actions of the cinematic narrator which go beyond those of perceptual enabling or lmic presenting, most notably in this context, narrative pointing through nondiegetic music. Whether there is a distinction worth making between lmic

Film Music and Narrative Agency

149

While I thus accept Chatmans postulate of narrative agency wherever there is narration, I do not endorse certain of his claims about the separation of narrators from the story worlds they are narratively presenting. Chatman says, for instance, that the [literary] narrator, by denition, does not see things in the story world; only characters can do that, because only they occupy that world.(120), and that the narrator cannot impinge on story space but must stay within the bounds of discourse space (123). It is, however, incoherent to postulate a narrator who offers us a window on or reportage concerning the doings of a set of individuals the narrator takes and presents as real, and yet insist the narrator is on a different plane, ctionally speaking, from those individuals, and in principle incapable of perceptual awareness of them. Chatman is confusing a narrators ctional level, which must standardly be the same as that of the other characters whose doings he/she/it is purporting to convey, and the narrators degree of story involvement, which is variable, often rather small, and in the limit, nil. A narrator and the events narrated by the narrator must be on the same ctional plane, otherwise cognitive relations posited between narrator and events would not make sense. The cinematic narrators logical status vis-`-vis a the lm world is to be distinguished from the narrators degree of involvementcausal, emotional, experientialin the story, i.e. what literary theorists mark as the narrators being either homodiegetic or heterodiegetic. Being heterodiegetic, or an outsider to the events being related, does not remove a lmic narrator ontologically from the characters he/she/it serves to offer us perceptual access to. Chatman fails to see that the narrator must perforce share the ctional plane of the characters, since they are apparently real and reportable to that narrator, and this is true whether the narrator is homodiegetic, i.e. involved in the story events, or heterodiegetic, i.e. uninvolved
presenter and perceptual enabler is less clear; if so, the former would include the latter but comprise in addition resources such as character voice-overs or mind-overs, affording access to the ctional world in a wider-than-perceptual vein. Except when the narrators relationship to the story being presented is clearly signaled, in the novel or lm, as one of relating a ction as such, e.g. as through a disclaimer like this is only a story, it never happened. But this is quite rare in ction lm; it is even rarer in literary ction, Thackerays Vanity Fair standing as a classic example, and John Fowless French Lieutenants Woman as a recent, though more ambiguous, one. Kendall Walton, in Mimesis as Make-Believe (Cambridge, MA: Harvard University Press, 1990), marks this distinction as one between reporting narrators and storytelling narrators (36872). The overwhelming majority of narrators in narrative ction are reporting narrators, and as Walton points out, in such cases narrator and events narrated necessarily belong to the same world. As in the event of a wholly effaced and omniscient third-person narrator, the norm for cinematic ctions.

150

Music

in them, standing to those events in merely a witnessing and transmitting capacity. I turn now to the notion of implied author in lm. The need for this concept is clear from the fact that lms, like novels
present phenomena that cannot otherwise be accounted for, such as the discrepancies between what the cinematic narrator presents and what the lm as a whole implies . . . unreliable narration presents the clearest but not the only case for the implied author [in lm] (1301) . . . in cinema as in literature, the implied author is the agent intrinsic to the story whose responsibility is the overall designincluding the decision to communicate it through one or more narrators. Cinematic narrators are transmitting agents of narratives, not their creators. (132) In short, for lms as for novels, we would do well to distinguish between a presenter of the story, the narrator (who is a component of the discourse), and the inventor of both the story and the discourse (including the narrator): that is, the implied author . . . (133)

So in lm we must generally distinguish between, on the one hand, the narrator or presenter of the story and, on the other hand, the ostensible inventor of (all of) the narrator, the story narrated, the narrative structure, and the cinematic entirety in which these are embodiedto wit, the implied lmmaker. The implied lmmaker is the agent who appears to have invented, arranged, and integrated the various narrative agents and aspects of narration involved in the lm, as well as everything else required to constitute the lm as a complete object of appreciation. The implied lmmaker, in short, is the image we construct of the lms makerbeliefs, aims, attitudes, values, and personalityon the basis of the lm viewed in its full context of creation. A lms narrator presents the events of the lms world from within it, whereas the implied author of a lm, if he or she can be said to present anything other than the lm itself, presents the world of the lm, at one doxastic remove, from a position external to it. For the implied lmmaker, as for the viewer, but in contrast to the lms narrator, the lms world is a ctional one, acknowledged as ctional throughout. The implied lmmaker cant be in the position of directly affording us, as with a silent gesture of behold!, the vision and audition of something that is only ctional with respect to himself, namely, the characters and their circumstances; that remains the prerogative of the lms narrator or presenter, who is, in a fundamental sense, and pace Chatman, one of them.

Film Music and Narrative Agency

151

V Before proceeding to my main concerns I will address some worries about the general postulate of a lmic narrator formulated by George Wilson in his penetrating study, Narration in Light. These worries are the most substantial and explicit of which I am aware, so if they can be allayed, the ground for such a postulate will be that much clearer. A reason Wilson offers initially for being wary of a standing postulate of cinematic narrator is that such a narrator is often conceived as an agent who is necessarily observing events, the image track being thus identied as the visual experience of that ctional observer. But this, as Wilson quickly notes, is unnecessary. The essential function of such a narrator is to show us what is to be seenor more broadly, to present to us what is to be seen and heardin other words, to enable perception, albeit ctional perception, of those events. The agent who shows, or permits us to see, need not be thought of as seeing as well. This is apparent in Wilsons own useful sketch of what a cinematic narrator would have to be: a ctional or ctionalized being, presupposed in any viewing of the lm narrative, who continuously provides to the audience, from within the general framework of the ction, the successive views that open onto the action of the lm. The alternative, then, is to conceive the cinematic narrator as a kind of perceptual pilot through the lm world, rather than as an observer of it whom we opportunistically inhabit.
Considered in this fashion, the narrator is a ctional gure who, at each moment of the lm, asserts the existence of certain ctional states of affairs by showing them to the audience demonstratively; that is, by ostending them within and by means of the boundaries of the screen. It is certainly part of our experience in lm viewing that we feel, usually subliminally, a constant guidance and outside direction of our perception toward the range of predetermined ctional facts which we are meant to see.

Having so well formulated this alternative, what problem does Wilson nd with it? Just this: that the entity described is analogous not to the narrator of a novel, but rather to its implied author. His reason for this reluctant conclusion seems to be that although the exact style and manner of this guidance the ne-grained articulation of the processes of showingmanifests traits of
George Wilson, Narration in Light (Baltimore: Johns Hopkins University Press, 1986). See also my review in Journal of Aesthetics and Art Criticism 47 (1989): 2902. Narration in Light, 132. Ibid. 1334.

152

Music

sensibility, intelligence, and character, these traits are ones we will naturally take to dene the persona of the lmmaker as expressed through the lm, or equivalently, the personality of the lms implied author, thus leaving no room for a lmic narrator as such. This seems to me too quick. First, Wilson gives no reason why such traits should not be assigned, in some cases, to a lmic narrator, and in others, to the implied lmmaker. Second, Wilson fails to consider that the assignment of traits to the implied lmmaker might very well interpretively depend on the assignment of traits, the same or different ones, to such a narrator, much as our image of the implied author of a novel is necessarily based, in large part, on our image of the narrator and how the narrator is managed or positioned by the author. But third, and most importantly, Wilson overlooks the fact that the implied lmmaker just cannot occupy the role of perceptual guide to the lms occurrences, and so, a fortiori, his particular mode of doing that cannot be what cues us to some of his traits. And that is because the implied lmmaker cannot logically be the presenter and ostender of events that are ctional with respect to him. That is to say, if we imagine anyone giving us access to those events, it cannot coherently be the lmmaker, in any guise. To be sure, the lmmaker can present representations of those events, i.e. the shots or images the totality of which constitute the lm, but he cannot offer us the vision and audition of those events themselves. If he be allowed a surrogate, however, a narrating agent presupposed by the process of narration, and ctionally on the same level as its subject matter, then this difculty disappears. A fourth and related reason why the lmmaker cannot do duty for the lms narrator in this connection is this. Often we want not only to attribute traits of character, sensibility, or intellect to some agent connected with the lm, but more specically, attitudes or views concerning the story that is unfolding. But the lmmaker is not in the right cognitive position for this; that is to say, he or she will not actually have attitudes or views towards the ctional personages or occurrences involved in the story, knowing they are merely ctional. I suspect that Wilson is unable to nd a place for, and thus underestimates the rationale for, the cinematic narrator, because he subtly conates the
Of course he can and does offer us the vision and audition of various events which took place during the lming, namely, the enacting of various roles by various actorsbut that is another matter. See Walton, Mimesis as Make-Believe, 366.

Film Music and Narrative Agency

153

devices or powers at the service of a director as crafter of a representation, and those in the command of an imagined perceptual guide to the world such a representation makes ctional; but these are different, and the ways they function to control our experience as viewers differ too. The lmic narrator allows us to perceive rst this person, then that, at such and such an apparent range, and for so long, clearly or not so clearly, etc., all of which manner of showing may give us a certain impression of the showers attitudes or motivations; the lmmaker chooses or stages the prolmic events that are to be lmed, decides on the camera distances and movements required for a shot, determines the lighting and length of shots, orders those shots in a certain fashion, etc., thus ultimately composing a narrative of a particular sort, with a particular sort of implied narrator, all of which manner of making may give us a certain impression of the makers personality or outlook. But the view we form of the narrating agency or intelligence, from the way it carries out its main charge as perceptual facilitator, need not coincide with the view we form of the human maker of the lm, from the way he or she fullls the demands of lmmaking. Curiously, Wilson allows that we can indeed imagine a lm where we would have
grounds for a distinction between a voyeuristic lmic narrator and a satirizing implied lm maker. In such a case, there would be enough of a personication of the manner in which the action is shown and enough of a contrast between the personication and what is implied about the lmmakers views of this to motivate the identication of [the former] as a narrator.

But to my mind, the very possibility of this kind of divergence is predicated on and presupposes the logical distinctness of the roles of lmic narrator and
Nor are supporters of the notion of a narrator internal to lm immune to this confusion. Consider the formulation of Nick Browne, quoted by both Branigan and Chatman: the authority which can be taken to rationalize the presentation of shots (The Rhetoric of Filmic Narration (Ann Arbor: UMI Research Press, 1982), 1). But since shots are constructional elements of lms as made objects, that authority can only be the implied lmmaker; the authority, or agency, that Browne is really after is that which appears to rationalize the presentation of views, or sights and sounds. There are other sources of Wilsons reluctance to embrace lmic narrators on a standing basis. He suggests at one point that a lmic narrator distinct from the implied lmmaker would, by analogy to literature, have to be a character that the text depicts directly or indirectly (Narration in Light, 136). But this is only half right; the narrator is indeed a kind of ctional character, but not one that need be depicted, as I understand that term. If purely dialogic short stories, or purely epistolary novels, have narrators, then such narrators are not depicted, either directly or indirectly. So the fact that cinematic narrators are standardly not depicted, i.e. nothing in lms shows or announces them as such, is not a principled impediment to acknowledging them. Ibid. 1367.

154

Music

implied lmmaker, even if in most lms, unlike the one Wilson conjures up, the personal traits ascribable to the occupants of the two roles tend to coincide. As we have seen, Wilson questions whether every standard lm narration must be understood to entail an implicit narrator distinct from the lmmaker. He maintains that in lms governed by the classical paradigm of transparency, which covers almost all narrative lm, we simply see the ctional events for ourselves, defeasibly taking the facts about them to be what we see them to be. But to my mind this sidesteps the question of how it is we are seeing what we are seeing, however reliable or unreliable it turns out to be. If this question, however, is not simply set aside then the only satisfying answer to it is that we are being shown such and such, by some agent, in some perhaps unspeciable manner. That is to say, the posit, however unvoiced, of an agency that is offering us sights and sightsan agency with certain powers, motivations, and limitationsseems inescapable if we are to justify our taking anything to be ctional in the lm world, on the basis of the moving images that are the only thing we are literally confronted with. It is not enough to just say that, with ction lm, the lms world is made visible to us, perhaps adding that there is a convention to that effect. Reasonalbeit reason operating in service of the imaginative understanding of ctiondemands an answer to how it is that a world is being made visible to us, and that demand, it appears, is only satised by the assumption of an agency responsible for that. One might still seek to avoid this conclusion by adopting the following stance: it is, indeed, as if we are being shown such and such, from a given perspective, by an agent within the lms world, with certain powers to make views of that world available to us, but we need not assume that there is such an agent. Here, though, we arrive at a distinction virtually without a difference. If it seems to us, at some level, as if we are being shown such and
My discussion so far may give the impression that I regard the role of the implied presenter of a narrative lm as conned to that of providing views on events understood as fully constituted independently of the showers activity. But while I think that that is indeed the dominant role of a lms presenter, it need not be the exclusive one. The cinematic narrator might in part be thought of as a fashioner or shaper of events that are only then presented, more straightforwardly, in certain ways (e.g. in certain lights, or in a certain order). This fashioning or shaping of ctional events thought of as existing, on a more basic level, prior to narrative attention, can be seen as a more subtle way of presenting events taken to belong to the underlying event structure with which narration is concerned. If so, then the narrative structuring of lms is a two-stage affair, and that which is effected through camerawork and editing is subsequent to that understood to be achieved in the staging of action and the manipulation of setting. (For more on this dimension of cinematic narration, and the rationale of its recognition, see Bordwell, Narration in the Fiction Film.)

Film Music and Narrative Agency

155

such, are we not in effect imagining that we are being shown such and such, and thus, nally, that there is, on the imaginative plane, something doing the showing? So I would claim. VI That nondiegetic music standardly serves to advance a lms narrative is something on which theorists of lm appear to agree:
Narrative is not constructed by visual means alone. By this I mean that music works as part of the process that transmits narrative information to the spectator . . . Voice-over is just one of many elements, including musical scoring, sound effects, editing, lighting, and so on, through which the cinematic text is narrated. The moment we recognize to what degree lm music shapes our perception of a narrative, we can no longer consider it incidental . . .

Another point widely agreed upon is that even if the primary purpose of nondiegetic lm music is the advancing of the narrative, there may very well be others. Here is a typical admonition concerning lm musics multiplicity of ends:
There is not one and only one function that music can perform in relation to movies. Aaron Copland suggested ve broad functions: creating atmosphere, underlining the psychological states of characters, providing background ller, building a sense of continuity, sustaining tension and then rounding it off with a sense of closure. These do not seem to be necessarily exclusive categories, nor do they exhaust the range of functions that music can perform in movies.

Not surprisingly, I am happy to join this double consensus: lm music often serves narrative in some way, but there is a range of other functions that such music sometimes performs. What I am concerned to demonstrate, however, goes beyond those two pieces of received wisdom. It is that the most fundamental division in the realm of lm music concerns the viewers assignment of responsibility for such music, i.e. the agency the viewer posits, usually implicitly, as responsible for the music being heard. It will turn out that there is a rough coincidence between lm music to which we intuitively accord
Kathryn Kalinak, Settling the Score: Music and the Classical Hollywood Film (Madison: University of Wisconsin Press, 1992), 30. Sarah Kozloff, Invisible Storytellers: Voice-over in American Fiction Film (Berkeley: University of California Press, 1988), 434. Gorbman, Unheard Melodies, 11. Nol Carroll, Mystifying Movies (New York: Columbia University Press, 1988), 216. e

156

Music

narrative signicance and lm music for which we implicitly hold an internal cinematic narrator accountable, and between lm music to which we do not accord narrative signicance and lm music that we implicitly assign directly to the implied lmmaker. When, though, can lm music be said to have narrative signicance? When does nondiegetic music function narratively? In order to answer this question we must have a plausible criterion of narrativity, or of actions within the purview of a narrator. In trying to arrive at one, it will be helpful to have before us a survey of the various functions that critics or theorists have observed lm music to perform. These functions include: (1) the indicating or revealing of something about a characters psychological condition, including emotional states, personality traits, or specic cognitions, as when the music informs you that the heroine is happy, or that the hero has just realized who the murderer was; (2) the modifying or qualifying of some psychological attribution to a character independently grounded by other elements of the lm, as when the music tells you that a characters grief over a loss is intense; (3) the underlining or corroborating of some psychological attribution to a character independently grounded by other elements of the lm, as when music emphasizes something about a situation on screen which is already fully evident; (4) the signifying of some fact or state of affairs in the lm world other than the psychological condition of some character, e.g. that a certain evil deed has occurred, offscreen; (5) The foreshadowing of a dramatic development in a situation being depicted on screen; (6) the projecting of a story-appropriate mood, attributable to a scene as a whole; (7) the imparting to the viewer of a sense that the happenings in the lm are more important than those of ordinary life, the emotions magnied, the stakes higher, the signicances deeper; (8) the suggesting to the viewer of how the presenter of the story regards or feels about some aspect of the story, e.g. sympathetically; (9) the suggesting to the viewer of how he or she is to regard or feel about some aspect of the story, e.g. compassionately; (10) the imparting of certain formal properties, such as coherence, cogency, continuity, closure, to the lm or parts thereof; (11) the direct inducing in viewers of tension, fear, wariness, relaxation, cheerfulness, or other similar cognitive or affective state; (12) the lulling or mesmerizing of the viewer, so as to facilitate emotional involvement in the ctional world to which the viewer would otherwise prove resistant; (13) the distracting of the viewers attention from the technical features of the lm as a constructed artifact, concern with

Film Music and Narrative Agency

157

which would prevent immersion in the lmic narrative; (14) the expressing by the lmmaker of an attitude toward, or view on, the ctional story or aspect thereof; (15) the embellishing or enriching of the lm as an object of appreciation. Without deciding, for each of these functions, which are properly considered narrative and which not, it would appear that some unequivocally are, and some unequivocally are not. What I will do at this point is explore a number of suggestions as to what the criterion of narrativity might be in regard to nondiegetic lm music, assessing them against the background of this array of observed functions, some of which, at any rate, would have to come out counting as narrative, some clearly not, and some having a status that might only be settled, clarifyingly, once a given suggestion were adopted. One possible criterion is this: (C1) does the music seem to issue from, be in service of, the agency one imagines to be bringing one the sights and sounds of the lms world? If so, then it can be reckoned part of the narration proper, and assignable to the cinematic narrator. Perhaps an equivalent formulation would be: (C2) does the intelligence one thinks of as bringing one the music seem to be the same as that charged with conveying the storyas opposed to that charged with constructing the lm? If so, then the music can be reckoned part of the narration, and assigned to the cinematic narrator. Though I think these criteria point in the right direction, there is an evident problem with them, insofar as we hope to look to them for guidance, especially in difcult cases. And that is that they are uncomfortably close to what they purport to analyze or elucidate, namely, whether a use of nondiegetic music is narrative or not. So if we are unsure whether a given cue is functioning narratively, we are likely to be almost equally unsure whether it feels as if it derives from the lms narrative agent. Thus, it would seem desirable to have some other mark, could we discover one, whose conceptual distinctness from the idea of narrative functioning was greater than that of C1 or C2. Such a mark might be that of making a difference in the narrative. Instead of appealing directly to an intuition of a connection of the music to a lms internal narrator, we can appeal instead to the notion of making-ctional, or generating ctional truths, in a lm. A criterion of nondiegetic music having
This would apply, note, even when such music is unforegrounded: if it appears to respond to the demands of storytelling, broadly understood, then it can be construed as something like musical musing, sotto voce, on the cinematic narrators part.

158

Music

a narrative function, and thus being attributable to a narrative agent, could be this: (C3) the music makes something ctionally truetrue in the story being conveyedthat would not otherwise be true, or not to the same degree or with the same deniteness. A counterfactual form of the suggestion is perhaps more transparent: (C4) would deleting the music in a scene change its represented content, i.e. what is ctional in it, or only how the scene affects viewers? If the former, then the music is an aspect of narration; if the latter, then not. We must briey discuss what it means to make something ctional in a work of ction such as a narrative lm. Something is ctional in a lm, according to a well-developed recent account, if it is to be imagined to be the case by viewers concerned to experience the lm properly. What thus makes somethinga proposition about the lms worldsomething that is to be imagined in the course of viewing is perceivable features of the lm, a public object, taken as a prop for guided imaginings. When we make-believe in accord both with the features of artistic props and the usually tacitly grasped principles for imagining that are in effect in a given artform, we are engaged in tracing out imaginary worlds, ones in which things are make-believedly, or ctionally, so. The ctional world of a representational art work, unlike that of a daydream or fantasy, is as it is because features of the associated proptext, canvas, lmproperly construed, are the way they are; not all is up to the imaginer. Props, through their existence and nature, generate ctional truths independently of what individual perceivers might choose to imagine. What does it mean for a proposition to be ctional, or true in a ctional world, in respect of a given work of art? Simply that there is a prescription to imagine it, a prescription encoded in the particulars of the artifact that serves as a prop for making-believe, and whose force derives from underlying conventions of construing works of the sort in question. Being ctional thus has an ineliminable normative dimension: it is what is to be imagined in a given context, rather than merely what may be imagined. For example, in Citizen Kane, Orson Welless image on screen being that of a large man makes it ctional that Charles Foster Kane is a large man; the opening shotsa series of lap dissolveshaving a certain visual content
See Walton, Mimesis as Make-Believe. For an entre into this important work, see my critical e notice, Making Believe, Dialogue 32 (1993): 35974 (reprinted in The Pleasures of Aesthetics).

Film Music and Narrative Agency

159

makes it ctional that at the beginning of the story one is shown Kanes estate, Xanadu, from a distance and shrouded in mist, and then at progressively closer range; Ray Collinss voice saying certain things on the soundtrack in the scene at Susan Alexanders apartment makes it ctional that Collinss character, Jim Gettys, has threatened Kane; the way the shot of Kane expiring is sequenced in relation to others which are understood as a ashback to Kanes childhood, makes it ctional that Kanes dying word, rosebud, refers to his beloved old sled, etc. Of course, much of this generation of ctional truths will be indirect, dependent on various conventions of the medium in effect and on other things taken provisionally as ctional, and accordingly, much of our knowledge of such ctional truths will be inferential. And sometimes, what is made ctional by a lms narration is orthogonal to, or even the opposite of, what rst appears to be the case, that is, what it initially seems we are to imagine is the case; unreliable, uninformed, or unforthcoming narrators, though not as common in lm as in literature, are still a signicant possibility. Applying this suggestion to the issue of narrativity in lm music, then, the question becomes, of a given cue, whether it generates, contributes to generating, or at a minimum, more rmly grounds, a ctional truth in the scene which it accompanies. Thus, lm music which, when interpreted in the light of prevailing conventions of the medium and the surrounding narrative context, indicated that a character was afraid or was remembering a past incident, or that a man had been executed or an agreement reached, or that a situation was fraught with danger or else full of hope, where these things would not be established, or not so denitely, without the music, would clearly count as narrative. We should note that nondiegetic music may, indeed, generate ctional truths even if only attended to with half a mind, or not consciously remarked at all while present. It will do this by causing a viewer, say, to perceive a scene as fraught with danger, even if the viewer is not aware of what is making her have that perception. Nevertheless, if such an imaginative perception is reliably produced in attuned viewers, and not undermined by subsequent aspects of the narration, then it may well be ctional that the scene is fraught with danger, even though the rest of the narrative indicators are insufcient to
Film music . . . often contributes subtly but effectively to the generation of ctional truthshelping to establish, for example, that ctionally a character is nervous or cocky or ecstatic (Mimesis as Make-Believe, 172).

160

Music

establish that and the viewer never realizes that it is the background music that in fact makes it so.

VII It is time to look at a range of illustrative examples of lm music. I begin with examples whose narrative functioning is obvious, and which conform, expectedly, to the making-ctional criterion proposed above. I then explore another range of examples, ones that exhibit a different sort of narrativity, and show how, on a more encompassing construal of the making-ctional criterion, these can be accommodated as well. Eventually, though, I turn to lms containing nondiegetic music that is not, by that criterion or any other, reasonably construed as narrative. The music in such lms instead serves other sorts of artistic function, ones attributable directly, I will argue, to implied lmmakers. One of the least ambiguous narrative uses of soundtrack music in mainstream lm occurs in Steven Spielbergs 1975 blockbuster, Jaws. I have in mind the shark motto devised by the composer, John Williams. This consists of an ostinato alternation of low staccato notes at the interval of a seconda kind of aural sawing. The motto has an unarguable informational mission, namely, the signaling of the presence of the shark. It is true that there is another, visual, indicator of the sharks presence when unseen, namely, shots from an offshore point of view, at the water line or slightly below it. But that indicator is not invariant in meaning, since it is sometimes employed when there is no shark about. The musical shark motto is the only reliable signier of the shark, and so has an ineliminable fact-conveying function. Correspondingly, it is clear that it is the presence of that motto on the soundtrack at a given point that makes it ctional that the shark, though as yet unseen, is in the vicinity of what is shown. David Raksins haunting score for Otto Premingers Laura provides some further instances of straightforward narrative use of lm music. The Laura theme, rst encountered diegetically on a record player in the apartment of the ostensibly (but not actually) murdered heroine, pervades critic Waldos (Clifton Webb) represented recollections of the early days of his relationship with Laura (Gene Tierney), and unmistakably signies his delight in her companionship. Subsequently we are treated to apprehensive versions of the Laura theme as detective McPherson (Dana Andrews), alone in Lauras

Film Music and Narrative Agency

161

apartment, studies the portrait of Laura over the replace; this cue then climaxes unsettlingly, revealing or underlining McPhersons frustration with his investigation at this point. The most striking cue, one much noted in the lm music literature, is a weird version of the theme produced by playing it on a piano but only recording the overtones of each note struck. This is heard as McPherson views Lauras portrait on a second occasion, before then drinking too much and falling asleep, and suggests the ghostly inuence Laura is beginning to exert over his poor detectives mind. In each of the foregoing cases, the music is plausibly viewed as making, or contributing to making, something ctional in the story: that Waldo delighted in Laura inordinately, that McPherson is (earlier) almost terminally frustrated with Lauras case, that McPherson is (later) succumbing to bewitchment by Lauras spirit. Another lm rich in narrative pointing of a theoretically unproblematic sort is Martin Scorseses Taxi Driver. Regarding a scene in which Travis (Robert De Niro), the lms semi-psychotic protagonist, is induced to move his cab away from the Manhattan workplace of a girl he is infatuated with and back into the grime and disorder of the city, one writer afrms that the music . . . here reveals that Traviss thoughts are not with the street but with Betsy. And of the bluesy, sensual saxophone tune itself, which stands for Betsy (Cybil Sheppard) in Traviss mind, the same writer has this to say: Traviss vision of idealized womanhood, the music implies, is strongly erotic. Thus, Bernard Herrmanns music does not serve merely to inform us about Traviss mental life, or to second redundantly what other elements of the lm establish about his mentality, but rather enters into making it ctional in the lm that Traviss mental life is a certain way at a certain time. Commenting on the blade-game ght scene in Nicholas Rays Rebel Without a Cause, Nol Carroll offers the following: The uneasy, unstable e quality of the music [by Leonard Rosenman] serves to characterize the psychological turmoilthe play of repression and explosive releasewith which the scene, and the movie, is concerned. If Carroll is right, the music of this scene, which intuitively seems an aspect of its narration, serves to underwrite as desired a ctional truth about the specic, highly volatile, character of the turmoil aficting the young protagonists. Another instructive example from Rebel Without a Cause occurs later in the lm, and consists of
See Kalinaks informative discussion in Settling the Score, 178. Graham Bruce, Bernard Herrmann: Film Music and Narrative (Ann Arbor: UMI Research Press, 1985), 68. Mystifying Movies, 217.

162

Music

a montage of two-way phone calls among various adults concerned with the whereabouts of the three main youngsters. This montage is covered by tense nondiegetic music, displacing the dialogue that would ordinarily be heard, the music thus signifying that the conversations, whatever their specic contents, are anxious ones. The opening of Elia Kazans On the Waterfront affords another illuminating example. An establishing shot of city docks, ocean liner in the distance, gives way to a street scene in which longshoreman Terry (Marlon Brando) becomes the focus of attention. Leonard Bernsteins jazz-inected score at this point involves a persistent drum tattoo overlaid with saxophone insinuations. Terry, in the darkening street, yells up to friend Joeys window, persuading him to go to the roof to recover one of his pet pigeons, where unbeknownst to Joey, two men are waiting for him. After Terry releases the pigeon he has been holding, and promises to join Joey in a moment, the score becomes loud, aggressive, and insistent, its rhythms more syncopated. The music telegraphs us that something bad is in store, that the men glimpsed on the roof are trouble; the music can be said to pregure Joeys fall, pushed off the roof by thugs of the corrupt union boss, though without dening precisely what is about to happen. The cue is clearly narrative, and just as clearly, makes it ctional that Joey is in danger, even before he leaves the window for his fatal visit to the roof. Later on, after the boss tells right-hand man Charley to straighten out his brother Terry or else, Charley leaves union headquarters to do something, we know not what. Bernsteins music at this point is very dramatic and tense: a series of rising notes in the brass, leading to a rhythmic explosion, the whole heard twice. The cue arguably conveys Charleys complex state of mind, faced with the necessity of keeping his errant brother, who is threatening to do the right thing, in line: a mixture of anger, shame, and angst. If it does not singlehandedly make it ctional that that is Charleys state of mind, the cue contributes ineliminably to making it so. A dissolve leads directly to the famous conversation between the brothers in the rear of a taxi. Consider, lastly, the nal sequence in Fellinis La Strada. Having ve years ago abandoned his erstwhile assistant, the childlike Gelsomina, after she becomes too withdrawn and depressed to work, Zampano the strongman discovers, by accident, what became of her. That evening he does his act perfunctorily, gets drunk, starts brawling, then goes down to the beach, which reminds us of where he rst acquired Gelsomina from her impoverished family. He walks into the water, goes back out, looks up at the sky apprehensively,

Film Music and Narrative Agency

163

then starts to bawl and grasp at the sand, on which he has ung himself in despair. At this point the La Strada theme on the soundtrack removes all doubt as to what it is Zampano is bemoaningnamely, the loss of Gelsomina and her innocent love. VIII Clearly, making something ctional in a lm is a sufcient condition of musical narrativity. Is it, however, a necessary one? Though providing the basic ctional truths of a story may be the central activity of a narrator, there are others that are almost equally paradigmatic of narration. One is the evincing of attitudes or feelings on the narrators part toward the story presented, in virtue of how the story is presented; another is the inviting of the viewer to adopt certain attitudes or feelings toward the story presented. In other words, in addition to giving access, in a particular manner, to the ctional states of affairs that constitute a story, a narrator generally manifests attitudes regarding the states of affairs to which access is afforded, and thereby suggests to the narratee attitudes to be adopted. In literature, for example, the narrator standardly tells us what happened, after his or her fashion, reveals, knowingly or unknowingly, his or her view of these happenings, and also suggests, explicitly or implicitly, how we should view what we are told happened. Now it seems plain that such narrational effects are often achieved by appropriate nondiegetic music: the music tells you how the presenter of the story regards the events being presented, or else how he would like you to regard them. But on the surface, this does not appear to be a matter of establishing, nuancing, or even conrming a ctional state of affairs in the story. So in light of that, can making-ctional be sustained as the effective mark of musical narrativity? I believe so. We need to make a distinction between what is ctional in a lms story and what is ctional in the world of a lm. The latter is a broader notion than the former. What is ctional in the lms world comprises, in addition to the facts of the story, the facts of its narration by the special, often almost effaced, ctional agent known as the narrator. All that is still within the sphere of the ctional, that is, of propositions to be imagined by a viewer
In some literary ctions, for example, Hemingways The Killers or Robbe-Grillets La Jalousie, this latter function may seem to have lapsed. But I would argue that even in such ctions there are attitudes the narrator implicitly invites the reader to adopt, precisely in virtue of so pointedly eschewing normal commentary.

164

Music

in comprehending the lm. The lms story consists of what is ctional about the characters who gure in the action; the lms world includes, as well, what is ctional about the narrator, in relation to either the story narrated or the implied audience of that narration. Returning to lm music, a plausible construal of some nondiegetic cue will often have the implication, not that it makes something ctional in the story, but that it makes it ctional either that the cinematic narrator has a certain attitude or feeling toward some event being presented, or that the narrator encourages viewers to have such an attitude or feeling toward it. In either case, musical narrativity will still correlate with musics making something ctional, only here it is a making-ctional in the lms world, as opposed to a making-ctional in the embedded story. Some examples will serve to clarify this more encompassing interpretation of musical narration in terms of making-ctional. Music functions narratively, by any intuitive assessment, in Hitchcocks Shadow of a Doubt, particularly at junctures when a scrap of Lehars Merry Widow waltz intrudes itself, suggesting the Merry Widow Murders that are central to the plot. Several characters are heard singing or humming the tune in the course of the lm, these occurrences being of course diegetic, but the tune is heard, in an altered form, as early as Dmitri Tiomkins title music, which accompanies a stylized shot of waltzing couples. Two notable nondiegetic occurrences after that are these. First, a few bars of the waltz theme in the cue that accompanies the familys greeting of Uncle Charlie at the train station, as they walk off to their car to take him home: a tracking shot of the group, heading toward the camera, is eventually reframed so that only Uncle Charlie is in view, and that is when the scrap of tune is heard. Second, a more prominent statement of the theme when Uncle Charlie gives young Charlie an emerald ring, and she notices it is already engraved inside with an unknown someones initials. In both cases, the music arguably serves to communicate something to the viewer about Uncle Charlies identity, connecting him in some as yet unexplained way to the waltzing image presented at the beginning. But does the music make, or even contribute to making, something true in the lms story as such, something that would not otherwise be the case? It is not clear that it does. To consider just the most obvious candidates,
Of course, when a narrator in a lm is also a character in the action, as with a homodiegetic voice-over narrator, then certain facts about such a narrator are also facts of the story.

Film Music and Narrative Agency

165

neither cue makes it trueeven viewed in retrospect, when a connection to Lehars tune is understood to import as well a connection to the Merry Widow murdersthat Uncle Charlie is the murderer, nor does the second make it true, say, that young Charlie suspects that he is. The reason is that those ctional truths are rmly established, and independently, by other elements in the lm. What, then, might they be doing? I suggest that the rst cue makes it ctional that the narrator is obliquely hinting to viewers with regard to Uncle Charlies identity, and the second makes it ctional that the narrator is, even more directly, connecting Uncle Charlie to something sinister in his past, though at that point viewers have no notion of what it might be. The second cue may, in addition, function as the narrators proposing of a deep psychic link between Uncle Charlie and young Charlie, one that her subsequent moral corner-cutting, in dealing with an uncle she then knows to be an unhinged killer, partially bears out. In any event, the status of these cues as narrative can be recovered in the guise of what is made ctional, not in the story as such, but in the narrators attitudes or actions with respect to viewers. But what of the curious musical image of waltzing couples rst encountered in the title sequence, which recurs nondiegetically and unchangingly at three crucial points in the story? In each case the image is superimposed over the action already on view, which continues underneath. The rst occurrence is after the interaction between Uncle Charlie and young Charlie over the emerald ring, as young Charlie goes off to clear the supper dishes, leaving only Uncle Charlie on screen. The second occurrence is at night in the town library, when young Charlie, after reading the newspaper account of the Merry Widow murders, gets up, almost reeling, as the camera tracks upward and away from her. The third and last is just as Uncle Charlie falls to his death beneath the wheels of a hurtling locomotive.
For example, that Uncle Charlie is the murderer is underwritten by his unexplained money in the opening hotel room scene, by his evident concern to keep an item in the daily newspaper unread, by his unreasonable aversion to being photographed, by his maniacal utterance at the dinner table about fat, wheezing, useless widows, by the already inscribed ring itself, etc. That young Charlie suspects him does not become true until she is informed about the manhunt by one of the two detectives who have been trailing Uncle Charliethough of course there have been signs, intended for and readable by the viewer, well before that. Their psychic kinship is adumbrated earlier in the lm, in the parallelism of our rst views of them both, reclining on beds with their hands behind their heads, in the worried, almost cynical remarks about family values that young Charlie makes when we rst hear her speak, and in the coincidence of young Charlie deciding to send her uncle a telegram just hours after he has, unbeknownst to her, sent one in her direction.

166

Music

The rst and second of these might be interpreted as the narrators display of the mental contents of the character then in frame, in the one case signifying Uncle Charlies meditation on his hidden identity, in the other, his nieces realization of that identity. But in addition to being implausible because it does not reect the very different emotional tones with which uncle and niece would have contemplated this identity, this sort of interpretation seems unavailable for the last occurrence, where ascription to the terried and soonto-be-obliterated Uncle Charlie of a contemplative thought about his past strains credulity to the breaking point. This suggests that the recurrent waltzing image should be construed as a form of narrators commentary: it is employed by the cinematic storyteller at crucial moments to underline in an intentionally jarring mannerbecause achieved through the elegance and innocence of a waltzUncle Charlies horric identity. Thus, what is made ctional by these musical cues is not that Uncle Charlie is the murderer, but that the narrator is adverting to that fact, almost sardonically, both before and after it is narratively established. Nicholas Rays Rebel Without a Cause provides another example whose analysis helps us to see our way here. The opening scene unfolds at a police station, where three juveniles whose lives will soon importantly intersect nd themselves separately in trouble. At one point Jim Stark (James Dean), who has been talking with a sympathetic counselor, bangs and kicks a desk in frustration, at the counselors explicit invitation. As his outburst concludes, dissonant music surges up briey on the soundtrack. This undoubtedly adds tension to the scene, but does it contribute to dening the ctional world in any way? That Jim is wildly and angrily frustrated is fully established by what the perceptual enabler of the lm has allowed us to see and hear of his outburst. What, then, is the music, which certainly seems to have narrative force, doing there in narrative terms? Perhaps this: it serves to get across the phenomenology of Jims feelings, giving viewers access to the quality of his outburst from the inside, supplementing the access afforded from the outside by the ordinary perceptual data of the scene. Suppose that is so. Then on the one hand, this could be construed as a subtle sort of making-ctional in the story, namely, making it ctional that the quality of feeling in Jims outburst was precisely such and suchthe quality the musical cue in question is expressive of. On the other hand, this could equally well be construed as a making-ctional concerning not Jim, whose emotional condition is perhaps overdetermined by other indicators in the scene, but instead the narrators stance toward the audience.

Film Music and Narrative Agency

167

That is, perhaps the cues cash value is that the narrator is inviting viewers to share in rather than merely observe what Jim was feeling, and as a consequence, encouraging viewers to adopt a sympathetic attitude to him. The cues narrativity, in other words, may be a matter of its denition of the ctional world of the lm, comprising both narrator and story narrated, rather than that of the story per se. Consider now the common use of background music to create atmosphere in a scene, but without attributing mental states to any character therein. Is there anything that can thus be said to be made ctional in the lm world? In cases where an appropriate atmosphere is created, i.e. one that seems consonant with the way the story is otherwise told, what is made ctional might be that the narrator wants the viewer to assume a particular mood or frame of mind as certain events are presented for perception. In cases, though, where the atmosphere created does not gibe with the style or tone of narration already established, then even indirect ctional generation of that sort may be absent. The musical creation of mood may then have to be understood not as a narrative action, but rather as one which aims to affect the viewer immediately in a way that has no ctional upshot. Where nondiegetic music adds atmosphere to a scene without plausibly making anything ctional in the lms world, simply producing a mood in viewers, it seems that responsibility for it, as for other nonnarrative, purely compositional elements of a lm, must rest directly with the implied lmmaker. Exploring the interpretive option just broachedof assigning musical cues to the implied lmmaker rather than the lms narrative agentwill be the focus of the remainder of this essay. But before turning to that I conclude this section with a brief look at narrative uses of nondiegetic music in Hitchcocks Vertigo. Vertigo boasts perhaps the greatest of classical lm scores, and its greatness as a lm is due, in no small measure, to that score and its masterful integration into the lm in almost every respect. The intrinsic interest and sophistication of Bernard Herrmanns score has been much discussed, but what is most striking about it in the context of the lm is how signicant a burden it bears for limning the mental states and traits of characters, by comparison with most other lms. Vertigo abounds in occasions where not only are viewers ctionally informed about the inner lives of the characters through soundtrack music, but the music is what in large part makes it ctional that their inner lives are to be so characterized. When Scottie (James Stewart) rst sees Madeleine/Judy (Kim Novak) at the rear of a restaurant in San Francisco, the music serves signicantly to

168

Music

characterize her for us and for him: if the camera movement toward Madeleine lets us experience the physical nature of Scotties immediate attraction to her, it is the music that most fully conveys the sensual mystery of the woman. This scene is instructive in other ways as well. Madeleine gets up to leave, comes toward Scottie, pauses momentarily, and is very noticeably framed and lit in proleshown, in effect, to best advantage. But who is doing that? The cinematic narrator, in order to indicate something about Madeleine and the overwhelming psychic effect she has on Scottie on rst encounter. The lmmaker, Hitchcock, cannot do thatthough he can do certain parallel things to Kim Novak and the set in order to bring about, on a ctional plane, the narrative result. The cinematic narrator is the one who, ctionally, showcases Madeleine, for our benet as trackers of the story, and then underscores this showcasing through the musical resources under its control, e.g. by crescendoing at the point of held close-up. After the crisis of the rst part of the lm, Scottie spends some time in a sanitarium, sunk deep in depression and aimless longing. Soon after his release, we are given a high pan over the front of Madeleines apartment building, as the love motifa four-note Tristan-like descending gureis sounded romantically by French horns. This foreshadows Scotties appearance in frame at the end of the camera movement, with Madeleine obviously in mind: he approaches a blonde woman in front of the building, about to get into what was Madeleines car, only to discover that it isnt her. The exact content of his hope and then disappointment is supplied by the musical cue. Scotties vertigo rst occurs in the lms opening scene, while he is hanging from a rain gutter, high above the city, having slipped in the course of pursuing a eeing felon. This is importantly recalled in the plots pivotal event, occurring halfway through the lm, which takes place at the Mission of San Juan Battista, from whose tower the real Madeleine, unwanted wife of Gavin Elster, will appear to have leapt to her death. As Madeleine rushes into the church, and Scottie begins to follow, Herrmanns music foretells the recurrence of Scotties vertigo: milder variants of the clash of tonalities which were heard in the [opening] rooftop sequence hint at the probable effect climbing the tower will have upon Scottie. The musical cue, it seems, generates
Bruce, Bernard Herrmann, 143. This scene illustrates nicely a narrative possibility mentioned above (see n. 21), whereby a cinematic narrator might be thought of as presenting story events, conceived of as already existing ctionally at a basic level, in a certain way, through a partial shaping of the event being viewed. Ibid. 173.

Film Music and Narrative Agency

169

the ctional truth, at the point it sounds, that Scottie is going to experience vertigo when he climbs, though he is not experiencing it now. In other words, that Scotties vertigo is coming becomes something that is to be imagined by viewers at that point in the lm. Alternatively, perhaps the truth is generated that Scottie knows it is coming, or is concerned that it might. In the lms nal scene, also set in this tower, the tremolo trills which are prominent during this, Scotties second ascent, suddenly cease, suggesting he has at that point overcome his vertigo and will be able to complete his trip to the top. At the start of the letter scene, the moto perpetuo string gures prominent in the opening rooftop scene recur, in an overwrought vein, accompanying Judys detailed recollection of the tower incident and her role in the deception perpetrated there. This underscores sonically how emotively charged the incident remains for her, and helps us understand why she is ultimately unable to carry through the writing of the letter of confession. In the famous nightmare sequence, the habanera music associated with Carlottaa dead woman with a tragic past with whom Madeleine appears to identifybecomes more discordant, almost parodic, through the addition of stereotypical castanets and tambourine, conveying unmistakably the intensity of Scotties oppression by Carlotta/Madeleine. But more specic psychological pointings yet have been laid at the door of the scoring in this lm, with some plausibility. According to one writer, the rather banal music that accompanies a walk taken by Scottie and Judy in the park adjacent to the Palace of Fine Arts, soon after he meets her and senses a kinship with the lost Madeleine, suggests Scotties feeling of dissatisfaction with this working-class version of the elegant, sophisticated woman of his memory. On a more general plane, Herrmanns music helps forge a connection between Scotties vertigo in the literal sense, that is, dizziness caused by heights, and vertigo in a metaphorical sense, that is, emotional and metaphysical disorientation, which in Scotties case results from his obsession with someone who in effect does not exist. While the dual nature of vertigoits involving both attraction and repulsionis realized visually through an unprecedented combined forward zoom and reverse tracking shot, musically it is realized through rolling arpeggiated seventh chords punctuated by harshly bitonal ones. In some subtle but undeniable way it is due to Herrmanns vertiginous musical cues
Even more conservatively, perhaps the only ctional truth generated is that the narrator is reminding us of the possibility of Scotties imminent vertigo, without it yet being ctional either that it is imminent, or that Scottie believes that it is. Bruce, Bernard Herrmann, 163.

170

Music

that it is ctional, in Vertigo, that the physiological and the psychic aspects of Scotties afiction are but two sides of the same coin. I have tried to show, through the varied examples in this section, the viability of a making-ctional criterion of narrativity for nondiegetic lm music. There is, I submit, an intuitive match between the concepts: any nondiegetic music we would regard as narrative in status is music that can be seen as contributing to making something ctional in the world of the lm and vice versa. IX Narration, though, however broadly construed and however subtly carried off, is not always the basic charge of nondiegetic lm music, and serving a narrative function not always the best explanation of its presence. I want now to consider lms where nondiegetic music is featured that appears not to be of a narrative sortwhere thus, in my terms, the music does not make anything ctional in the world of the lm and is not reasonably assignable to the lms internal narrator. Instead, the music seems best understood as directly at the service of the implied lmmaker. I begin with some lms that are in different ways intermediate or borderline in regard to the contrast I want eventually to draw. In Fellinis semi-autobiographical 8 1/2, Guido, a famous but oundering director, has gone to a fashionable spa to try to recover his mental equanimity and decide on a direction for his new lm. We nd him in a spacious bathroom, as Wagners Ride of the Valkyries begins on the soundtrack. There is a cut to masses of people taking the waters at the spa, walking in rows and carrying parasols, among whom Guido eventually takes his place and receives his allotted glass. We see a conductor conducting, though with no orchestra in sight, and later see that he is leading a small salon groupone that could not be the source of the music we hear in the form we hear it. That cue ends and Rossinis overture to The Barber of Seville immediately starts up, but with a robustness, once again, that surpasses the resources of the musicians visually established as present. The effect of both cues, it seems, is one of gentle mockery of the behavior and attitudes of the spas clientele. The musical soundtrack during this sequence is what one might call quasidiegetic. That is to say, the music can be thought to be audible in the world of

Film Music and Narrative Agency

171

the story, because it is ctionally grounded in an observable source, and even conrmed later as something heard by a character (as by Guidos subsequent whistling of snatches of the Rossini)but not in the precise form heard by the viewer, in respect of volume, instrumentation, or performance quality. The same quasi-diegetic status attaches to the music in the nal scene, the press conference-cum-party, designed to launch Guidos supposed lm, at the extravagantly erected Spaceship site. We hear Nino Rotas excited music, which begins with a variant on Khatchaturians Saber Dance, and eventually brings in almost all the other motives heard earlier in the lm, as Guido is mobbed by impatient questioners and alternately shielded or prodded by his handlers, all captured in swooping, restless camera movement. Once again, we are shown a small band set up on a platform, and can even observe at one point the synchronization of the soundtrack with the rhythm, visually apparent, of the bands drummer, but there is still a discrepancy between what we can hear of Rotas marvelous score and our sense of what sort of sound the band visually in evidence could have produced. So, does such quasi-diegetic music serve a narrative function? To the extent the music is considered nondiegetic, its function seems to be, in the rst scene, satirical commentary, and in the second, mood enhancement, both arguably from a point of view internal to the lm. So despite their peculiar status, these cues, insofar as they are nondiegetic, are plausibly ascribable to the cinematic narrator. They make things ctional: in the rst instance, that the narrator views the spa goings-on satirically, and in the second, that the narrator wants to induce a certain mood in viewers with regard to the nal episode. The soundtrack musics equivocal status as diegesis does not thus seem to yield anything correspondingly intermediate as regards narrative assignability. Another example of intermediate status occurs in the rogue auto scene in Hitchcocks North by Northwest, in which the villain Van Dams henchmen attempt to kill the hero, Thornhill, by forcing him to drive down a dangerous cliff road while completely inebriated. I would claim that the music of this scene not only generates tension and underlines the drivers state of drunkenness, but at the same time signals, through its jokey style and lighthearted character, the absence of any real danger for Thornhill. Is this then a communication from the cinematic narrator, or from the implied lmmaker? That is to say, is it ctional that Thornhill is not truly in peril, or at least that the narrator knows he is not? Or is it rather that Hitchcock is telling us, on the sly,

172

Music

that he does not intend to do away with his main character at this point? It is hard to say which, but in a lm whose borderline self-conscious or modernist character has often been remarked, this is perhaps not surprising. Most of the music in Peter Weirs Witness, composed by Maurice Jarre, functions in the by now familiar mood-setting, character-delineating, attitudeevincing, or thought-specifying way, and is unproblematically categorizable as narrative. It begins with oating, gently pulsating synthesized chords, as images of Amish farmers looking up out of elds, and buggies traveling down roads, occupy the screen. What is conveyed is a sense of harmony and awe, a sense of the homogeneous spirituality of the world inhabited by the Amish, especially as compared with the vulgar and violent world of Englishers (the Amish term for their secular neighbors). During a sequence in which an Amish boy in Philadelphias 30th St. Station gazes high above him at an erotic statue of two mythic gures in some sort of embrace, the pulsating music, in voicelike chords, comes back, suggesting his bewilderment and wonder at the statue and what it depicts. A variation of this gently pulsing music is prominent during detective John Books night of healingwith Rachel, a beautiful Amish widow, at his bedsideat whose farmstead he has ended up with a gunshot wound. The music serves to suggest the growing intimacy and spiritual bond between them. After the violent climax, in which Book manages to dispose of his corrupt pursuerswith the help of some Amish corn, providentially stored in a silothe pulsing music underscores the long, silent glances of farewell between the two protagonists, reafrming the essential goodness of their interaction, which stops poignantly short of actual love-making. In all the foregoing, the music is naturally construed either as establishing something about the characters or else as evincing the attitudes of the cinematic narrator towards themattitudes we are clearly invited to share. However, there is the virtuoso Barn Raising scene, located roughly in the center of the lm, to consider. This provides the occasion of the lms main musical cue: an extended piece, lasting about four minutes, on the order of Pachelbels Canon (that is to say, variations on a ground bass). The image track shows us wagons laden with supplies, coming together, people on foot congregating, getting ready to work, and then, in stages, the raising of the
See Wilsons discussion in Narration in Light, ch. 4. The particular visual look of these images is due to Weirs trademark use of idealizing telephoto shots.

Film Music and Narrative Agency

173

barn, beginning with walls assembled at an earlier time, and nishing with the whole superstructure in place. The music, by means of its unity, solid ow, and arching sureness of direction, admirably symbolizes the strength of the Amish and the spirit of life-afrming communitarianism exemplied in the activity of cooperatively building a newlywed couple the barn they will need to sustain themselves. What, then, gives any pause in regarding this cue as wholly narrative? Only this: the meter and rhythms of the music in this scene are largely and signicantly, though not slavishly or mechanically, synchronized with the actions visually depicted. The pace and pattern of the visual editing seem to respond, not so much to any internal narrative demand, but rather, to the steady progression of the music. The cue is not so much designed to esh out the scene as the scene seems designed to illustrate the cue. All told, this suggests assignment of the cues music to the implied lmmaker, as opposed to the internal narrator, since the artful synchronization noted is most naturally taken as an aspect of the aesthetic construction of the lm as the conjunction of an image track and a sound track, rather than an aspect of how the narrator is presenting, through resources available to him, the story. It seems plausible to regard the music of Barn Raising as attributable, at least in part, directly to the implied lmmaker. The main cue in Hugh Hudsons Chariots of Fire occurs near the beginning of the lm, accompanying a scene of athletes in training: two dozen or so men running along the ocean in gym whites, represented as the fty-yearold memory of one of the runners. Vangeliss synthesized music, a tune of simple nobility over a throbbing bass with snare drum-like accents, is heard throughout, as the credits roll. The cue lasts a few minutes, and the scene ends visually with the group of men cutting inland and returning to the grounds of a building in Kent, where they have gone to train in preparation for the 1928 Olympics. Now this cue may contribute in part to narrationunderstood as makingctionalby making more precise the state of the runners as exhilaration, as opposed to mere determinedness, or by evincing a narratorial attitude, e.g. one of condent control, or by indicating a mood the narrator would like to impose on the viewer, e.g. one of heroism. But there still seems to be a certain surplus value, as it were, to the cue. Those narrative ends do not appear to exhaust the functioning of the cue; its scale and expressiveness seem more than is called for with respect to those ends, imparting to the activity

174

Music

of jogging on the beach an almost godly aspect, without it becoming ctional in the story that such activity really has such status, or even that the narrator believes that it does. Instead, it seems tempting to regard it as attributable, at least in part, to the implied lmmaker directly: it appears to testify to the almost religious regard in which he holds the athletic efforts of those young Britishers of yesteryear. The emotive surplus value of this cue, as far as plausible narrative functioning is concerned, is what points, it seems, to the implied lmmaker as a locus of attribution.

X Having uncovered some cases of lm music with equivocal or partial narrative status, we are now ready to contemplate cases of substantially, perhaps wholly, nonnarrative lm music. My claim is that such music, which I characterize as additive (or juxtapositional) lm music, is attributable directly to the implied maker of the lm. Such music alters, often powerfully, the artistic content or effect of the complete lm, but it does not do so by nuancing narration, i.e. by making or helping to make things ctional in the lms world. As a rst example, consider Robert Bressons Mouchette. There is only one signicant musical cue in the lm, a segment of Monteverdis Magnicat. It is heard very near the opening, during which the titles are projected, and again at the end, when Mouchette, an abused country girl of 13 or so, commits suicide by rolling in a sheet into a pond and drowning. Lindley Hanlon gives a sensitive reading of the music in this lm that supports, I think, a largely nonnarrative understanding of it, an understanding that connects it rather more closely with the lmmaker than with the lms internal storyteller:
From Mouchette on, Bresson uses music only at the beginning or the end of a lm unless the source of the music can emanate from the space and situation of the lm narrative . . . It is a more subtle, less intrusive means on Bressons part of authorial commentary on the action of the lm . . . Recurring after Mouchettes death, the Monteverdi music seems to function as Bressons requiem for the girl, who has wrapped herself in shroudlike vestments . . . The words of the Magnicat afrm the possibility of another life after death and sanctify Mouchettes decision to escape from the despair of her own life.
Sound in Bressons Mouchette, in Elisabeth Weis and John Belton (eds.), Film Sound (New York: Columbia University Press, 1985), 32930.

Film Music and Narrative Agency

175

The music here is most plausibly assigned to the implied lmmakeras afrming the general possibility of grace as exemplied in the tale of Mouchette rather than to the lms relatively effaced internal presenter, especially as it seems to frame the ctional narrative from without, like a pair of musical bookends, as opposed to shaping it from within. Terrence Malicks extraordinary lm, Badlands, provides an outstanding example of an appropriated score, consisting mainly of extracts from Carl Orffs Musica Poetica and Erik Saties Trois Morceaux en Forme de Poire. This score also serves as one of my key examples of nondiegetic lm music that is not, in the main, usefully construed as narrative. Badlands, based loosely on the Charles Starkweather shooting spree of 1958, contains a partially unreliable narration, since two components of it, the image track and the voice-over narration by one of the main characters, Holly (Sissy Spacek), are at odds with one another (in some respects, only at certain points, and in other respects, throughout). Here, as is customary, the visual representation is taken to be the more truthful, on the convention that seeing is believing, and so when what is shown conicts with what is told, we are inclined to credit the former. Orffs and Saties music, I maintain, is characteristically employed in Badlands in a mode of distanced and reective juxtaposition to the story narrated, by an intelligence standing just outside that narration. It is not, in general, attributable to the lms narrating agent, but only to the implied lmmaker. To make this point I examine at length one particular cue. Fairly early on, we are shown Kit (Martin Sheen), the lms other main character, working cattle in a feedlot, after having been red from his job as a garbage collector. On the soundtrack is a striking, far from inaudible, portion of Orffs score, consisting of sharply rhythmic xylophone or marimba music, built on an exotic scale, having no obvious connection with, or ttingness to, gritty scenes of cows being force fed, almost expiring in the heat. That is to say, there is nothing in the character of the states of affairs depicted that the music could plausibly be thought to second, nor anything indeterminate about those states of affairs that the music might plausibly be thought to specify. Could it be narrative in the sense of expressing the cinematic narrators view of the situation depicted? This seems unlikely, if only because it is rather
In identifying this theme as of grace I of course rely on a knowledge of Bressons oeuvre as a whole, and of the artist implicit in that oeuvre, one with a deeply Catholic vision of the world. Chatman, Coming to Terms, 136.

176

Music

unclear what sort of attitude could be signaled by such music in relation to the events shown. In addition, the cinematic narrator, who often visually corrects or gainsays Hollys romantic and simplistic notions of what has transpired in her time with Kit, comes across as an agency too sober and straightforward, almost nonhuman in its detachmentconsider the odd montages of nature shots that occur occasionally during the lm, giving the impression of an iguana-eyes point of viewto be credited with a sentiment as quirky and mischievous as that expressed by this musical cue. Might the music be narrative in virtue of acting to characterize Hollys recollective impression of Kits job at the feedlot? Such a hypothesis is multiply problematic. First, we havent been given any reason to think the nondiegetic music is in the service of the voice-over narrator, but at most, the cinematic narrator operating from the point of view of or on behalf of some character; that is to say, there must be rather special indications, not here present, before we will think of nondiegetic music as a resource belonging to, rather than applied in elucidation of, a character in the story. Second, since there is reason not to regard the image track as an accurate version of Hollys memoriesit regularly outstrips, and occasionally contradicts, her verbal narration of what happenedthe ground for thinking of the soundtrack music as signifying Hollys impression of those sights seems lacking. There is little reason to think, in particular, that she ever visited Kit at the feedlot or witnessed the kinds of scenes we see on screen. Third, whatever attitude we found such music to connote, it seems not to be one we would ascribe to hazy-minded Holly while the thought of Kit at the feedlot was before her mind. This leaves as the only interpretively live possibility the assignment of the music to the implied lmmaker who, from a point outside both the story and its narration, has apparently added this music as a kind of counterpoint to the ctional drama. To what end? It is hard to say, especially without an interpretation of the lm as a whole, but possibly one of aesthetic embellishment, or derangement of the viewers moral compass, or refraction of the storys content in a distorting mirror, or external meditation on the lms happenings.
The sequences that make it clearest that the image track is not to be thought of as a reliable representation of Hollys occurrent memories are one in which we see Kit shoot a football and then hear Holly, a minute later, recount this event, and another in which we see Kit, trying to outrun his police pursuers, suddenly stop his car, get out, shoot its left front tire at, and then blatantly await capture, while Holly alludes to the incident, never observed by her, in a mode of speculation rather than reportage: Many times Ive wondered about why Kit didnt get away. He said he had a at, but from the way he kept coming back to that, I doubt it.

Film Music and Narrative Agency

177

Now the music of Orff and Satie is characterized in general by an intentional simplicity, a primitiveness of musical materials, and a studied directness of effect, and that employed in this lm is no exception. Thus perhaps the function of this music in the lmon a global plane, rather than scene-byscenecould be said to be a reection of the basic childlikeness and obliviousness to social reality of the two principals, and especially that of Holly, the verbal narrator. I think that is so, but for the reasons given above this music, and that aspect of the lms content, is best laid at the door of the implied lmmaker, rather than any agent internal to the narrative. My next examples come from Woody Allens Love and Death, whose appropriated score is derived entirely from the suites to Lt. Kije and Alexander Nevsky by Prokoev. The sleigh-like music from Lt. Kije starts up after Natasha (Diane Keaton) announces her engagement to a herring merchant, and extends through her riding off in a carriage and subsequent shots of Russian troops in training, marshaled to protect Russia against Napoleon. This music has a satirical effect, more properly attributed to Allen as auteur, than to Boris, Allens character, as narrator, or even the cinematic narrator conceived as encompassing Boris verbal narration. A farcical battle scene between Russian and French troops, shortly thereafter, is accompanied by the grim and heavy music for the Battle on the Ice from Alexander Nevsky; the mismatch is palpable, and the implied equation of the two battles laughable. Both the satirical intent inherent in this juxtaposition, and the frame of cultural reference with which it operates, seem to put it beyond ascription to either Boris or the cinematic narrator. The last example I discuss of a lm much of whose musical soundtrack is best seen as additive or juxtapositional, rather than narrative, is Stanley Kubricks A Clockwork Orange. The opening creditthe words A Clockwork Orange on a garish orange eldis followed by a close-up of Alex and his pals (droogs) in a bar that dispenses drugged milk (moloko), disposing its consumers toward acts of ultraviolence. Soon Alexs voice-over is heard, which establishes what we will soon see as Alexs recollections of his recent past. Walter Carloss synthesized music here is a slow-moving, quasiHandelian progression, with a hint of Dies Irae. It functions narratively in setting an appropriate mood, in suggesting something of the effect of molokodrinking, and in perhaps foreshadowing some of the grim doings the narrator, acting on Alexs behalf, has in store to present to us, in due course. But the appropriated music employed in the lm, notably that of Rossini and Beethoven, functions rather differently.

178

Music

Rossinis La Gazza Ladra Overture begins on the soundtrack as an old man is being beaten by Alex and his droogs, continues over a cut to another gang of youths assaulting a naked girl on a stage, leading to a ght between the two gangs, and covers the escape of Alex and his droogs from the scene by car, fading out only as they approach a house in the country whose occupants they are going to terrorize. There is no obvious narrative appropriateness to the music: it seems neither to convey information about the events shown, nor to suggest the narrators perspective on those events, nor to suggest an attitude that viewers should plausibly adopt toward them. I take it that the rst and third points will be granted without dissent; the second, though, might be supported further, as follows. If the claim of narrative function is to be sustained on that ground, it seems we would have to posit either a perversely inhuman cinematic narrator, whose lighthearted view of the proceedings is reected in the music, or else a psychologically more normal one who merely signals to us, through the music, Alexs perversely comic perspective on the violence he is perpetrating on others. The rst possibility strikes me as unmotivated, while the second, though more promising, faces the problem that it casts Alexs reactions on perhaps too high a level of sophistication. Thus we arrive, once again, at the assignment of this music directly to the implied lmmaker as interpretively the most reasonable option. As such, how does it function? Pasted onto the scenes of violence presented by the lms internal narrator, it invites us, at least initially, to see them as a joke, thus making us complicit in the mindless pleasure of Alex and his pals in inicting pain, in the expectation, presumably, of getting us to be even more horried when we realize what weve been duped into. Kubrick, and not the cinematic narrator, is addressing us directly through this odd and unsettling juxtaposition of music and story. A similar scene takes place in Alexs room at home, with two girls he has picked up in record shop. It is lmed in extremely fast motion, to the accompaniment of Rossinis William Tell Overture. Here both the fast motion lming and the superimposed frenetic music seem to reect the activity of the implied lmmaker, as opposed to that of the lms perceptual enabler or internal commentator. A related, though distinct, use of music occurs in a scene also set in Alexs room, to which he has repaired after the rst nights round of ultraviolence. He deposits things in his booty drawer, checks on his pet python, and puts the scherzo of Beethovens Ninth Symphony on his sound system. The music,

Film Music and Narrative Agency

179

here diegetic, is synched to a montage of close-ups of statue parts, as Alex imagines acts of sex and violence, recounted in voice-over. But this intrastory perversion of Beethoven by the protagonist echoes and parallels the implied lmmakers supercially warped overlaying of Alexs recollections of occasions of torture and fornication with Rossinis diverting scores. Near the very end of the lm, Alex is being questioned by a few intellectuals, including the writer he crippled earlier in the lm, about behavioral conditioning via background music. It is not too much to suggest that this scene obliquely raises within the lm the issue of lm musics legitimacy and role, and of its possible subversive effects, e.g. the undermining of autonomy or the blunting of rationality. This self-consciousness in the lm about what we may call nonaesthetic or incidental uses of music reinforces the assignment of additive, as opposed to narrative, status to the Rossini overtures appropriated by Kubrick for A Clockwork Orange. XI Though in many cases where nondiegetic lm music is more reasonably assigned to the implied lmmaker rather than the lms narrative agent, we nd that such music is being used ironically or satirically, e.g. as with Love and Death and A Clockwork Orange, it is important to remember that that is not the only possibility. The examples of Mouchette and Badlands, and in a partial vein, Witness and Chariots of Fire, illustrate as much. And we may also observe, at this point, that nondiegetic music is not the only music in a lm responsibility for which may redound, without intermediary, to the implied
A contrasting, rather more cynical, view of the mode of lm-scoring of which A Clockwork Orange was perhaps the pioneer is provided in this recent commentary: Faced with the task of differentiating their scenes of brutality and mayhem from all the other scenes of brutality and mayhem, lmmakers are using music to distance the viewer from violenceor to comment ironically on it. As the images get more explicit, the accompanying tunes seem to get more frothy. Everything from Bach to hook-laden pop-rock songs provides background for images of st ghts, shootings, stabbings and torture (Kenneth Chanko, Its Got a Nice Beat, You Can Torture to It, New York Times, Feb. 20, 1994). The scene may in fact be what Wilson calls a rhetorical gure of narrative instruction, something offered by the lmmaker to the viewer as a key to interpreting the lms narration generally. See Narration in Light, 4950. Another intriguing case is Slava Tsukermans Liquid Sky (1983), whose soundtrack employs an overmodulated synthesized harpsichord version of eighteenth-century composer Marin Maraiss hypnotically repetitive Sonnerie de Ste. Genevieve du Mont. It is not clear to meon the basis of a single viewing, many years agowhether that music belongs in the satirical or the non-satirical subcategory of additive lm music.

180

Music

lmmaker, and which may be read by us as a direct reection of authorial stance or personality. Jane Campions lm, The Piano, offers a case of lm music commissioned and composed for diegetic insertion in a lmMichael Nymans music for mute protagonist Adas pianismrather than nondiegetic accompaniment. The musics characterization of its ctional originatorAdais a function that can only be assigned, it seems, to the implied lmmaker, as the agent who has chosen the characters, their actions, and their traits, in constructing and arranging the elements of the lmic object as she has. How does the distinction we have been exploring, of lm music as additive versus lm music as narrative, relate to another standard classication, namely that of lm music as commentative? The answer is: not simply. The equation of narrative and commentative will not do, for two reasons. First, some music of clearly narrative function is not reasonably thought of as commentative, unless all information-conveying counts as commentary. Second, some additive music seems to supply a commentary, if oblique, on matters with which a lm is concerned. In light of this, we might distinguish between externally commentative music, assignable to the implied lmmaker, and internally commentative music, assignable to the cinematic narrator. Still, it is important to stress that musical commentary on the events of a ctional story as such, or the characters guring in those events, remains a possibility only for the cinematic narrator internal to the ction. Additive music, assignable to an implied lmmaker, might generate, as noted, a kind of commentary as well, but it could not be on the ctional events themselves, from a perspective internal to the ctional world, but at most on the representation of those events or on the signicance of events of that type. The implied lmmaker of a ction lm is not on the same plane as the events of the lms worldwhich are for him, as for us, ctionaland so his direct commentary on those events is not a coherent option. For instance, if the Magnicat cue at the end of Mouchette expresses Bressons attitude of consolation toward Mouchettes suicide, this has to be understood not as an
One critic has remarked on the music for this lm as follows: Both the orchestral and solo keyboard music suggest a modern minimalist gloss of Chopin and Liszt but spun off plain, abrupt folk tunes . . . the pianism suggests someone doggedly trying to speak through the keyboard . . . As distinctive as it is, the music is strangely cramped and emotionally arid . . . the solo piano passages sound too much like elementary practice exercises to soar into the stratosphere (Stephen Holden, New York Times, Jan. 30, 1994).

Film Music and Narrative Agency

181

attitude literally directed on the suicide of Mouchettean event in which Bresson presumably does not believebut instead as an attitude bound up with the lms representation of that event, or directed toward events of the sort represented by the lm. A standard function of nondiegetic lm music, we have observed, is to reveal, conrm, or make precise a characters feelings or attitudes toward something or other in the story. Such a function makes most sense in connection with a narrator, rather than an implied lmmaker, since it presupposes an agent on the same plane, ctionally, as the characters, whose existence the narrator credits, and whose lives the narrator selectively presents to us. The deliverances of narrative lm music seem to come from one who shares a world with the characters, rather than one who has invented them, and everything else in the ctional world, from whole cloth. On the other hand, another standard function of nondiegetic lm music is to bathe the incidents of a lm in a common atmosphere. The thematic, instrumental, and stylistic continuities typical of lm scores help to create a consistency of tone or feeling across the span of a lm, especially where the events presented are not very tightly connected in a dramatic sense. Thus this, rather than any narrative task, seems to be the main function of Rotas score for Fellinis Amarcord. When nondiegetic lm music has this function, it is more naturally ascribed to an implied lmmaker than to an internal cinematic narrator. Nondiegetic lm music bridging scenes of different character, say, or smoothing over large lapses of time, is of this sort. Such music, like the presentational, voice-over, and mind-over narrators in a lm, is understood primarily as constructed or arranged by the implied lmmaker in putting together the aesthetic object that is the total lm, rather than as something used or employed by the cinematic narrator in its different narrative capacities. Returning to the ve functions of lm music recognized by Copland, I would suggest that only twounderlining characters psychological states and sustaining and releasing tensionare clearly assignable to the cinematic narrator. The othersensuring continuity, providing background ller, and creating atmospherecan with equal, or more, justice be regarded as activities of the implied lmmaker, in that they seem aimed directly at the viewer
A function highlighted by Nol Carroll in Mystifying Movies, 21623; Carroll labels lm e music of this familiar type modifying music.

182

Music

as an aesthetic subject, at causing his or her experience to be a certain way, rather than at dening or delineating the lms ctional world. If we consider, similarly, the list of functions drawn up by Gorbman in her study of the operation of classical lm music, I would suggest that twothe signifying of emotion, and the referential and connotative cueing of narrativeare assignable to the cinematic narrator, while the remaining twothe provision of continuity and the achievement of unitymake most sense as assignable to the implied lmmaker. What of my own list of fteen functions of lm music, drawn up earlier, in Section VI? By present lights, I think they sort out as follows: functions (1), (2), (3), (4), (5), (6), (7), (8), and (9) are arguably narrative, in that they involve making something ctional in the lm, and so music functioning in such ways is assignable to the cinematic narrator. Functions (10), (11), (12), (13), (14), and (15) are arguably nonnarrative, and are often achieved through music of additive status, assignable only to the implied lmmaker. There is not, however, a perfect correspondence between the division of functions as either narrative or nonnarrative, and the categorization of cues as either narrative or additive, because a cue can have signicant functions of both sorts. What is true is roughly this: if a cue has signicant narrative function, whether or not it functions in addition nonnarratively, then it is a narrative cue, whereas if a cue has no signicant narrative function, then it is an additive cue. The question I have been exploring in the latter part of this essay can be put as follows: when is nondiegetic lm music primary a compositional element in a lm, at the command of the implied lmmaker, and when is it instead, or in addition, an instrument we imagine as at the service of the cinematic narrator, generating truths in the world of the lm, either about the story as such or about the act of its narration? But perhaps the same question poses itself, on close examination, for a number of other lmic elements viewed initially just as compositional, e.g. lighting or camera angle. When is the dim or ltered quality of light in a sceneas in Vertigo, when Judy reemerges into Scotties presence as Madeleinemerely a directorial choice and when a manifestation, as well, of narrative activity on the part of the lms internal presenter, showing things in a light they would not otherwise appear in? When is an off-kilter view of a man running across a squareas in The Third
Though in regard to the last of these, creating atmosphere, it was suggested earlier how this can in many cases be understood as having narrative status, if the atmosphere involved is one the lms narrative agent can be plausibly thought of projecting. Unheard Melodies, 73.

Film Music and Narrative Agency

183

Manjust a matter of the directors tilt of the camera in relation to the actor being lmed, and when is it to be regarded as well as connoting an intervention of the cinematic narrator, as showing us the character from an oblique perspective, with whatever that suggests about either the character or the narrators view of him? The issues addressed here concerning the interpretation of nondiegetic lm music resonate, I suspect, across the whole spectrum of meaning-making elements in lm.

10
Evaluating Music
Wagners music is better than it sounds. (Mark Twain) If it sounds good, it is good. (Duke Ellington)

The above opposed epigrams about musical worth neatly serve to introduce the sorts of issues I want to explore. Is Twain right, in terms of what he implies about the grounds for evaluating music, or is Ellington? If they are both right, how can we reconcile the apparent conict between the principles suggested by their respective observations? One possibility for reconciliation is that Ellingtons epigram is to be taken straight, while Twains is to be understood as tongue-in-cheek. Another is that Twain is alluding to secondary or sophisticated aspects of musical worth, while Ellington is focused on primary or elemental ones. A third is that their epigrams simply apply to different, and disjoint, spheres of music, Twains to the classical tradition, and Ellingtons to that of jazz. A fourth is that Twains observation reects the fact that, for some music, evaluations evolve over time,
First published in P. Alperson (ed.), Musical Worlds: New Directions in the Philosophy of Music (University Park, PA: Pennsylvania State University Press, 1998), 93108. An earlier version of this essay appeared in Revue Internationale de Philosophie 198 (1996): 593614. These epigrams are not, of course, strictly inconsistent with one another. For Twains epigram does not even entail that Wagners music is good. But suppose, for the sake of argument, we take Twain s remark to have the force, that there is music that is good, but does not sound good. Even so, there is no conict per se with Ellingtons dictum, but only with its converse, to wit, if music is good, it sounds good. However, if we construe Ellington as implicitly committed to the biconditional, which seems reasonable, then a contradiction emerges. Otherwise put, if we elaborate the principle behind Ellingtons remark to be music is good just insofar as, and to the degree that, it sounds good, and that behind Twains remark, taken straight, to be music is not always good just insofar as, and to the degree that, it sounds good, then the logical opposition of these principles is plain.

Evaluating Music

185

so that impressions formed on rst exposure are replaced, on deeper acquaintance, by opposite ones, while Ellingtons observation underlines that, for much music, the ears rst impression is a pretty reliable guide to musical worth. I will forbear trying to decide among these possibilities for reconciliation. I will, however, assume there is some truth to Twains remark underneath its display of wit. We shall see a reection of this shortly in our discussion of whether all the value of music as music can be encompassed under the rubric of how good it sounds, or even that of how rewarding it is to experience.

I When confronting the issue of musical value, two questions must be distinguished at the outset. One question is that of the value of music generally. Why is any music valuable, and how does music, of any sort, add to or enrich human life? A particular form of this rst question, though one that takes it perhaps closer to the second question, asks what makes music distinctively valuable, as opposed to other arts; that is, what does musicmore or less any musicoffer that other arts or activities do not, or at least, not to the same degree? A second question is that of the value of particular pieces, genres, or styles of music, and is more obviously inherently comparative in nature. What makes a given piece (genre, style) of music valuable, or alternatively, what makes this piece (genre, style) of music more valuable than that one? In even pithier guise, the two questions are, in effect, (a) what makes music a good, or contributory to human good, and (b) what makes this music good, or better than other music? Once these questions are roughly distinguished as above, an immediate concern is what is the relationship, if any, between acceptable answers to them? Does knowing the answer to (a) help with the answer to (b), or vice versa? Is there a constraint in either direction, so that, for example, what makes music valuable generally is also essentially what makes particular pieces valuable, or that what makes particular pieces of music valuable is, writ large or in some way summed, what makes music valuable generally? One possible relationship might be that a work that displayed in high degree the properties
See Stephen Davies, Musical Meaning and Expression (Ithaca: Cornell University Press, 1994), 2756.

186

Music

that make music valuable generally, or one that fullled well the function fullling of which makes music of value generally, was therefore a relatively valuable piece of music. But is that so? It seems unlikely that things are as simple as that. In any event, a clear case of divergence between the value of music as a wholeof the phenomenon or practice of musicand the value of an individual piece of music would be the opportunity music affords to bring people together in a social setting for shared experience and interaction. Call that the social value of music. This appears to be a value of music generally without being a differentiating value of any individual piece of music. Other candidates for values of music generally that are yet not such as to add much, if anything, to the comparative value of an individual work of music, would be these: serving as a vehicle of relaxation; serving as a distraction from practical concerns; serving as an accompaniment to and facilitator of activities involving bodily movement, such as dancing, marching, exercise, or physical labor. II In an important study devoted to the second of our questions, namely, that of the artistic value of individual works of art, Malcolm Budd offers a completely general conception of such value, meant to cover all artworks in all artforms. Budd aims to identify an artworks artistic value in such a way as to differentiate it from other values it may possess, e.g. as social record, religious artifact, nancial investment, or totem of prestige. Budds straightforward proposal is that the artistic value of a work of art, its value as art, is determined by or is a function of the intrinsic value of the experience the work offers. By the experience the work offers Budd means an experience in which the work is fully and correctly understood, its individual nature grasped for what it is. By the intrinsic value of the experience Budd means the value of having such an experience for its own sake, rather than for the sake of any effects or consequences the experience may engender. What is excluded from the artistic value of an artwork is thus (a) anything about the work not directly reected in the right sort of experience of it, and (b) any value of such experience that is not purely intrinsic, i.e. is in some way instrumental.
Values of Art: Pictures, Poetry, and Music (London: Penguin, 1995).

Evaluating Music

187

Budds proposal, applied to music, is thus in line with, if obviously a renement of, Ellingtons dictum that musics value lies, above all, in how it sounds when you listen to it. There is little point in gainsaying the healthy intuition behind both Ellingtons dictum and Budds proposal. Still, though Budds proposal seems roughly acceptable, there are at least three grounds on which it can be challenged. One concerns the presupposition of a viable division between the effects or consequences of an experience, and the parts or elements of an experience, a division that may be difcult to sustain. The second concerns the restriction to the intrinsic value of the experience a work offers as the sole gauge of a works value as art, a restriction that appears inadequately justied. And the third concerns the connement of artistic value to that which is manifested in or through experience of a work, a connement that seems at odds with certain rmly grounded judgments of artistic value. I elaborate on these in turn. The distinction between an experience and its effects, though unproblematic on its face, has some tendency to dissolve under scrutiny. Experiences often have no unequivocal beginning and ending points. They characteristically do not start up with the sharpness of a pistol crack, nor do they characteristically close with a full stop. Many experiences have indeterminate beginnings, and take shape slowly. Often, rather than ceasing abruptly, they simply fail to continue developing or ramifying, though exactly where and when may remain elusive. This blurriness-around-the-edges is evident enough with traumatic experiences, such as losing a loved one, but attaches, if less blatantly, to many more ordinary experiences, appreciative ones among them. Turning to the case at hand, the endpoint of the experience of a musical work in audition, for instance, is fairly fuzzy, with no clear cut-off between the experience itself and what might be called, ttingly enough, its echoes and reverberations. A second, more important, challenge to the formula Budd offers us is this. Even if we assume the artistic value of a work of music to lie wholly in the value of the experience, suitably demarcated, that the work offers, it may not be defensible to restrict that value to the intrinsic, as opposed to instrumental, value of the experience, that is, the value of the experience itself as opposed to its effects. Granting for arguments sake the workability of the division in questionbetween effects of, and elements in, a given experiencethere
It may be that the broader value these instrumental goods contribute tothat a life is a certain way or possesses certain featuresis intrinsic, even if they remain instrumental goods of the experience. (For discussion see Ch. 24 of this volume.)

188

Music

is insufcient reason to hold that only the intrinsic value of the experience a work offers is a measure of its artistic value, rather than what accrues in virtue of the experiences effects. Suppose a given work of music is valued in part because it gives insight, when properly experienced, into the character of romantic love or the inevitability of suffering. If so, would this count as an intrinsic or an instrumental value of experiencing the music? The best answer would seem to be that it is both. For it is intrinsically valuable to have such insights while listening, or in the course of subsequent reection on the music, but also instrumentally valuable, because the insights are thus, after all, acquired, something one can summon up for further use or benet. If something counts as an insight, and if music affords it, then it is then an enduring asset, something whose value goes beyond the connes of the experience in which it is acquired. The same could be said about music that gave one access to a point of view that had not previously been available to one; the value of this would seem to transcend the value of the experience of achieving such access through the music, and thus be partly instrumental. It is also natural to consider here the possible moral effects of music. If there are any such, and if they could be shown to issue with some regularity from the comprehending experience of certain music, then they would constitute an instrumental value of the experience of such music that would seem to be at least a candidate for inclusion in the musics artistic value. But is there not a Catch-22 of sorts lurking in regard to the putative improving tendency of certain music? It seems that either you already have a rened nature or developed moral sensibility allowing you to appreciate such music, in which case you will not improved by it, or else you do not, in which case you will not be able to properly appreciate the music, and so it will fail to have its proper effect. In other words, that you appreciate such music presupposes that you are a person of some moral capacity; thus, such music would seem to be powerless to transform you into that. A partial answer to this quandary is as follows. Even if you have to have certain minimal moral capacities in order to adequately appreciate great music,
Two essays on this topic are Colin Radford, How Can Music Be Moral?, Midwest Studies 16 (1991): 42138, and Donald Walhout, Music and Moral Goodness, Journal of Aesthetic Education 29 (1995): 516. The topic is also explored at length in Kathleen Higgins, The Music of Our Lives (Philadelphia: Temple University Press, 1991). Higgins makes a good case for recognizing a moral dimension to music, though her focus is on plausible ethical effects of whole genres and practices, as opposed to individual musical works.

Evaluating Music

189

calling as it does on emotional repertoire and practical insight as much as on perceptual ability, exposing yourself to such music plausibly helps to develop or reinforce such capacities, through providing a controlled arena in which such capacities to respond are exercised in a specic way, serving as a kind of touchstone of what it is to be human. In other words, it is true that you might already have to be disposed, in some measure, to a moral form of life, in order for great music to be of benet to you, but immersing yourself in it might still function, ceteris paribus, to make you more human than you were beforewithout, of course, guaranteeing any such result. The example of Hannibal Lecter from Jonathan Demmes Silence of the Lambs is usefully recalled here, for that unparalleled lm villain notoriously displays both the most malevolent cannibalism and the most cultivated appreciation of J. S. Bach. If one is so unkind as to observe that Hannibal Lecter is a ction, and a meretricious one at that, it remains true that the coexistence of moral turpitude and aesthetic renement in a given person is more than merely imaginable: Richard Wagner approached this, and many of the Nazis who later gloried in his music exemplied it rather fully. But such cases, ctional or actual, are largely a red herring in regard to claims of a moral dimension to music and the possible relevance of such to some musics artistic value. No one would claim that great music can, entirely on its own, make appreciators of it better people, nor that great music, however supplemented and seconded, is likely to make all appreciators of it moral. The most anyone can sensibly propose is that such music, properly grasped, exerts, through the attitudes or states of mind the music projects or the complexes of feeling it evokes, a humanizing and moralizing force, though one easily enough overridden or neutralized, and thus that, all things being equal, people exposed to such music tend to be morally better, more humane, than they would otherwise be. Of course even this remains unproven, but it is not to be dismissed out of hand. At the least, good music may help to remove barriers to moral education by increasing an individuals awareness of the subjectivity of others, which is
The examples of Hannibal Lecter and Nazi doctors enjoying their chamber music in the evening strike us so forcefully, I suggest, precisely because they are, in fact, exceptional. They violate an empirically grounded regularity of artistic taste comporting with some degree of moral awareness. While it seems almost natural for purveyors of gangsta rap, say, to be engaged in shady or illegal activities from time to time, given the thrust of the music and the implausibility of attributing it all to a persona unrelated to the rapper, it is more surprising when professional pianists and violinists are found to be so engaged. The exceptionof immoral behavior from devotees of ne musiccalls attention to the presumption that there is likely some degree of correlation between aesthetic renement and moral awareness.

190

Music

clearly a prerequisite to treating others as ends in themselves and taking their interests into account in deciding how to behave. Appreciation of good music may plausibly lead to a more vivid imaginative grasp of the mental life of others, a necessary condition for regarding those others in a morally appropriate manner. But does great music simply inform us of morally relevant data or acquaint us with morally relevant perspectives, thus supplying a necessary condition of acting morally, or does it in addition motivate us, to some extent, to act morally? If it does not, then it will not be, even all things being equal, a morally improving force, but only something that lays the groundwork for acting in a moral way. However, I think there is reason to believe that some music, at any rate, is not only of epistemic value, but also motivating in relation to moral life. In sum, then, it is reasonable to take a musical work to be greater, and greater as art, on the assumption that it has moral effects or tendencies of the sort postulated among those who experience it fully and correctly, leaving aside whether it can be shown to have them. And this constitutes a ground of value, or reason for valuing, distinct from that of the specic musical merits in virtue of which the work possesses such moral force as it does. This is not, of course, to license every demonstrable benet of the experience of a work as contributory to its artistic value. Being instrumentally benecial in the waysmoral, cognitive, and emotionalI have been discussing is arguably a part of arts proper purpose or missionunlike, say, the capacity of experience of a work to alleviate mental illness, induce sleep in the weary, or promote a sense of self-satisfaction. The instrumental benets to which I draw attention are consonant with historically prevalent intentions of composers for their works, as well as being implicitly involved in received judgments about greatness in art. Such benets rightly enter, it seems, into the assessment of a musical works goodness as art. There are surely other instrumental benets of the experience of music that could reasonably be reckoned relevant to its worth as art. Music can be instrumentally valuable in virtue of providing, through its sounding form, a paragon or practicum of how to move or to behow, in effect, to go on. Good music, adequately experienced, can serve as a highly abstract, though suggestive, design for living.
This relates to Monroe Beardsleys suggestion that music might reasonably be held to symbolize or exemplify general patterns of continuation, growth, or development. See his Understanding

Evaluating Music

191

Some music, correctly appreciated, may be instrumentally valuable in virtue of the inuence exerted on ones general outlook, enabling one to think or feel the world differently, thus enlarging ones life as a result. Such music does so by embodying a process of thought or a frame of mind, one that through attentive and sympathetic listening one is allowed to enter, and that might not be otherwise communicable, or communicable in as vivid and effective a way. Some music, nally, may be instrumentally valuable through contributing to the sense of self and the formation of individual personality. Musical works can arguably help to constitute and dene the self that attends to them, internalizes them, and identies with them. Some music may even have the disposition to produce such effects on personality as to count as transgurative. It is important to note that the objective value of a piece of music in this respect is distinct from its personal value in that respect to a given individual. Both are comparative, but only the latter is relative in a strong sense. The former concerns a pieces power or potential, among pieces of music generally, to contribute to self-denition and the like in virtue of its formal and expressive qualities; the latter concerns the actual historical contribution of a given piece of music to shaping some persons identity. The latter may be due, in part, to the former, but there are usually idiosyncratic factors at work as well. That some piece of music had a profound effect on Sam, and became a touchstone of his emotional and intellectual life thereafter, means that it is, unquestionably, a piece of music with personal value for Sam. But it may not, for all that, have much objective value in this regard, that is, it may lack signicant potential to so affect prepared listeners generally. III I turn now to my third challenge to Budds proposal about artistic value. Even if the artistic value of music is acknowledged to be centrally a matter of the value of experience of it, whether intrinsic or instrumental, that does not exhaust what such value comprehends. Why? Because certain things enter

Music, in Kingsley Price (ed.), On Criticizing Music (Baltimore: Johns Hopkins University Press, 1981). See Leonard Meyer, Some Remarks on Value and Greatness in Music, in Music, the Arts, and Ideas (Chicago: University of Chicago Press, 1967), 2241; Anthony Savile, Kantian Aesthetics Pursued (Edinburgh: Edinburgh University Press, 1993), ch. 6; and Higgins, The Music of Our Lives, ch. 6.

192

Music

into an artworks artistic value that are not reducible to the value of experiencing the work in the prescribed manner. The artistic value of a work of music, in other words, may quite reasonably outstrip its experiential value, even broadly understood. An important component of musical artistic value is what may be called inuence-value, the impact a musical work has for the better on the future course of music. Examples of works in whose artistic value there is a signicant component of inuence-value include Beethovens Eroica Symphony, Debussys Prelude to the Afternoon of a Faun, Schoenbergs Piano Pieces, Op. 23, and Stravinskys Rite of Spring. This component of the artistic value of music goes beyond the value, intrinsic or instrumental, of the experience it offers. It is reected in the works of value that the aforementioned works have spawned, the new avenues of musical composition they have opened up, or the unforeseen sorts of musical experiences they prepare the ground for, but do not themselves afford. When we identify musical works as seminal, revolutionary, or ground-breaking, and praise them as such, we are in the realm of artistic value I am labeling inuence-value. It may be suggested, though, that actual inuence-value is one thing, and potential inuence-value another. The former would consist in actual positive effects on the future of music, whereas the latter, by contrast, would consist in being such as to give rise to such effects, conditions of reception being favorable. In other words, a works potential inuence-value would amount to its having the capacity to benecially inuence the future of music through its directly appreciable artistic features, to wit, its form, expression, or technique. Having distinguished actual from potential inuence-value in this manner, one might then go on to suggest that only potential inuence-value, which ows directly from a works having the right stuff, is relevant to claims of artistic value. Yet it seems to me that, despite the validity of the distinction between them, actual inuence-value, and not just potential inuence-value, is properly accounted part of the artistic value of works of music, especially when such works are seen not in isolation, but as part of an ongoing tradition of music-making and musical thinking. No doubt having the right stuff, and
I am ignoring, for simplicity of discussion, positive effects on other arts or spheres of culture. But in principle there is no reason to exclude these as irrelevant to an assessment of artistic value. Also, in speaking of inuence-value I shall, unless otherwise indicated, have in mind positive inuence value, that of inuencing the future of art for the better.

Evaluating Music

193

at the right time, gures in a works artistic value, but so, it seems, does what becomes of that historically. Since actual inuence on the history of music depends, as we know, not only on the nature of the work and the relations it bears to its antecedents, but on a contingent degree of receptivity to and uptake of what it offers, we may need to recognize a measure of artistic luck in how much artistic value accrues to a work. For that is what actual, as opposed to merely potential, seminality, revolutionariness, and so on, require. But what of inuence that is freaky, undeserved, or even a result of repugnanceas when an inferior work, in an unpromising mode, prompts an artist to angrily create something superior in a wholly other vein? The sort of inuence that would seem germane to artistic value is where earlier art prompts emulation, adaptation, or further exploration in the same or related directionswhere it serves as an example and inspiration to later art, rather than merely a negative spur to it. Perhaps, then, what is defensible is only that some actual inuence-value contributes to a works value as art; perhaps only when such actual inuence-value is coupled with, or rests on, a works potential inuence-value does it do so. Still, in those cases, contribute it does. But suppose that, contrary to what I have just been arguing, it is really only potential inuence-value that can be held germane to artistic value, a works particular fortunes in the subsequent history of music being discounted as irrelevant. The existence of such inuence-value would still constitute a challenge to Budds proposal. For a works potential inuence-valueits power or propensity to alter the stream of musical culture for the betterremains something over and above the features of the work in which that power or propensity inheres, and is distinct from and not reducible to the value, whether intrinsic or instrumental, of experiencing the work correctly. Having said this much, inuence-value may now rightly be qualied as a secondary sort of artistic value, though a real one nevertheless. The reason is as follows. Such value is clearly parasitic on primary, experience-based artistic value, in the sense that it is value that accrues to a work in virtue of its issuing in or paving the way for other works that have a value beyond inuence-valuepresumably, experiential value. To have led to the creation of other works, but ones without notable experiential value, or ones with
Note also that inuence-value, of either sort, might very well reside naturally in a set of pieces, rather than any particular one of the set, though those individual pieces would then share or participate in the inuence-value of the whole. A good example would be the last ten or so of Mozarts piano concertos.

194

Music

only inuence value, would seem not to amount to artistic value of any kind. Inuence-value is thus like a promissory note that needs to be redeemed by the exhibition of subsequent induced works of positive experiential value. But such notes can be redeemed, and have been, time and again. Finally, are there other varieties of artistic value possessed by individual works of music, apart from inuence-value, ones that go beyond the value, intrinsic or instrumental, of experiencing the work comprehendingly? Here are some candidates: (a) Problem-solving value: part of the artistic value of a musical work might reside in the problems it solves subject to various constraints, formal or expressive. It is true that the way a work answers to problems set by its predecessors might enter into how it strikes an informed ear, with awareness of such solutions being intrinsically rewarding to sustain, but the fact of constituting a solution to a preexisting artistic problem of some importance would seem to be a ground of value in itself, not reducible to the value of a listeners awareness of such a solution having been arrived at. (b) Originality-value: part of the artistic value of a musical work might lie in its originality or innovativeness relative to its tradition. Now although this might again be encompassed in the experience the work offers to an informed ear, in that originality, a backward-looking characteristic, comes across when one holds a work up against its context of emergence, its prototypes and predecessors, while listening. However, the originality of a work vis-`-vis its backa ground, a complex relational property, is of value in a way that goes beyond the value of appreciating such originality. (c) Performance-value: part of the artistic value of a musical work might be as a source of pleasure for performers in negotiating its difculties, or as a vehicle for a performers display of emotionality or taste. Clearly, music might be especially enjoyable to play, or pose challenges to execution that are exhilarating to overcome, or provide unusual opportunities for selfexpression, without being particularly rewarding to listen to. Furthermore, some music might be said to be more valuable in virtue of allowing, more
See Stephen Davies, Musical Understanding and Musical Kinds, Journal of Aesthetics and Art Criticism 52 (1994): 6981; Anthony Savile, The Test of Time (Oxford: Oxford University Press, 1982). A similar brief might be lodged for constructional features of a musical work that are not appreciatively accessible, yet are partly causally responsible for features that are. One would have to make the case that there was value here that outstripped both that of the features made accessible to the ear and that derived from appreciative reection on the role played by such constructional features.

Evaluating Music

195

than most, for differently revealing interpretations from performance to performanceperformances that work out differently the relation between the musics form and content. So it seems that there certainly are other non-appreciative-experience-based sources of artistic value, apart from inuence-value. I will not, however, try to settle here exactly how many, or how important, those sources are. IV My aim in the remainder of this essay is to explore the primary, that is to say, intrinsic-experiential, value of a piece of music for a listener. I want to try to pinpoint what that fundamentally consists in, and what it might, most generally, be gauged by. As the preceding discussion has made clear, I hold that a signicant part of the artistic value of a piece of music, that is, its value as art, may be non-experiential, consisting in such things as originality, or inuentiality, or being a solution to a standing problem in a musical tradition. In addition, I hold that the experiential value of a piece of music as art may in principle go beyond what is intrinsically valuable in such experience, and properly reect extrinsic benets of certain kinds. My target here, however, is precisely that central dimension of musics value for listeners, namely, that which is intrinsic to the listening experience itself, wherein such experience is deemed worth having for its own sake. Even if we agree that the artistic value of a work of music is primarily given by the intrinsic value of experiencing the work with understanding, our criteria for assigning musical value are naturally apt to be considerably more concrete than that. Asked to defend the judgment that some piece of music was good as music, a listener will rarely, and certainly not only, submit that experience of it is of high intrinsic value. There seems surely to be a place for, or a role played by, various low-level criteria, such as attractive melody, interesting rhythm, intense expression, pleasing timbres, inventive harmony,
Another sort of value that might seem to gure in artistic value may be denominated composervalue. The composer-value of a musical work would reside in its functioning as a model for other composers of forms, techniques, or procedures for realizing valuable artistic ends. However, such value would appear to be pretty clearly parasitic on the value of works for listeners and performers, the intended recipients of music. That is to say, a musical work will be valuable in virtue of guiding and motivating other composers to further composition only insofar as it itself embodies value of some sort for listeners or performers; otherwise, that it prompts and enables further composition along the same lines would be, if anything, to its discredit. Note that this might rightly be taken to include, in addition to apprehension of music as it sounds, retrospective reection on music after audition is complete.

196

Music

intelligible overall form, and so on. Such criteria are, of course, invoked more frequently on the ground than is the inherent rewardingness of an understanding experience of music. Yet clearly, such low-level musical merits lack generality, and fail to be even presumptively good-making, taken by themselves. Music can be good, and easily so, without attractive melody (for example, Stravinskys Rite of Spring or Weberns Five Movements for String Quartet), without signicant harmonic invention (for example, much of Palestrina or Handel), without substantial rhythmic interest (witness Saties Gymnopedie No. 1), without pleasing timbres (witness Bachs keyboard music), without evident emotional expressiveness (as with Conlon Nancarrows studies for player piano), and so on. What is interesting is whether there are any defensible intermediate principles, so to speak, ones less abstract than that of the intrinsic rewardingness of the experience of listening, but not so concrete as those just recalled. Between the most abstract condition that might be taken to analyze the core of musics artistic valuethat experience of it be intrinsically worthwhileand the most concrete experientially satisfying merit features appealed to in casual justication of verdicts of goodness of musice.g. attractive melody, varied orchestrationmay lie midlevel principles that capture the ground on which such verdicts and justications rest, while providing a specication of general validity of what experiential musical value centrally consists in. One approach popular in philosophical aesthetics for almost fty years though it has somewhat wilted of late under attacks from particularists about evaluationis to identify general properties of works of art that invariably conduce to or underlie artistic goodness across the artsand presumably do so because such properties are inherently satisfying to experience, alone or in combination, under appreciative conditions. The modern source of this tradition of theorizing is Monroe Beardsley, who proposed that there were three primary canons of criticism: unity, intensity, and complexity. The claim is that insofar as a work exhibits one of these features it is ipso facto better as art than it would otherwise be, and that any of its features that appear to contribute to its value as art can be shown to be forms of, or to rest on, those three primary criteria: unity, intensity, and complexity.
These were rst enunciated in his Aesthetics: Problems in the Philosophy of Criticism (New York: Harcourt, Brace & World, 1958).

Evaluating Music

197

But there has been signicant skepticism over whether those three are always, in all circumstances, positive-tending features; whether there arent sometimes interactive effects among them, ones that prevent an increase in one such dimension from always being value-enhancing; whether those three are all the independent positive-tending features with regard to artistic goodness; whether there is any effective way to sum such criteria, so as to justify comparative rankings of works exhibiting the three primary criteria to different degrees; and whether the presence of such features, even in high degree, is ever sufcient by itself to support a judgment of high comparative artistic value. So rather than attempt rehabilitation of that approach, I want to pursue another tack, one that tries to bring into relief the distinctive grounds of value for individual musical works, as opposed to works of art generally, while at the same time hewing more closely to the intuition that artistic value is centrally a function of the worthwhileness of the experience it offers. What I have in mind are proposals that might naturally be regarded as specications of that intuition as it applies in the musical case. V Consider these two approving responses to music: (a) I like how it sounds; and (b) I like how it goes. While neither of these responses, which strictly speaking are simply subjective judgments of approval, constitutes an adequate basis for judgments of the artistic value of a piece of music, (b) both says more than (a), and also comes rather closer to being such a basis than does (a). And that is because (b) captures a feature fundamental to musics being good, at a level less abstract than that of experience of it being intrinsically rewarding. Someone offering response (a) commits himself only to approving, rst, the mere sonic appearance of the music, and second, an aspect of the music manifestable in the smallest perceivable doses. Someone offering response (b),
Some writers, notably George Dickie, have expressed this skepticism by doubting whether there are any strong, as opposed to weak, principles of criticism involving such criteria. Weak principles get one only to judgments to the effect that a work has some artistic valuewhich is, of course, compatible with its being a bad work of art. See his Evaluating Art (Philadelphia: Temple University Press, 1988). But see Frank Sibley, General Criteria and Reasons in Aesthetics, in J. Fisher (ed.), Essays on Aesthetics (Philadelphia: Temple University Press, 1983); and John Bender, General but Defeasible Reasons in Aesthetic Evaluation: The Particularist/Generalist Dispute, Journal of Aesthetics and Art Criticism 53 (1995): 37992.

198

Music

on the other hand, has gotten closer to the musical heart of the matter, since such a response reects enjoyment taken in musical progression, in how music evolves from point to point, whose most obvious manifestation is perhaps rhythm, and the way musics parts, and especially those of small span, are joined together from moment to moment. I like how it goes is a more telling specication of I nd experience of it intrinsically rewarding than I like how it sounds, because the latter, strictly speaking, indicates nothing of the essentially kinetic nature of musics basic appeal. I like how it sounds, one may observe, would apply equally well to an imaginary art of short-duration sound bursts, carefully synthesizer-designed for aesthetic delectation. The appeal of the products of such an art overlaps only slightly, I would claim, with the appeal of music, which is indicative of why how it sounds is less apt as an epitome of what makes music good than how it goes. I like how it goes connects with the absolutely crucial notion of following music. There is in fact a rough equivalence between I like how it goes and I nd the experience of following it intrinsically rewarding. To take satisfaction in some music is, above all, to enjoy following it, and its value as music is plausibly quite centrally its enabling an experience of following its evolution over time that is intrinsically rewarding. Leonard Meyers central idea on the topic of musical worth is very much in this vein. It is that music is valuable insofar as it sets up expectations in the listener for how it will proceed, and then subsequently fullls or frustrates those expectations in various ways, to varying degrees, and with more or less delay. The most valuable music, for Meyer, provides a structure in which there is delayed, but not indenitely postponed, fulllment of expectations, the gratication that results ultimately being the more valuable for the delay involved in attaining it. Clearly this is a further concretization of the idea that music is valued insofar as one relishes how it goes, or takes satisfaction in following it, one that gives it more specic content still. Unfortunately, and especially as it seems to privilege the unlikely or deviant continuation, it may be too specic, too limiting, to cover all ways that music may go, i.e. unfold or develop over time, that people nd intrinsically rewarding. Sometimes a continuation that just seems right, beautiful, or cogent, may not be particularly unexpectedmay not, in terms Meyer favors in a later formulation, impart a large amount of information relative to alternatives.
See Some Remarks on Value and Greatness in Music.

Evaluating Music

199

A related suggestion, in this middle terrain between the conceptually unassailable and the empirically false, which concretizes further the idea of enjoying following music as the key to its central experiential artistic value, was offered by Edmund Gurney. The idea is that music, or at least music of the sort Meyer has labeled teleological, that is, giving an appearance of purposiveness or goal-directedness, has experiential value almost entirely in virtue of the satisfyingness of its individual parts experienced sequentially, i.e. in virtue of the satisfyingness of its small-scale evolution or progression, from moment to moment. Now another such suggestion, on roughly the same level of concreteness, would bring in, nally, that on which I have so far been silent, namely, the expressive side of music, and locate the experiential value of music centrally in the satisfaction of apprehending and responding to musics expressive aspect. If the previous suggestion aims to capture most comprehensively what music offers on a formal level, as a process uninterpreted in other than musical terms, then the present suggestion aims to capture most comprehensively what music offers on the level of content, or as a process interpreted in terms of human life. For surely much of the interest of music is wrapped up in what it intimates of human gesture, feeling, and agency. But a principle more adequate than either of these may be arrived at by, as it were, putting them together. More than merely enjoying following the music in its concrete particularity or nding satisfying precisely how it goes, and more than enjoying perceiving and responding to the gestural, affective, and agential qualities that emerge as the music unfolds over time, what one nds intrinsically rewarding in the experience a good piece of music offers, and what perhaps most importantly determines its artistic value, is its very particular wedding of its form and content. That is to say, with a good piece of music one enjoys how it goes, to be sureits individual, temporally evolving formand again one enjoys what it conveysthe attitudes, emotions, qualities, actions, or events it suggestsbut above all, one enjoys
See his major work, The Power of Sound (1880). I here make use of conveys as a blanket term for musics relation to its content, broadly speaking, i.e. anything beyond the properties it possesses as merely a sequence of sounds. Conveys is thus intended to cover expresses, exemplies, represents, symbolizes, signies, suggests, even evokes; that is, it is being used to stand for the whole array of meaning relations music may exhibit. Much the same, by the way, is true of expressiveness and expressive, as those terms gure in later formulations; they too should be understood as covering dimensions of musical meaning beyond that which can strictly be denominated musical expressiveness, as I would analyze that notion. (See Chs. 5 and 6 of this volume.)

200

Music

and nds intrinsically rewarding the fusion of how it goes and what it conveys, the precise way in which what it conveys is embodied in and carried by how it goes. If we equate the how it goes of a piece of music with its congurational (or kinetic) form, and the what it conveys with its expressive (or interpretive) content, then the what it conveys in relation to how it goes of such a piece might reasonably be identied with its form or its content, more comprehensively viewedthat is, with its signicant form, or immanent content. Call this picture of things Model 1.

Model 1
(i) How it goes: how note follows note, chord chord, motive motive, phrase phrase, and passage passage: congurational/kinetic form. (ii) What it conveys: gesture, action, feeling, mood, emotion: expressive/ interpretive content. (iii) What it conveys in relation to how it goes: signicant form/immanent content. Of course, marking off congurational/kinetic form from expressive/interpretive content in this manner is not meant to suggest that these are standardly distinguished as such in concrete listening experience. Usually one attends to content-infused form, or formally-embodied content, rather than either form or content per se, though it remains possible, with effort, to focus abstractly on such form or content as such. The progression or movement of music is usually heard as both congurational and expressive, with those aspects fused together. Indeed, the boundary between purely intramusical relations, motions, and tensions grasped at the level of conguration (or kinesis) and gestural, affective, or actional contents grasped at the level of expression (or interpretation), is perhaps essentially blurred. The above model, then, just makes explicit the two poles, so to speak, of the object on which musical appreciation properly focuses.
Of course this is a familiar idea in the annals of aesthetics, if not put in precisely these terms. Certainly Croce, Collingwood, and Dewey defended versions of the intimacy of form and content in art, and Budd endorses the idea at several points in his book. A congenial development of the idea can be found in an essay of Richard Eldridge, Form and Content: An Aesthetic Theory of Art, British Journal of Aesthetics 25 (1985): 30316, though it is there elevated, mistakenly in my view, into an account of the notion of art itself.

Evaluating Music

201

What is here proposed as the main focus of musical appreciationthat is, what music conveys vis-`-vis how it goeshelps explain, I think, why the a prevailing attitude in relation to music one loves is the desire to experience it repeatedly and fully in actual hearing. For only through such experience, or the simulacrum of it that mental simulation can provide, can one access the specic fusion of human content and audible form my working formula points up. Merely abstractly recalling the form, or even reviewing it concretely but unrespondingly, does little to satisfy the distinctive hunger for a sorely missed piece of music. Even less does abstractly recalling the musics content, referenced in some manner or other, without retrieving and rehearing its specic sonic embodiment or vehicle. The above model, and the formula from which it is derived, is perhaps most adequate to the experience of music on the basic or ground level, as a succession of events of relatively short duration, each with a signicance of its own, more or less absorbing. But another, more complex, model also recommends itself to us, in which it is acknowledged that what I have designated the side of content itself admits of formthat is to say, a how it goesat a higher level, which form then generates in turn a further dimension of content, thus providing ultimately for the relationship between that content and its underlying form. Call this Model 2.

Model 2
(i) How it goes on an expressive level: how an episode of one expressive character follows an episode of another such character, and the pattern of this succession as a whole: expressive form. (ii) What it conveys in virtue of how it goes on an expressive level: dramatic content. (iii) What it conveys dramatically in relation to how it goes expressively: (global) signicant form/immanent content.
For simplicitys sake I am leaving out the fact that high-order, or dramatic, content depends on and emerges out of not only high-order, or expressive, form but also large-scale congurational form, that is, a piece of musics architectonic structure. But to acknowledge that in this model would reduce its transparency even further. (For more on such complications, but a defense of the primacy, nonetheless, of small-scale congurational form, see my Music in the Moment (Ithaca: Cornell University Press, 1998).) Dramatic is not an ideal name for all varieties of global expressive content, but dramatic here, like conveys and expressive earlier, is intended broadly, covering all sorts of content of a global sort that music may convey, in virtue of its overall form and its pattern of expressiveness as a whole.

202

Music

To complete the picture we would need to recognize that relationships between higher-order forms and contents and those on the ground level might also enter into the appreciation of music, that is, serve as potential foci of intrinsically rewarding attention. Thus, if from Model 1 we have three elements, congurational form, expressive content, and their resultant low-order signicant form/immanent content, and from Model 2 three elements as well, expressive form, dramatic content, and their resultant high-order signicant form/immanent content, then in theory there are another nine relationships that become candidates for musical apprehension or contemplation, and hence, loci of evaluative assessment. Of those, the relationships between expressive and congurational form, and between dramatic and expressive content, would seem the most relevant. But at this stage of reection we surely begin to lose our grip on what the experiential value of music mainly consists in. Thus I think we may safely ignore any further relationships of this sort, and rest with the two models already articulated, in which the central object of musical appreciation is identied as the relationship of content to form, on lower and higher levels, with such form and content themselves functioning, as well, as secondary and supporting objects of attention.

VI Conning our sights to Model 1, the simpler and more fundamental of our two models, it might seem that there is an asymmetry between the component labeled how it goes (congurational form) and that labeled what it conveys (expressive content), as potential loci of musical value. For the formermusical formappears to have an independent claim on listeners, that is, it is of interest to apprehend on its own, while the latterextramusical contentmay appear, taken by itself, to lack any such claim. Budd, for example, adopts such a viewpoint: our experience of the expressive aspect of music is not separable from our experience of the music, and we value the expressive aspect not in itself but as realized in the music. In other words, that a work of music has a congurational form that is absorbing, expressive meaning aside, is already a ground of musical value, such form being an appropriate, if limited, object of musical appreciation; but not so,
Values of Art, 152.

Evaluating Music

203

it seems, that a work has this or that expressive content, identied apart from the musical form that realizes it. But is this really so? I think not. What is true is that there is an asymmetry in degree in that a works expressive meaning abstractly identied is of less artistic interest, and less contributory to artistic value, than the quality of its congurational form or the manner in which the expressiveness emerges from said form, but it is false that such meaning, considered apart from its vehicle, is of no such interest, or that its nature has no bearing on the artistic value of a work. Unusual, rare, subtle, deep, profound targets of expression, if attained, make a musical work better, because such contents are themselves more rewarding to contemplate, engage with, or respond to. For instance, a musical work expressive of bittersweet melancholy, or communicating the gestures of resignation, is arguably artistically more valuable, all things being equal, than one expressing simple cheerfulness or conveying an ordinary sort of anger. Recall that for Budd, the question about a musical works artistic value is simply this: Is the experience the work offers intrinsically valuable? The question I have highlighted, by contrast, is rather: What is the nature of the experience musical works typically offer, so far as the intrinsic value of such experience is concerned, and what more specically are the dimensions of the experience in which such value resides? I have ventured to generalize about the value-relevant dimensions of the experience of music, and have concluded that they are irreducibly three: experience of congurational or kinetic form, experience of expressive or interpretive content, and experience of the embodiment or realization of the latter in the former. So if the experience a work offers possesses intrinsic value, this will be because the work is found inherently rewarding expressively, or congurationally, or in terms of its fusion of the expressive and the congurational, or in all three ways. Thus we can, if we like, now identify three prima facie criteria of experiential goodness in music: one, how rewarding it is to experience how the music goes, that is, how rewarding it is to follow as tonal process; how rewarding it is to register or respond to what it conveys; and how rewarding it is to experience what it conveys in relation to, or as embodied in, how it goes. Does it then follow that a work that rates highly in all three dimensions of the experience it affords is thus necessarily better than a work that rates highly in only two such dimensions? For instance, that a work with a distinctive representational content that exhibits a musically absorbing form and weds the two in a successful manner is necessarily better than a work with a more

204

Music

commonplace expressive content exhibiting musically absorbing form successfully wed to that content? The answer, I believe, is that it does not follow. This is because of the possibility, even likelihood, of interaction or interference between the dimensions of such experience. For example, it may typically be the case that works of markedly representational character, even when musically absorbing and exhibiting good integration of the representational and congurational, do not afford an experience on the whole as inherently rewarding as much music of nonrepresentational and only middling expressive character; the value of the appreciative experience of such a work is not the sum of the values of the three dimensions of appreciation considered in isolation, because in trying to realize them all in a single appreciative experience one nds that attention to the representational dimension harmfully competes with attention to the congurational, more so than when music has extramusical but not representational content. Still, though the value of musical experience will not, if this is right, be simply the sum of the value of the three dimensions of such experience that I have identied, those dimensions would seem to be the ones in which all such value resides, and the satisfyingness in those dimensions taken separately the only prima facie reasons to regard a work as musically good. VII How do the formulas and models of experiential value proposed above fare when held up against a piece of unquestionable and surpassing musical worth? By way of conclusion I choose, for brief examination in this light, Schuberts Piano Sonata in A major, op. post., D. 959. I suggest that this sonatas experiential value, that is, its central value as music, ts comfortably under the umbrella provided by those formulas and models. For, in short, its goodness as music is a matter of how satisfyingly it goes as a purely musical process, of how satisfying it is to engage with the content it conveys, and above all, of how satisfying it is to experience the way in which what it conveys is embedded in, intertwined with, and borne by how it very precisely goes. But let us try to identify, in regard to this particular musical composition, some of what, in particular, makes it so good. Now the schematic midlevel answer to the question Why is Schuberts sonata musically so good? is that the experience it affords, on a congurational level, on an expressive level, and on the level of their interrelation, is so inherently satisfying or worthwhile. On

Evaluating Music

205

the purely kinetic level, it is highly absorbing to follow, in its small-scale and large-scale movement; from beginning to end the ear is regaled with beautiful, cogent, and original forms, or ways of sounding in time. On the level of content, what it expresses or conveys is distinctive, intense, and well worth engaging with. And on the level of fusion, the specic manner in which the work musically conveys its content seems wholly compelling, exhibiting throughout an elegance and rightness of means to ends. Still, it will not be amiss to go into somewhat more detail in regard to the expressive and dramatic dimensions of Schuberts musical essay. I begin with the former. This A major sonata possesses an intensity and variety of emotional expression quite out of the ordinary, covering almost the full range of human feeling. It displays perhaps the greatest range of moods and affects of any of Schuberts works, yet manages to tie them all together into such a satisfying whole that listening to it is like living a human life in microcosm. What are salient loci of expressive goodness in this sonata? The following are four instances of distinctive expression, distinctively achieved, in most cases through music that is, in addition, itself entirely satisfying on a congurational level: (1) The uneasy stasis, suggestive of restless anxiety or obsession, at the beginning of the rst movements development, achieved through harmonic oscillation between C major and B majoras opposed to the more usual nonreversing journey from key to key. (2) The unique nostalgia of the rst movements coda, in which the movements opening rhetorical theme is recollected as if through a hazeits vital, almost peremptory force drained off, but its essential identity intact. (3) The unparalleled violence, approaching chaos, of the slow movements middle section, a remarkable evocation of someone going desperately out of control. (4) The charming perkiness of the scherzo movements main theme, which charm is based in part on the piano writings suggestion of string pizzicati.
The means here would include the congurational form basis of the musics local expressiveness, and the expressive form basis of the musics dramatic content, or global expressiveness. Some expressive forms, or patterns of expressiveness, are less satisfying than others, whether as forms per se, or in virtue of the sort of higher-order content they generate or fail to generate. The specic image of a man distraught to the point of tearing out his hair is, for me, almost inescapable.

206

Music

I turn now to dramatic, or global expressive, content. A striking example is afforded by the intense, soul-searching dialogue in which the musics persona is engaged at the climax of the second episode of the sonatas rondo nale, at measures 180-210. In this passage the musics persona seems to pose, then reluctantly answer, questions it would rather not face, but that can no longer be avoided. And a sympathetic listener cannot help but feel they are the life and death questions of his or her own existence as well. As for the source, formally speaking, of the passages dialogic quality, it is the stark alternation of treble and bass in the sounding of the passages sharply etched, individually expressive motifs. Another example of global expressiveness in this sonata is something almost impossible to put into words: it is reected in the impression a listener receives during its rst movement, and perhaps most pointedly, its opening statement, of being addressed by an adult, and as an adult, of reectively and unhurriedly being given the benet of someones wisdom and maturity. This is one of several quite singular attitudes of mind somehow communicated by this sonata in the large, attitudes that give it the sort of moral force I earlier speculated may belong to certain pieces of music. A nal example of global expressiveness in this sonata concerns a special sort of unity that some pieces of music attain. One kind of most valuable music is that which displays unity in such manner and degree as to stand as a powerful emblem of the unication of opposites and the reconciling of the diverse, not through the subduing or overpowering of one element by another, but through the evincing of a deeper relatedness despite supercial differences. Call unity of this sort transcendent unity, in contrast to the merely formal unity that is, in most cases, its substantial underpinning. Such emblematicity of wholeness as I have in mind is not just an abstract relationship that one may reect on intellectually, but rather something one feels or registers in the course of listening. Schuberts sonata displays that sort of unity, if any piece of music does. There is a striking thematic, rhythmic, and harmonic unication effected in subtle ways throughout the sonata, despite the rich surface variety of its four movements, which helps to generate the pronounced transcendent unity the piece evinces. Here I will just note some aspects of the sonatas formal unication, ones that seem to play some part in its achievement of the other sort of unication.

Evaluating Music

207

The opening six bars of the rst movement are echoed, in loose retrograde, by the closing eight bars of the fourth and last movement; the rst movements development section is derived almost entirely from the opening gure of the movements second theme; there is a clear reminiscence of the agitation of the slow movements middle section in the interpolated repeated-note triplet gures in the reprise of the slow movements main theme; the tensions resident in the rst and second movements reappear in the second half of the scherzo movements main section, employing a variant of the gure from the rst movements development; the opening motif of the trio section of the scherzo is a loose inversion of the rst movements opening motif; rising or falling semitone motion is an important element in the main themes of the rst, second, and fourth movements, and includes the mysterious alternation of A major and B-at arpeggios at the very end of the rst movement; nally, there is a clear locus of harmonic gravitation in the sonata constituted by keys a third up from the tonic. Of course elements of formal unity in a musical composition, even when widespread, do not inevitably issue in the sort of global expressiveness I have labeled transcendent unity, whose ultimate mark is a sense of transguring oneness in the listener, but I submit that in this case, they do. It would be quixotic to think that this incomplete and selective survey of loci of goodness in one very ne musical composition can conclusively establish the validity of the schematic midlevel formulas proposed earlier for that in which musical value residesto wit, congurational, expressive, and expressive/congurational satisfyingness, or at a higher level, expressive form, dramatic content, and dramatic content/expressive form satisfyingnessbut at least it has turned up nothing that fails to fall handily under them. In any event, it is only through such midlevel formulas, I think, that we may perspicuously bring together on this subject the view from above, in which musical value is understood in terms of the intrinsic rewardingness of experience of a work, and the view from below, in which musical value is seen in terms of familiar featuresones of sound, melody, rhythm, or mooddirectly cited by listeners as grounds for approving a work.

This page intentionally left blank

11
Musical Thinking
If a lion made music, would we get what he played? (pseudo-Wittgenstein)

1. It has sometimes been remarked that making musicthat is, composing, performing, or improvising itinvolves thought, or is a form of thought. If so, what is the nature of the thinking that goes on in making music? And what of listening to music? Is the experience of the comprehending listener also a kind of thinking? How does musical thinking differ from the paradigm of thinking, that is, the formulation and manipulation of thoughts in words? Can musical sequence itself, rather than the activity of producing or auditing it, be regarded as a kind of thinking? In short, is music thought? In the course of trying to shed light on these issues I will take as a springboard various remarks of Wittgenstein on music that are to be found here and there in his writings. I will also yield to the temptation to emulate, in a small degree, Wittgensteins elliptical, oracular manner, a manner particularly apt to the exploratory stages of a philosophical investigation, which is certainly the case here. Whether what results should be considered an homage, a parody, or some mixture of the two, I leave to my readers to decide. 2. It seems clear from a number of Wittgensteins remarks, especially ones directed to particular composers, that he was indeed inclined to regard music as thinking. In one place we nd the following invocation: The strength of the thoughts in Brahmss music (CV, p. 23). In another place we are told that one can point to particular places in a tune by Schubert and say: look, that is the point of the tune, this is where the thought comes to a head (CV, p. 47).

First published in Midwest Studies 27 (2003): 5968. The following abbreviations are used for the citations in the text: PI = Philosophical Investigations, 3rd edn. (New York: Macmillan, 1953); BBB = The Blue and Brown Books, 2nd edn. (Oxford: Blackwell, 1958); CV = Culture and Value (Chicago: University of Chicago Press, 1980).

210

Music

What is most striking about these observations is how natural it seems for Wittgenstein to think of music as a kind of thinking, how little in need of defense he appears to take that view to be. What if one invoked, by contrast, The strength of the thoughts in the cuisine of les fr`res Troisgros, or The e strength of the thoughts in Michael Jordans basketball playing? Would this seem as natural? Could we easily speak of a moment in Jordans progress to the basket, or of a dish in a ten-course meal at Troisgros, where the thought comes to a head? I suggest not. 3. In the Investigations and elsewhere, Wittgenstein remarks that one might describe the effect of a passage of music by saying Here it is as if a conclusion were being drawn (PI, p.182) There are a couple of things to note about this. First, Wittgenstein does not say that, in such a passage, a conclusion is being drawn; rather, it is as if a conclusion were being drawn. So far, then, we are in the realm of analogy or metaphor, or perhaps of the dawning of an aspect. Second, the character of some passages of music to which Wittgenstein is calling attention is specically that of seeming to draw or reach a conclusion, as after a period of reection; it is not the idea of merely concluding, in the sense of stopping or terminating. Compare the endings of Beethovens Piano Sonata op. 110 or Dvoraks Seventh Symphony, which seem to sum up and crystallize what has gone before, with the endings of, say, minuet movements from symphonies of the Classical period, even great ones such as Mozarts Fortieth or Forty-rst. The former have this special rhetorical character of concluding, whereas the latter have only the mundane character of coming to a closehowever satisfyingly. 4. Wittgenstein treats the phenomenon further in another place: If I say, for instance: here its as though a conclusion were being drawn, here as though someone were expressing agreement, or as though this were a reply to what came beforemy understanding of it presupposes my familiarity with conclusions, expressions of agreement, replies (CV, p. 52). What Wittgenstein is underscoring here about the appreciation of music is this. Music is not understood in a vacuum, as a pure structure of sounds fallen from the stars, one which we receive via some pure faculty of musical perception. Music is rather inextricably embedded in our form of life, a form of life that is, as it happens, essentially linguistic. Thus music is necessarily apprehended, at least in part, in terms of the language and linguistic practices that dene us and our world. But by the same token, should we not expect that our understanding of linguistic phenomena will sometimes be inected by our musical

Musical Thinking

211

understanding, especially in light of the fact that our musical capacities are awakened at least as early as our linguistic ones? For example, we may describe certain speech as sing-songy, a conversation as not having the right rhythm, and the papers at a conference as not harmonizing. Furthermore, in tonal languages, such as Japanese or Indonesian, the distinction between speaking and singing is to some extent effaced. Though language may be essential to the human form of lifewhereas music, though universal, arguably is not, since we can presumably imagine human life without music, but not without languageonce both are present their interpenetration is assured, and we cannot help interpreting the one in terms rooted in the other. 5. It is true that the question What are you thinking? most often elicits a verbal answer, such as Its going to rain, or I need to buy milk soon, or She is very attractive. But why not, on some occasion, a musical phrase, or even a particular rendition of a musical phrase? If someone asks me what I was thinking, can I not sometimes truthfully say the opening of Mendelssohns Violin Concerto? Could I not, in response, even whistle that opening, and in a particular way? Note that the former response would not be the same as saying I was thinking of the opening of Mendelssohns Violin Concerto. For of course anything might be an object of thought. But that doesnt make it into an example of thought. No, the Mendelssohn opening is what I was thinking, not what I was thinking of. 6. We say of some music that thought went into it, or there was thought behind it, and mean to contrast that with cases we might rather describe as thoughtless note-spinning. Is the distinction between thoughtful and thoughtless musicor thought-lled and thought-free musiccoincident with that between good and bad music? If not, it is probably not too far removed from it. When we estimate the quality of music we often refer to the mind that is revealed in it, the mind one comes into contact with in listening to it, the mind that is reected in it, and so on. Granted that there is more to mind than thinking, can there be less to mind than that? If not, then can we easily deny the label of thinking to music of any worth, given the mind that stands before us in sound when such music is played? 7. Wittgenstein remarks in several places that it is common to experience a musical phrase as a question. It is also not uncommon to experience another phrase as an answer. And experiencing music in such ways seems part of what it is to understand music. Note that this is a matter of phrases striking us as questions and answers on more than the purely musical plane; that is,

212

Music

we are here speaking of more than the sense in which one phrase can serve as musically the answer to another, in terms of completing its melodic arch or balancing its harmonic movement. Now, if two phrases of music strike us as having a more-than-musical question-and-answer relation, must there be a content to the question that the rst phrase seems to embody? In other words, must it be possible to say what exactly the rst phrase is asking? If not, then what does the claim that the phrases have a more-than-musical question-and-answer relation amount to? Perhaps just that they convey the character or physiognomy of questioning-and-answering, though without constituting a specic question-andanswer. 8. In The Brown Book Wittgenstein observes: . . . if repeating a tune to ourselves and letting it make its full impression on us we say, This tune says something, and it is as though I had to nd what it says. And yet I know that it doesnt say anything such that I might express in words or pictures what it says (BBB, p. 166). As we have already noted, very often music makes on us the impression of a communicative act, and more specically, a speech act or utterance. There is nothing more common than the sense that expressive music is speaking to us, and though the embedded claim is perhaps not to be taken literally, neither is it merely a weak metaphor whose cash value would be simply that the music seems meaningful, or that one gets something out of it. As Wittgenstein says, the impression of speech from music is so strong that we often feel impelled, however misguidedly, to try to ascertain exactly what is being said. And we are not satised, it seems, unless we can exhibit what is said in other than musical terms, and preferably verbal ones. But should we be thus dissatised? Is there a communicative medium that should be privileged above all others which help to constitute the lived world? One is reminded of the anecdote in which Beethoven, having played for some visitors his latest piano sonata, was asked, But what does it mean, Herr Beethoven?, to which his response was just to play the sonata over again. To require that musical thought, if it is to truly deserve that label, must be such that it can be rendered articulate or verbally paraphrased, would seem to smack of a double-standard. It would not impugn the claim to being thought of a stretch of discourse to note that what it conveyed could not, so far as we could see, be put into music. Why, then, should it be held to impugn the claim to being thought of a stretch of music that what it conveys cannot, in general, be put into words?

Musical Thinking

213

9. It is instructive to draw up an illustrated catalog of thoughtful actions that we can hear in musical passages, or that we can hear musical passages as instantiating: asserting (e.g. the opening of Schuberts Piano Trio, Op. 100 No. 2), questioning (e.g. the opening phrases of Beethovens Piano Sonata No. 18, Op. 31 No. 3), musing (e.g. Schumanns Des Abends), imploring (e.g. the ute introduction to Bellinis aria Casta Diva), angrily despairing (e.g. the opening of Mahlers Second Symphony), menacing (e.g. the opening of Stravinskys Symphony in Three Movements), defying (e.g. the opening of Beethovens Fifth Symphony), cajoling (e.g. the sixth part of Vaughan Williams Job: A Masque for Dancing), comforting (e.g. the rst section in moderate tempo near the beginning of Faurs Requiem), disapproving e (e.g. the orchestral interjections in the rst part of the nale of Beethovens Ninth Symphony), and even nose-thumbing (e.g. the opening of the nale of Beethovens Second Symphony). And there are passages in which can be heard meditating, applauding, bemoaning, heaven-storming, and so on. Can a medium capable of summoning up such a range of mindful actions be a domain in which thought is absent? 10. So one kind of musical thought is this: musical passages wearing an appearance of thoughtful acts, such as questioning, concluding, searching, and the like. But another kind, surely, is this: musical passages giving evidence of thought processes in their creator. Let us bring the contrast between these two senses of musical thinking into clearer relief. One is thinking that seems as if it is embodied in musical process, that is, thinking that the music itself strikes us as being engaged in, or perhaps, that we are induced to imagine that the music is engaged in. Two is thinking in the composer that we take to be implied by musical process, that is, thinking that the music betokens on the composers part. We might even go so far as to say, as Wittgenstein would urge us to do, that we directly hear the composers thought in the musical process. For we are confronted with compositional choices at every turn that we cannot but regard as manifestations of mind. Some examples of the rst kind of musical thinking, of which we had a number of illustrations earlier, would be where music seems to be embarked on reection, or to be lost in wonder, or where one musical phrase seems to answer the question posed by a preceding one. Examples of the second kind of musical thinking would be the assessment we infer Bach must have made in devising a fugue theme combinable with itself in counterpoint, or the judgment we suppose Mozart to have exercised, in composing a piano sonata, in

214

Music

designing a second theme whose character would contrast suitably with that of the rst theme, or the vision we understand Beethoven to have displayed in opting for a C-sharp rather than a C-natural in the fourth bar of the opening theme of the Eroica Symphony, setting up a tension exploited signicantly later in the movement. Again: embodied thinking in music is thinking we ascribe to the music, as something it appears to be doing, and has no identiable object, whereas implied thinking in regard to music is thinking we ascribe to the composer, and has a quite denite object, namely the evolving composition itself. 11. Yet possibly the most important way in which music is a kind of thought does not reside either in musics frequent suggestions of thoughtful actions, in or its implications of thoughtful fashioning on the composers part. It may reside instead in the mere succession from chord to chord, motive to motive, or phrase to phrase at every point in any intelligible piece of music, whether or not there is any suggestion of recognizable extramusical action, or any implication of specic compositional deliberation. Call such musical thinking intrinsic musical thinking. But why call such succession thinking? Obviously this is not enjoined by any rule of language. Still, musical succession has features that set it apart from succession in general. It is a purposive-seeming, goal-directed temporal process, an intelligent form of continuation in time, and one naturally subject to assessment in cognitive terms, such as coherence or logicality or making sense. In addition, it is succession that we know emanated from a human mind, and that we hear under the inuence of that postulate. If one insists that that is not enough for thinking, is one not just assuming that thinking is necessarily in words? And why should one assume that? Of course, if music be admitted to be thinking on the grounds just offered, the door is also open for dance, mime, and abstract lm to be considered thinking as well. But such an implication is not, I think, to be feared. 12. Let me attempt to trace the process of embodied thought in the rst movement of Beethovens Tempest Sonata, Op. 31 No. 2, one of the most rhetorical pieces of music in all Beethovens oeuvre. The movements opening gesture, a four-note rising motif in largo tempo beginning with an arpeggiated A major chord, has about it a pronounced air of uncertainty and wonder (ms. 12). It is followed by a descending allegro
The details of my account follow in broad measure the analysis given in Charles Rosen, Beethovens Piano Sonatas: A Short Companion (New Haven: Yale University Press, 2001), 16870.

Musical Thinking

215

motif in D minor which anxiously frets, ending in an adagio turn of questioning character (ms. 26). Next the largo motif returns, to be followed by a more excited variant of the earlier allegro D minor music (ms.716), whereupon the music gathers resolve in a passage in octaves (ms. 1718), before issuing in a full cadence on the tonic and the rst episode of pure afrmation, a declaration in the bass and in allegro tempo of the opening four-note rising motif, now wearing a minatory cast, but rounded off in the treble by a new motif, plaintive and supplicating. Move now to the beginning of the development, where the rhetorical character of the movement becomes even more pronounced. The opening fournote largo motif returns, with its initial arpeggiated chord extended, and is heard three times, each time outlining a different chord, ever more removed from the tonic (ms. 938). It is hard not to hear this sequence of returns as a deepening of the uncertainty and wonder expressed by the largo motif on its rst appearance, and the agitated minor key music that succeeds it as an exacerbation of the minatory proclamations of the exposition (ms. 99118). Consider, nally, an episode at the end of the development prior to the recapitulation, where the music takes on even more unmistakably than before the appearance of a mindful agent. After six measures of sustained chords and a descending passage in bare octaves (ms. 13342), which strike one as clearing a space for reection, there follows a recitative with all the earmarks of a soliloquy, in part due to the thinning out of the musical texture and the starkness of the melodic line that remains (ms. 14358). This utterance is at turns meditative, questioning, and anxiousthe last of these due to an eruption halfway through of the fretting motif from the sonatas beginning, though in muted form. As one attends closely to this movement, one cannot fail to be struck by the mind manifested in its progression, and more specically, by the series of communicative acts incarnated in the music itself. Yet whose mind is so manifested, one may ask? In one sense, it is the mind of the imagined agent of those acts, what one may call the persona of the music. In another sense, it is the mind of the composer, who has in effect constructed, or caused to emerge, the persona to whom the communicative acts heard can be directly attributed.
This is a good point at which to note that the opening arpeggiated chord of the movement, which recurs in various guises, is a sixthree chord, the type of chord that typically introduces recitative in opera, and which thus might be said to adumbrate the recitative that occurs later in the middle of the movement.

216

Music

But at this point an objection may be raised. Let us grant that the music of the rst movement of the Tempest Sonata exhibits a series of images of thought-lled actions. Does that show that the movement literally constitutes thought, or is literally a thought process? Here we must add that this movement of the Tempest Sonata is, of course, not a random or accidental concatenation of such images. It is, rather, a meaningful concatenation of them, one that makes sense to us, and one that induces us to imagine a mindful agent of those acts, of whose mental life the music then appears as the narrative. However, the objector continues, all that shows is that music, or at least some music, is the narrative of an imaginary thought process, not a thought process itself. But how much of a difference is that? 13. Consider now, in a change of gears, how the sorts of thinking involved in the activities of composing, performing, improvising, and listening to music saliently differ. One way they differ is this: following is a key idea in the last of theselisteningbut not the others. To understand music to which one is listening is, at bottom, to follow it, that is, to experience its evolution in an involved way, exercising certain perceptual abilities and emotional sympathies, anticipating and projecting that evolution, responding appropriately in the moment to each twist and turn. That following musicas opposed to mere listening, or half-listeningis a form of thinking is evidenced by the near impossibility of doing any other thinking, of an unequivocal sort, at the same time. Musical process absorbs and effectively lls the mind that attends to it with any seriousness. But following is not, it seems, of the essence of composing, performing, or improvising. Rather, determining that is, the determining of notes as constitutive of a workwould seem to be the essential activity of composing; interpreting that is, the interpreting in concrete sound of notes already giventhe essential activity of performing; and generating that is, the creating of music on the spot, subject only to relatively loose constraints, the essential activity of improvising. Now, on the one hand, these activities of determining, interpreting, and generating music might all be classed as productive, whereas that of following music might by contrast be classed as receptive, though that should not make one lose sight of the anticipatory and constructive element in the activity of following music by ear. On the other hand, the activities classed as productive
For further description of the activity of following as a core component of the appreciation of music, see my Music in the Moment (Ithaca: Cornell University Press, 1998).

Musical Thinking

217

in a sense also involve actions of following: the composer follows one measure with another as he composes, the performer follows his reading of one phrase with his reading of the next, the improviser follows what he has just played by playing something else. But those sorts of following are manifestly not the same as that involved in listening. In the one case what is central is the tracking of what already exists, whereas in the other case what is central is a bringing into being at each step. 14. As an illustration of the thinking involved in improvising, I turn to Stan Getzs solo on his famous recording of Antonio Carlos Jobims The Girl from Ipanema. Knowing Jobims basic tune, and hearing Getzs treatment of it, we marvel at the melodic, rhythmic, harmonic, and timbral possibilities that Getz brings out that we didnt suspect the tune possessed. I would single out for remark just the high melodic leap Getz takes about midway through the repeat of the rst strainwhere the lyric, signicantly, has the words sways so gentleand the playful one-long three-short rhythm he introduces in the refrain in place of the original dotted one. If this is not thinking in sound, then what is it? Surely music that in the span of a mere forty bars manages to suggest a whole way of beingfor my part, I have often wished to live some of the time as that solo soundscannot be music in which thought fails to be present. Note that with improvised music a distinction invoked earlierbetween embodied thinking in music, wherein music presents us with images of thoughtful actions of an imaginary persona, and implied thinking in music, whereby musical process betokens or signies thought in the composerhas almost no purchase. Is it Getz or is it the musics persona who exults in that high-ying turn mentioned a moment ago? Is it Getz or is it the musics persona who gets down in that one-long three-short rhythm? And is there much importance to deciding? Music, we may note, and improvised music especially, stands as one of the supreme exemplars of the fusion of inner and outer in mental life that Wittgenstein was at pains to underline, a fusion that in the last analysis invites us to transcend the opposition between inner and outer, a transcendence that was perhaps the ultimate goal of Wittgensteins philosophy of mind. 15. If music is in some sense thinking, then, as noted earlier, bad music should tend to have the character of bad thinking. One of the ways we mark out bad thinking is by the epithets dumb and stupid. Hence, if music is thinking, at least some bad music should wear the appearance of dumbness

218

Music

or stupidity. One clear example, I would say, from an early work of a great composer, Franz Schubert, is the beginning of the nale of his String Quartet No. 7 in D. Now why is this music so stupid, and thus bad? The problem, zeroing in on just the opening six bars, which are even repeated, is that it consists in two largely unrelated ideas, and more specically, of a rst idea of utter banality followed by a second idea which is an emphatic closing gesture, one entirely unjustied by the meager four-bar ditty that precedes it. Compare this, though, to a little known piano sonata by Beethoven, the Sonata No. 16, Op. 31 No. 1, in G. This music once also struck me as somewhat stupid, because of its quirky premature use of a closing gesture similar to that featured in the Schubert, and the similarly unpromising character of its melodic materials. On longer acquaintance, though, the music seems anything but stupid. The Beethoven is, despite its modest materials, miles beyond the Schubert in development, ow, and organicity. Of course, not all bad music is stupid, or bad because it is stupid. Some music is bad because it is bland, or bombastic, or bathetic, or lacks balance just to stick with b words. This is to be expected, since even if music is thinking, and thus sometimes bad in the way thinking is generally, music is other things as wellmovement, gesture, pattern, expression, narration, depictionand can thus exhibit failures in those respects, and not just fallings-off from the cardinal virtues of thought, such as cleverness or cogency. 16. Finally, what is the connection between Wittgensteins views on understanding musicthat it is manifested by a complex of behaviors, such as illustrative gestures, apt comparisons, suitable hummings, and appropriate movings to music, that its criteria are neither inner acts of comprehension nor articulate paraphrases of musical content but a range of outwardly demonstrable responses and capacitiesand the claim that Wittgenstein also endorsed, which has been my focus here, that music, no less than language, incarnates thought? Here is one way to articulate the connection or connect the dots: Both music and language are forms of thought. Understanding music should therefore be analogous to understanding language. The former, like the latter, is a matter of use, that is, of knowing how to operate with the medium in question in particular communicative games, in particular contexts. But knowing how
I here use the expression communicative games, instead of language games, to avoid privileging language over other forms of communication or meaning-making.

Musical Thinking

219

in regard to music, as with knowing how generally, does not consist in propositional knowledge but rather in behavioral and experiential abilities and dispositions. Hence if music is thought we should naturally come to understand it as we come to understand thought in words; not by learning how to decode or decipher it, but by learning how to respond to it appropriately and how to connect it to our lives. 17. Parallel to the question at the heart of this essayIs music thought?would be the question, Is speech thought? In other words, one might wonder whether a stretch of intelligible verbal discourse was literally thinking, or was instead only the expression of literal thinking, that is, certain occurrences or processes in the mind of the speaker. Wittgenstein, of course, argued that there is no reason to think of thinking as a purely inner process, of which our observable behavior, however intelligent, can be no more than the outer shell, and thus no reason not to recognize as thinking the normal deployment of language. But for those who balk at the idea even that intelligible verbal discourse is thought, that is, that thought has outer as well as inner forms, and who claim that such discourse only manifests thought, the central claim of this essay can be suitably recast. It becomes this: intelligible music stands to literal thinking in precisely the same relation as does intelligible verbal discourse. If that relation is not exemplication but instead, say, expression, then music and language are, at any rate, in the same, and quite comfortable, boat.

12
Musical Chills
1. Although the value of music goes beyond, in a number of respects, the pleasure it provides when appropriately attended to, the capacity to provide such pleasure is clearly a signicant part of that value. To be sure, there are values that attach to both individual pieces of music and music as a whole that do not cash out in terms of pleasures afforded listeners, but such values are not the focus of my present reections. When I speak of pleasures properly afforded a listener by music, a particular sort of attention to music should be understood to be involved. First, such attention is close and concerted. Second, such attention is locally focused, though global-context-sensitive. Third, such attention is aesthetic, or appropriate to music as art; in other words, attention carries to the role played by what is precisely heard in the music in the generation of any pleasure that results, so that the music serves, at least in part, as the object of such pleasure. Finally, such attention is stylistically and historically informed, if only on a tacit level, thus allowing each musical characteristic, whether formal, expressive, rhetorical, or representational, to be registered for what it is. 2. I recall here some obvious dimensions of difference among musical pleasures. Musical pleasures differ in how active they are, how intellectual they are, and how essentially physiological they are. Musical pleasures differ with
Earlier versions of this essay were Musical Frissons, Revue Francaise dEtudes Amricaines 86 e (2000): 6476; and Musical Chills and Other Delights of Music, in J. Davidson (ed.), The Music Practitioner (Aldershot: Ashgate Publishing, 2004), 33551. For both sides of the issue, see my essays Pleasure and the Value of Works of Art, in The Pleasures of Aesthetics, and Evaluating Music (Ch. 10, this volume). For a defense, inspired by Edmund Gurney, of the primacy of local focus in the comprehension of music by ear see my Music in the Moment (Ithaca: Cornell University Press, 1998). For more on this aspect of aesthetic attention, see my What Is Aesthetic Pleasure?, in The Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996). For further admonitions in this vein, see my Musical Literacy, in Pleasures of Aesthetics, and Stephen Daviess Musical Understanding and Musical Kinds, Journal of Aesthetics and Art Criticism 51 (1993): 6981.

Musical Chills

221

respect to intensity, e.g. they may be acute or mild; with respect to duration, e.g. they may be passing or long-lasting; with respect to durability, e.g. they may be one-shot affairs or eminently repeatable; with respect to communicability, e.g. they may be highly esoteric or widely shared. Musical pleasures arguably differ in the different values thereof, e.g. elevated or triing, in the different moral qualities thereof, e.g. humanizing or dehumanizing, and in their different social imports, e.g. solidarizing or exclusionary. Musical pleasures differ, further, in when they are taken, so to speak, relative to when the music to which they refer is heard. That is to say, though most musical pleasure, I would claim, arises from the real-time following of music in its formal and expressive evolution, some musical pleasure is anticipatory, preceding audition, and some musical pleasure is recollective, occurring after audition. One dimension of difference among musical pleasures corresponds to two contrasting modes of listening, or perhaps two contrasting stances towards the listening that is going on. On the one hand one may, without losing contact with the music in its full particularity, let a piece of music enfold one, envelop one, wash over one, so that one gives oneself over to it in a personal way, as to a lover, or perhaps a trusted therapist. On the other hand one may undertake to keep music at a distance, so to speak, observing its lapidary details, its emotional maneuverings, its dramatic gestures as something external to and apart from the self that listens. Each mode carries with it distinct sorts of pleasure, ones which, manifestly, are not easily combined on a given occasion. A question that naturally presents itself is whether there is a one-to-one correspondence between musical pleasures and pleasurable musical features. That is to say, is it the case that for every musical pleasure there is a musical feature such that the pleasure is a pleasure in that feature? To me this seems unlikely. Of course many instances will conform, and obviously so, to the proposition under consideration. For example, pleasure in the intricacies of Bachs counterpoint, or pleasure in the melliuousness of phrase in Mozarts
On the social import of jokes, which serve at the same time to bind together members of a given group while putting at a distance those of other groups, see Ted Cohen, Jokes, in Eva Schaper (ed.), Pleasure, Preference, and Value (Cambridge: Cambridge University Press, 1983); the lesson applies, with modications, to musical communities that form around a given work, genre, composer, or performer. On possible moral aspects of music see my Evaluating Music, and also Kathleen Higgins, The Music of Our Lives (Philadelphia: Temple University Press, 1991), Colin Radford, How Can Music Be Moral?, Midwest Studies in Philosophy 16 (1991), and Anthony Savile, Kantian Aesthetics Pursued (Edinburgh: Edinburgh University Press, 1993), ch. 6. See my Music in the Moment.

222

Music

late piano concerti, or pleasure in the sheer amplitude of Schuberts C major symphony. But what of the pleasure of being simply transporteddazzled, blown away, knocked off ones feetby music? Naturally one can always say that it was the music as a whole, in all its concreteness, that transported one. But that doesnt imply that there is any particular featuresay, formal perfection, sensual beauty, or expressive depththat has claim to being the object of the pleasure in question. Alternatively, one could postulate a sort of transportative virtueevidently possessed by the music, given it has transported oneand maintain that the pleasure of being transported is a pleasure taken specically in that power of the music. But that would convince no one. I conclude that even if most musical pleasures readily reveal themselves to be pleasures taken in, or turning on, particular musical features, logically this need not be true of all musical pleasures. 3. I come now to the primary concern of this essay, namely, a strange and strangely pleasurable response to music that is, I assume, familiar to most serious music lovers, and one that has interested this music lover for the longest time. It is the singular phenomenon of music-induced chills, or as I will also denominate them, frissons. How do such chills or frissons arise, why do we take pleasure in them, and what broader signicance or value might they have? My point of departure will be a recent empirical study of the phenomenon conducted by the neuroscientist Jaak Panksepp. The pleasure associated with musical chills is clearly of a sort we can label physiologically centered. That is to say, musical pleasure in such cases revolves around a particular physiological effect, in the present instance, the skinsuffusing chill in question, where such effect is an integral part of the pleasure experienced. And one reason the musical chill phenomenon is philosophically interesting is this. How can a mere physiological tingle or shiver, so to speak, a mere bodily disturbance, be of appreciative signicance? Any number of philosophers of art, most famously Nelson Goodman, have accustomed us to view as ridiculous, through the ridicule they have heaped on it, the idea that sensations might as such have a legitimate role in aesthetic response. What good is a mere sensation, even an agreeable one, in the context of art? What does it tell of or testify to? Does it inform us of some matter of artistic fact?
Recall Schumanns apostrophe of the symphonys heavenly lengths. The Emotional Sources of Chills Induced by Music, Music Perception 13 (1995): 171207. All page citations to this article are given in parentheses. See his Languages of Art (Indianapolis: Bobbs-Merrill, 1968).

Musical Chills

223

Does it illuminate some artistic relation of ideas? If neither, then consign it, if not to the ames, then at any rate to the dustbin of appreciative theory. Such is the prevailing Goodmanian wisdom on this score. But it is not one I completely share, hence my interest in musical chills. I turn now to Panksepps study, beginning with some of his attempts to characterize the target phenomenon: the tingly somatosensory feeling that can be evoked by certain kinds of music . . . (172); the provocative and often delightful bodily experiences that deeply moving passages of music arouse in many people (173); a bodily rush commonly described as a spreading gooseesh, hair-on-end feeling that is common on the back of the neck and head and often moves down the spine, at times spreading across much of the rest of the body (173). Panksepp notes that, despite its intriguing nature, the prickly skin response usually called shivers, thrills, or chills in English has not received the experimental attention it deserves. He observes further that people rarely discuss the experience, and there is no unambiguous referent for it (173). Taking up that last point, there is indeed a terminological problem for what we wish to discuss, in that none of chill, thrill, or shiver seems entirely apt to denote the phenomenon under investigation, each of those terms carrying connotations, whether of coldness (chill), or risk (thrill), or tremor (shiver), that are in some degree undesired. Possibly the term frisson, a partly nativized immigrant from French, is the best of the designations available for the phenomenon. At any rate, I will often speak in what follows of musical frissons, though in discussion of Panksepps paper I will call them musical chills in deference to his preferred term for the experience in question. Experiments were conducted on undergraduates at a small midwestern university in the United States. Panksepp employed as his test material popular music of the 1970s and 1980s, items having been proposed by his subjects themselves as chill-inducing. Thus with few exceptions, the test selections were songs, mainly of the soft and hard rock variety. Ideally, of course, one would have preferred that the experimenters had used textless selections devoid of programwhat Peter Kivy calls music alonebut it is no surprise that the design of the experiment, in which selections were elicited from the student population, did not conduce to that. Panksepp seems oddly unconcerned about the possible collateral effects of song lyrics, with the articulate ideas and sentiments they contain, on the phenomenon under study, but naturally it is a possible source of reservations about some of his results.

224

Music

Here is one specic experimental result, rather typical of the study as a whole:
The highest rate of reported chills was .5 chills/min/person for the beginning 3-min segment . . . from Pink Floyds album Final Cut, which, on average, yielded essentially the same number of chills as ones own song [that is, the selection provided by the subject himself/herself] . . . . . . it was clear that the majority of the chills to this piece occurred in response to the dramatic crescendo at the beginning of the second minute. (1789)

As regards the most important general result of the study, it would seem to be this: Overall, the data support the thesis that sadness or melancholy is an emotional dimension more signicantly related to the production of chills than is happiness (187). In other words, it is negative, rather than positive, emotion in music that appears more efcacious in inducing the chill experience. Yet clearly some positively toned music in the study was found capable of inducing chills. This prompts Panksepp to the following speculation, one that would, if sustained, preserve a role for negative emotion in the generation of chills in all cases: it will be worth considering whether the chills provoked during happy music are caused by segments where happiness and sadness are inextricably entwined in bittersweet feelings (187). I will later return to this conjecture, which I believe to be on the right track. 4. Panksepps studies targeted a number of different factors plausibly thought to bear on the incidence of musical chills. These included (a) the gender of the listener, (b) the degree of familiarity with the music, (c) the degree of liking for the music, (d) the emotional quality of the music, and (e) the dynamic and tonal contour of the music. Panksepp observed a strong correlation between chills and degree of both familiarity and liking, a strong correlation between chills and both rise in volume and rise in pitch, a fairly strong correlation, which we have already noted, between chills and music of sadmelancholynostalgic character, and a weak, but statistically signicant, correlation between chills and being female.
Panksepp expands on this thought later on in the following manner: happiness and sadness work together, and the most moving music allows the two processes to be blended in such a way as to magnify our sense of ourselves as deeply feeling creatures who are conscious inheritors of the tragic view (198). That is, female subjects were generally more susceptible to musical chills than were male subjectsor at any rate, they were more likely to report having them.

Musical Chills

225

In my view there are probable determinants of the chill experience that Panksepps experiments did not target, though they are ones that would admittedly be hard to investigate quantitatively. Two worth mentioning are (1) the musics neness of expression and (2) the musics temporal expressive shape. Panksepp seems not to have considered the possibility that neness of expression, whether glossed as depth or intensity or exquisiteness of expression, may be crucial in triggering the chill experience, rather than the expressing of negative emotion per se. Panksepp seems also not to have considered the likelihood that a pieces expressive structuring in time, that is, the pattern of succession of its individually expressive parts, what one might loosely label its expressive narrative, contributes importantly to its chill potential, with some sorts of succession, some kinds of narrative, being more apt to elicit chills than others. I elaborate on this observation later on. 5. Recall now Panksepps suggestion that music with the greatest capacity for inducing chills may well be of an emotionally hybrid naturethat it is music in which positive and negative affects are in some manner or other interwoven or combined. This connects to what I am inclined to propose, on the basis of my own musical experience, as perhaps the crucial determinant of chill-inducingness, namely poignancy of expression, or perhaps equivalently, expression of poignancy. A profound truth about life is that almost all situations or conditions encountered are of mixed character. One is cognizant of the bad, if only peripherally, even when rmly engaged with the good, and one glimpses the good even when caught up in the bad, intermingled as they are in virtually anything. Nor is this necessarily regrettable. For the mutual focusing of positive and negative elements that results arguably ends up enhancing the appreciation of whatever good is being enjoyed. The essential poignancy of human life, one may suggest, resides in its mixed nature, in the indissociable union of its joys and ills, the inescapable commingling of its pluses and minuses. Thus, were we to assume that the prime determinant of musical frissons was poignancy of expression/expression of poignancy, it would not be surprising to discover that the music most reliably able to induce such frissons
See Stephen Davies, Musical Meaning and Expression (Ithaca: Cornell University Press, 1994), ch. 7, for related reections. Similar thoughts are to be found in Nietzsche, as Anthony Storr observes: Nietzsche realizedno one more vividlythat the only life we know is constituted by opposites. Pleasure is inconceivable without pain; light without darkness; love without hate; good without evil . . . . This is why the greatest art always includes tragedy (Music and the Mind (New York: Free Press, 1992), 158).

226

Music

was not that of unmitigated despair, nor that of untroubled gaiety, but that in which there was some admixture of the two. 6. Here is a somewhat haphazard list of pieces containing passages conducive to the production of frissons or chills, at least in my experience: Brahms, String Quintet in G, Op. 111, rst movement; Brahms, Piano Trio in B, last movement; Brahms, Intermezzo, Op. 118 No. 1; Schubert, Piano Sonata in C, D. 958, last movement; Chopin, Prelude Op. 28 No. 6, Mazurka Op. 17 No. 4 and Etude Op. 25 No. 1; Scriabin, Etude Op. 42 No. 5; Sibelius, Symphony No. 5, rst movement coda; Faur, Violin e Sonata in A, rst movement; Franck, Violin Sonata in A, last movement; Schumann, Piano Concerto, rst and third movements; Saint-Sans, Piano e Concerto No. 2, rst movement; Poulenc, Sonata for Flute and Piano, rst movement; Mahler, Symphony No. 5, fourth movement; Mahler, Symphony No. 6, rst and third movements; Shostakovich, String Quartet No. 8, rst and second movements; Ravel, Gaspard de la Nuit, rst movement; Nielsen, Symphony No. 2, rst movement coda. From that list it might appear that only highly charged music of the Romantic or early Modern period is capable of inducing frissons in this listener, but that is not the case. Here are some other pieces, of earlier vintage, that have this power: Beethoven, Piano Sonata in E, Op. 109, andante; Mozart, Piano Concerto No. 20 in D minor, rst movement; Mozart, Piano Concerto No. 23 in A, second movement; Vivaldi, Concerto for Four Violins in A minor, Op. 3, rst movement; Vivaldi, Concerto for Two Violins in A minor, Op. 3, rst movement; Albinoni, Adagio for Organ and Strings in G minor; and somewhat surprisingly, the rst movement of Haydns sunny Piano Sonata No. 60 in C, where certain chromatic bridge passages in the exposition, recapitulation, and development usually produce the effect in my hearing of them. Passages with the capacity to induce chills need not even be heard, strictly speaking, for that capacity to be realized: it sufces in many cases for them merely to be run through vividly in aural imagination, courtesy of the mental
There is clearly a resemblance between this idea of the poignant as reecting the essentially mixed nature of what life has to offer and the traditional idea of the sublime as the correlate of an aesthetic response distinct from that of the beautiful and involving a fusion of pleasure and pain. Yet the experience of the sublime and the frisson-centered experience of poignancy cannot be simply identied, rst, because not all sublime experiences involve frissons, and second, because not all frisson-centered experiences exhibit the specic cognitive character of the sublime, in particular, the aspect of awe.

Musical Chills

227

CD player, for frissons to be produced. And it seems generally true that when one actively seconds or parallels music one is auditing, by a sort of inner singing, frissons are more likely to occur. If so, that stands as yet another of the many rewards of active, rather than passive, involvement with music, though the reward itself, as has been noted, might perhaps be described as passive in nature, being a sort of sufferance or submission. The opening movement of Brahmss String Quintet in G, Op. 111, affords me one of the most sustained chill experiences of any piece that I know. But why, exactly, do I particularly relish that chill, as it steals up my spine and pervades my body, suffusing it with a sort of oxymoronic warmth? It is, perhaps, inherently pleasurable, but arguably no more so than having ones hair stroked, or settling into a well-stuffed armchair, or consuming a nice mousse au chocolat. So why do I value it above, or at any rate, differently from, those other delights? What makes the musical chill I receive from this piece particularly welcome must be more than the mild corporeal pleasure it affords. That something more, in this case, and as far as I can understand it, seems to be an accompanying feeling of surrendering control, of letting go, of delivering oneself to a powerful force, a guide to the terror and mystery of existence. More generally, I suggest, chills of the sort in question announce themselves as the mark of a confrontation with some fundamental truth of life, bodied forth by the music that so moves us. Such chills are received not as mere physiological disturbances, but as ones fraught with signicance. The echoes of religious experience here are intended. As has often been observed, the greatest music seems to provide a passable substitute for the sacred, for those who nd themselves doxastically challenged in regard to the traditional demands of religion. For many music lovers the listening room is a kind of chapel, at least when certain items are on the order of service. Anyway, contrast that movement by Brahms with a roughly contemporaneous piece, the Russian Easter Overture of Rimsky-Korsakov. Though colorful, imbued with feeling, well put together, and possessing a comparable degree of forward momentum, the Russian Easter Overture is not, I suspect, a
According to anecdote, in composing this resplendent music Brahms was thinking of time spent at the Prater, an amusement park on the outskirts of Vienna, and the simple joys of life available there, including, as he was supposed to have remarked, the pretty girls. But that sounds perhaps too earthbound a note for the almost superhuman afrmation and exhilaration bodied forth in the music. Nietzsche realized that, for many people, the concert hall and the art gallery have replaced the church as places where the divine can be encountered (Storr, Music and the Mind, 155).

228

Music

piece likely to induce musical chills. Though enjoyable, and even absorbing, as music it is simply too supercial: its energy does not speak to or tap into anything profound in human nature. And so it does not summon from the depths, as it were, those frissons that seem so full of import. 7. I return now to Panksepps study, and specically, to Panksepps evolutionary speculations on the underlying cause and biological signicance of musical chills. My interest here is not so much in whether these speculations are well-founded but in whether, if they are, this must affect the way in which those who are susceptible to musical chills need regard them.
It is clear that people are most likely to have chills to music that has moved them in the past . . . However, since unfamiliar sad music was more likely to provoke chills than unfamiliar happy music, the evidence suggests that there are more primitive instinctual neuropsychic components that underlie the phenomenon . . . . I will argue that the chill ultimately reects a property of ingrained neural systems of our old mammalian brain that monitor emotions related to social proximity and separation. (195) . . . we presently know a great deal about the neural circuits for separation distress that lead young animals to cry out when they are lonely and lost . . . . Internal feeling of coldness and chills when parents hear separation calls may provide increased motivation for social reunion. Thus the separation call may have been designed, during the evolutionary construction of the brains emotional systems, to acoustically activate a thermally based need for social contact . . . Sad music may achieve its beauty and its chilling effect by presenting a symbolic rendition of the separation call (e.g. a high-pitched crescendo or a solo instrument emerging from the background) in the emotional context of potential reunion and redemption. (1989)

The issue I want to raise is this. Once we have scientic insight into the causesneurophysiological, biochemical, evolutionary, or what have youof musical frissons, what impact should this have on our pleasure in experiencing them or in being subject to them? Need such knowledge have a deationary effect, serving to undermine our satisfaction? Once we realize that the responses in question are just, supposing Panksepps speculations to be on the mark, a legacy of our evolutionary past, an artifact of a mammalian brain still sensitive to the separation calls of errant young, must we rationally cease to regard them as bearers of signicance beyond the biological? I think not. Supposing the underlying cause of musical frissons to be an approximation to the separation call of lonely adolescent mammals, this need not invalidate the other dimensions of such frissons in which their value for us seems to reside. For rst, the pleasure in being so affected

Musical Chills

229

by the music, whatever its remote causes, is real, and can be rationalized in terms of the beauty or depth or poignancy of the music to which the pleasure seems most directly a response. Second, that a response has certain underlying causes of an evolutionary sort does not preclude its also possessing for us a certain signicance, and possibly reecting recognition of something important about life as some music seems to embody it. But in fact the evolutionary account of our susceptibility to frissons proposed by Panksepp seems to me somewhat unlikely, at least as applied to the sort of musical frissons I have primarily in mind. For there are really two kinds of musical frisson that need to be distinguished. On the one hand there are those, of relatively short duration, that are essentially timbrally and/or dynamically induced, that is, produced by sound quality as such, and typied by the effect on many persons of a clear and strong soprano voice. On the other hand there are those, of relatively long duration, that are for the most part melodically/harmonically/rhythmically induced, that is, produced by sound structure as such. Structural features that appear to conduce to frissons of this latter, more extended, sort include certain kinds of melodic sequence, certain kinds of chord progression, chromatic intensications, pedal points, suspensions, delayed cadences, sustained tremolos, and melodic leaps. Thus, that the former sort of frisson rests on the precipitating musics resemblance to piercing mammalian calls of separation may have some plausibility, but that the latter sort of frisson has its roots there as well is rather less plausible. For the latter sort of frisson exhibits more in the way of temporal shape, of tension accumulated and discharged, of emotion built up and released. Frissons of that sort thus seem harder to ascribe merely to a particular color and volume of sound. 8. One researcher apart from Panksepp who has interested himself in musical frissons is the prolic psychologist of music John Sloboda. In one study Sloboda attempts to identify the structural features associated with various

Of course there is also another sort of frisson instrumental music is capable of inducing, that which relies on the force of external associations, e.g. of a patriotic or sentimental sort. Such would be the frisson produced in someone by the hearing of his national anthem or the waltz that was playing when he rst laid eyes on his future wife. But I am not concerned with that sort of frisson here, since clearly it is not rooted in musical sound or structure themselves. Of course most such features will not by themselves produce frissons, if only because the specic musical embedding of such features, and their interaction with other features constitutive of a given passage, is absolutely crucial.

230

Music

pronounced physical responses to music. The results of his study point to the following six features as strongly associated with musical frissons or, as he prefers to call them, shivers: melodic appogiaturas; melodic or harmonic sequences; enharmonic modulation; unprepared harmonic change; sudden dynamic or textural change; and early arrival of expected events. Appropriating a central idea of music theorist Leonard Meyer, Sloboda conjectures that what ties such devices together is that they all involve frustration of musical expectations. Though undoubtedly containing some truth, this conjecture seems an over-generalization. Consider just melodic sequences, which appear on Slobodas list of conducing features, or sustained tremolos, which appear on mine, neither of which seems to contribute to shivers primarily by countering expectations. In a more recent study, Sloboda suggests that it is not unexpected musical turns per se that conduce to emotional peaks such as shivers, but rather both the degree of unexpectedness of such events and the density of unexpected events in a given stretch of music. That improves the explanatory power of the conjecture, no doubt, but without making it wholly adequate. The reason, as I earlier hinted in discussing Panksepp, is that the explanation accords no place to the registering of the expressiveness of the music and the character of that expressiveness. Slobodas explanation of musical chills, in other words, though of a cognitive rather than an evolutionary sort, is insufciently cognitive, or more exactly, insufciently cognitively complex. No formula that is couched in terms of structural and expectational variables alone, and thus fails to acknowledge the role of perceived expressiveness, can be entirely predictive of frissons, especially the type of frissons that are of the most appreciative importance. To illustrate this point I venture an explanation of this more cognitively complex sort of my experience of the Scriabin Etude p. 42 No. 5, mentioned earlier. The tempo of the piece is fast, and the expressive marking is affannato, or breathless. The overall shape of the piece is roughly AABABA, and though the piece is passionate and agitated throughout, there is a signicant contrast in emotional tone between the A and B sections, with the A section being troubled and despairing in character, and the B section projecting, when heard in context, a more lyrical and hopeful state of mind.
Music Structure and Emotional Response: Some Empirical Findings, Psychology of Music 19 (1991): 11020. Musical Performance and Emotion: Issues and Developments, in S. W. Yi (ed.), Music, Mind, and Science (Seoul: Western Music Research Institute, 1999).

Musical Chills

231

Now when in this marvelous etude B rst succeeds A, my spirits momentarily lift, as I detect a will to cast off the cloud of doom conveyed by A, a striving toward something more positive, though with an undercurrent of anxiety that does not depart. But when B again gives way to Aor more exactly, when B is on the verge of giving way to a version of A even more desperate than before (ms. 2730)I am struck by a palpable sense of the hopelessness of the aspiration eetingly perceived in B. B then appears to me to be overcome by A, to be recaptured by it, and I realize that all is indeed lost, that there was never really any hope for this doomed passion, that despair has now uncontested domain. This is the juncture at which the occurrence of chills is for me almost inevitable; and then again, though less powerfully, at the second, less psychologically crushing shift, in the second half of the piece, from a more febrile version of B back to A (ms. 467). What I want to suggest is that only a quasi-narrative account of this sort, positing a sequence of half-conscious, semi-articulated thoughts taking the form of an emotional scenario, can adequately explain the power of music such as this Scriabin etude to induce frissons in a range of listeners. And though it would be quixotic of me to expect all listeners strongly moved by this music to confess to precisely the scenario I have sketched, I believe that the sense of the music that most such listeners would extract from their own experiences would conform, at least roughly, to such a scenario. This is a good point at which to acknowledge the important role that specic performance of a piece plays in the generation of musical chills, given a structurally based potential for that in the music itself. Undeniably, certain performances exploit the chill potential inherent in a piece better than others, through their specic shaping and pacing of the sequence of musical events in which chill potential evidently resides. In the case of the Scriabin etude just discussed, for example, the performance by Ruth Laredo is by far the most effective I am acquainted with in that regard, the performance of the Brahms quintet by the Juilliard Quartet is more effective than that by the Guarneri Quartet, and the performance by Emmanuel Ax of the bridge passage in the Haydn sonata noted above is decidedly more conducive to chill production than the equally ne but more hurried performances by Glenn Gould and Mikhail Pletnev. To investigate further the whys and wherefores of the differential chill efcacy of different manners of performing a given piece of music would, however, take us too far aeld.
Scriabin: Complete Piano Sonatas, Nonesuch Records.

232

Music

9. What lessons about the pleasurability and value of musical frissons might we now draw from our consideration of their causes and conditions? It will help to begin by comparing the musical chills with which we have so far been concerned, and which are ordinarily experienced with pleasure, with some other species of bodily chill. Consider the chill that shoots through you when you suddenly think, rightly or wrongly, that you have deleted a computer le on which you have been working for hours, or the chill that runs down your spine when you sense that a burglar has just broken into your house through a ground-oor window. Chills of this sort are obviously far from pleasurable, and not ones we take satisfaction in, though of course they have a certain practical value. What makes for the difference in pleasure or satisfyingness between such chills and the ones typically had from music? In broad terms, it appears to be the cognitive construal of the frisson, or the way the frisson is interpreted, that makes most of the difference. More narrowly, what makes the difference is the sort of perception underlying the chill in question and bound up with it. The computer-glitch and burglarentry chills are unpleasant because they reect the perception of something bad, whereas chills had in response to music are pleasant because they reect the perception of something good. However, that may not account for all the difference in the cases. For it is not clear that the chills in question are precisely the same, physiologically and phenomenologically speaking. If we may generalize, those which arise in the course of engagement with music tend to blossom more slowly, to suffuse the organism more gently, and to have softer edges, whereas those prompted by sudden perceptions of danger or loss seem more sudden and piercing. These latter are more truly chilling, as it were. Thus, even among bodily chills to which we are subject there may be qualitative differences that suit certain of them to a role in experiences of positive value, while others are excluded, on the same grounds, from so taking part. In any case, whatever their distinctive qualitative nature, the pleasurability of musical chills seems at least in large part dependent on what a listener takes such chills to signify. For example, if you take your chill to signify that you
At least they are so for the vast majority of music lovers who are subject to them; that is, they are experienced as hedonic in tone. I have discovered, however, that there are listeners who in fact do not nd chills from music delectable, but on the contrary, unpleasant. My speculation is that these are listeners of a certain personality type, who resist letting go or surrendering control in even the smallest measure. It is unclear whether for such listeners there is any scope for cognitive framing of received chills rendering them more palatable.

Musical Chills

233

have recognized the musics poignancy or neness of expression, then you will likely take pleasure in it, whereas if you take it to signify, say, that the treble is turned too high on your stereo, you likely will not. Well and good. But might the chill not in fact be valuable apart from the pleasure you take in it when you regard it as signicant, in virtue of what it might actually signify? The question before us, in other words, is that of the epistemic value of musical chills. Do they really have any such value, or are they merely physiological responses that we take pleasure in when we regard them a certain way, namely, as having epistemic value? Musical chills obviously have one kind of epistemic value, but it is not of a sort that answers to our purposes. That is, they indicate that the music being heard is, so to speak, chillogenic, or capable of inducing chills in listeners under certain conditions. Clearly, what we are interested in under the rubric of the epistemic value of musical chills is not that, but rather what such chills might possibly indicate beyond themselves, concerning either the music or the listener. There are, it seems, two possibilities for what musical chills might reliably signal. One, they might signal the perception of some notable feature of the music, such perceptions often occurring without the subject being clearly conscious of them. Two, they might signal the presence of the notable musical feature itself. In other words, they might directly testify to something happening in the subject, namely, perceptions at a more or less conscious level concerning features of the music; or they might directly testify to the presence of such musical features, somehow functioning as immediate registrations of them, in something like the way pains function as indicators of bodily damage. I am inclined to think that musical chillsor at any rate, many such chills had by experienced listenershave epistemic value in at least the rst sense. Such chills are the upshot of perceptions of expressiveness in music, and reliably signal those perceptions to perceivers, who are often only dimly aware of them and their contents. Such musical chills thus have intentionality of a complex sort: as corporeal markers of perceptions directed on something distinctive in the music, they are, in the rst instance, about the music-directed perception that is occurring, yet also, in the second instance, about the feature of the music on which the perception is directed.
Why not take the line according to which such chills directly represent valuable properties of the music? For it certainly seems as if such chills register something about the music rather than

234

Music

Earlier I gave an account of the sort of thoughts and perceptions involved in my hearing of a particular Scriabin etude, and which I maintained lay behind and were crucial to the chills I experience in listening to it. I now briey do the same for the Haydn and Brahms examples offered before, in further illustration of the likely epistemic value of at least some musical chills. As with the Scriabin, this will not be to say that the thoughts and perceptions adduced were fully formed on rst hearing of those pieces, or even on the rst hearing of them in which chills occurred. It is rather a matter of plausible reconstruction and articulation of what, prior to that, is more inchoately or obscurely apprehended. Naturally, however, once reconstructed and articulated in a form that seems to square with the underlying, more inchoate apprehension, such thoughts and perceptions are likely to become part of the explicit content of subsequent experiences of the music. Anyway, my hearing of the Brahms quintet includes a recognition, at some level, of the intense vitality embodied in that incomparable opening passage, and especially in the cello part, which strives heroically to assert its individual nature against the generalized, insistent sawing of the other strings. And my hearing of the Haydn sonata includes a recognition, at some level, of the subtlety and suppleness of the shifts in harmonic color in the course of that modulating passage, which displays the marvelous seamlessness of an organic process. These are highly distinctive features of the music in question, and my body, according to the present hypothesis, signals to me my emotionally fraught perception of them by the chills in which those perceptions eventuate, ultimately prompting me to articulate those perceptions on a more conscious level. But the chills, I suggest, are usually the rst sign that one has registered something of depth or signicance in the music. In other words, chills of the sort I have been discussing thus serve as focusers of attention, as direct aids to appreciation, drawing attention to expressive aspects of musical structure that might otherwise escape notice. A few more words are in order on the nature of the chills exemplied in my experiences of these pieces by Haydn, Brahms, and Scriabin. The chills in question are plausibly the upshot of perceptions of emotionally signicant patterns in music, which perceptions have themselves an emotional character.
something about me. The problem is coming up with a plausible mechanism as to how chills could be directly sensitive to such complex properties as neness or poignancy of expression. A two-stage model seems more promising, as it does for, say, cold-induced chills, whereby shivering from cold directly indicates coldness of the body (internal temperature), and then indirectly, coldness of the environment (external temperature).

Musical Chills

235

These emotion-imbued perceptions reach a level of intensity sufcient to issue in a physiological responsethe famous chillwhich we can now understand as a kind of overow of the emotional perception developed to that point. Such a chill is thus not merely an indication of an emotional perceptionas would be, say, a twitching of ones left earlobe or a blue ash in ones visual eld, if regularly conjoined with such perceptionsbut the natural culmination of the emotional perception itself. 10. What position have we arrived at, then, on the appreciative value of musical chills? It would seem to be this. First, musical chills can be sources of aesthetic satisfaction insofar as we take them to signal perceptions of something signicant about the music to which we are listening; and second, musical chills can have epistemic value insofar as they actually signal that we have indeed perceived something signicant about the music, often serving to rst alert us that we have done so. But the Goodmanian qualm may now reassert itself. For it may seem that it is not, after all, chill sensations as such that have appreciative value, but only the perceptions to which they are connected in favorable cases. And if so, one might be tempted to insist that it is really the music-directed perceptions that count appreciatively, not the chill sensations, thus ultimately acceding in, rather than diverging from, the Goodmanian position according to which chills as such are both without value and irrelevant to value. If musical chills are only appreciatively valuable when functioning as signalers of perceptions regarding the music, then does not all the value in fact reside in those perceptions, whether or not chills accompanying them take place? Not quite. For what is important, and what registering chills while listening underscores, is the response to music of, so to speak, the whole person. Responses to music of that sortones that are cognitive, emotional, physiological, and behavioral at onceare arguably of greater value than more limited or restricted responses to music being heard. Thus even if such chills are valuable only when had in connection with certain perceptions regarding the musics content or structure on the part of those who have them, the chills remain essential to the full value of the experience in question. The chills are essential to the full value of the experience, one might say, because they represent a bodily conrmation of what is registered at the same time, usually in an obscure manner, by the mind. Arguably, the marking of perceptual apprehensions by felt frissons imparts a kind of added value to musical experience. Such frissons stand as corporeal endorsements of what is concomitantly grasped in more cognitive terms. A central appreciative value

236

Music

of such frissons, I suggest, lies precisely in the afrmation of wholeness that they afford, of mind and body resonating together in response to a given musical utterance. There is thus something right about the idea that not to feel frissons or chills at certain junctures in a piece of music is to be missing something appreciatively. For not to register such effects is to have an experience of the music that is not as responsive, in some sense, as it should be. The underlying rationale may be that, as embodied creatures, the responses we most admire are not those in which only mind, or only body, are involved. To be most human, it appears, is to react to things, and perhaps especially works of art, with our whole selves.
On the account that has the chill directly signaling the perception of a valuable musical property, such as the musics embodying of a particular emotional scenario, the force of the idea that the chill in seconding the perception serves to make ones response to the music more whole may appear undercut, since on that account the perceptions object is the music, while the chills object is the perception and not the music, so that perception and chill are not clearly responses to the same thing. But this appearance can be mitigated. For the chill, though directly signaling the perception, still indirectly signals the musical property, and thus the idea can be preserved that there is effectively a combined cognitive and physiological response to the music. There may be a parallel here in the moral sphere, where a person who can, say, classify acts into good, bad, and heinous, but does not experience any revulsion in contemplating acts of a heinous sort, is regarded as somehow morally lacking.

PART I I I PI C T U R E S

This page intentionally left blank

13
Wollheim on Pictorial Representation
I In On Pictorial Representation Richard Wollheim offers us an elegant prcis of the account of pictorial representation he has developed over the past e thirty years. In addition, he comments on competing views of the matter, and responds to criticisms or requests for elaboration that his own view has elicited. Though I eventually express some reservations on the view myself, the extent of my accord with Wollheim on this topic is rather large, as I now indicate. First, I agree with Wollheim that the concept of pictorial representation, or depiction, cannot be explicated without appeal to a characteristic sort of experience, the sort of experience Wollheim has denominated seeing-in. Sustaining an appropriate seeing-in experience, that is, a seeing-in experience that conforms with the artistic intention governing a given picture, is what is criterial of such representation, and not anything else. Second, I agree with Wollheim, as against Budd, that seeing-in is generally prior to, and not to be analyzed in terms of, the perceiving of resemblances as such, whether between objects or experiences. The fundamental rationale for so insisting is this. Though perception of resemblance, or more narrowly, structural isomorphism, between object aspects or visual elds, may be a concomitant, trigger, or consequence of

First published in Journal of Aesthetics and Art Criticism 56 (1998): 22733. Richard Wollheim, On Pictorial Representation, Journal of Aesthetics and Art Criticism 56 (1998): 21726. Notable bulletins in that development include Seeing-as, Seeing-in, and Pictorial Representation, in Art and Its Objects, 2nd edn. (Cambridge: Cambridge University Press, 1980); Imagination and Pictorial Understanding, Proceedings of the Aristotelian Society, Suppl. Vol. 60 (1986): 4560; Art, Interpretation and Perception, in The Mind and Its Depths (Cambridge, MA: Harvard University Press, 1993); and above all, Painting as an Art (Princeton: Princeton University Press, 1987).

240

Pictures

seeing-in, it is not equivalent to seeing-in. Seeing-in can occur without such perceptions, and vice versa, and so there can be no identifying them. The experiences of perceiving resemblances and seeing-things-in-other-things are different, and irreducibly so; the former is inherently relational and comparative, the latter not. We may observe, in addition, that were seeingin to be identied with perception of structural isomorphism, then since the latter is clearly a notion of degree one would expect the former to be as well. But seeing-in is not evidently a notion of degree, nor is that of depiction, which seeing-in underwrites; seeing-in and depiction are closer, if anything, to being on-off or all-or-nothing affairs. What is likely true in this matter is that a nonzero degree of structural isomorphism between a representation and its subject is required for seeing-in to take place, that is, that some such isomorphism may be a causal precondition of seeing the subject in the representation; the mechanisms whereby seeingina kind of seeing, after allis enabled to occur seem to require as much. But even if that is so, the perception of such isomorphism, as opposed to its mere existence, remains strictly unnecessary to the occurrence of the distinct experience of seeing-in. Third, I agree with Wollheim, as against Walton, that seeing-in is generally prior to, and not to be analyzed in terms of, imagined seeing. A reason for that insistence, beyond those hinted at by Wollheim, is as follows. If seeing-in is equated with imagined seeing of a certain kind, that is, if every case of the former is made out to be a case of the latter, then we lose a resource for explaining some of the special character, whether of immediacy, intimacy, absorbingness, or emotional impact, of some pictures as opposed to others (or alternatively, of some occasions of experiencing pictures as opposed to others), by appeal to the idea that although all pictures in being perceived as such induce seeing-in, only some pictures induce (or only some occasions of experiencing involve) actually imagining seeing the object that a picture
To elaborate: perception of resemblance between visual eld 1 and visual eld 2 explicitly involves relating and comparing those items, while seeing object X in painting Y does not involve a parallel relating and comparing of those items. The second term in such an experience, that is, Y, does not enter into the content of the experience involved, though naturally it is involved in generating the experience and in xing what experience it is. The content of the experience, consisting as it does in seeing X, in a manner of speaking, is basically just X . (I say basically, since it might be held that the content in question is not precisely X but something more like image of X. The point would remain that such content was non-relational, or at least, not such as to involve a relation to Y .)

Wollheim on Pictorial Representation

241

represents. Imagining seeing X in viewing Y implies, as a default, imagining you are face-to-face with X; but it seems doubtful whether one is standardly doing that merely in virtue of seeing X in Y, that is, recognizing the look of X in the design of Y. II The basic shape of Wollheims position on pictorial representation is thus one I nd congenial. But I have come to have various qualms about its specic articulation, qualms that prompt me to a friendly interrogation of some of its constitutive elements. As a result I am led to venture certain claims that Wollheim would, I am sure, be reluctant to embrace. Still, the picture of picturing that I uphold remains, in broad outline, a clearly Wollheimian one. The elements of Wollheims position that I will examine are these: the treatment of trompe-loeil; the status of twofoldness in seeing-in; the recognitional aspect of seeing-in; the scope or range of seeing-in; and the appeal to the artists fullled intention as a standard of representation. At more length, the questions I want to pursue are as follows. (1) Is trompe-loeil precluded from being understood as representational because it is designed to forestall an apprehending experience characterized by twofoldness? (2) Is the experience of seeing-in in fact necessarily characterized by twofoldness, that is, simultaneous awareness of medium and of subject, such that seeing-in has always a congurational as well as a recognitional aspect?; (3) What can be said about the recognitional awareness that is arguably at the core of seeing-in, especially if congurational awareness, or awareness of medium, is not always present as well?; (4) Is seeing-in really the same phenomenon or mental state across all the sorts of things it is said can be seen in pictures? (5) Is the artists fullled intention to depict such and such an apt criterion of what it is correct to see in a picture, and so of what it depicts?
On my conception of it imagining is necessarily in some degree active or contributory, though not necessarily something one is aware of initiating, and not necessarily something under complete control of ones will. By contrast, seeming to one as if what I propose captures, as well as anything can, the experience at the heart of pictorial seeingis passive or receptive, not something one brings about and actively sustains, but something that, in the last analysis, simply occurs. Seeing X in Y is something that happens to one, even when deliberate mental actions of various kinds, for instance, framings, thinkings, or suggestings, serve as triggers to such happening.

242

Pictures

III That trompe-loeil pictures pose a problem for the seeing-in theory of depiction is, I think, undeniable. If being a depiction requires inviting and sustaining seeing-in, and if seeing-in is an experience that necessarily involves twofoldness, and if twofoldness necessarily implicates awareness of and attention to pictorial surface, at some level, then it seems that trompeloeil pictures cannot be depictions. Though Wollheim is content to accept this consequence of his seeing-in account, it strikes me, as it has others, as counterintuitive. Now there is in fact a way to understand trompe-loeil pictures as supporting appreciative experiences with something like twofold character, and thus as thereby having clear claim to depictive status, before addressing the question of whether simple seeing-in is necessarily characterized by strict twofoldness. It is this. When we see trompe-loeil pictures as pictures, that is, when we are aware that they are pictorial contrivances, when we are past the point of being taken in by them, when we recognize them as trompe-loeil while allowing them to continue to fool the eye, then something like twofoldness, or simultaneous awareness of subject and medium, is present, even though the medium is, in a way, transparent. In such cases there is a kind of awareness, perhaps even visual awareness, of the surface, in the sense that visual attention is carried to it, despite the fact that with a perfect trompe-loeil the surface remains invisible. Once you grasp that something is a trompe-loeil you can attend to its surface, and in its visual aspect, even though you cannot by hypothesis see the surface as such. What you can do with a trompe-loeil painting, as with any painting, is mentally focus on the surface before you at the same time as you register its pictorial content, notwithstanding the fact that in such cases the surface does not end up arresting your vision. But let us put aside that resolution of the difculty, appealing as it does to an exceedingly liberal construal of twofoldness, and consider again the problem generated for the theory of picturing by trompe-loeil. It seems there are two options open to us. We can either allow that seeing the pictorial content of a trompe-loeil painting without realizing it is such, and so a fortiori without any awareness of the paintings surface, is still an instance of seeing-in, and
See, for example, Dominic Lopes, Understanding Pictures (Oxford: Oxford University Press, 1996), ch. 2.

Wollheim on Pictorial Representation

243

thus that such seeing does not always involve twofoldness (rst option), or else deny that seeing the pictorial content of a trompe-loeil painting without realizing it is such is an instance of seeing-in, thus retaining twofoldness as a necessary feature of such seeing (second option). One might argue in favor of the second option that naively registering the pictorial content of a trompe-loeil does not involve seeing the picture as a picture, and for that reason should not be accounted a case of seeing things in the picture. In addition, since attention to form concurrently with content, or to content-as-embodied-in-form, is often taken to be the heart of what it is to carry aesthetic attention to an object, one might further argue, against the rst option, that by its lights seeing-in would not necessarily exhibit aesthetic character. However, as is probably apparent, such an argument would be weak, since the considerations on which it turns seem more convincingly deployed in the opposite direction. Plausibly not all seeing-in or registering of pictorial content is aesthetic in character, or even informed by the awareness of pictures as pictures; for instance, that directed to or had in connection with postcards, passport photos, magazine illustrations, comic strips, television shows, or movies. Thus, any view that builds aesthetic character, or even awareness of pictures as pictures, directly into seeing-in would seem to have something amiss. It seems perfectly reasonable to hold that one can be seeing things in pictures, in virtue of looking at pictures, even when one is not seeing them as pictures, and a fortiori, without appreciating them aesthetically. I propose, then, that we embrace the rst option, whereby simple seeing-in, and what we might call pictorial seeing proper, are distinguished, with only the latter denitionally implying twofoldness. Pictorial seeing, or seeing pictures as pictures, is indeed a sine qua non of aesthetic appreciation of pictures, but the fact is that there can be seeing-in in connection with pictures that is not even pictorial seeing, that does not involve any awareness of pictorial properties or the medium in which they are embedded. If you see a woman in a picture in virtue of visually processing a pattern of marks, then of course in some sense you are thereby perceiving the medium
See Richard Eldridge, Form and Content: An Aesthetic Theory of Art, British Journal of Aesthetics 25 (1985); Malcolm Budd, Values of Art (London: Penguin, 1995); and Jerrold Levinson, Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996), chs. 1 and 2. I am not, of course, denying that we may often be cognizant of such pictures as pictures, or carry aesthetic attention to them, only that we must or even usually do so. I have been inuenced on these points by Lopess critique of seeing-in theory in Understanding Pictures.

244

Pictures

in which those marks inhere or consist. But it is far from clear that when you see the woman in the picture you must in some measure be attending to, taking notice of, or consciously focusing on the pictures surface or patterning as such. Yet that does appear to be part of the import of twofoldness as Wollheim construes it: Looking at a suitably marked surface, we are visually aware at once of the marked surface and of something in front of or behind something else. I call this feature of the phenomenology twofoldness. That twofoldness as Wollheim understands it means that the experience of seeingin involves, in its congurational as well as its recognitional aspect, some level of conscious apprehension and not, say, merely unconscious registering, is conrmed by this more extensive passage from Painting as an Art:
The twofoldness of seeing-in does not, of course, preclude the one aspect of the complex experience being emphasized at the expense of the other. In seeing a boy in a stained wall I may very well concentrate on the stains, and how they are formed, and the materials and colours they consist of . . . and I might in consequence lose all but a shadowy awareness of the boy. Alternatively, I might concentrate on the boy, and on the long ears he seems to be sprouting . . . and thus have only the vaguest sense of how the wall is marked.

A crucial issue, then, would seem to be what, exactly, being visually aware of a picture surface amounts to. Not, surely, receiving information from the surface, or being sensitive to changes in features of the surface; such construals are too weak for the purpose, since they are too easily satised by mental states, for example subdoxastic ones, that lie below the level of consciousness. Not, surely, thinking or reecting that one is seeing the surface as one sees it; such a construal would be too strong, collapsing visual awareness per se and selfconscious visual awareness. Perhaps, then, something like this: attending to the surface as one views it and is affected by it. But if anything like that construal is adopted, it is indeed doubtful that the seeing-in involved in grasping pictorial content always entails or includes visual awareness of the surface as well. At any rate, Wollheim has not indicated an intermediate notion of awareness that might be apt to the needs of the case but that does not import any degree of attention whatsoever.
On Pictorial Representation, 221. Painting as an Art, 47 (my emphases). We might, echoing Death of a Salesman, underline that in order for something to count as awareness attention must be paid, at least in same degree. Note that if this critique of seeing-in in respect of whether it necessarily displays twofoldness is correct, we may still rest with Wollheims desired characterization of a depiction as a marked

Wollheim on Pictorial Representation

245

IV The task remains, though, of saying what simple seeing-in consists in, given that it does not inevitably involve visual awareness of medium, that is, attention in some degree to medium, and yet is not just seeing in the ordinary sense. As has rightly been observed, clarifying what Wollheim calls the recognitional aspect of seeing-inand what we may now take to be the very core of seeing-inseems incumbent on a supporter of the seeing-in approach to pictorial representation. Here, then, is a stab at what such recognition amounts to. In looking comprehendingly at a picture of a woman, say Kees van Dongens engaging and mildly fauve canvas, La chemise noire, one does not necessarily perceive an isomorphism between experience of the picture and experience of a woman (Budd), nor does one invariably imagine seeing a woman (Walton), nor, in all probability, does it seem to you that a woman is actually before you (Gombrich). Rather, I suggest, it seems to you as if you are seeing a woman (alternatively, you have an impression of seeing a woman), in virtue of attending visually to portions of the canvas. The core of seeing-in, in other words, is a kind of as-if seeing that is both occasioned by visually registering a differentiated surface and inextricably bound up with such registering. Of course more needs to be said about the tight relation required here between the registering of visual information and the perception of pictorial
surface intended for seeing-inonly that should now be understood as simple seeing-in, it being granted that in most cases the surface is also clearly intended for pictorial seeing, with its inherent twofoldness, as well. To say it does not involve visual awareness of the medium is not, of course, to say that it does not involve visual processing of the information embodied in the medium. See Kendall Walton, Seeing-in and Seeing Fictionally, in James Hopkins and Anthony Savile (eds.), Psychoanalysis, Mind and Art (Oxford: Blackwell, 1992). Note that seeming to one as if P, or having the impression that P, are not locutions that entail believing that P or even thinking it probable that P. For example, It seems to me as if I am falling unsupported, said in a rapidly descending elevator. There is a real question whether the experience I have continued to refer to as simple seeing-in should in fact be so called. Two reasons give pause. The rst is that the association of seeing-inwhich is, after all, a term of art introduced by Wollheimand twofoldness is so entrenched that an experience of seeing-in sans twofoldness sounds almost oxymoronic. The second is that conceiving such experiences as the seeing-in label encourages one to do, as a matter of seeing things in surfaces, does undeniably occasion strain where trompe-loeil pictures are concerned, since in such cases the surfaces are, by hypothesis, neither seen nor seeable. One might thus concede that the visual experience of pictures I have been calling simple seeing-in, and that is present even when twofoldness is not, might in certain cases with more justice be called rather seeing-from.

246

Pictures

content. The relation has to be such as to rule out non-standard causal routes by which a pictures visual array might lead one to have an impression of seeing a woman, for example, one where such an array triggered, at a subperceptual level, a chemical change that in turn issued in a localized hallucination of a woman just like the woman seeable in the picture. The impression of seeing, or as-if seeing, at the core of seeing-in is thus one intimately bound up with the registering of the visual data afforded by the picture, whereby the latter in a sense constitutes or realizes the former. V None of this is to deny that much of the interest and appeal of seeing-in lies in the possibility of twofoldness in ones experience of a picture, that is, simultaneous awareness of both picturing pattern and pictured object, where ones seeing-in thus becomes seeing pictorially, properly speaking. Yet equally important, I would suggest, is the option, in which one might at turns indulge, of switching back and forth between awarenesses or focusings of attention of those two kinds, seeing sometimes only pure pattern, sometimes only pure object. In fact it would seem reasonable to include, within the ambit of pictorial seeing, that is, seeing of the kind normative for pictures understood and appreciated as pictures, both seeing
It is a virtue of Waltons account of depiction, of course, that it secures the desired intimacy in the most direct fashion, by making the act of perceiving the picture that which the viewer imagines to be an act of actually seeing the subject of the picture. But it seems to me that what necessarily happens in such a case of seeing-in is at most that one takes ones apprehending of a surfaces forms and colors to be a seeing of a woman, in a sense that does not imply that one believes or suspects that one is seeing a woman, but not that one imagines of such apprehending that it is a seeing of a woman. One might worry, nally, that if simple seeing-in is construed so as not to necessarily involve awareness of a pictures surface, then simple seeing-in and simple seeing will collapse. But this worry is unfounded. In the case of simple seeing-in you seem to see X, that is, you have an as-if-seeing experience of X, in virtue of visually registering certain congurations of a surface, rather than in virtue of being in the visual presence of X. In the case of simply seeing X, it is true as well that you seem to see X, that is, have an as-if-seeing experience of X, but then there are differences. With simply seeing X there is, rst, the belief, or tendency to believe, that X is before you, and second, it is X, and not merely a surface congured to afford an impression of X, that indeed is before you. This is all admittedly rough and ready, not intended as careful analysis. The point is just that the experiences of simple seeing-in and simple seeing can surely be discriminated, though once twofoldness is abandoned as a sine qua non of seeing-in such discrimination may not be a wholly internal matter, but may instead rest on matters such as what is precipitating the experience and what sort of mechanisms are involved in its doing so.

Wollheim on Pictorial Representation

247

where there is simultaneous awareness of design and content (i.e. twofoldness), and seeing in which there was alternation back and forth between phases of simultaneous awareness of design and content and phases of exclusive or near-exclusive focus on one or the other. It seems that our knowing engagement with pictures does in general display an alternation between phases of simultaneity, often sustained without deliberateness, and ones of switching, often occasioned by deliberate reection on what ones experience is like. Pictorial seeing might thus conveniently be stretched to cover such activity in all its phases. It is hard to overestimate the keen interest that viewers of painting naturally take in bringing simultaneously into relation, or alternating systematically between, the recognitional and the congurational, or the pictured and the picturing, in different styles of depiction. This is one of the obvious, but nevertheless deep, sources of fascination with the differences among NeoClassical, Impressionist, Post-Impressionist, Expressionist, Cubist, Surrealist, and Abstract-Expressionist treatments of what is in some sense the same subject. We are endlessly amazed with the variety of ways there are to pictorially construct familiar objects, so that patterns or designs that would seem to have little in common, compared as such, are revealed to have an afnity in supporting equally a visual impression of, say, a cow. That a cow can be made, visually speaking, out of dots, dashes, lines, angles, masses, smears, or mere chiaroscuro, is something we delight in bringing home to ourselves through this activity of regularly correlating design and content in our apprehension of a painting. Each time, after absorption in the represented world, that we attend primarily to the congurations afforded by a painting, we derive anew the pleasure of seeing of what the objects of that world have been made. But there is yet more than that. Different styles of representational painting arguably give us access to unique kinds of beings, allowing us to see things not encountered in the real world at all, rather than merely allowing us to see familiar things in a new way. What I have in mind are beings such as these: Ingres-women, Picasso-women, and De Kooning-women; Kirchnermen, Beckmann-men, and Grosz-men; and nally, Miro-dogs, Klee-dogs, and Dubuffet-dogs. Paintings of the respective artists familiarize us with extraordinary creatures of that sort, ones that can enter importantly into ones imaginative and interpretive repertoire; such paintings do more than simply show us how those artists, or their implied personae, may be said

248

Pictures

to have viewed ordinary women, men, and dogs. Of course, after making the acquaintance of Ingres-women, Kirchner-men, or Miro-dogs, one may then be in a position to spot their instantiations, or near-instantiations, in the world around one, hors de peinture. That is to say, we achieve acquaintance with kinds of beings whose exemplars are not all of them, or not necessarily, ctional. VI I have concentrated so far mostly on seeing-in as it applies to objects. But as Wollheim urges, seeing-in may be held to range over actions and events as well, and even over individuals-merely-of-a-certain-kind as opposed to particular individuals. Concern arises, however, as to whether the range of seeing-in is usefully taken to be as wide as Wollheim proposes. The concern might alternatively be expressed as one of whether the seeing-in involved in all such cases is sufciently of a piece as to merit the single label. Let us look to what Wollheim says about the outer limits of seeing-in as he descries them, as illustrated in the example of the classical landscape with ruins. Wollheim suggests that a suitably prepared and prompted viewer plausibly can see, in such a painting, all the following: columns, columns-as-havingcome-from-a-temple, columns-as-having-been-thrown-down, columns-ashaving-been-thrown-down-hundreds-of-years-ago, and columns-as-havingbeen-thrown-down-hundreds-of-years-ago-by-barbarians. But such a viewer cannot, Wollheim submits, see in the painting columns-as-having-beenthrown-down-hundreds-of-years-ago-by-barbarians-wearing-the-skins-ofwild-asses. But why not? Why are all those other qua-objects seeable in the picturewhich I take to be roughly interconvertible with the seeability therein of corresponding states of affairsbut not the last? What principle of cut-off for the qua-objects or states of affairs that can be seen in a picture
A story from the golden era of The New Yorker is relevant here. James Thurber, one of the early great cartoonists of the magazine, was once the subject of discussion at the weekly art meeting being presided over by the then editor, Harold Ross. The point at issue was the seal perched on the headboard of a bed in one of Thurbers most famous cartoons, in which the wife is vocally skeptical of her husbands claims to have heard a seal bark. Someone at the meeting, noting the somewhat loosely drawn character of the seal, asked Do seals look like that?. To which Rosss reply was: Thurbers seals look like that. (From an interview with New Yorker Cartoon Editor Robert Mankoff, The Washington Post, Dec. 7, 1997.) On Pictorial Representation, 224.

Wollheim on Pictorial Representation

249

does Wollheim have in mind? What he offers is an operational test: propose for seeing-in a candidate state of affairs and note whether it makes a difference in the suitable spectators experience. Yet in the absence of a clear idea of the bounds of seeing-in, it is hard to know how one would interpret the results of such a test. It is not entirely obvious what would rule out seeing the columns as having been thrown down by barbarians wearing the skins of wild asses. After all, we no more see the vandals and their destructive acts than we see the equine pelts they may very well have sported. Of course we can speculate on what it is that makes a non-manifest state of affairs or condition a reasonable candidate for seeing-in as Wollheim conceives that. Possibly a condition being such that perceptual inference to it is highly compelling, or a condition possessing visual traces of a relatively unequivocal sort, at least for a properly backgrounded viewer, makes such a condition something that can be seen in a picture. But I am not concerned to worry further about these or any similar suggestions. The real problem, I think, is that seeing non-manifest states of affairs in pictures, seeing occurrent actions in pictures, and seeing objects in pictures may be importantly different phenomena, whose differences, say as regards spatial localization or permeability by thought, may be more obscured than illuminated by considering them together as members of a species. Seeing-in may not be univocal across its putative instances, especially if, as suggested earlier, twofoldness is not even an invariant feature of the experience of seeing one thing in another. It is not clear that the same sort of activity or perception is involved when going from seeing-in of objects to seeing-in of events to seeing-in of conditions, or from seeing-in of physical events or conditions to seeing-in of psychological events or conditions. For example, localization, the property of such-and-suchs being seeable more or less right where the relevant pictorial design is, may be characteristic of the seeing in paintings of physical objects, but somewhat less so of the seeing of events, and very much less so of the seeing of psychological entities, whether objects or events. And the permeability of seeing-in to thought, reection, or conceptualization seems progressively more pronounced, in general, as one moves from the seeing of objects to the seeing of events to the seeing of only indirectly evidenced states of affairs in a painting. And nally one might add, for good measure, that a role for imagination in the robust sense appears considerably more plausible in regard to seeing-in of this latter sort than to seeing-in of the former two sorts. These divergences make suspect, at the least, the assumption that seeing-in in all the cases claimed by Wollheim is of a uniform nature.

250

Pictures

VII Lastly, there is to my mind a problem about Wollheims appeal to fullled representational intentions as a standard for what a picture represents. Here is a formulation, though from an earlier paper: Very roughly, P represents X if X can be correctly seen in P, where the standard of correctness is set for P by the fullled intentions of the artist of P. A difculty lurks here that Wollheim and other actual intentionalists about meaning have some tendency to gloss over. It is this. What it is for the pictorial intentions of the artist of P to be fullled cannot be specied apart from what suitable viewers are enabled to see in P. Such intentions are fullled if viewers are in fact enabledand enabled without undue thematic prompting or inordinate mental contortionto see in P what the artist intended be seen there. The artists fullled intention cannot be thought of as an independent condition to which viewers responses can be held accountable, but can only be understood in terms of the responses of appropriately primed and backgrounded viewers being the ones they were intended to be. Another way of making my point would be to say that the standard of correctness for depiction is not, as Wollheim sometimes puts it, the fullled intentions of the artist, but merely the intentions simpliciter of the artist for a certain sort of seeing-in, given that they are capable of being complied with by the pictures intended viewers. The artists representational intentions only are fullled if suitable viewers are enabled, on reasonable prompting, to see-in the painting in accord with the artists representational intentions. Thus it arguably makes little sense to say they comply with the artists fullled intentions in this regard, since such do not, as it were, preexist such compliance. VIII Richard Wollheims theory of pictorial representation is the fruit of long reection, deep insight, and an intense love of painting. In the course of
Imagination and Pictorial Understanding, 46. To be fair to Wollheim, the formulation of the intentional condition on representation in the present essay almost entirely escapes the problem, highlighted here, to which earlier formulations were subject: Representational meaning, indeed pictorial meaning in general, is, on my view, dependent, not on intention as such, but on fullled intention. And intention is fullled when the picture can cause, in a suitable spectator, an experience that tallies with the intention (On Pictorial Representation, 226).

Wollheim on Pictorial Representation

251

this essay I have criticized that theory in a number of respects, notably, its treatment of trompe-loeil, its conception of seeing-in, and its appeal to the artists fullled intention as a standard. Observe, however, how much remains of what Wollheim has urged in what I am willing to afrm on this vexed topic: pictorial representation involves the intentional marking of a at surface so as to elicit a distinctive sort of visual experience in appropriate spectators, an experience we may continue to call seeing-in as long as we understand that this sometimes amounts only to what may be more transparently labeled seeing-from, where such experience is indeed elicited from those spectators in virtue of their attending to the surface as marked.

14
What is Erotic Art?
Here is an answer to the question of my title: erotic art is art which aims to engage viewers sexually through explicit sexual content, and that succeeds, to some extent, in doing so. In addition to developing and defending that answer, in this short essay I explore the boundaries of the concept of erotic art, and some of the psychological and social implications of erotic art. The scope of this essay is restricted in two ways. First, the essay concerns visual art exclusively, and then, almost entirely, visual art of two dimensions. But of course, that is not to deny the existence of erotic literature, dance, cinema, or even, less obviously, music. Second, the examples are drawn only from Western art, and mainly Western art since the Renaissance. But that is not to deny the prominence of erotic art in non-Western traditions of art, for instance, those of India or Japan. There are arguably strict and less strict senses of the term erotic art. If the denition offered above captures the strict sense of the term, applicable to art that is unquestionably or unequivocally erotic, we may also recognize art that is erotic according to a looser construal of the term, including art that is only accidentally erotic, art that is only instrumentally erotic, art that is only covertly erotic, and art that is only erotic, a bit paradoxically, in virtue of being anti-erotic. These different cases will be addressed in detail in what follows. I M AIN QUE ST IONS The chief philosophical questions regarding erotic art would seem to be these. (1) What is the distinction, within art, between erotic and non-erotic art, and how sharp is this distinction? (2) What are the normative implications, if any,

First published as Erotic Art, in E. Craig (ed.), The Routledge Encyclopedia of Philosophy (London: Routledge, 1998), 4069.

What is Erotic Art?

253

of the different forms and modes of erotic art, and what, in particular, is the distinction between erotic art and pornography? (3) How can erotic art in fact be art, that is, something properly eliciting an aesthetic response, traditionally characterized as disinterested, while also aimed at provoking sexual desire, the very paradigm of an interested reaction? (4) In what way, if any, do the criteria for assessing erotic art differ from those appropriate to assessing art of other sorts, and how does the degree of eroticness of a work of erotic art relate to its goodness as art? This essay will be devoted almost exclusively to the rst and second of these questions.

II THE CONCEPT OF EROT IC ART A good proportion of the work of many great visual artistsRubens, Ingres, Delacroix, Degas, Rodin, Gauguin, Matisse, Magritte, Munch, Klimt, Schiele, Picasso, Modiglianiis unquestionably erotic. But what is it, precisely, for art to be erotic? It seems that, at a minimum, it must have sexual content. Though sexual content may be either overt or covert, let us consider rst such art as has overt sexual content. Typically, this takes the form of depictions of unclothed or semi-clothed human beings, alone or accompanied, at rest or performing actions of a sexual nature. But for art to be accounted erotic, it must do more than represent the naked human body or otherwise make reference to sexual matters: not all art concerned in some way with sexuality counts as erotic. An anatomical sketch of private parts by a D rer or a Leonardo, a realistic study of a gynecologists u examining room, a comic strip featuring pneumatic bimbos, are none of them erotic, despite their inclusion of sexual content. Rather, erotic art is art that treats its sexual content in a particular way or that projects a certain attitude toward it. Erotic art is art aimed at arousing sexual interest, that is, at evoking sexual thoughts, feelings, or desires in viewers, in virtue of what it depicts and how it is depicted, and which achieves some measure of success in that regard. The intent to awaken and reward sexual interest through what is depicted can be taken as criterial of at least central cases of erotic art. The erotic work of art does more than merely refer to or acknowledge human sexuality; rather, it expresses an involved attitude toward it, whether of fascination, obsession, or delectation, and in addition, invites the viewers imaginative engagement, along similar lines, with what is shown.

254

Pictures

Erotic art not only aims at engaging the sexuality of the viewer, but typically also reects that of the artist. That is to say, erotic works usually embody a perspective on what is depicted that suggests sexual interest, and of a particular sort, on the makers part. Furthermore, the sense of sharing in what at least appears to have been sexually stimulating to the artist often plays a causal role in the viewers own stimulation by what is depicted. It is worth emphasizing that the sexual response occasioned by erotic art occurs largely on the plane of imagination, consisting primarily of thoughts, images, and feelings, and rarely goes as far as full physiological arousal; the upshot of engagement with erotic art is imagined desire as often as it is real desire. This is not unrelated to the distinction between erotic art and pornography, which is touched on below. As suggested above, the term erotic art in its central usage covers art that aims at, and that at least minimally succeeds at, stimulating sexual thoughts and feelings in its target audience. But this leads, easily enough, to two looser usages, according to which, roughly speaking, meeting either the intentional condition or the success condition independently qualies a work as erotic. On the rst such looser usage, a work counts as erotic if it is apparently aimed at stimulating sexual thoughts and feelings even when it is not successful in doing so, while on the second such usage, a work counts as erotic if it succeeds in stimulating viewers sexually even when not intended or even apparently intended to do so. Art of the former sort might be labeled nominally erotic art, while art of the latter sort might be labeled accidentally erotic art. Finally, perhaps some artworks reasonably accounted erotic neither aim at nor achieve viewer arousal as such, that is, sexual thoughts, feelings, or sensations directed towards what is depicted, but instead are erotic in virtue of facilitating the imagining of erotic states of others, without unequivocal erotic involvement on the viewers part, that is, without the viewer identifying with or entering into those states, in either reality or imagination. Such art might merit the label of obliquely erotic art.

III INSTRUMENTALLY EROTIC ART AND ANTI-EROT IC EROT IC ART With some erotic art the evocation of erotic feelings, rather than being the main order of business, is a secondary aim, and is employed or manipulated by the artist primarily in order to achieve some further end. For example,

What is Erotic Art?

255

Tom Wesslemanns caricatures of female pulchritude and Mel Ramoss exaggeratedly voluptuous pin-ups use erotic images to achieve a kind of wry humor, Degass monotypes of brothel scenes serve purposes of social commentary, and the recombinant sexual imagery of Magritte and Dali is a means to psychological disorientation of the viewer. We might label such art instrumentally erotic art. As a result of these secondary aims, the excitatory tendency of such works is generally weakened, and is sometimes wholly neutralized. In some limiting cases works are in effect about erotic artthey are commentaries on, or satirical appropriations of the conventions and mechanisms of, ordinary erotic artbut without being erotic in the central sense, that is, ultimately aimed at sexually engaging the viewer. Such art may be accounted erotic in virtue of leading the viewer to question the presuppositions and consequences, social and otherwise, of erotic responses, without inviting or even permitting viewers to have such responses. Some other cases of works representing sexual matters without appearing clearly erotic will serve to illuminate further the boundaries of the category: 1. Lysippuss sculpture Aphrodite, Botticellis Birth of Venus, and Cranachs Eve occasion some hesitation if classied as erotic. Probably this is because we take the primary intent of the artist to have been to embody ideals of the human form, male and female, and not to prompt imaginative erotic engagement on the part of viewers of either gender. But this may be ingenuous; at any rate, such a line could not plausibly be extended to exclude from the erotic Donatellos sensuous, almost coquettish, David. 2. Picassos Demoiselles dAvignon occasions hesitation of a different sort. Though the painting presents women who are not only nude but in fact prostitutes, they are depicted in a highly non-realistic mode, which shortcircuits erotic involvement, as well as drawing attention primarily to the paintings formal and expressive dimension. 3. Judy Chicagos The Dinner Party, an elaborate sculptural installation, uses female genital imagery, and in a celebratory fashion, but probably not in an erotic way; its sexual content is of the purely symbolic, rather than sexually involving, sort. 4. Lucian Freuds paintings of naked subjects, though displaying some of the hallmarks of erotic art, are not obviously erotic, being more evocative of the boucherie than the boudoir an observation even truer of the images of nudes in Francis Bacons paintings. Philip Pearlsteins super-realist

256

Pictures

gure paintings or, in another vein, Dubuffets quasi-paleolithic images of squashed and splayed humanity, belong here as well. The works of these latter artistsBacon, Freud, Pearlstein, Dubuffetare not aptly described merely as non-erotic, as are, say, a Corot landscape or a Chardin still life, but rather as anti-erotic. But in a broader sense they, unlike the Chardin or the Corot, are erotic after all, in the sense of being concerned with sexuality in a way that reects the sexual interests of the maker and engages those of the viewer, if not in a positive manner. De Koonings raw and primitivist images of women come naturally to mind in this connection as well, though the case can be made that those images project an even more ambivalent attitude to human sexuality than those of Bacon and Freud, with their admixture of awe, terror, and admiration. IV COVERTLY EROTIC ART It is relatively easy to give plausible examples of erotic art with no explicit depictions of sexuality or nakedness: Georgia OKeeffes landscapes and still lifes, with their oblique evocation of female anatomy; Caravaggios paintings of Bacchus or St John the Baptist, with their coded references to homosexual experience; or Berninis marble of St Teresa in spiritual ecstasy, a state readily translated by the viewer into its profane cousin. What the criterion of covert sexual content is, however, remains unclear. Depiction of objects recognized as sexually symbolic, such as umbrellas or fruit, especially when they are juxtaposed with human subjects, may be a typical indication of such content, but can hardly serve as a general mark. According to some writers, virtually all art has covert sexual content in virtue of being the expression of unconscious wishes or fantasies of a sexual sort. For Richard Wollheim, for instance, Ingress history paintings, Bellottos landscapes with buildings, and Poussins landscapes with water are as substantially imbued with sexuality as Goyas Naked Maja or Titians Venus of Urbino. Even so, it seems that not all covertly sexual art is usefully considered erotic, but only that which is plausibly aimed, if unconsciously, at exciting sexual thoughts or feelings in target viewers, and which succeeds in doing so. In putative cases of covert sexual content, the arousal of the appropriately
Richard Wollheim, Painting as an Art (Princeton: Princeton University Press, 1987).

What is Erotic Art?

257

backgrounded and oriented viewer may be what in fact signals the presence of such content, and justies its ascription. V T HE REL ATIONALIT Y OF E ROTIC ART If a painting is erotic, this is in virtue of being aimed at and to some extent eliciting an erotic response from a certain class of viewer, the paintings intended or target audience. Such classes may be delimited not only by requirements of sensitivity and background knowledge, such as are appropriate to art of any kind, but also by ones, less acquirable, of physiological makeup or sexual orientation. Thus, a painting may be erotic in virtue of being designed to produce, and succeeding in producing, an erotic reaction in heterosexual males, elderly homosexual males, young heterosexual girls, homosexual women, or bisexuals of either gender. There is a fact of the matter, if a hazy one, about whether a given painting is erotic, but it is an inherently relational one, whose nature is only fully evident when the group targeted for response is identied. Indeed, according to Linda Nochlin, the very term erotic art is understood to imply the specication erotic-for-men . Still, once such implicit indexing has been made explicit, it may then be cancelled, so as to recognize art that is erotic relative to other target groups. VI SOCIAL AND POLITICAL DIMENSIONS OF EROT IC ART Recent writers on erotic art stress the way in which entrenched genres and conventions of representation embody dominant ideas and assumptions about the nature of men and women and their proper relationship. Paintings such as Delacroixs Death of Sardanapalus, Gr mes Oriental Slave Market, eo or the Turkish Bath and Jupiter and Thetis of Ingres lend themselves readily to such analysis. For example, Linda Nochlin speaks of the power relations obtaining between men and women inscribed in visual representation as a focus of her investigations. Equally frequently noted is the element of voyeurism in erotic art. It is said that the spectator is a voyeur, at least ctionally, with the artwork often
Linda Nochlin, Woman, Art, and Power and Other Essays (New York: Harper & Row, 1988). Ibid.

258

Pictures

seconding or echoing this by containing a depicted spectator who, together with the viewer, regards the erotic object. The implicit or explicit voyeurism of erotic art is, in addition, sometimes held to reect the necessary impotence of the artist in respect of the imaginary, and thus unattainable, individuals depicted within his art. Finally, the relationship of erotic art and pornography has been much queried. The latter may be distinguished from the former in at least two ways. One, pornography might be said to have, by denition, no signicant artistic aspect. That is to say, pornography in the strict sense makes no credible appeal to viewers to consider the mode and means of depiction, as opposed merely to what is depicted; pornography, unlike art of any kind, is wholly transparent in both aim and effect. Two, pornography might be said to have, as a central intent and characteristic result, not merely the stimulation of sexual feelings or fantasies in viewers, but the devaluation or degradation of its subjects, usually women. By such criteria, although Courbets Sleep, which depicts two beautiful nude women in the arms of Morpheus and each other, or Schieles Reclining Woman, which presents its subject provocatively spreadlegged and scarlet-nippled, perhaps court categorization as pornography, on reection they remain at some distance from it; though the images in question are starkly arousing, even exploitative, the technique of their construction, the style in which they are rendered, the preceding art history they encapsulate, and the access they afford into their makers psyches, are at least as absorbing as what those images atly represent, and conspire to redeem them as art.
For a more nuanced account of the distinction between pornography and erotic art, see my Erotic Art and Pornographic Pictures, Philosophy and Literature 29 (2005): 22840 (Ch. 15, this volume). Some instructive writings on erotic art, apart from those of Wollheim and Nochlin already noted, are these: John Berger, Ways of Seeing (London: Penguin, 1972); Kenneth Clark, The Nude: A Study in Ideal Form (Princeton: Princeton University Press, 1956); Edward Lucie-Smith, Sexuality in Western Art (London: Thames and Hudson, 1972); Linda Mulvey, Visual and Other Pleasures (Bloomington: Indiana University Press, 1989); Lynda Nead, The Female Nude: Art, Obscenity, and Sexuality (London: Routledge, 1992); Griselda Pollock, Vision and Difference: Femininity, Feminism, and the Histories of Art (London: Routledge, 1988); Roger Scruton, Sexual Desire: A Moral Philosophy of the Erotic (New York: Free Press, 1986); Guy Sircello, Love and Beauty (Princeton: Princeton University Press, 1990); Leo Steinberg, Picassos Sleepwatchers, in Other Criteria (London: Oxford University Press, 1972).

15
Erotic Art and Pornographic Pictures
Only in primitive art, with its urgent need to evoke the sources of fertility, are the phallus and the vulva emphasized, as it were innocently. By ancient Greek and Roman times there already existed the special category of the pornographicgraphic art or writing supposed, like a harlot, or porne, to sexually stimulate.

I As regards philosophical analysis of the opposition between the erotic and the pornographic, there are a number of reasonable goals one might have: to preserve as many considered intuitions about the opposition as possible; to present the opposition in a clearer light than it enjoys when casually invoked; to propose modest sharpenings to the standard opposition that either account for our experience in this domain more fully, or allow us to organize our thinking about the domain more perspicuously. I hope in this essay to make progress toward some of those goals. Though the scope of the opposition of erotic and pornographic goes beyond the visual, my focus here will be the opposition as it exists in the visual sphere, and even more narrowly, in the sphere of two-dimensional images. In addition to preserving and clarifying a distinction between erotic art and pornography, I hope also to make an intelligible place for erotica, as something intermediate between the other two. Here, then, are some intuitions on the erotic and the pornographic: 1. The erotic and the pornographic are both concerned with sexual stimulation or arousal.
First published in Philosophy and Literature 29 (2005): 22840. John Updike, Can Genitalia Be Beautiful?, review of Egon Schiele exhibition at the Museum of Modern Art, New York Review of Books, Dec. 4, 1997.

260

Pictures

2. While the term erotic is neutral or even approving, the term pornographic is pejorative or disapproving. 3. While erotic art is a familiar, if somewhat problematic, notion, pornographic art seems an almost oxymoronic one. 4. Whereas pornography has a paramount aim, namely, the sexual satisfaction of the viewer, erotic art, even if it also aims at sexual satisfaction on some level, includes other aims of signicance. 5. Whereas we appreciate (or relish) erotic art, we consume (or use) pornography. In other words, our interactions with erotic art and pornography are fundamentally different in character, as reected in the verbs most appropriate to the respective engagements. In what follows I try to accommodate all of those intuitions. As I will need a distinction, invoked in the rst of them, between sexual stimulation and sexual arousal, let me spell that out before proceeding. By sexual stimulation I will mean the inducing of sexual thoughts, feelings, imaginings, or desires that would generally be regarded as pleasant in themselves. By sexual arousal I mean the physiological state that is prelude and prerequisite to sexual release, involving, at least in the male, some degree of erection. And by sexual release I mean something that I take it needs no spelling out. How to differentiate erotic art from pornography, and from erotica as well, is of course not the only important philosophical question about erotic art. Here are two others: (1) How is the erotic aspect of erotic art compatible with the disinterested or distanced frame of mind that seems required for aesthetic engagement with or appreciation of a work of visual art? (2) How is the degree of eroticness of a work of erotic art related to its artistic value? I will briey address those two questions toward the end of the essay, but most of my effort will be devoted to the prior question, that of the differentiation of erotic art, erotica, and pornography. II How, then, to effect that differentiation, in the domain of visual images to which I have restricted my inquiry? One possibility would be by the specic kind of response aimed at. Thus erotic art might be said to involve images intended to stimulate sexually but also to reward artistic interest, erotica to involve images intended to stimulate sexually but not to reward artistic interest, and pornography to involve images intended to arouse sexually in

Erotic Art and Pornographic Pictures

261

the interests of sexual release. Another possibility would be by the character of the sexual representation involved, so that pornography would involve sexually explicit images, erotic art sexually inexplicit images, and erotica images intermediate in their explicitness, irrespective of the ends for which the images were fashioned. And a third possibility would be by the moral status of the images involved, with pornography offering morally objectionable sexual images, erotic art morally unobjectionable sexual images, and erotica sexual images of borderline moral status. For reasons that will emerge shortly, the rst tack is the only viable one for making out the threefold distinction we are after. What makes the difference among the three kinds of sexual image, in other words, is what they are for, what response they are designed to evoke, what they are meant to do to us or we with them. The stark contrast is between erotic art, which invites artistic interest, and pornography, which positively deects such interest, whatever degree of artistic interest it might, as it were accidentally, sustain. So formulated the contrast clearly presupposes a satisfactory gloss on what artistic interest amounts to, but that will be provided shortly. What we have, then, is a subdivision of the broad category of erotic, or intentionally sexually interesting, images, into three subcategories: erotic art, erotica, and pornography. But I maintain neither that the boundaries between these subcategories are sharp, nor that there are no examples of erotic images that perhaps t into none of those subcategories. Though there is little need, I think, to offer examples of erotic art or pornography, it will be useful to mention an example of what to my mind counts as erotica, namely, classily provocative lingerie ads of the Victorias Secret variety. Note also that the broad category of erotic images, at least roughly exhausted by the three subcategories I have detailed, is not as broad as the category of sexual images generally, that is, images of any sort depicting sexual phenomena, such as sexual organs or acts or conditions. Thus, illustrations in textbooks of gynecology or eld guides to baboon mating behavior, though they depict sexual phenomena, do not count as erotic images, since they are not intended to interest viewers sexually; nor does security camera footage of a rape or molestation, even if in fact sexually stimulating or arousing to certain sorts of viewers. III Let me return briey to the other options mentioned earlier for demarcating the erotic from the pornographic, namely by degree of explicitness of

262

Pictures

representation involved or by moral status of the representations involved, in order to show their unworkability. The rst of these is the more easily dispatched. The fact is that, although as a rule erotic art involves representations of organs or acts that are less explicit than those of pornography, erotic art also embraces representations that are equally, if not more, explicit, than those in many instances of pornography. Much of the erotic art of artists such as Schiele, Klimt, Picasso, or Dali is as explicit as any Penthouse pictorial, and this is true also of certain traditions of Japanese graphic art. Pass on to the idea of distinguishing pornography from other varieties of erotic image on moral grounds. Sometimes the dening mark of pornography is taken to be its implied attitude toward its subject, an attitude, perhaps, of demeaning or degrading them. But even if such an implied attitude is typical of pornography, its not clear it can be held to be an invariable feature of it, unless the mere fact that such subjects are depicted for the purposes of sexual fantasizing and arousal is taken to be demeaning or degrading. As for the idea that pornography might be conceived more loosely as erotic images that are in some way or another morally objectionable, there are at least two problems with that. First, even if a moral case could be made against most pornography, it is unlikely that it can be made against it all; surely there must be morally acceptable ways of deploying sexual imagery to aid persons in achieving solitary sexual release. Second, even if such a case could be made against all pornography, its morally objectionable qualities would seem to be a consequential feature of it, rather than a dening one. Compare embezzlement, or the unauthorized taking or diverting of funds for private gain in a business context. That may very well be immoral, but its immorality is not reasonably made part of the denition of embezzlement. In addition, there are plausibly cases, such as when embezzlement is undertaken to save ones family from starvation, where embezzlement is not immoral, even though it is as a rule. IV Now, what do I mean by the artistic interest or dimension of an erotic image? Roughly, its form and the relation of that form to its content; the way the content has been embodied in the form, the way the medium has been employed to convey the content. We can speak almost equivalently of the artistic intent
For reections on the morally worrisome dimension of pornography for consumers of pornography, see my Sexual Perversity, Monist 86 (2003): 3054.

Erotic Art and Pornographic Pictures

263

of an image as of its artistic interest or dimension, if that is understood in a hypothetical manner, meaning the intent the image appears to manifest to have viewers attend to its artistic dimension or artistic interest. I will thus loosely alternate artistic interest, dimension, and intent in what follows, since although in another context one might want to tease them apart, for present purposes there is no need to do so. So characterized, an image that has an artistic interest, dimension, or intent is one that is not simply seen through, or seen past, leaving one, at least in imagination, face to face with its subject. Images with an artistic dimension are thus to some extent opaque, rather than transparent. In other words, with artistic images we are invited to dwell on features of the image itself, and not merely on what the image represents. Both erotica and pornography predominantly aim at sexually affecting the viewer, one with an eye toward stimulation, the other with an eye toward arousal, and accordingly do not seek to have attention rest on the vehicle of such stimulation or arousal, the medium through which the sexual content is communicated or presented. Erotic art, though aimed in part at sexually affecting the viewer, at stimulating sexual thoughts and feelingsthats what makes it erotic art, after allalso aims in some measure to draw the viewers attention to the vehicle, inviting the viewer to contemplate the relationship between the stimulation achieved and the means employed to achieve it, and more broadly, the relationship between the erotic content of the image and its other contents, such as expressive, dramatic, or religious ones. Its thus no accident, but highly telling, that photography is the prime medium for pornography, that which has displaced all other such media in that connection. For photography is the transparent medium par excellence, that is, the medium that comes closest to simply presenting the requisite objecttypically, a woman or a man or combinations thereofdirectly, as material for sexual fantasy and gratication. Though photographs of course can be art, and more specically, erotic art, they also lend themselves extremely well to non-artistic employment, which makes use of their inherent transparency, whereby they serve, if we let them, as mere aids to seeing. As we all know, pornography is essentially a kind of substitute or surrogate for sex, whether a poor one or not we can leave aside. That is why it is appropriate to characterize it narrowly in terms of the facilitation of sexual fantasy
The locus classicus in philosophy on the transparency of photography is Kendall Waltons Transparent Pictures, Critical Inquiry 11 (1984): 24677.

264

Pictures

in the name of arousal and release. And that is why to fulll that purpose its images should be as transparent as possiblethey should present the object for sexual fantasy vividly, and then, as it were, get out of the way. Nothing does that better than photography. V I turn now to a provocative essay on our subject by Matthew Kieran, whose conclusions are diametrically opposed to those I have been trying to establish. Kieran attempts to make a place for pornographic art, refusing to accept that the extension of that concept is by denition the null set, that pornographic art is indeed the oxymoron it appears to be. His attempt, however, simply shows that the works he champions are not pornographic art, or even erotica, as I use those terms, but instead erotic art of a distinctive kind, in which pornography or erotica are themselves subjects of the art in question, or in which pornography or erotica have been turned to artistic ends, and so transformed into art. One of Kierans examples is the novel Vox, by Nicholson Baker, which Kieran argues counts as pornographic art, though of course we are here dealing with literature rather than visual art. But I maintain that Vox is not pornographic art, if that means it is both art and pornography, though I grant that it is, in a sense, pornographic. I will remove the air of paradox from that assertion a bit further on. Whats true of Vox is that it mimics pornography, and in particular, phone sex, appropriating its gestures, tropes, and outer appearance, but does so in order to produce a work of literature, and thus, art. And one can admit that it is, at many places, sexually stimulating, even arousingafter all, a simulacrum of something often has many of the same properties and powers. But the point is that that is not all it is, nor all it is intended to be, nor what it is ultimately aimed at producing in readers by way of experience. Vox resembles pornography, to be surethat is obviousbut is not identical to pornography, because its paramount aim is not that of producing sexual arousal and release. Yet it is, of course, a mild turn-on, owing to the effective simulation of verbal pornography it presents throughout its length, and couldnt,
See his Pornographic Art, Philosophy and Literature 25 (2001): 3145. Much the same case was made earlier, though with less philosophical sophistication, by Susan Sontag, in The Pornographic Imagination, Styles of Radical Will (New York: Farrar, Straus & Giroux, 1969).

Erotic Art and Pornographic Pictures

265

in fact, achieve its artistic aims, which are variously psychological, parodistic, and pyrotechnic, if it werent. Another of Kierans examples is the ensemble of erotic drawings by Gustav Klimt. His discussion of those drawings is very insightful, showing as it does how Klimts mastery of artistic techniques helps to focus attention on the sexual parts, features, and states of the women depicted, making those drawings all the more erotic. But it is one thing to say that certain artistic devices, masterfully deployed, can enhance the erotic charge of a representation. It is quite another to say that a viewers focusing on those devices will enhance the representations erotic charge for the viewer, that is, render it more stimulating or arousing. There is every reason to think it will not, that it will rather temper the stimulation or arousal involved, replacing what is thus lost, however, with a portion of aesthetic pleasure. So what Kierans discussion shows, to my mind, is that Klimts drawings are not pornography, but rather art, albeit art that might be mistaken for pornography by inattentive viewers, or that might be used as pornography by viewers happy to lose sight of its artful fashioning and just enjoy the erotic upshot thereof. Making room for pornography that is also art, I suggest, is a bad idea. First, allowing that something can be pornographyand not just resemble pornography, or mimic pornography, or have a pornographic avor, or be quasipornographicand art at the same time, leaves no place for the category of erotic art as distinct from pornography. Second, the aims of true pornography and the aims of art, erotic art included, are not compatible, but war against one another, in the way that has already been sketched. One induces you, in the name of arousal and release, to ignore the representation so as to get at what is represented, the other induces you, in the name of aesthetic delight, to dwell on the representation and to contemplate it in relation to the stimulatingness or arousingness of what is represented. Now to remove the air of paradox from my assertion above that Vox is art, perhaps even pornographic art, but not pornography. We can say that there is pornographic art, if we just mean that there is art that has a pornographic look or character, but then such art is not yet pornography. That is, it is not both art and pornography. Analogously, there is art, for example, certain kinds of contemporary painting, such as that of Richard Estes or Alex Katz, which might be described as photographic. This means the art has a certain photographic look or character, not that it literally is photography. And there is writing, such as that found in certain kinds of newspaper articles, which can be called telegraphic, but that doesnt literally make that writing telegraphy.

266

Pictures

There is, in effect, a strong and a weak sense of the term pornographic in the expression pornographic art. In the strong, or conjunctive, sense, something is pornographic art if it is both art and pornography; in the weak, or modifying, sense, something is pornographic art if it is art and has a pornographic character, look, or aspect. It is only in the weak sense of pornographic, I submit, that there is pornographic art. And in that sense, of course, the inference from x is pornographic art to x is pornography fails. Recognizing that the expression pornographic art has a strong and a weak reading helps explain why many astute and progressive thinkers are as open to the idea as they are, for the ease of satisfying the weak reading of the expression creates the illusion that the strong reading is satised as well. Kieran tries to dispel three considerations against pornographic art understood in the strong sense. The rst consideration is that pornography is by denition non-artistic, or without artistic interest, to which Kieran replies that pornography is just highly explicit erotica, and since all admit that erotica can be art, there is thus no conceptual bar to pornography being art as well. But we have already seen that degree of explicitness cannot be the distinguishing mark of pornography, both because some erotic art is more explicit than some pornography, and because the essence of pornography arguably has something to do with what its images are for and not just what they show. The second consideration is that pornographys central aim, namely, to arouse sexually through explicit means, militates against its achieving artistic interest, even if that is not precluded by denition, to which Kieran responds that such an aim is not in fact incompatible with achieving such interest. Kieran is right about that; there is nothing to prevent pornography from having artistic interest, even though it doesnt aim at having it. What remains to be shown, though, is that as pornography it can be art, not just that it can have artistic interest. The third consideration Kieran addresses comes closest to that on which my brief against pornographic art is based. It is that there is an appreciative problem about pornographic art, which makes it impossible to appreciate an object as art and as pornography at the same time, because attention to its artistic aspect entails inattention to its pornographic aspect, and vice versa. Kierans response to this, naturally, is to deny the conict; however, he supports this denial only by arguing against views, such as those of Martha Nussbaum and Roger Scruton, according to which pornographic interest is necessarily objectifying or depersonalizing, and showing that it need not be. But this seems like a red herring. What needs showing is that attending to

Erotic Art and Pornographic Pictures

267

an image in order to be sexually aroused by it does not conict with attending to an image for its artistic features, not that attending to an image for its arousal value is not in conict with regarding the depicted subject as a person. Properly characterized, and not simply as highly explicit erotica, pornographys central aim, to facilitate sexual arousal in the name of sexual release, though it does not, as Kieran notes, preclude artistic interest from being present in an image, does, contrary to what Kieran says, militate against a viewers artistic engagement with the image, because it enjoins treatment of the image as transparent, as simply presenting its subject for sexual fantasizing, thus entailing inattention to the form or fashioning of the image. Hence if something answers to the central aim of pornography it cant at the same time answer to the aims of art. Thus, at the least, nothing can be coherently projected as both pornography, in the strict sense, and art.

VI It might be objected to my claim that the status of art and the status of pornography are mutually exclusive that pornography, understood as erotic imagery aimed at facilitating sexual arousal, fantasy, and release, does not preempt artistic interest on a viewers part, even when it is recognized as pornography, since some pornography works precisely by engaging the artistic interest of the viewer. The idea, in other words, is that some images are more arousing for some viewers when such viewers attend to or concentrate on aspects of the image as such, such as its form or style or embodied point of view, rather than merely being affected by them unwittingly in various ways. I see no harm in granting this, for perhaps there are viewers whose arousal is enhanced by attending explicitly to aspects of the vehicle of arousal. But even in such cases, so long as the image is being regarded as pornography, aspects of the image are not being appreciated for their own sakes, but only as instruments to more effective arousal, fantasy, and release. If such images are intended to be so regarded, then they constitute a complex mode of pornography, aimed at a cognitively atypical viewer, rather than instances of erotic art per se. That
For more on the alleged objectication of depicted subjects by makers or consumers of pornography, see my Sexual Perversity. Or at least, cognitively atypical male viewer. For it has been suggested to me that what I here label atypical is closer to typical for female viewers.

268

Pictures

sort of pornographywhat one might call artful pornography, consisting of images that invite attention to their artistic aspects precisely so as to enhance sexual arousal or fantasy involvement on the viewers partis perhaps the only serious challenge to the claim that art and pornography are necessarily disjoint. But that challenge fails nevertheless. Though for reasons already given the strategy of artful pornography is generally self-defeating, even when it is successful, and arousal is achieved precisely in virtue of the viewers attention having been drawn to the artistic aspect of the image, if such drawing of attention is entirely in the service of arousal aimed at, then the image remains pornography, however artful, and not art. So there is art that has a pornographic, or intentionally arousing, dimension. And there is pornography that has an artistic, or aesthetically interesting, dimension. But the former is not thereby pornography, and the latter is not thereby art. What usefully denes and differentiates pornography and art are their central aims, and those aims are incompatible. One requires form/ vehicle/fashioning to be transparent, while the other requires them to be, at least in some measure, opaque. What is arousing in pornography, generally speaking, is imagining interacting with or doing things to the depicted, usually unclothed, person, not pausing on the manner or means of depiction. Transparency of medium is all to the good of arousal, and is thus a virtual sine qua non of pornography. Opaqueness of medium is all to the good of art, but invariably weakens, and sometimes even wholly undermines, arousal. It is instructive in the present connection to consider some famous examples of European painting which clearly irt with the status of pornography. Consider Courbets notorious The Origin of the World, which graphically displays the midsection of a nude reclining woman, or in a less agrant vein, Ingress Turkish Bath or Velsquezs Toilet of Venus or a Bronzinos An Allegory. Were these, perhaps, the pornography of their day, despite the fact that they now grace the walls of our nest museums of ne art? Only, I suggest, if we are speaking hyperbolically. The arousal of male viewers was undoubtedly part of the intention with which they were
In terms of a formulation offered recently by Dominic Lopes in order to capture what it is to have an aesthetic interest in a photograph, when one consumes a pornographic photograph one is decidedly not appreciating seeing the photographed object through the photograph (Lopes, The Aesthetics of Photographic Transparency, Mind 112 (2003): 43348). Rather, one is relishing the sight of the photographed object simpliciter, the better to be aroused by and to fantasize about it. The vehiculing of that sight by photographic means is not an object of interest at all.

Erotic Art and Pornographic Pictures

269

painted; in the case of the Courbet, made on commission for a pashas private viewing, such arousal was perhaps the overriding intention. But surely the makers of those paintings intended as well that the attention of viewers be directed to features of the works themselvesthe handling of paint, the arrangement of forms, the play of internal perspectivesand the relation of those features to manifest sexual content. Thus those paintings count as erotic art, notwithstanding the fact that they can be regarded or employed pornographically, as perhaps they were by some of their owners. The work of Egon Schiele, perhaps the greatest of erotic artists, is also problematic for my thesis, but can be handled similarly. The inconvenient fact is that Schieles manifest intention for many of his sexually themed drawings was indeed pornographic, since they were created expressly for male patrons with precisely that sort of use in mind. There are, from my perspective, two ways to deal with this fact. The rst is simply to accept that, on the conception defended here, those drawings must be accounted pornography, but pornography that it is uncommonly aesthetically rewarding, and otherwise justiable, to treat as erotic art. The second is to posit for those drawings an implicit artistic intention as robust as the explicit pornographic one, as evidenced by the unmistakable aesthetic features of those works, in virtue of which they can be accounted, though uneasily, erotic art after all.

VII Viewed in a certain light, it might seem odd that there is even such a thing as erotic art. The appeal of art is fairly clear, as is that of pornography, but why should there be something that, so to speak, straddles the two, given their inherent opposition? A key to the distinctive appeal of erotic art, I think, lies in this. In normal circumstances, sexual stimulation leads beyond itself, to sexual arousal and sexual activity. Sexual stimulation is normally a preliminary, an antechamber, a way station in relation to what follows. It is thus not something to which one attends, or on which one concentrates, for itself. But with erotic art, the stimulation denitive of the category is hitched to a concern for the formal and
That said, were one to maintain that the Courbet, in particular, really does fall on the side of pornography, so prominent is its intention to arouse its target audience, I would not demur. Uneasily, again, because the pornographic projection of those works is in conict with their artistic projection.

270

Pictures

other features of the stimulating representation that anchor it in the broader category of art, a concern which thus inhibits such stimulation from taking its usual path towards sexual arousal and activity. We are thus constrained to dwell in and on our state of stimulation without going beyond it, appreciating instead its relation to and interaction with the other aspectsformal, expressive, social, politicalof the representation that occasions it. Part of what we enjoy in a work of erotic art can thus be described as a kind of tensionone between life and art, to put it simplya tension that generates an edgy pleasure akin to, though not identical with, that of the sublime. VIII Before concluding let me return very briey to the other philosophical questions concerning erotic art formulated at the beginning of this essay. One concerns how, with erotic art, stimulation or arousal are compatible with aesthetic appreciation, where this is understood to entail some degree of disinterestedness or capacity for contemplation. A short answer is that stimulation and arousal have to be held in check, neither suppressed nor given completely free rein, and attention made to focus on the erotic qualities of the picture in relation to the formal means that achieve or underlie it. The other concerns how, with erotic art, the erotic quality of a work and its artistic value are related. A short answer is that the more erotic a picture is, while not becoming effectively pornographicthat is, such as to induce fullblown sexual arousalthe better as art, provided the pictures erotic dimension is interestingly and intimately related to the other dimensions of the pictures content, something that it may take sensitive interpretation to establish. Beyond that it would be unwise to generalize. IX The central argument of this essay can be stated as follows: 1. Erotic art consists of images centrally aimed at a certain sort of reception, R1. 2. Pornography consists of images centrally aimed at a certain sort of reception, R2. 3. R1 essentially involves attention to form/vehicle/medium/manner, and so entails treating images as in part opaque.

Erotic Art and Pornographic Pictures

271

4. R2 essentially excludes attention to form/vehicle/medium/manner, and so entails treating images as wholly transparent. 5. R1 and R2 are incompatible. 6. Hence, nothing can be both erotic art and pornography; or at least, nothing can be coherently projected as both erotic art and pornography; or at the very least, nothing can succeed as erotic art and pornography at the same time.

This page intentionally left blank

PART I V I N T E R P R E TATI O N

This page intentionally left blank

16
Two Notions of Interpretation
I. There are many roads into the contested domain of interpretation. One road, which I will pursue here, centers on an opposition between two notions or modes of interpretation, which correspond roughly to two questions that might be asked about an artifact, event, or object. These questions are What does it mean? and What could it mean?. I will refer to the rst of these modes as the DM (does mean) mode and to the second of these as the CM (could mean) mode. The rst might also be labeled the determinative mode of interpretation, the second the exploratory mode of interpretation. An issue concerning interpretation in the CM mode that confronts us at the outset is how to construe the could in the question what could X mean? There are several possibilities here. We can construe could epistemically, the question then becoming what could X mean, given what is known?; or we can construe could logically, the question then becoming what is it logically possible for X to mean?; or we can construe could more pragmatically, yielding a question such as given both what is known and what is logically possible, what might X reasonably be taken to mean?. I will refrain from stipulating a construal of the could involved in CM interpretation, but will assume that the force appropriate in a given case will make itself evident. Another preliminary matter requires attention. I will here be concerned only with interpreting as an activity which seeks the meaning, signicance, purpose or role of that on which it is directed, and which issues in an interpretation stating or formulating some such meaning, signicance, etc. Call
First published in A. Haapala and O. Naukkarinen (eds.), Interpretation and its Boundaries (Helsinki: Helsinki University Press, 1998), 221. There is an obvious afnity between what I am calling exploratory interpretation and what Arthur Danto has dubbed deep interpretation (Deep Interpretation, in The Philosophical Disenfranchisement of Art (New York: Columbia University Press, 1986)), and even more between my notion and what Umberto Eco has dubbed overinterpretation (Interpretation and Overinterpretation (Cambridge: Cambridge University Press, 1992)). I will forgo, however, any attempt to pin down the respects in which the three notions differ.

276

Interpretation

this semantic interpreting. Obviously, semantic is here to be understood fairly broadly. I am thus putting to the side, most notably, the activity labeled interpretive in the performing arts which consists in playing, singing, reciting, dancing, staging, or enacting a pre-existing work in a particular way. Call that performative interpreting. This is not to deny that there are important relations between performative interpreting and semantic interpreting, nor to deny that performative interpretation may reect or contribute to semantic interpretation of something (such as the work being performed). It is just to insist that performative interpreting is not literally the same activity as semantic interpreting; that is to say, it does not, as such, seek and propose answers to the question of what something does or could mean. I will also put to one side the sort of interpreting that is done, in realtime, by translators of spoken language for the benet of those who do not speak that language. Call that translational interpreting. Though such activity indeed aims at rendering the meaning of a stretch of discourse, in most cases it involves little that is either decipherative or conjectural; that is, the meaning involved is evident to the interpreter, and its reformulation in the other tongue is a largely, if not wholly, mechanical affair. II. Before proceeding I recall some received wisdom about semantic interpretation. 1. Interpretation standardly involves the formation and entertaining of hypotheses, the weighing of possibilities of meaning, signicance, purpose, or role in regard to a given phenomenon or thing. 2. Interpretation standardly involves conscious thought, deliberate reection, explicit reasoning, or the like. Not all perception or understanding or apprehension is properly viewed as interpretive; some such is clearly preinterpretive, and serves as that on which interpretation rests, or that from which it departs. 3. Interpretation standardly presupposes the nonobviousness of what is being interpreted; if one simply and securely sees that X is F, if there is no question of choosing or deciding to do so, then remarking that X is F is not a matter of interpreting it.
See Ludwig Wittgenstein, Philosophical Investigations (London: Macmillan, 1958). See Richard Shusterman, Beneath Interpretation, in Pragmatist Aesthetics (Oxford: Blackwell, 1992). See Annette Barnes, On Interpretation (Oxford: Blackwell, 1988). See also G ran Hermern, o e Expression, Meaning, and Nonverbal Communication, in Jeanette Emt and G ran Hermern o e

Two Notions of Interpretation

277

I will more or less accept these three features as denitive of any activity worth labeling interpretive: such activity is hypothesis-involving, thoughtinvolving, and concerned with nonobvious attributions of meaning, signicance, purpose, or role. I do not claim, however, that these features are sufcient to mark an activity as interpretive. They may not be. For it seems that at least many cases of scientic inquiry and philosophical theorizing display those features without being what we would comfortably categorize as semantically interpretive activities. III. The following is a more or less random list of semantically interpreted things, in the rather broad sense sketched above: (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) a persons facial expression a persons gesture or bearing a persons action or behavior a rock in the desert tea leaves at the bottom of ones cup lines on the palm of ones hand an ambiguous phrase or sentence an ambiguous visual gure (e.g. a duck-rabbit) a situation (e.g. a swinging door or a missing car) an ofcial act (e.g. the closing of an embassy) a medical symptom (e.g. a rash or a fever) an x-ray or radiograph a Rorschach blot a work of literature a philosophical argument the starry heavens animal markings readings and measurements jokes and witticisms metaphors and aphorisms a stain on the sidewalk an unexpected natural event a hand shadow gure movements on the stock exchange the deliverances of an oracle

(eds.), Understanding the Arts: Contemporary Scandinavian Aesthetics (Lund: Lund University Press, 1992).

278

Interpretation

(26) (27) (28) (29) (30) (31) (32) (33)

a social custom a graph or diagram a chess move of ones opponent a tablet of cuneiforms a crossword puzzle clue dreams and daydreams a passage in the Bible a book in the Library of Babel

I am inclined to think, in reviewing this list, that the idea can be upheld that there are basically two different modes of interpreting, and corresponding to them, two different motivations for interpreting. For each of the above items, either the DM mode or the CM mode of interpreting recommends itself immediately, or else the two modes strike us as equally appropriate. Consider some examples. Querying a rash on a childs chest (11), say, seems straightforwardly a matter of DM interpreting; we want to know what it does mean, in the sense of what underlying medical condition it betokens. Querying an ambiguous gure (8) is just as clearly a matter of CM interpreting: we want to know in what ways it could be seen. A puzzling situation in real life (9) prompts us to discover its actual, and not merely possible, signicance, and thus calls for DM interpreting. Rorschach testing (13), a psychoanalytic procedure aimed at eliciting construals of any sort, with none counting as correct or incorrect, is thus a paradigm of CM interpreting. On the other hand, the starry heavens (16) may well be seen as calling both for explanation, whether cosmological or theological, and so DM interpreting, and also scanning with an eye for portents, and so CM interpreting. A hand shadow gure on a wall (23) may be regarded as an occasion for seeing in a number of waysCM interpretingor else as simply pointing to what the shadowmaker had in mind to projectDM interpreting. A passage in the
In commenting on this essay in its earlier incarnation Peter Lamarque (Objects of Interpretation, Metaphilosophy 31 (2000): 96124) suggests that to classify Rorschach blots as phenomena appropriate for CM rather than DM interpreting may be to misclassify them, on the grounds that Rorschach blots are used diagnostically, as signifying something about the subject who responds to them. But I remain unconvinced. It may be true that what a person P sees in a Rorschach blot R means a number of things as regards Ps personality, recurrent fantasies, and unconscious conicts, at least if the theory behind the diagnostic use of such blots is valid, but it does not follow that Rorschach blot R means those things. In other words, that reactions to a given blot are psychologically revealing doesnt show that there are in fact answers to what a given blot means, nor that we are seeking such answers in reacting to them. In short, that responses to Rorschach blots mean something doesnt show that the blots themselves mean something.

Two Notions of Interpretation

279

Bible (32) may be taken by some, in DM mode, to have just a single meaning, xed by the intentions of the divine author, while others, operating more in CM mode, take the passage as a springboard for tracing numerous at least possible meanings. But what are the two motivations that correspond to the two modes of interpretation we have identied, and which different phenomena call forth in differing degrees? I suggest that behind DM interpreting, in any sphere, lies a spirit that might be qualied as scientic, practical, and knowledge-seeking. Part and parcel of this spirit is a desire for understanding, explanation, discovery, or communication. Moved by this spirit we strive to establish or further our grasp of or contact with the real, whether in the form of nature in general or some particular human nature, that is, some individual mind or minds. Behind CM interpreting, by contrast, lies a spirit that might be qualied as ludic, liberated, and freedom-seeking. Central to this spirit is a desire for cognitive play, much like that which Kant located at the core of the aesthetic, without a concern for cognitive payoff of a concrete sort, a fascination with possibilities of understanding, explanation, discovery or communication, but no care for their actuality. Moved by this spirit we strive to deepen our appreciation of alternatives, of the spaces in which meaning is formed, but without privileging the actual or the crystallization of meaning that takes place therein; rather than seeking secure knowledge of the world, natural or man-made, we prefer to glory in our imaginative apparatus for doing so. One thing that can be expected to affect the sort of interpretive mode we adopt in confronting a given phenomenon, of course, is whether we understand it as having been designed for interpreting, or offered for interpretive activity, rather than as something that simply exists or occurs. But contrary to what one might at rst think, there is no automatic inference from an assumption of intentional projection to the appropriateness of a DM mode of interpretation, nor from an assumed absence of intentional projection to the appropriateness of a CM mode of interpretation. The relationship between the presence or absence of intention behind a phenomenon and the mode of interpreting aptly brought to bear on that phenomenon is a complicated one. On the one hand, jokes, radiographs, and grimaces are products of intention that invite DM interpreting, while metaphors and ambiguous gures and moves in chess are products of intention that invite CM interpreting. On the other hand, tracks, traces, and earthquakes invite DM inquiry even though they are not the products of intention, while certain equally non-intentional phenomena, such as sidewalk blotches,

280

Interpretation

tea leaf patterns, or the constellations of the night sky, strike at least some of us as an invitation to CM inquiry. Interpretation may be the characteristic method of the human sciences, and especially in its exploratory mode, but not all exploratory interpretation aims at the reconstruction or decipherment of intention. Now a reason someone might have for refusing to recognize what I call exploratory interpretation as interpretation at all is this. Interpretation, it might be said, presupposes the possibility of misinterpretation: misinterpretation could be claimed to be a necessary correlate of interpretation. But if so, then exploratory interpretation might be disallowed on the grounds that nothing under that rubric counts as misinterpretation. Certainly in the case of determinative interpretation misinterpretation has an obvious place. If you interpret something to mean what it does not, and you are operating in determinative mode, then you are misinterpreting it. But it is not clear that misinterpretation can get no toehold in the case of exploratory interpretation. Admittedly misinterpretation cannot there be glossed as it is for determinative interpretation, or else virtually all exploratory interpretation would count as misinterpretation, since it does not supply what a given item does mean. But misinterpretation in exploratory mode might still have an application, applying in those cases in which what is offered in response to the question of what something could mean is something that the given item could not conceivably mean. For instance, muddy footprints on the doorstep just could not mean that 3 is the cube root of 27, and a pained expression just could not mean the Statue of Liberty. Of course, for misinterpretation of an item to be possible it must in some sense be possible that the item has a meaning, but exploratory misinterpretation as just glossed requires no more than that. Thus, in sum, even if interpretation does presuppose the possibility of misinterpretation, that is no obstacle to counting exploratory interpretation as interpretation. IV. I will shortly examine more closely a number of other items on my list of semantically interpretable things. I will continue to be interested in the question of whether all varieties of interpretive activity can be illuminatingly categorized as aiming at answering either the what does it mean? or the what could it mean? question, but it is fair to say that a positive answer to that question serves as a working postulate of my investigation. In addition to that postulate of two basic modes of and motivations for interpretingwhat might be called Thesis 0 of this essayI offer

Two Notions of Interpretation

281

for consideration some additional theses, which will nd support in the subsequent discussion. Thesis 1: DM or determinative interpretive inquiry presupposes, at least defeasibly, that there is a single answer to the question of an items meaning or signicance, while CM or exploratory interpretive inquiry presupposes, at least defeasibly, that there are many such answers. Thesis 2: In some cases CM inquiry is only heuristic or instrumental to engaging in DM; in other cases CM inquiry is engaged in for its own sake. When engaged in for its own sake, CM inquiry has no predetermined stopping point, and might in principle continue indenitely. Thesis 3: DM inquiry invariably involves a phase or moment of CM inquiry engaged in heuristically or instrumentally. That is, in the course of determinative inquiry one will invariably consider what an item could mean or signify, in the interests of turning up reasonable candidates for what the item does mean or signify. I turn now to a more detailed look at certain items on my list. V. An X-ray or radiograph. Consider some typical radiological judgments: Thats a tibia fracture, This means intercranial bleeding, or That looks to be an ovarian cyst. What I have in mind are cases one might describe as easy calls. At rst blush, such cases appear to have nothing of the conjectural or hypothetical about them. That is to say, they are cases in which identication of what is shown on the X-ray occurs automatically and unreectively, cases in which, at least to the radiologist, what is shown is perfectly apparent. Is this sort of radiology then not interpreting after all? I would argue that it still is. Why? Here are some possible reasons. (a) Because such cases are continuous with other cases, that is, most or at least many occasions of radiology, which are more clearly interpretive. (b) Because such cases are instances of a certain activity, namely, reading a radiograph, that has an inherently interpretive status. (c) Because such cases involve applying an expertise, which allows radiologists to discern what is not open to normal vision. Though these reasons point in the right direction, none of them seems quite right. We will arrive at a more satisfying answer shortly. Of course not everything that is perceived is interpreted, as we have already noted. When I see cups, bowls, or spoons on the kitchen table I dont interpret those objects as cups, bowls, or spoons, nor arguably do I even see them as cups, bowls, or spoons. I simply see cups, bowls, or spoons, as the case may be, and perhaps that they are such. Ordinary (that is, non-artisanal) tableware is not as such intended to be read or construed, is not designed to carry information,

282

Interpretation

does not call for deciphering; a cup is a cup, but does not mean cup. By contrast, a given radiograph may show, indicate, indeed mean, such and such a medical condition, e.g. fracture, tumor, or murmur. Just seeing that something is a radiograph is not normally the end of cognitive engagement with it. Return now to the problem of the easy call, such as that provided by a practiced radiologist on an unproblematic radiograph, and the claim, ventured above, that this should still be classied as interpretive, by contrast with the unreective seeing of spoon or bowl at breakfast. The reason is that though the radiologist concludes more or less immediately that the X-ray indicates a fracture, cyst, or whatever, this is still arguably a matter of concluding, as opposed to simply registering. And part of the reason for so viewing it is that the radiologist has the responsibility to rule out other possible readings, either explicitly, through actual consideration of alternatives, or else implicitly, in virtue of his ability to tell one condition from another in radiographic terms, an ability acquired via training in which explicit consideration of relevant alternatives would have gured prominently. In other words, the radiologists reading of the X-ray is an act in which alternative construals are either manifestly or latently involved, and is thus not to be assimilated to cases of unmediated and unproblematic perception. A eld of alternatives is part of the activity of reading a radiograph in a way it is not part of everyday unselfconscious seeing. This serves, I believe, as an illustration of Thesis 3 above, that DM interpreting involves, at least defeasibly, a phase of CM interpreting. For clearly unproblematic radiography, like radiography in general, is to be understood on the DM model of interpreting if understood as interpretive at all. But as we have just seen, it may be right to so classify it precisely because it implies a space of possibilities acknowledged, in a prior and exploratory moment, even if this is a space which has been collapsed down to one possibility in the judgment actually issued. VI. A move in a game of chess. If approached in the DM, or determinative mode, one attempts to infer from ones opponents last move, in conjunction with the general position of the board, what ones opponents plan is for the next stage of the game. Doing this presupposes that there is such a plan, that ones opponent is not simply moving randomly or instinctively, and that that plan is singular and determinate. With that assumption in effect, one attempts to discern what specic succeeding moves ones opponent envisages will ensue. Interpretation in this guise is a matter of ascertaining what the given move means, in the sense of inducing or inferring the concrete scheme in which it gures in the mind of ones opponent.

Two Notions of Interpretation

283

By contrast, if one adopts the CM or exploratory mode of interpretation, one will seek to ascribe different plans to ones opponent, imagining different possible justications for the move in question, but without assuming psychological reality for any of them. One might interpret a chess move in CM mode either for the intrinsic interest of doing so or in order to be better prepared for various eventualities in the working out of the game or the various paths it might take. In CM mode, which evolution of the pieces on the board is most plausibly ascribed to ones opponent as the plan he has actually in mind to carry out is but one focus of interest in the move that ones opponent has just made. Equally important is the whole eld of possibilities a given move opens up, possibilities that can be viewed as meanings of the move in question, in the sense of representing alternative objectives that ones opponent might have had in view, whether or not he actually did. VII. A metaphor. In interpreting a metaphor, is one primarily seeking what it does mean, or what it could mean? That is, does one naturally adopt DM mode or CM mode? Here is an example I came across recently, a saying attributable to the novelist Balzac: Fame is the sun of the dead. Now, what could that mean? It could mean that as the sun sustains the living, fame sustains, or gives life to, the dead. Or it could mean that as living things naturally turn toward the sun, those contemplating death naturally turn toward the prospect of posthumous fame, for the consolation it may provide. Or it could mean that the dead constitute a kind of solar system, one centering on the property of fame, viewed as some kind of a Platonic entity. Or it could mean that too close an approach to fame by the dead will cause their wings to melt. But what Balzacs metaphorical remark does mean, I take it, is some subset of the things it could, with no restrictions on appropriateness or plausibility, mean. Thus, in the present case, what it means is arguably given, at least in part, by the rst two suggestions, but not by the last two. The principle of such selection, roughly speaking, is something like what meanings it would be plausible to hypothesize that the utterer of the metaphor intended to convey to hearers by uttering the metaphor in a given communicative context. In such cases, one asks initially what it could mean, but with the ultimate aim, in most cases, of arriving at an answer to what it does mean. Of course
This reects a hypothetical intentionalist perspective on literary meaning. (See Hypothetical Intentionalism: Statement, Objections, and Replies, Ch. 18, this volume. See also Whos Afraid of a Paraphrase?, Ch. 17, this volume, for further discussion of metaphor.)

284

Interpretation

what a metaphor does mean may very well, and usually does, exhibit multiplicity and ambiguity, but if so, there is a sense in which this simply enters into the meaning, broadly viewed, that it determinatively has. This is thus another illustration of Thesis 3. VIII. A clue in a crossword puzzle. Were one to regard a crossword puzzle as a sort of found object, as an unpremeditated conundrum posed by a fortuitous constellation of lexical elements, a conundrum which perhaps has an intelligible solution and perhaps does not, nothing would be more natural than to adopt toward it a CM interpretive stance. Since on that assumption there would be no question of the solution to the puzzle, the most one could do would be to hypothesize various possible readings of clues with an eye to making answers to them t together with answers to other clues, arrived at provisionally in the same vein. Now on the usual and well-founded assumption that an ostensible crossword puzzle is an actual crossword puzzle, devised by human hand, we will also entertain different ways in which its clues might be takenfor example, as straight, ironic, joking, elliptical, or self-referential. But note that we will do this only insofar as it is necessary to pin down the sense intended by the puzzle maker, one guaranteed to lead to an answer that works in solving the puzzle as a whole. So the ordinary exercise of completing a crossword puzzle, it seems, is one that involves a determinative stance toward the clues provided, however exploratory one may wax in the name of such determination. For our interpretive querying has then a clear target: namely, what was in the puzzle makers mind in offering us the clues that he did. For only in ascertaining that will one arrive at the correct solution to the puzzle in question. IX. A volume in the Library of Babel. Jorge Luis Borgess marvelous work of ction The Library of Babel, the lesser known and less discussed companion piece to Pierre Menard, Author of the Quixote, elaborates a conceit that is, in my opinion, at least as intriguing and as fertile as that which gures in Pierre Menard. As described by Borges, the Library of Babel is an immense, cyclical, and unending repository of all books possible within certain xed parameters. More specically, the Library contains all possible books of the following format: 410 pages in length, 40 lines to a page, 80 or so characters to a line, and employing 25 orthographic symbols, to wit, 22 alphabet letters plus the comma, period, and space. The number of such books, while not strictly innite, is mind-bogglingly large, as carrying out the permutational calculation reveals: it is greater than 25 to the 1,300,000th power. Moreover,

Two Notions of Interpretation

285

the length limitation on individual volumes is no real obstacle to the Librarys containing books of length greater than 410 pages, such as Prousts A la recherche du temps perdu or Gibbons Decline and Fall of the Roman Empire; they too exist there, only spread out over several volumes of this sprawling universe of books. Now of course only a minuscule fraction of this googolplex of tomes will appear to be in readable English, only a minuscule fraction in almost-readable French, and only a somewhat larger, though still tiny, fraction will appear to be written in any known human language. But so long as there are hands willing to separate the wheat from the chaff, there would seem to be much worth harvesting. Thus the rst reaction to such a cornucopia, as reported by Borgess narrator, is spontaneous euphoria at the idea of all possible books, in which all conceivable human knowledge, wisdom, insight, and literary achievement would thus be contained. But this is soon succeeded by the realization that such a Library must prove entirely worthlessmuch inferior, in fact, to that of the average junior college. For texts stripped of intentional projection, of context of generation, of even a specied language in which they are to be understood, are not works of literature of any kind, and have no determinate content. As pure texts, that is, as strings of syntactic elements that can be thought of as deriving from a process of permutation on a xed set of such elements, there is nothing to choose among them. What seems, at rst glance, to provide all the meaning in the world in fact provides no meaning at all. The prospect of an innity of meanings, which results when interpretation of an item is entirely unconstrained, turns out to issue in an absence of meaning. Borges strikingly underlines the uselessnessfrom the point of view of knowledge or discovery, at leastof even the seemingly most cogent volumes in this most disillusioning of libraries, by pointing out that every such volume has a multitude of near-doppelgngers that would appear to negate, contradict, or call into a question whatever it was that the given volume might be taken to be saying or conveying. The Library of Babel, Borges tells us, is literally a sphere. But it is also, guratively, a sphere for the purest expression of the CM interpretive impulse. Pick up any volume, posit a language or scheme of translation in which it is, at a ground level, to be understood, then posit, in the manner of the renewing
Actually less than a googolplex, but still much more than a googol. (A googol is 10 to the 100th power, while a googolplex is 10 to the googolth power.) That would seem a fair description of French without any diacritical marks to individuate vowels.

286

Interpretation

readings recommended at the end of Pierre Menard, an authorial intention and historical situation for the resulting text, and then seek to discern what further meaning or signicance might reasonably be ascribed to such a text. Next, begin again, with new values of the two variables. In this activity, whose game-like character is evident, there is no truth, no fact, about what the given volume does mean. But as it turns out, this is all the more liberating, allowing us the greatest possible freedom to exercise our conjectural and combinatorial powers concerning what something could mean. From the point of view of CM interpretation, at any rate, the Library of Babel is a valuable resource, an inexhaustible springboard to hermeneutical high jinks without limit. It also provides, incidentally, a conrmation of Thesis 2 above. X. Though we have only tested its mettle on a fraction of the thirty-three items on the list in section III, we can now be said to have some reason to regard the DM/CM opposition as a fruitful lens through which to view what can seem a bewildering variety of interpretive situations. We can also be said to have some reason to regard as sustainable as well the three theses enunciated earlier in connection with that opposition. Yet it is perhaps time to observe that DM interpreting and CM interpreting are, in a sense, not as separable as my discussion to this point may have suggested. That DM interpreting invariably involves, if passingly, a phase of CM interpreting is of course something already enshrined in Thesis 3. But CM interpreting also exhibits an involvement in the reverse direction, in the following sense: Asking what a given phenomenon could mean presupposes, for its intelligibility, at least the idea of what the phenomenon does mean, if not of course the existence of such meaning. Thus our two modes, though from one angle contrasting and alternating, can from another angle be seen as
It is a short step from this to the idea of virtual artworks as the product of willful abandonment of all intentional and contextual constraints on interpretation in regard to a given actual artwork, holding xed only manifest or perceivable structure on some construal of that. It can be argued, however, that virtual artworks, though not without interest, are ultimately much less valuable and important to us as human beings than are actual artworks. (See Jerry Fodor, Deja vu All Over Again: How Dantos Aesthetics Recapitulates the Philosophy of Mind, in Mark Rollins (ed.), Danto And His Critics (Oxford: Basil Blackwell, 1993), for development of the related idea of virtual etiologies for actual artifacts generally.) A possible objection to this is as follows: It may be that in most cases it could mean X can be glossed as it is possible it does mean X. But in some cases, for example, those of Rorschach blots, it is arguable that it cannot be so glossed. A response: it is true that when inviting interpretations of Rorschach blots you are interested in what someone just might see in the blot. But if such an interpretation is conceived as what the blot could mean, as opposed merely to what one might see in it, then it seems the usual gloss, reecting the claimed conceptual link between CM and DM interpreting, would have to be allowed.

Two Notions of Interpretation

287

mutually implicating. Neither is logically whole, so to speak, without the other. That interpretation in CM mode in some sense presupposes interpretation in DM mode might be seen as an illustration of the Aristotelian point that, at least conceptually, potentiality presupposes actuality.
It should be admitted, nally, that in many cases of interpretation we seem to be concerned neither with what something could mean, nor with what it does mean, but rather with what it is, seeking what Georg von Wright called explicative interpretation. But whether explicative interpretation falls within the ambit of semantic interpretation, even broadly conceived, is another question.

17
Whos Afraid of a Paraphrase?
1. Having never written on the topic of metaphor before, which rightly or wrongly I had always considered a fairly marginal one as far as aesthetics was concernedso much fuss for one little trope!I decided to begin where many modern discussions begin, that is, with Donald Davidsons seminal essay on the topic. I attempt to orient myself in this debate by xing on certain of Davidsons assertions in that essay and offering my reactions to them. As will be seen, I come down pretty clearly on the anti-Davidsonian side of the fence, in that I regard the idea of metaphorical meaning as ultimately defensible, as long as one correctly identies what has such meaning, correctly locates how such meaning is acquired, and acknowledges that such meaning is perhaps not all there is to a metaphor, depending of course on how broadly or narrowly one chooses to deploy the notion of meaning. Certainly if metaphorical meaning is restricted to what is in principle capturable by paraphrases, there is indeed more to metaphor than that. But there is also, and undeniably, it seems to me, that as well. There is indeed a tendency, when the object is to challenge the adequacy of paraphrase for the elucidation of metaphor, to construe the idea of paraphrase as implicitly importing an ambition of paraphrase without remainder, thus implying that paraphrases, if sufciently elaborated, give the whole of what a metaphor is about and can thus do duty for them. Evidently, insofar as such an unstated importation is not remarked, the pretension of paraphrase to a role in the elucidating of metaphor will be unfairly denied, being held to be necessarily more sweeping than it in fact need be. The fact that the task of exhibiting in literal language the metaphorical meaning of a metaphor might not, perhaps, ever be completely dischargedthe
First published in Theoria 67 (2001): 723. What Metaphors Mean, in Inquiries into Truth and Interpretation (Oxford: Oxford University Press, 1984). Page references to the edition cited are given in parentheses.

Whos Afraid of a Paraphrase?

289

fact that it might always be possible to expand or supplement the paraphrases with which one seeks to cash out such meaningshould not be thought to license the inference that therefore the task cannot be carried out, and thus that the paraphrases offered at any given point necessarily fail to articulate any part of the meaning that a metaphor possesses. 2. But to begin, now, with Davidson. (1) The concept of metaphor as primarily a vehicle for conveying ideas, even if unusual ones, seems to me as wrong as the parent idea that a metaphor has a special meaning . . . if I am right, a metaphor doesnt say anything beyond its literal meaning (p. 246). What this assertion most clearly overlooks is the fact that there are phases in the comprehension and reception of a metaphor, in the acceptance of it as apt or just, once proposed. Perhaps oversimplifying, we might posit a rst phase in which the metaphor, typically an evident falsehood if construed literally, generates cognitive dissonance, shakes up associations, induces seeing-as of an unfamiliar sort, and so on, all of which is by design, and a second phase in which, once the dust settles, certain paraphrasable meanings, if the metaphor is minimally effective, precipitate out of the rst, intentionally disorienting phase.* That the meaning of a metaphor only emerges at this second phase, while evading us during the rst, is no reason, of course, for failing to recognize the existence of such meaning. (2) The supposed gurative meaning of a simile explains nothing; it is not a feature of the word that the word has prior to and independent of the context of use, and it rests upon no linguistic customs except those that govern ordinary meaning. (p. 255). This pronouncement fails to recognize that even if a metaphor is based on nothing but the pre-existing literal meanings of its constituent words, there may yet be a subsequent, metaphor-specic interpretive custom or mini-practice that forms around a successful metaphor, so that it becomes, even before eventual pasturage as a dead metaphorand one might remark, parenthetically, that all metaphors should be so luckyan available descriptive resource of the language, one capable of right or wrong employment.*
Typically but not necessarily, as is shown by examples of twice true metaphors, such as No man is an island, brought to our attention notably by Ted Cohen. (See his Metaphor and the Cultivation of Intimacy, Critical Inquiry 5 (1978): 113, and Metaphor, Feeling, and Narrative, Philosophy and Literature 21 (1997): 22340). Here is another such metaphor: Life is no bowl of cherries.

290

Interpretation

(3) . . . what we attempt in paraphrasing a metaphor cannot be to give its meaning, for that lies on the surface; rather we attempt to evoke what the metaphor brings to our attention (p. 262). First of all, if it were really true that the meaning of a metaphor lies on its surfaceand one presumes, in construing this metaphor, that this is a smooth, unbroken surface, free of nooks, crannies, or potholes in which things can hidewould anyone ever need to have the meaning of a metaphor explained? But more importantly, this assertion ignores the fact that what is evoked by a metaphor often crystallizes in a fairly pronounced mannerthat is to say, signicant intersubjective convergence as to what the metaphor recommends to our attention manifests itselfthus allowing to paraphrase the job of articulating, at least partially, the meaning of the metaphor. Ted Cohen remarks much the same thing in a recent essay on our subject: The metaphorical content of a metaphorical expression is more or less specic . . . [while] there is no function that will calculate the metaphorical content of an expression from its literal meaning . . . there is a content that it is correct to take from the expression. Back to Davidson: (4) . . . in fact there is no limit to what a metaphor calls to our attention, and much of what we are caused to notice is not propositional in character. When we try to say what a metaphor means, we soon realize that there is no end to what we want to mention (p. 263). This declaration, nally, is blind to the fact that there are many things which, while they are indeed resemblances or similarities between the terms of the metaphor, are arguably no part of what the metaphor means, or is plausibly taken as conveying, as the history of subsequent interpretation of the metaphor would establish. For example, it is no part of the meaning of the metaphor No man is an island that no man is made of sand, whereas it is determinably part of such meaning that no man is inherently isolated from the society of other men.
Metaphor, Feeling, and Narrative, 227. A similar observation is offered by Anders Engstrom, in Metaphor and Ambiguity, Danish Yearbook of Philosophy 31 (1996): If there exists an inclination towards certain interpretations within a community of speakers, it should be clear that a basis for prescribing these would be possible, and that a notion of metaphorical meaning could be maintained (pp. 1213). Of course the determinacy of this meaning is partly a function of the metaphors embedding in Donnes poetic discourse as a whole, but that does not affect the point at issue.

Whos Afraid of a Paraphrase?

291

Richard Moran has made the point as follows:


It may be well be true, as Davidson says, that a metaphor makes us attend to some likeness, but it is not true that attending to just any of the innite aspects of likeness between the two things counts as understanding the metaphor . . . the process of interpretation couldnt even begin without some sense of which are the relevant dimensions of the comparison.

Understanding metaphor, in other words, involves hitting on and attending to the right likenesses, which the specic juxtaposition of the terms not customarily brought together brings to the fore, in the given context in which they are juxtaposed. 3. But the central claim in Davidsons brief against metaphorical meaning, and that on which his argument seems almost entirely to turn, is the claimI spare you a quote conrming that this is indeed the core of the argumentthat in a metaphor the constituent words carry only their original, literal meanings, and do not acquire new, metaphorical ones. Now with this claim I am inclined to agree: being used in a metaphor, however successful, does not affect the meaning of a word as a term in the language, even passingly. The problem, though, is that this in no way yields the conclusion desired, that metaphors, that is, metaphorical utterances, lack a meaning, acquired in context, a meaning that one may as well call metaphorical meaning. That in a metaphor the constituent words do not acquire new, metaphorical meanings, ones that might gure eventually in the dictionary, in perhaps fth or sixth place, and that the metaphor would not work, would not perform as a metaphor at all, were its constituent words not to retain their ordinary, preexisting, meanings, does not entail that the metaphorical sentence, or perhaps better, the sentence taken as a metaphor, does not acquire in situ a metaphorical meaning, one that paraphrases can be charged with exhibiting. It is hard to underestimate the rhetorical benet that Davidson draws from this usually
Moran, Seeing and Believing: Metaphor, Image, and Force, Critical Inquiry 16 (1989): 106. See also, and earlier, David Novitz, Knowledge, Fiction, and Imagination (Philadelphia: Temple University Press, 1987), ch. 7. David Novitz, however, argues against Davidson that in a successful metaphor some constituent words must in fact acquire new, metaphorical meanings, if the metaphor as a whole is granted to have a new, metaphorical meaning (Knowledge, Fiction, and Imagination, 1546). But I think that Novitz is impelled to this conclusion by not distinguishing sharply enough between sentences in a language, whose meanings are perhaps solely a function of the meanings of their constituent words, and utterances on an occasion, whose meanings are not.

292

Interpretation

unremarked shift from words or terms to sentences or utterances. The justice of his position in regard to the former redounds undeservedly, and somewhat surreptitiously, to the credit of the latter. 4. It is useful to compare the operation of metaphors to that of something more homely, such as exclamations. Consider the exclamation Fire!. Exclamations have meanings, in context, which go beyond the meanings of their constituent wordsor in this case, wordand often in a stable and persisting manner. This meaning is partly propositional, and thus paraphrasable, e.g. there is a re in the vicinity and everyone is urged to leave, and partly non-propositional, consisting in an expressive, illocutionary force, though one that can also be described well enough. Would it be reasonable to argue that since the constituent word of this exclamation retains its meaning in the exclamation itself, and since the exclamation would not function as such were that word not to do so, therefore there is no further, as it were, exclamational meaning to that exclamation, one that outlasts its use on a given occasion? I think not. Consider also more closely the issue of dead metaphors. Here are two from roughly the same sphere, as regards their literal roots: you are the light of my life and never again darken my door. Obviously such expressions now have quasi-literal meanings themselves, as not very extended paraphrases of them would conrm, such as, respectively, you impart meaning and joy to my life and are central to it and you are not welcome in my home and I hope never to see you again. But now try to picture those dead metaphors in their youth, a youth shrouded in the mists of time and no doubt unrecoverable.* Is it yet plausible to think that what is captured in the paraphrases offered a moment ago, those homespun distillates, was no part of the meaning of those metaphors when newly minted, that those metaphors acquired such meaning, all of a sudden, only after being laid denitively to rest in the graveyard of spent expressions?* To take such a line, it seems to me, is to adopt a sort of doctrine of semantic creation ex nihilo as regards metaphor, the prospects of which seem no brighter than those of the parent doctrine in theology. 5. It is sometimes remarked that, even granted the differences between individual words in isolation and complete utterances in context, Davidson is still right to deny metaphorical meaning to metaphors. And that is because the supposed meanings of metaphors, unlike the literal meanings of words in a language or the literal meanings of sentences composed from them in rulegoverned ways, are entirely bound to occasions of use, are simply a matter of

Whos Afraid of a Paraphrase?

293

what speakers intend to convey on such occasions, and so belong entirely to pragmatics rather than semantics. The nal upshot of this, then, is that that they do not deserve to be called meanings at all, since they are not possessed of any generality, stability, or exportability outside the specic circumstances in which they see light, and thus have no explanatory value as far as semantic theory is concerned. But this sort of vindication of Davidson, however guarded, mistakenly exaggerates the occasion-boundedness of metaphors. Although born on given occasions, and acquiring concrete, if never completely paraphrasable, signicances in specic contexts of utterance, successful metaphors retain such signicances in enduring fashion, ones they carry with them on future occasions of use, once successfully constituted. The acquisition of signicance by a metaphor on the occasion of successful use is often a relatively permanent one, and the more so, one may suppose, the more just or compelling the metaphor. The signicance of the metaphorical ascription of sunhood to a young woman by her lover is now, some ve hundred years after Shakespeares Romeo and Juliet, hardly less stable, less exportable, less available for meaningmaking in the sense of speakers meaning, than the signicance of a literal ascription of sunhood to a particular heavenly body. That metaphors are born in context and acquire their content therein, against a complex background of shared understandings, assumptions, and dispositions, should not lead us to think that metaphors necessarily perish once those originating circumstances have past, and with them the content then acquired. No, the content of many metaphorsthat is, metaphorical assertions, understood as species of utterance is as general, stable, and exportable as that of most literal assertions, and so would seem to count as semantic content of some sort. If describing some thing as the opiate of the people, or the hobgoblin of little minds, or a wound in the side of ones country, or a tale told by an idiot and full of sound and fury, or someone as a snake in the grass, or a utensil, or an Emma Bovary, or no Jack Kennedyand I have tried here to avoid metaphors that would be accounted irretrievably deaddoes not
Qualied support of Davidson in this vein, preliminary to advancing a theory of metaphorical meanings that would not be subject to such objection, can be found in Josef Stern, What Metaphors Do Not Mean, Midwest Studies in Philosophy 16 (1991): 1352. I was about to add, or an affair between a predicate with a past and an object that yields while resisting, but then I had to admit that I couldnt think of a reasonable re-employment of this wonderful metaphor of Nelson Goodmans, that is, to characterize something other than metaphor itself.

294

Interpretation

qualify as employing an expression with an available, paraphrasable, occasiontranscendent meaning, then I begin to be unsure what does. There is something called cultural literacy, in the sense E. D. Hirsch has made familiar, and having a feeling for the sense of such metaphors, utilizable outside the site of their original application, is a part of it if anything is. 6. Another source of skepticism as regards metaphorical meanings has its roots not in Davidson, but in the later Wittgenstein. Wittgensteins discussion of secondary uses of language, presupposing yet diverging from those which can be considered primary, is taken by some to show the pointlessness of postulating secondary senses borne by such language in its secondary uses, in addition to the senses such language carries in virtue of its primary uses, uses that ground the language game in question. But cases of the sort discussed by Wittgenstein in his remarks on secondary uses of language, designed to forestall the postulation of parallel secondary senses, seem to me very different from those of successful metaphor. Wittgensteins favored examples, recall, concern the colors of vowels, e.g. e is yellow, or degrees of stoutness of the days of the week, e.g. Tuesday is thin. Now such examples of, let us call it, quasi-metaphorical assertion have an evident whimsicality, idiosyncrasy, its-a-free-country-ism about them that is quite foreign to cases of genuine metaphor. In genuine metaphor, there is a rightness, or click, in the disparate things that the metaphor brings into conjunction, which unleashes a meaning for the sentence that was all along a potential for specic illumination, if an unnoticed one, of the literal meanings of the terms employed once conjoined. But even were a certain predisposition to view e as yellow as opposed to other colors, or to view Tuesday as thin as opposed to fat, to be observable in a given linguistic community, that would hardly be enough to make yellow or thin a permanent descriptive resource of the language of that community in regard to vowels or days of the week. Thus, the fact that secondary uses of language of this sort do nothing to prompt recognition of corresponding senses of the terms involved does not
As David Hills suggests in a recent insightful essay on our topic, bona de metaphors possess some degree of poetic power (Aptness and Truth in Verbal Metaphor, Philosophical Topics 25 (1997): 119). Hills usefully identies and defends two positions on metaphor that he labels aestheticism (metaphors are rightly assessed aesthetically, that is, for aptness) and semanticism (metaphors have semantic content, that is, a distinctive paraphrasable meaning). As should be clear, these are positions to which I also subscribe. The idea that Tuesday might generally be regarded as thin, incidentally, has always struck me as especially misguided, given the conicting force of the concrete and, one might add, public association of Tuesday and fatness in the holiday of Mardi Gras.

Whos Afraid of a Paraphrase?

295

cast much, if any, suspicion on the idea that bona de metaphorical uses of language might be said to underpin graspable metaphorical meanings corresponding to them. 7. Elsewhere I have developed a view of literary interpretation, or at least of the central meaning of a text offered as literature, that I call hypothetical intentionalism. Insofar as metaphors are utterances, and insofar as they can be seen in particular as literary utterances, albeit of a small-scale sort, then a hypothetical-intentionalist view of literary meaning would seem naturally to recommend itself for getting a handle on metaphorical meaning as well. What, then, does such a view maintain? According to hypothetical intentionalism a literary work is to be construed as an utterance, one produced in a public context by a historically and culturally situated author, where the central meaning of such a work is thus a form of utterance meaning, as opposed to either textual meaning, the meaning of the brute text as a string of words in a language, or utterer meaning, the meaning the utterer, speaker, or author had in mind and intended to get across. Utterance meaning, in turn, is understood on a loosely Gricean model according to which what an utterance means is a matter, roughly, of what an appropriate hearer would most appropriately take an utterer to be trying to convey in employing a given verbal vehicle in the given communicative context. As applied to literature, and eshed out in certain ways, what it amounts to is roughly this: the core meaning of a literary work is given by the best hypothesis, from the position of an informed, sympathetic, and discriminating reader, of authorial intent to convey such and such to an audience through the text in question. In the light of that explication, then, we can give a fairly straightforward hypothetical-intentionalist account of the fact, emphasized by a number of writers on metaphor, that understanding a metaphor requires identifying or homing in on the right likenesses or connections between terms in a given case, putting aside those that are not to the purpose. For correctly
See my Intention and Interpretation in Literature and Messages in Art, both in The Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996), and Hypothetical Intentionalism: Statement, Objections, and Replies, in M. Krausz (ed.), Is There A Single Right Interpretation? (University Park, PA: Penn State University Press, 2002), and Ch. 18, this volume. I am not alone in advocating such a view: William Tolhurst, Alexander Nehamas, Gregory Currie, and Stephen Davies subscribe to something similar, and the broadly Gricean outlines of the view will, at any rate, be familiar. In addition, certain current competing views, such as the actual intentionalisms defended by Nol e Carroll, Robert Stecker, and Paisley Livingston, are perhaps more hypothetically intentionalist than they appear to be on the surface.

296

Interpretation

understanding a metaphor in context, by hypothetical-intentionalist lights, involves arriving at the best hypothesis, in epistemic and aesthetic senses, as to what a speaker would likely have intended to convey or draw attention to, what similarities a speaker would plausibly have wanted to highlight or bring into focus, bearing in mind that this will naturally open out into resonances that a speaker would have endorsed or welcomed in connection with the metaphor offered, if not ones it is plausible to conjecture were foreseen or perceived in advance. Now even if this is accepted, and the identication of metaphorical meaning, a species of utterance meaning, is understood to run along hypothetical-intentional lines, it remains true, in my opinion, that whether or not a given utterance is in fact a metaphorical one is irreducibly an affair of actual intentions, to be determined by whatever means serve generally to determine the actual intentions of utterers. That is, in suggesting the plausibility of a hypothetical-intentionalist view of metaphorical meaning, as a species of literary meaning, I am not thereby proposing, and in fact I would explicitly disavow, a hypothetical-intentionalist view of metaphorical status. For indeed anything of a certain linguistic form, roughly, that of an assertion, positive or negative, might be being projected as a metaphor or with a metaphorical intent, and thus the most plausible construction we might put on an utterance might categorize it incorrectly. On the other hand, in the absence of knowledge of whether something is in fact a metaphorical assertion, it would be reasonable to invoke a hypothetical-intentionalist principle as a justication for revisably ascribing or withholding metaphorical status to a given utterance. 8. Until now I have been highlighting the character of metaphors as possessors of relatively stable and graspable meanings, and as reusable on occasions other than those in which they rst arise. I want at this point, however, to underscore an aspect of metaphors that runs in rather the opposite direction, that is, as against their, as it were, complete exportability. What I call attention to now is the inseparability of content and form in metaphor,
For discussion, see Intention and Interpretation in Literature. Thus I can concur with this remark of Max Blacks regarding assignment of metaphorical status, but only in the sense of defeasible assignment absent knowledge of actual intention: Our recognition of a metaphorical statement depends essentially upon two things: Our general knowledge of what it is to be a metaphorical statement, and our specic judgment that a metaphorical reading of a given statement is here preferable to a literal one (More About Metaphors, in Andrew Ortony (ed.), Metaphor and Thought, 2nd edn. (Cambridge: Cambridge University Press, 1993), 356).

Whos Afraid of a Paraphrase?

297

something which, reasonably construed, I take to be valid for all artistic phenomena. It can be asked whether metaphor should be conceived as primarily a conceptual matter, as primarily a linguistic matter, or rather as one that essentially straddles the divide. Is the essence of a metaphor a sentence in a language, that is, a sequence of words, or is it rather a constellation of concepts, which different strings of words make available? If the latter, how can those concepts be identied apart from a given linguistic formulation, given the presumed impossibility of exact translation of almost any term in a natural language into any other? Does the possibility of neutral designation of the concepts involved in a given metaphor require us to presuppose something like a language of thought, or at least some identity of concepts across users of different languages, however that comes about? I am inclined to the view that regards the concrete vehicle of a metaphor more or less those words, in that language, in that order, with just those precise rhythms, resonances, and prosodic propertiesas ineliminable, and thus as not after all the mere vehicle, but the very body and soul of the metaphor.* That is to say, even once the cognitive content of a metaphor is approximated through acceptable paraphrases, and its imagistic force, that in virtue of which we are made to see one thing in terms of another, is identied as well, there seems to remain a residue that attaches to the specic feel of the words employed to invoke that paraphrasable meaning and put that imagistic force
I do not assume that verbal metaphors are literally speaking works of art, but only that they can be appropriately treated in many respects as if they were, e.g. they can be assessed for their aesthetic merit. See Arnold Isenberg, On Dening Metaphor (Aesthetics and the Theory of Criticism (Chicago: University of Chicago Press, 1973), 10524), who aptly observes that metaphors are always strokes, if not always works, of art. The view that metaphor is essentially a conceptual matterthat what are called conceptual metaphors are the fundamental ones, and that ordinary verbal metaphors are just the verbal expression or externalization of such underlying conceptual metaphorsis a view that at present enjoys considerable currency, owing principally to the work of George Lakoff, Mark Johnson, and Mark Turner. (See Lakoff and Johnson, Metaphors We Live By (Chicago: University of Chicago Press, 1980); Johnson, The Body in the Mind (Chicago: University of Chicago Press, 1987); Lakoff and Turner, More Than Cool Reason: A Field Guide to Poetic Metaphor (Chicago: University of Chicago Press, 1989); Lakoff and Johnson, Philosophy in the Flesh (New York: Basic Books, 1999).) I dont wish to deny that there may be, on some psychological level, what these thinkers call conceptual metaphors. I only wish to insist on the at least equal claim to reality of metaphors as specic pieces of language, and the prima-facie irreducibility of the latter to the former. (For similar cautions see Hills, Aptness and Truth in Verbal Metaphor.) See Morans two-part analysis of metaphor, in Seeing and Believing, according to which metaphors have, on the one hand, a paraphrasable cognitive content, and on the other, an evidently non-paraphrasable imagistic force, consisting in a prescribed, asymmetric viewing of one thing through the lens of another.

298

Interpretation

in play. In short, the verbal substance of a metaphor bids fair to be considered essential to it, and not just its conceptual structure. But perhaps we should recognize two notions of metaphor, or two criteria of identity for metaphors, a restrictive and a permissive. On the permissive criterion, the metaphor is detachable from its linguistic formulation, and is a sort of thought, or way of thinking, with respect to certain things. On the restrictive criterion, the metaphor is undetachable from its linguistic formulation, and is thus like a miniature work of poetry, akin to a haiku. Which of these notions or identity criteria is more apt would seem to depend on the context of inquiry into metaphor. Thus, it seems advisable to recognize both. At any rate, with a metaphor considered as something like a work of literary art, if a tiny one, we must experience the content of the metaphor, including both its paraphrasable meaning and its imagistic force, through its specific verbal form, if we are to fully appreciate the metaphor. That outstanding metaphors can usually be translated from one language to another should not lead us to forget that the result is a translation, in which some part of the metaphor has been lost, in that part of the metaphorical charge of the original is indissolubly bound up with specic words, specically deployed. Un bon mot in one language is not necessarily un bon mot in another, to which that very phrase bears witness; for a good word neither really means the same thing as, nor has the same uid sound as, un bon mot, which sound contributes to its being the expression it is and to its conveying what it does. Metaphors are certainly not guaranteed to survive largely intact when translated, even faithfully, from one language to another. Man is a wolf to man is one metaphor, in English, but another, Lhomme est un loup pour lhomme, in French. For notice that the metaphor in French suffers, in comparison to the metaphor in English, in possessing to a slightly lesser extent, because of the unavoidability of articles in French, the quasi-palindromic quality which is a feature of both versions, and which reinforces the idea of reciprocity of behavior that is at the conceptual core of the metaphor. Assonance, alliteration, symmetry, syncopation, and so on are all part of a metaphor as a verbal entity. Translations of metaphors from one language to another, in short, might with justice be considered metaphorical cousins of one another,
In Max Blacks terms, I am suggesting that metaphors are paradigmatically emphatic: A metaphorical utterance is emphatic to the degree that its producer will allow no variation upon or substitute for the words used (More About Metaphor, 26.) The Latin original of the metaphor, ascribed to St Francis of Assisi, is homo homini lupus, which, again, has its own peculiar linguistic avor.

Whos Afraid of a Paraphrase?

299

rather than simply the same metaphorical individual in different linguistic dress.* 9. In approaching the writing of this essay, my rst idea, I confess, was to compose it entirely in metaphor, as a way of demonstrating rhetorically that effective metaphors, construed in context, have a paraphrasable meaning operative within a linguistic community, whatever non-paraphrasable force or charge they may have as well, and can thus communicate a set of thoughts, advance a position, perhaps even ignite a revolution. Since that seemed, on reection, rather too hard to carry off, I decided instead to cast an interested eye on the novel I had just begun rereading, David Lodges Small World, combing it for gures with which to test the validity of the convictions about metaphor I have been airing here. As it turns out that novel, being written in a casually elegant but not especially lyrical style, is not overly rich in metaphor, generally achieving its effects by other means, notably similes, which in comparison are fairly thick on the ground.* In any event, here is an almost complete listing of metaphors from the rst fty pages: 1. 2. 3. 4. 5. Dismay had already been plainly written on many faces (p. 3). . . . the narrow beds, whose springs sagged dejectedly in the middle (p. 3). He exhaled rather than pronounced the syllables (p. 9). . . . he bent forward over his roast shoe-leather (p. 11). The rest of the audience was performing the same tableau of petried boredom as before (p. 16). 6. . . . you have succumbed to the virus of structuralism (p. 27). 7. . . . the individual campus is . . . the heavy industry of the mind (p. 43). I take it that few of us would have much trouble in discerning what is conveyed by these gures, in elaborating the connotations the metaphorically
I have, however, employed a larger than usual number of at least somewhat live metaphors in this essay, which I can now reveal is what the unexplained asterisks after certain sentences the observant reader will have noted are meant to mark. But, and this was the point of my little conceit, I take it that none of those metaphors impeded understanding of the cognitive content of this essay, and may even have facilitated that understanding. David Lodge, Small World (London: Penguin Books, 1984). Page references are in parentheses. The qualication almost is required partly because the borderline between metaphorical senses and second-order or third-order literal senses is irredeemably blurry. So for example, the list given does not include His face darkened as he added . . . (p. 7), since I regard darkened there as literal, but the case might be made that darkened in such a context is simply a very tired metaphor. Full disclosure also prompts me to add that, in those rst fty pages of Lodges novel, two characters end up discussing a metaphor, that of a woolly fold as it occurs in Keatss poem The Eve of St. Agnes.

300

Interpretation

invoked predicate carries in relation to the literal subject to which it is, with a characteristic degree of unusualness, applied. For example, as regards the last and most ambitious of these metaphors, the individual campus is therein painted as unwieldy, as on its way to obsolescence, as perhaps even harmful to the health of its familiars.* Note that the specic novelistic context is not even necessary for the construal of these metaphors, except in the case of the third, where it perhaps helps to know that the syllables exhaled form the name of a girl, Angelica, whom the exhaler has fallen in love with moments after making her acquaintance. But rather than devote any more time to metaphors of middling quality, I look briey to the opposite ends of the metaphorical spectrum, addressing rst two metaphors which are in my opinion especially ne, as well as being, as it happens, good examples of humor, and then those that have, as it were, fallen off the scale, in the sense that it is impossible, or virtually impossible, to construe them as metaphors at all, despite their partaking of the canonical form A is B where A is B is manifestly false.* Here are two candidates for excellence in the category of metaphor, the one encountered recently, the other an old favorite: Lotteries are a tax on the mathematically challenged. A wife is an umbrella; sooner or later one hails a cab. Ill spare you any paraphrase of these metaphorical assertions, in the rst case because its sense is fairly transparent, in the second case because, though its sense is not immediately transparent, it is more delicious if deciphered on ones own. I turn now to the outcasts of the society of metaphors, those sorry wouldbe tropes consisting in manifestly false identity sentences that simply resist metaphorical redemption.* Note rst that it is in fact rather difcult, and perhaps impossible, to nd such entirely metaphor-resistant sentences. It is instructive to consider Richard Morans throwaway example of a supposedly completely protless yoking of two items, the taste of sugar and the discovery of America, which, so he claims, must fail to strike any metaphorical sparks because the things brought together have inherently no relation to one another.* But the statement The taste of sugar was the discovery of America would actually not be too bad a way of conveying, metaphorically,
Offered by Freud in his Jokes and Their Relations to the Unconscious (New York: W. W. Norton, 1963). Seeing and Believing, 106.

Whos Afraid of a Paraphrase?

301

that the exploitation of cane sugar in the Caribbean islands was a driving force in the further exploration of the Americas. Nevertheless, it seems to me that almost metaphor-resistant sentences can be found, though perhaps not at random. One formula for balking metaphoricity would seem to be to equate assertively things from the same family or category, at roughly the same level of specicity. Thus it appears difcult, if not impossible, to make anything metaphorically of gold is lead, a fox is a wolf, or every woman is a man. On the other hand, given a bit more context and a crucial third element in the mix, a metaphor equating things on the same level from the same family can sometimes work remarkably well, as in this curious witticism: Wagner is the Puccini of music. Now, what does this mean? What is the cognitive content or imagistic transformation that, once caught sight of, causes us to laugh? It is admittedly hard to say. But here is a stab at it. Puccini, through being invoked to position Wagner in relation to music, seems to have been placed outside of music entirely, a rather grave insult to a composer, one has to admit, while Wagner, for his part, comes off hardly better, having been guratively equated with a composer who was not, so to speak, even a composer at all. In any case, to revert in closing to what is probably my main theme, note how this metaphor, once effectively essayed, becomes an available resource, one deployable, with suitable substitutions, on other occasions, and with what one sees no good reason, ultimately, not to regard as a metaphorical meaning. Thus, Rachmaninov and Respighi might also be called, of course unfairly, Puccinis of music, Queen the Puccini of rock music, George Winston the Puccini of jazz, Renoir the Puccini of painting, Gore Vidal the Puccini of American literature, and so on. For all I know, I may well be the Puccini of metaphor studies. At any rate, Ive made a start toward the title.
As I say, difcult, but not impossible. Thus gold is lead might be employed, in the right context, to express disapproval of earthly riches, or admiration for golds specic gravity. Work in linguistics on the permeability and interconnectedness of semantic elds would in fact argue against there being any cases of sentences whose metaphorical deployment was to be precluded absolutely. (See Eva Kittay, Metaphor: Its Cognitive Force and Linguistic Structure (Oxford: Oxford University Press, 1990).) Max Black offers this as an example of a likely metaphor-proof form of words: a chair is a syllogism (More About Metaphor, 23). Black is silent on why such a predication resists metaphorical interpretation, but one may speculate that it turns on both the utter dissimilarity and the categorial remoteness of the terms involved. What are the musical failings for which Wagner directlyand Puccini indirectly, if more harshlyare here being reproached? It is hard to say for sure, but I imagine windiness of expression, cheapness of effect, and dramatic overblownness are among the failings targeted. (I hasten to add that I am not here endorsing those charges!)

18
Hypothetical Intentionalism: Statement, Objections, and Replies
1. In an earlier essay I have defended at some length a view on literary interpretation which I call hypothetical intentionalism. The view centers on the idea that a literary work should be seen as an utterance, one produced in a public context by a historically and culturally situated author, and that the central meaning of such a work is thus a form of utterance meaning, as opposed to either textual meaning or utterer meaning. Utterance meaning, in turn, is understood on a pragmatic model according to which what an utterance means is a matter, roughly, of what an appropriate hearer would most reasonably take a speaker to be trying to convey in employing a given verbal vehicle in the given communicative context. As applied to literature, and eshed out in certain ways, what it amounts to is this: the core meaning of a literary work is given by the best hypothesis, from the position of an appropriately informed, sympathetic, and discriminating reader, of authorial intent to convey such and such to an audience through the text in question. Thus hypothetical intentionalism is a perspective on literary interpretation which takes optimal hypotheses about authorial intention, rather than actual authorial intention, to provide the key to the central meaning of literary works. Since the key notion here is that of a best hypothesis on the part of readers, which would seem to entail the consideration by them of a variety of hypotheses, it might seem that hypothetical intentionalism was committed to a picture of literary interpretation as a species of what I have elsewhere
First published in M. Krausz (ed.), On the Single Right Interpretation (University Park, PA: Pennsylvania State University Press, 2002), 30918. See Intention and Interpretation in Literature, in The Pleasures of Aesthetics (Ithaca, NY: Cornell University Press, 1996): 175213. The view is further deployed and developed in my Messages in Art, also in The Pleasures of Aesthetics: 22441. For an initial statement of the view, see William Tolhurst, On What a Text Is and How It Means, British Journal of Aesthetics 19 (1979): 314.

Hypothetical Intentionalism

303

characterized as CM, or could mean, interpretive activity, as opposed to DM, or does mean, interpretive activity, the latter aimed at identifying what, if anything, something does mean, with the former concerned rather with what it is possible, in a manner of speaking, that something might mean. But this is where a distinction between heuristic (instrumental) and nal (intrinsic) engagement in CM interpretive activity is important. For hypothetical intentionalism, CM engagement with a work is purely instrumental, and not undertaken for its own sake. The consideration of various possibilities of construal serves only to identify what is in fact the best hypothesisthat is, the most explanatorily plausible and, to a lesser extent, aesthetically charitable construction we can arrive atregarding a works intended import. What a work in fact means, however multifaceted that may be, remains the focus of inquiry. 2. Does a view like hypothetical intentionalism allow for a multiplicity of distinct, at least nominally incompatible interpretations, when interpretation is being conducted ultimately in a DM or determinative spirit? The answer is yes, in virtue of the existence of ties or draws, roughly speaking, among distinct and competing hypotheses concerning a works import. That is to say, nothing precludes there being, in a given case, two or more informed hypotheses frameable as to authorial intent that are explanatorily and aesthetically optimal. Even so, there are ways in which to view such multiplicity so that it becomes, from a certain angle, unitary. I suggest that it is always possible, in principle, to combine competing reasonable rst-order interpretations of a work into a totality that is embraceable from a more encompassing perspective. The fact that we have no logical notion handy for representing the ensemble of acceptable interpretations taken together does not show that the most correct and comprehensive interpretation of a work of art is anything other than that ensemble. The logical notions that naturally suggest themselves, conjunction and disjunction, are both, in different ways, misleading or unsuitable. We dont wish to say that the overall correct interpretation of W, where R1, R2, R3 are individually acceptable interpretations of W, is just R1 or R2 or R3, nor do we wish to say that it is just R1 and R2 and R3. Rather, it is each and all of R1, R2, and R3, and yet not their simple conjunction or disjunction. The best, most correct and comprehensive, interpretation of a work of
See Two Notions of Interpretation, in A. Haapala and O. Naukkarinen (eds.), Interpretation and Its Boundaries (Helsinki: Helsinki University Press, 1999), and Ch. 16, this volume.

304

Interpretation

art subject to multiple individually justiable readings must be an interpretation which enfolds all such readings: a kind of global or subsumptive reading, so to speak, which acknowledges all the individually acceptable readings and puts them into relation with one another. That is the sort of perspective I am inclined to adopt toward those examples of multiple acceptable yet incompatible readings that have gured so prominently in discussions of artistic interpretation, prompted by works such as Kafkas Castle, Jamess Turn of the Screw, Becketts Waiting for Godot, and De Koonings Woman paintings. To make this a bit more concrete, my global or subsumptive interpretation of, say, Kafkas The Castle might run roughly as follows. The Castle reasonably admits of theological, bureaucratic, psychoanalytical, existential and epistemological readings, and in ways that can be related to one another, mostly reinforcingly; on the other hand it does not reasonably admit, say, of entomological (castle as teeming beehive), chivalric (castle as prison of beautiful damsel), or oneiric (castle narrative as dream report) ones, nor do such readings relate reinforcingly to any of those already acknowledged as admissible. There can be little doubt that if the individual interpretations of The Castle invoked above, e.g. the theological or bureaucratic ones, have merit, then the best interpretation, tout court, of Kafkas novel will be an inclusive one having more or less the form indicated. If that is borne in mind, the narrowly logical problem of reconciling multiple distinct and individually insightful readings of such a work, with which philosophical discussion has been excessively preoccupied, will seem of reduced importance, or even to have disappeared. 3. But what, more precisely, is the form of global or subsumptive interpretations of the sort I have invoked above? The form of such an interpretation, I*, would seem to be something like this: Ws meaning is such that it is partly given by/aptly viewed under interpretation 1, partly given by/aptly viewed under interpretation 2, . . . and partly given by/aptly viewed under interpretation n, where those embedded interpretations, I1, I2, and so on, are understood as rst-order sub-interpretations subject to the higher-order
A similar perspective is advanced by Robert Stecker, Art Interpretation, Journal of Aesthetics and Art Criticism 52 (1994): 193206. Though all true statements are conjoinable, that may not be the best way to hook up a pair of true interpretations into a more comprehensive true interpretation . . . If The Turn of the Screw is intentionally ambiguous, it doesnt represent the governess as battling with hosts and having hallucinations . . . Better, the novella is such that it can be correctly read either as representing the governess as battling ghosts or as representing the governess as having hallucinations (p. 201).

Hypothetical Intentionalism

305

interpretation, I*, which subsumes them, though not simply disjunctively or conjunctively. We might further want to distinguish among such global or subsumptive interpretations those which, like the above, are simply collective (or enumerative), and those which are integrative (or hierarchic), and thus include an account of the relations of importance or centrality obtaining among the subinterpretations brought together in the global interpretation. Admitting global/subsumptive interpretations of an integrative/hierarchic, rather than simply collective/enumerative, sort does open the door to possible ties among competing such interpretations, ones that put different weights on or differently position the sub-interpretations they acknowledge in common. For from a hypothetical-intentionalist perspective, two different ways of organizing in relation to one another the individually attractive subinterpretations of a work might be equally plausibly hypothesized, in light of all the appreciatively relevant data, to be what the contextually understood author ultimately wanted readers to grasp. Still, given how subtle the differences are likely to be between two such integrative/hierarchic readings acknowledging all the same individually acceptable sub-interpretations, multiplicity at this level, even if not ultimately eliminable, must surely strike one as not much of a qualication on the idea of there being such a thing, grosso modo, as the meaning of a literary work. 4. It will be no surprise that the doctrine of hypothetical intentionalism has failed to win universal acceptance among theorists of interpretation. Textualists, deconstructionists, and actual intentionalists still abound. What is more, the unconverted, and particularly actual intentionalists, have not been shy to voice their criticisms of the doctrine. I thus here outline some brief responses to objections that have been raised to a hypothetical intentionalist account of literary meaning such as I have proposed, whose bare bones were sketched above. The basic idea, recall, is that on such an account literary meaning, the object of literary interpretation in a determinative mode, is constitutively bound not to what a historically untethered text might be saying, nor to an authors actual, psychologically real semantic intentions in composing the texteven ones that might be said to have been successfully realized in the textbut to our best hypothesis, as ideally comprehending readers, as to what the concretely situated and publicly available authors semantic intentions were in composing the text he or she did.

306

Interpretation

Objection 1: Drawing a veil of ignorance across some aspects of a works actual creative history and not others, as a hypothetical intentionalist approach to work meaning enjoins, is unacceptably arbitrary. Response: Erecting a rough cordon around essentially privatewhich is not to say, epistemically inaccessibleinformation is hardly arbitrary from a literary point of view. The making of literature is an individual, largely interior endeavor, but it is also a public, convention-governed one, bound by mutually understood rules for producing and receiving literary offerings. These rules might quite naturally specify that facts related to context of origin beyond what an ideally prepared and backgrounded reader could generally be expected to know were irrelevant to xing or constituting the meaning of the work as an utterance in that context. The artists state of mind is not our ultimate goal as interpreters of literary works, but rather what meaning can be ascribed to those works, albeit as the indissociable products of those very particular communicative agents; thus not all obtainable evidence as to the artists state of mind is automatically germane to the project of delineating what a work issuing from that mind and presented in a literary setting arguably means. Objection 2: Hypothetical intentionalism is committed to a communication model of the literary domain, but such a model does not sit well with the appeal to idealized, as opposed to actual, audiences that is a feature of sophisticated versions of hypothetical intentionalism. Response: The communicative model arguably presupposed by literary activity does not commit us to authors projecting their works for specic and specically envisaged audiences, ones contemporaneous with the author, rather than, less restrictively, for whatever audiences, present and future, are well suited to receive and understand the work in its historical, cultural, and authorial context. Call the former the narrowly communicative model of literary activity, if you like, and the latter the broadly communicative model. The point is that communication with appropriate readerswhoever, whenever, wherever they might beis still communication, even when such readers are not narrowly identied or targeted in advance.
See Anthony Savile, Instrumentalism and the Interpretation of Narrative, Mind 105 (1996): 55376; Paisley Livingston, Arguing Over Intentions, Revue Internationale de Philosophie 198 (1996): 61533; and Robert Stecker, Artworks: Denition, Meaning, Value (University Park, PA: Pennsylvania State University Press, 1997), ch. 10. See Savile, Instrumentalism and the Interpretation of Narrative.

Hypothetical Intentionalism

307

Objection 3: Allowing optimally hypothesizable intentions to trump actual ones, where the basic nature or status of a work is concerned, opens the door to an unacceptable level of indeterminacy as regards work content. Response: Hypothetical intentionalism regarding work content or meaning is rightly coupled with actual intentionalism as regards both the status of works as literature and their category or genre location within literature. Note that category or genre specications can be taken as tantamount to or generative of very general semantic, or perhaps metasemantic, intentions, indicating what sorts of meaning, at the least, are to be sought in a given work, which helps to dissipate the worry sounded by opponents of hypothetical intentionalism that hypotheticism regarding authorial semantic intentions will issue too readily in indeterminacy of meaning. Furthermore, coupling actual and hypothetical intentionalism in this manner strengthens the claim of literature so understood to be communicative, in almost the narrower sense distinguished a moment ago, and for two reasons. First, readers who attempt to arrive at meaning by hypotheticist lights are entitled from the outset to know, and so ideally do know, at least what category of offering they are dealing with. Second, the actual author, in being obliged to show his opening hand, that is, vouchsafe to readers directly the approximate nature, if not the precise import, of his work, thus does not remain entirely behind a veil as far as the constitution of meaning is concerned. Objection 4: The best hypothesis about authorial intention must, logically speaking, be that which is correct; thus there can in fact be no divergence between actual authorial intention and our best hypothesis as to what that intention is or was. Response: This is a simple misunderstanding. Obviously, best hypothesis in the formulation of hypothetical intentionalism cannot be taken to mean that which in fact happens to be correct, and so best in the sense of true. Rather, the best hypothesis by hypothetical intentionalist lights is that which we would have most reason to accept or adopt given the totality of evidence that is both available and admissible, i.e. given the totality of what is derivable from the text and its legitimately invoked surrounding context.
Ibid. On the notion of categorial vs. semantic intentions in relation to a work of literature, see my Intention and Interpretation in Literature, and in a more critical vein, Paisley Livingston, Intentionalism in Aesthetics, New Literary History 29 (1998): 83146.

308

Interpretation

Objection 5: Even if there is a distinction, regarding authorial intention, between a best hypothesis, in the sense invoked by hypothetical intentionalism, and a true hypothesis, why should we ever favor the former over the latter, once we have arrived at the latter, by whatever means we have at our disposal? Surely in science we would not prefer our methodologically soundest hypothesis regarding some state of affairs over what was in fact the case, were we to learn what that was. Response: This objection misunderstands the goal of literary interpretation as conceived by hypothetical intentionalism, which is not to discover, for its own sake, the authors intention in writing the text, as if criticism were at base a matter of detective investigation, but to get at the utterance meaning of the text, that is, what it not the author is saying, in its author-specic context. Utterance meaning just is constitutively tied to a most reasonable projection of utterers intent in the given context, and does not collapse into utterers meaning. Thus even when the latter is available it does not displace the former as the object of literary interpretationas opposed to biographical sleuthing. Objection 6: Hypothetical intentionalism that acknowledges the necessity for interpretation to ascertain actual intentions of a categorial or constitutive sort has already betrayed the vaunted autonomy of the literary workits independence in a fundamental respect from its creatorthat it claims to safeguard. Response: This is not so. A restricted autonomy, to wit, as regards resultant meaning, is still autonomy; furthermore, it is arguably the only sort of autonomyas opposed, say, to that requiring detachability from
See Nol Carroll, Interpretation and Intention: The Debate Between Hypothetical and e Actual Intentionalism, Metaphilosophy 31 (2000): 7595. Carroll charges that to proceed in this way would appear to be fetishizing our method over what the method is designed to secure (p. 83). Thus the charge that hypothetical intentionalism simply substitutes warranted assertibility for truth where literary interpretations are concerned (see Carroll, Interpretation and Intention, 84) is similarly unjustied. Hypothetical intentionalism, at least when advanced with a background commitment to metaphysical realism, retains that distinction, but relocates it with respect to the items involved. For hypothetical intentionalism, a true literary interpretation of a work W by an author A writing in context C is one given by what is, so to speak, optimally warrantedly assertible about the intention with which A, writing in C, composed Ws text. But if a literary interpretation is thus true, then it is more than just warrantedly assertible. (For further relevant discussion, see Gregory Currie, Interpretation and Objectivity, Mind 102 (1993): 41528.) See Gary Iseminger, Actual Intentionalism vs. Hypothetical Intentionalism, Journal of Aesthetics and Art Criticism 54 (1996): 31926.

Hypothetical Intentionalism

309

generative contextthat it seems important to insist on where literature is concerned. Objection 7: Hypothetical intentionalism, which identies the core meaning of a work with a best projection of authorial intentionwhere such projection may not in fact coincide with any meaning actually intendedthus appears in the last analysis to be not really concerned with either the author or his achievement. Hypothetical intentionalism unjustiably severs the work from the agent who has created it. Response: Again, this is not so. Hypothetical intentionalism accords the semantic intentions of the actual author a crucial role, only it is a heuristic rather than a nal one. Authorial intention is what truth-seeking interpretive activity necessarily aims at, the idea being that what one would most reasonably take to be that intention, on the basis of the text and a full grasp of its author-specic public context, yields a true interpretation of the literary work, understood as an artistic utterance, which is embodied in the text. As to severing a work from its author, hypothetical intentionalism pleads not guilty; it simply insists that the meaning of a literary work, however informed its interpretation must be by the authors public identity, and even by certain of the authors actual intentions, is not constrained to being just what the author intended it to mean, even where that intention is fully compatible or consonant with the contextually situated text. And even though a literary work is inextricably the work of just that author, in that precise context, the author is not the ultimate arbiter of what his or her work means, i.e. what it appears to convey or communicate to an appropriately backgrounded reader. Finally, hypothetical intentionalism doesnt deny authors their achievements, it just locates those achievements in the utterance meanings their uttered texts attain, owing for the most part to the ingenuity with which those texts have been contrived, and not in the utterers meanings, which those texts also, in favorable cases, subserve. Objection 8: The defense of hypothetical intentionalism ultimately rests on the claim that it accords better with current interpretive practices than does actual intentionalism, alleging that critics in framing interpretive hypotheses do observe the proposed ban on inherently private information, such as direct
See Iseminger, Actual Intentionalism vs. Hypothetical Intentionalism, and Livingston, Arguing Over Intentions. Nor is the author in the best position to discern that, in any event. Authors, because of their unique perspectives on their own works, are generally very far from being ideal readers of them.

310

Interpretation

but hidden authorial proclamations of what a work is intended to mean, and so on. But in fact this is not the case. Response: Hypothetical intentionalism does not ultimately rest on an empirical claim about actual interpretive practices, taken in their full and motley variety, but rather on what are arguably norms underlying the most defensible of such practices, understood as ones that truly answer to our interests in literature as literature. It is on that elusive and highly contestable terrain that the dispute about the merits of hypothetical intentionalism must be conducted, rather than that of statistical conformity or non-conformity with current practice. Admittedly, a full case for hypothetical intentionalism on those grounds remains to be made. 5. A literary work is an utterance, of course, but it is a sort of grand utterance, one governed by different ground rules of interpretation than are ordinary utterances. Our interests in literature are communicative ones, where communicative is understood broadly, but they are not, pace certain recent writers, more narrowly conversational ones. This means, in part, that in literary contexts, unlike conversational ones, we have a prior and independent interest in utterance meaning entirely apart from whatever utterer meaning may stand behind or parallel that utterance meaning as constituted. In conversation, if we dont understand what someone has said we may quite properly get him to explain further what he meant, or to qualify or retract his words. In literature, if we dont understand what a contextually situated text is saying, we cannot legitimately demand explication from the author, or
See Carroll, Interpretation and Intention. I note a further objection that is often raised against hypothetical intentionalism, one I admit to nding more troubling than those reviewed above. The objection is that the distinction presupposed by hypothetical intentionalism between essentially public and essentially private information regarding an author, where the former enters into the appreciatively relevant context for the work while the latter does not, is fundamentally untenable, or fatally blurry. (See Stecker, ArtWorks; Livingston, Intentionalism in Aesthetics; and Carroll, Interpretation and Intention.) Clearly, however this distinction is made out, it cannot be equated with that between published and unpublished information, if only because that would have the implausible consequence that a works meaning, i.e. what is given by a correct interpretation of it, would change upon publication of certain appreciatively relevant facts about how a work came to be that it just happened were not known outside of the authors immediate circle. This is not the place to attempt a full reconstruction of the needed distinction, but one might begin to rene the concept of a works appreciatively relevant public context by focusing on what the author appears to have wanted his or her readers to know about the circumstances of a works creation, beyond what is implicit in the authors previous work and the authors public identity. At any rate, such a thing will not uctuate with the contingencies of actual publication of the information in question. Finally, that the distinction, however reconstructed, might remain blurry is not fatal to its utility. See Nol Carroll, Art, Intention, and Conversation, in Gary Iseminger (ed.), Intention and e Interpretation (Philadelphia: Temple University Press, 1992).

Hypothetical Intentionalism

311

instruct him to modify his offering; at most we can ask him to conrm that the text he has given us is indeed the text he wants us to have, as we set about to interpret it as literature.
For further defense of hypothetical intentionalism and criticism of actual intentionalism, see Saam Trivedi, An Epistemic Dilemma for Actual Intentionalism, British Journal of Aesthetics 41 (2001): 192206, and Gregory Currie, Interpretation in Art, in Jerrold Levinson (ed.), Oxford Handbook of Aesthetics (Oxford: Oxford University Press, 2003), 291306.

This page intentionally left blank

PART V A E S T H E T I C P RO PE RT I E S

This page intentionally left blank

19
Aesthetic Properties, Evaluative Force, and Differences of Sensibility
In this essay I revisit a position on aesthetic attributions I have held for some time, a position rooted in some seminal essays of Frank Sibley, and which can be labeled aesthetic realism. I reect on some challenges which have emerged to the viability of that position, and determine what accommodations, if any, are called for. First I sketch the position, borrowing with modication from an earlier short essay of mine. I then formulate a number of worries about the position which have lately come into view, and try to see where they lead. I Aesthetic attributions to works of art, and the terms used to effect such attributions, are largely descriptive; that is to say, they are based on, and obliquely testify to the occurrence of, certain looks, impressions, or appearances which emerge out of lower-order perceptual properties. Insofar as an aesthetic attribution is intended as objective, that is, as the attribution of a property of intersubjective import, such looks or impressions or appearances are relativized to
First published in E. Brady and J. Levinson (eds.), Aesthetic Concepts: Essays after Sibley (Oxford: Oxford University Press, 2001), 6180. Aesthetic Concepts, Philosophical Review 68 (1959): 42150, of course, and other essays mentioned in subsequent notes. The view defended also owes something, in broad measure, to writings of Monroe Beardsley and Kendall Walton. Being Realistic About Aesthetic Properties, Journal of Aesthetics and Art Criticism 52 (1994): 3514. An earlier and fuller statement of the position, though with different emphasis, can be found in my Aesthetic Supervenience (1983), reprinted in Music, Art, and Metaphysics (Ithaca: Cornell University Press, 1990). See Frank Sibley, Aesthetic and Non-aesthetic, Philosophical Review 74 (1965): 13793; Frank Sibley, Objectivity and Aesthetics, Proceedings of the Aristotelian Society Supplement 42 (1968): 3154; Monroe Beardsley, The Descriptivist Account of Aesthetic Attributions, Revue Internationale de Philosophie 28 (1974): 33652. See also G ran Hermern, The Nature of Aesthetic o e Qualities (Lund: Lund University Press, 1988).

316

Aesthetic Properties

a perceiver who views a work correctly, and thus approaches the condition of what has been called, following Hume, a true critic or ideal judge. That means, in particular, someone who properly situates a work with respect to its context of origin, including its place in the artists oeuvre, its relation to the surrounding culture, and its connections to preceding artistic traditions. It is true, of course, that some terms used in aesthetic discourse are completely, or almost completely, evaluative. That is to say, they are terms with no or almost no descriptive content, meaning they do not imply anything about the kind, category, or nature of the object to which they are applied. Sibley called these solely evaluative terms. Such terms either just denote a degree of aesthetic value or disvalue believed by the speaker to be present, or else serve to express the speakers attitude, approving or disapproving, toward the object in question. On either construal such terms have evaluative force, if any do. Striking, splendid, excellent, mediocre, miserable, execrable, and so on, are terms of this sort. Most aesthetic terms, however, clearly have a substantial descriptive content; they cannot, whatever evaluative force they may have aside, be applied to just anything. There are, for instance, formal terms, e.g. balanced, chaotic, unied; expressive ones, e.g. melancholy, anguished, cheerful; metaphorical but non-psychological ones, e.g. delicate, steely, brittle; and natively aesthetic ones, e.g. graceful, gaudy, garish. It is sometimes proposed that the commonest aesthetic terms, such as those just mentioned, are of mixed character, having both a descriptive and an
For elaborations of objectivity along these lines, see Richard Miller, Three Versions of Objectivity: Moral, Aesthetic, and Scientic, and Peter Railton, Aesthetic Value, Moral Value, and the Ambitions of Naturalism, in Jerrold Levinson (ed.), Aesthetics and Ethics: Essays at the Intersection (Cambridge: Cambridge University Press, 1998). See Kendall Walton, Categories of Art, Philosophical Review 79 (1970): 33467; Mark Sagoff, Historical Authenticity, Erkenntnis 12 (1978): 8393; Philip Pettit, The Possibility of Aesthetic Realism, in Eva Schaper (ed.), Pleasure, Preference, and Value (Cambridge: Cambridge University Press, 1983), 1738; Gregory Currie, Supervenience, Essentialism, and Aesthetic Properties, Philosophical Studies 58 (1990): 24357; and Levinson, Aesthetic Supervenience. See Frank Sibley, Particularity, Art, and Evaluation, Proceedings of the Aristotelian Society Supplement 48 (1974): 121. Nick Zangwill, in a useful essay, The Beautiful, the Dainty, and the Dumpy, British Journal of Aesthetics 35 (1995): 31729, calls them verdictive aesthetic terms. I am not inclined to put beautiful and ugly in this category without comment, as I think that such terms, in their primary employment in regard to visual objects, imply particular kinds of phenomenal impression, ones respectively of pleasing harmoniousness and displeasing disharmoniousness, rather than simply merit or demerit in the abstract. I recognize, though, that beautiful is also used as a non-specic term of aesthetic appraisal, with the sense merely of aesthetically excellent. Zangwill denominates all such terms substantive aesthetic terms. See The Beautiful, the Dainty, and the Dumpy.

Aesthetic Properties, Evaluative Force, and Sensibility

317

evaluative component. It would be a rule of the use of such terms that in applying them one would be indicating both that a thing had a given property and that one valued or disvalued that property or ascribed value or disvalue to it. Sibley called these evaluation-added terms. However, as Sibley noted, there is little reason to think that these common aesthetic terms, ones that feature prominently in art criticism, are any more inherently evaluative than those characterizing human conduct in the context of moral assessment, terms such as courageous, honest, or merciful. Whether someone is courageous, honest, or merciful would seem to be open to straightforward observation, though observation of a complex sort, and by an observer versed in the human form of life. Thus, there is at least reason to doubt that parallel terms of everyday art criticism, denoting generally desired qualities of appearance, are even evaluation-added terms. They seem rather to be basically descriptive terms, naming properties which, as it happens, constitute merits or defects in certain spheres of assessment. Sibley labeled such terms descriptive merit-terms. We may grant that many common aesthetic terms, for example, gaudy, appear to entail evaluations on the part of the speaker. Still, most cannot be held to do so strictly. First, it seems possible to approve a work for its gaudiness, say, or despite its gaudiness. This suggests that the essence of gaudiness is not a judgment of disapprobation on the speakers part but instead a kind of appearance: a perceptually manifest effect one can register independently of any evaluative assessment of or attitudinal reaction to that effect. Second, the evaluative implications, loosely speaking, of terms like gaudywhich perhaps derive, in part, from past histories of use in connection with particular canons of criticism or tastecan be explicitly cancelled or disavowed, without semantic anomaly. Thus, terms of this sort, despite their air of evaluativity, are such that they can nevertheless be ascribed without strictly entailing anything about the speakers evaluative attitudes. Nick Zangwill makes a helpful observation along the same lines, proposing that for most substantive aesthetic attributions it is only a conversational implicature, in Grices sense, that an evaluation is being made.
Particularity, Art and Evaluation, 6. In moral philosophy such terms are usually said to generate thick descriptions, ones that evaluate as well as describe. But see Zangwill, The Beautiful, the Dainty, and the Dumpy, for some disanalogies between the aesthetic and the moral cases. Particularity, Art and Evaluation, 6. See on this Hermern, The Nature of Aesthetic Qualities, ch. 5. e The Beautiful, the Dainty, and the Dumpy, 322. Zangwill later denies, however, that this line is sustainable for at least one such attribution, that of gracefulness. He claims that although

318

Aesthetic Properties

But suppose for the sake of argument that there is a purely evaluative element in certain common aesthetic terms of the substantive and not solely evaluative sort, ones such as gaudy, or maudlin. Suppose, in other words, that it is correct to regard such terms as evaluation-added terms, implying an evaluative stance toward or estimation of an object on the part of a user of the term. There would, I claim, still remain a purely descriptive, distinctively aesthetic content in such an attribution, consisting roughly in an overall impression afforded, an impression that cannot be simply identied with the structural properties that underpin it. But what evidence is there of such higher-order perceptual impressions? What reasons are there to accept their existence? First, one can often nd alternative descriptions, sometimes requiring several words, of the distinctive experiential contents involved, in which the evaluation-added element of the original attribution, if any, has been removed. For instance, one might approximate the descriptive content of gaudy by bright, non-harmonious, eye-catching color combinations. Second, one can often get disputing critics to focus on the common perceptual ground in their aesthetic responses. For instance, a critic might be brought to admit that he is aware of the look or appearance another critic has remarked on with evident relish, reserving his right to dislike it, that is, to exercise his taste in the sense of personal preference with regard to it. Here is a musical example. The opening of Bachs Concerto for Three Harpsichords and Strings in D minor, BWV 1063, offers a vivid expression of grimness; it might even be described as starkly grim. Now, some competent listeners like grimnesslike that character, like being confronted with itand others do not. It is easy to imagine those latter folk simply labeling the opening depressingly dour and having done with it, but it also seems
it is context that makes merits or demerits of most aesthetic properties, and that most are thus not even pro tanto merits or demerits, gracefulness appears to count only and always to the good: Can the grace of something ever be a demerit? It is hard to see how it could (p. 324). But I am inclined to disagree. Grace would seem to be aesthetically contraindicated in an expressionist painting or sculpture of the mass executions at Babi Yar. If so, then even gracefulness may not be, tout court, a pro tanto merit in works of art, and the positive evaluative overtones of its attribution to a work may be only a matter of conversational implication. For further complications regarding aesthetic attributions, see Iuliana Corina Vaida, The Quest for Objectivity: Secondary Qualities and Aesthetic Qualities, Journal of Aesthetics and Art Criticism 56 (1998): 28397. More precisely, by the look characteristic of bright, non-harmonious, eye-catching color combinations.

Aesthetic Properties, Evaluative Force, and Sensibility

319

more than likely that they could be brought to acknowledge the aptness of the characterization starkly grim as well, only adding under their breaths, if you like that sort of thing. Thirdly, unless one assumes there are core aesthetic impressions of a qualitative sort, distinguishable from reactions of approval or disapproval per se, it becomes difcult to explain what competent critics with evaluative differences of opinion could really be talking about. Surely its not just that one approves a certain arrangement of lines and colors, or pitches and rhythms, or words and phrases, and the other not. Rather more likely is that each registers the overall effect of the arrangement in question, that there are descriptions, reasonably neutral ones, they could even agree upon to characterize it, but that one favors it and the other does not, or one thinks it makes the work good and the other does not. In addition, failing to acknowledge distinctive aesthetic impressions as the core descriptive content of common aesthetic attributions makes a mystery out of what the aesthetic experiences of perceivers of any sort could possibly consist in. So suppose, for the sake of argument, that many aesthetic terms, for example gaudy or maudlin, include an evaluative component irreducibly. This would hardly show that there are no objective aesthetic properties in the wings when such terms are being correctly applied. Such properties supply the purely descriptive content of any evaluation-added terms as there are in the critics repertoire, as well as guring as the entire content of others, such as sorrowful, frenetic, serene, passionate, balanced, delicate. Even were some of the pairs of aesthetic terms of the type highlighted by Sibley, e.g. delicate/anemic, bold/gaudy, or heartfelt/maudlin, to differ primarily or only in the implied attitude or reaction, pro or con, of the ascriberon his or her taste in the preferential sensethe presence of an aesthetic property to which they advert in common, and not just complexes of non-aesthetic ones, seems strongly indicated. The fact that we often lack a single word of neutral cast to refer to the experiential terrain descriptively shared proves little.
In Being Realistic, the same point was illustrated by the bumptious excitement of the opening of the fourth movement of Tchaikovskys Fourth Symphony and the distraught longing of the second theme of the rst movement of Francks Piano Quintet, to whose aesthetic qualities evaluative reactions clearly differ. Regarding the latter piece, we may note this recent testimony from Roger Scruton: The unctuous narcissism of Csar Francks Piano Quintet is certainly an expressive e feature, but not a virtue in the work that possesses it. Nevertheless, it is part of the power of this work, that it so successfully conveys this somewhat disreputable state of mind (The Aesthetics of Music (Oxford: Oxford University Press, 1997), 148).

320

Aesthetic Properties

As we have already remarked, whatever evaluative force is carried by such terms as we have just been examining, there are clearly descriptive limits on their application. Not just any visual pattern can be disapproved of by calling it, say, gaudy, chaotic, or amboyant. I suggest that such limits are xed, in the last analysis, by distinctive phenomenal impressions or appearances associated with such terms, not by sets of structural properties that disjunctively serve as their occurrence conditions. Thus the existence of certain aesthetic propertiesroughly, the dispositions to afford such impressions or appearancescan hardly be denied, even if such properties are not always all that is conveyed by standard terms of aesthetic description, if an evaluative component in them be admitted. Attitudes toward the impressions afforded by artistic structures may evolve over time, even uctuate back and forth, but surely something unitary of a broadly perceptual sort often remains constant throughout. If an aesthetic predicate P is admitted to have, in some sense, an evaluative vector, perhaps it follows that there is not then a straightforward or purely descriptive property being P. Still, it seems there will always be, in the ofng, an associated evaluatively neutralized aesthetic property, being Pn , a disposition to afford a certain phenomenal impression to adequately positioned perceivers. More generally, we can say that there is in regard to a given object almost always a descriptive aesthetic content such that ideal judgers who would not apply to the object all and only the same aesthetic predicatesbecause they have, by assumption, different reactions or attitudes towards that contentcan still agree on what that content is. Furthermore, we should not rule out that there might be such aesthetic content even if no suitably neutral terms were readily available to evince the agreement on it that there might be.
Though for convenience I here adopt the usual dispositional analysis of perceptual properties, I believe that strictly speaking such properties are not precisely dispositions, but are rather just strongly supervenient on such dispositions. (Colin McGinn has recently argued this persuasively, so far as color properties are concerned, in Another Look at Color, Journal of Philosophy 93 (1996): 53753, while Gregory Currie argues in the same vein, as regards aesthetic properties, in Supervenience, Essentialism, and Aesthetic Properties.) What, then, are perceptual properties themselves? In my view color properties are roughly manifest ways of appearing visually, related to but not identiable with dispositions to appear visually in such and such ways, while aesthetic properties are roughly manifest ways of appearing phenomenally, where the ways involved are of a higher order than those involved in the basic sensory properties. Unlike a dispositional conception of perceptual properties, a way-of-appearing conception is not in conict with the intuition that, at least under favorable conditions, one just looks and sees what color property or aesthetic property something possesses. (For further elaboration, see What Are Aesthetic Properties?, Proceedings of the Aristotelian Society. Supplement 79 (2005): 21127 (and Ch. 20, this volume).

Aesthetic Properties, Evaluative Force, and Sensibility

321

II So what, then, is there to give pause in this picture of aesthetic attributions and aesthetic properties? What reasons might there be to reconsider some aspects of this picture? Here are some of the worries that have been raised in this connection. One: it is not really possible to separate the descriptive and evluative components in an attribution or property. Two: the descriptive component of such an attribution or property is not given by overall gestalt impressions but by structural features of a non-aesthetic sort. Three: there just are no qualia, no phenomenal properties of any sort, so a fortiori there are no aesthetic qualia. Four: even if these other worries are laid to rest it is inescapable that there are irresolvable differences in aesthetic judgments among even ideal judges, yet it is unclear that a realist perspective on aesthetic properties can properly accommodate that fact. In what follows I give each of these worries a hearing. III First worry. Despite what was urged above, many will still be inclined to insist that most aesthetic terms just do have an evaluative component of one sort or another, at least in practice or in context. Very well: simply focus on the evaluatively neutral phenomenal core of such terms and you will arrive at bona de aesthetic properties. But as Hamlet famously complained, theres the rub. Perhaps it is not really possible to identify the purely evaluative component of an aesthetic term so as to allow its subtraction from the import of the term as a whole, leaving a purely descriptive component whose cash value is an aesthetic property; perhaps what is descriptively conveyed by an aesthetic term cannot be isolated, cannot even legitimately be presumed to exist, apart from what the term conveys in its concrete use. If that is so, then even modest aesthetic
One philosopher who is clearly skeptical of the conceptual surgery required is Alan Goldman: A different question is whether we can always analyze evaluative properties into evaluative and non-evaluative components. Since we have viewed these properties as relations between objective properties and evaluative responses, it might seem that the answer must be afrmative. But I have also pointed out that many of the higher-level properties of this sort are unspecic on their objective sides. Although it should be possible in principle to analyze specic references to such properties into objective and subjective components, we cannot do so for the properties themselves (Aesthetic Value (Boulder, CO: Westview Press, 1995), 26).

322

Aesthetic Properties

realism, afrming the existence of evaluatively neutral aesthetic properties at the root of aesthetic attributions, whether evaluatively charged or not, would be too optimistic. What to say? Well, one way it might turn out that the phenomenal impressions associated with aesthetic terms might not be exhibitable denuded of all evaluative aspect would be if such impressions were themselves inherently pleasant or unpleasant. The impressions that go with nding something graceful or harmonious, say, are plausibly of that sort. To the extent that phenomenal impressions are inherently pleasant or unpleasant, they would seem of necessity to bring in their train corresponding reactions of favor or disfavor. Thus, to that extent, isolating neutral phenomenal impressions at the core of aesthetic attributions would be a chimerical pursuit. But two points bear making in response. First, it is surely not the case that all aesthetic impressions are inherently hedonically valenced. It rather seems that most such impressions are, for most perceivers, more or less hedonically neutral for instance, those of the peacefulness of a landscape, the angularity of a design, or the urbanity of a passage of violin music. Second, even were all aesthetic impressions hedonically valenced, it would still be possible rationally either to approve or disapprove the affording of such an impression at a particular point in a particular work and to report its occurrence by an appropriate evaluatively charged term. That is to say, the inherent pleasurability of the impression at the core of an aesthetic property would still be distinguishable from the evaluative component of that property, so to speak, when picked out by an evaluatively charged substantive term. There are other things to say in regard to the problem of the distinctness or isolatability of aesthetic impressions, but they will be evident in my discussion of the third worry noted, that concerning the existence of qualitative phenomena generally. IV Second worry. What exactly does the descriptive component of an aesthetic property consist in? Is it in fact a unitary impressiona look or appearance that an object is tted to afford, as I have argued, or is it rather a plurality of combinations of non-aesthetic features that an object might possess? Call the former the phenomenological account of aesthetic properties and the latter the

Aesthetic Properties, Evaluative Force, and Sensibility

323

structural account. Are there additional reasons that can be given in favor of the phenomenological account of aesthetic properties as opposed to the structural one? I am not sure that I have any more to offer than I have already advanced, here and elsewhere, or that can be found in the writings of Sibley, Beardsley, and Walton. What argues above all for the phenomenological conception is its comporting better with the evident semantics of aesthetic attributions: when we ascribe an aesthetic property it seems that what we are ascribing, at base, is an emergent way of appearing, and not a range of ensembles of disparate traits that, it so happens, sustain such a way of appearing. In any event, the phenomenological account presumably gains in convincingness the more light we can throw on exactly what an overall phenomenal impressionthe suggested core of an aesthetic attributionmight involve. Though I have until now stressed the perceptual character of such impressions, we need not deny that such impressions might be partly affective as well. In other words, some aesthetic impressions might be bound up with feelings consequent on apprehending an objects non-aesthetic features. Registering such an impression from an object might, as Derek Matravers has recently suggested, involve apprehending the objects sensible qualities and forms with a certain feeling, or having a certain feeling in apprehending them. Furthermore, such impressions, though basically perceptual, might be importantly mediated or inected by conceptual activity at some level; for example, the impression of a pattern as unied might partly consist in a tendency to construe the parts of the pattern under a concept such as that of <tting or working together>. Clearly, more work needs to be done on this matteron
One might alternatively label the latter the reductive account. See my Aesthetic Supervenience. This is marked by calling them, as is often done, gestalts. Aesthetic Concepts and Aesthetic Experiences, British Journal of Aesthetics 36 (1996): 26577. Matravers proposes that some aesthetic property terms, such as delicate, have both a non-aesthetic, condition-governed use and a properly aesthetic, experience-tied use. With the latter, there is no getting around a phenomenological account of the property being attributed: Consider the vase again; experiencing it causes an observer to believe it is uted, blue, made of very thin glass, about six inches tall, and so on. By itself this may be enough to ground the non-aesthetic judgement that it is delicate. If, in addition, the observer gets a feeling (perhaps akin to loss, or anxiety about the objects fragility) this may prompt an aesthetic judgement [that it is delicate]. It would be senseless to use a different term; what needs to be communicated is the response to just those properties of the vase that are to do with its being (non-aesthetically) delicate (p. 273). For suggestive proposals on this score, see Monique Roelofs, Aesthetic Experiences and Their Place in the Mind (Ph.D. dissertation, University of Maryland, 1997). See also Berys Gaut, Metaphor and the Understanding of Art, Proceedings of the Aristotelian Society (1997): 22441,

324

Aesthetic Properties

what exactly may be comprised in the phenomenological impression or experiential complex at the core of a given aesthetic property. We should also not overlook the role that paradigm objects or patterns might play in pinning down the descriptive content of aesthetic attributions on a phenomenological view of them. Perhaps, with such a paradigm at hand, one could indicate the target overall impression or feeling as the impression or feeling one gets from this, ostending the object or pattern in question. Of course, there is always the problem of whether people will go on, or generalize, in the same waya problem which Wittgenstein brought to philosophical saliencebut there is no reason to think it inevitable that they will not. For one, corresponding to the innate quality-spaces that Quine once posited for the sensory modalitiesin order to account, for example, for why we see the visible spectrum as falling into basically six colors, rather than, say, either three or thirty-threethere may even be innate aesthetic qualityspaces, which would help to ensure that many of our more basic aesthetic impressions were at some level mutually conformable, registered at roughly the same levels of specicity. V Third worry. A phenomenological account of aesthetic properties naturally faces general skepticism about the existence of phenomenology. There are a number of philosophers who deny the existence of qualia across the board, that is to say, intrinsic, qualitative properties of conscious experience, more familiarly known as the ways things seem or appear to us. The implications for the aesthetic sphere are plain. If there are no qualia of low order, such as sensory qualia, then there are presumably no qualia of higher order, such as aesthetic qualia. If there are no taste, smell, color, touch, or timbre qualia, there are certainly none, we may be sure, for grace, delicacy, melancholy, garishness, effeteness, or balance. And thus a phenomenological account of aesthetic properties would be precluded from the outset. Probably the best-known basher of qualia is Daniel Dennett.
for insightful remarks on the doxastic versus imaginative involvement of concepts in aesthetic experience. For illuminating discussion of the role of such paradigms in aesthetic discourse, see Vaida, The Quest for Objectivity. However, the relativism that Vaida claims follows from variability in paradigm preference among aesthetic judges is not as unavoidable as she makes it out to be.

Aesthetic Properties, Evaluative Force, and Sensibility

325

I want to take a brief look at his best-known bash, which occurs in a paper called Quining Qualia. Dennett tries to show that people can be mistaken about their own qualia in such manner as to suggest that we arent really entitled to think there are qualia at all to be right or wrong about. One of his central thoughtexperiments in support of this is the story of Chase and Sanborn, coffee tasters at Maxwell House. Chase and Sanborn both used to like the house coffee, and now neither does. Chase claims it tastes the same to him, though he now grades it lowhe has apparently become a more sophisticated drinkerwhile Sanborn claims it tastes different to him, though were it still to taste the same to him he would grade it as high as everhis taste buds have apparently deteriorated. The claims of both claimants depend in part on the reliability of their memories. Now Dennett asks whether we can take their claims at face value. Might one or both of them simply be wrong? Might their predicaments be importantly the same and their apparent disagreement more a difference in manner of expression than in experiential or psychological state? That is, might it not just be a matter of how they interpret their conditions and histories, rather than a difference in the conditions and histories themselves? Focusing on Chase, Dennett sketches two possibilities. One, his coffeequalia have remained the same, while his reactions toward them have altered. Two, his coffee-qualia have altered gradually over the years, so that his memory of the original quale has become unreliable, while his reaction to the original quale, could he recover it, would be the same. Testing might not be able to resolve which was the case, Dennett maintains, because the alternative hypotheses might not differ in any observable consequences. This is especially so if we consider hypotheses midway between these two, with admixtures of both. Dennetts argument concerning this and similar examples is thus that the postulation of qualia is without justication because there may be no way to distinguish empirically, in a case of qualia ostensibly altered over time,
Daniel Dennett, Quining Qualia (1988), reprinted in William Lycan (ed.), Mind and Cognition (Oxford: Blackwell, 1990); all page references are to that reprinting. For other criticisms of Dennetts case for quining qualia, ones largely orthogonal to mine, see Georges Rey, Dennetts Unrealistic Psychology, Philosophical Topics 22 (1994): 25990; Eric Lormand, Qualia! (Now Showing At A Theater Near You), Philosophical Topics 22 (1994): 12756; and Bredo Johnsen, Dennett on Qualia and Consciousness: A Critique, Canadian Journal of Philosophy 27 (1997): 4781. Quining Qualia, 527.

326

Aesthetic Properties

between actually altered qualia and alteration in the memory of qualia previously had. Since there may be no way in which the difference between the hypotheses could ever be empirically manifested, even for the subject, the hypotheses cannot really differ in content, and thus the notion of qualia in terms of which they are framed must be bogus. I have two comments on this argument. My rst comment is that the argument seems at base a vericationist one, albeit of a sophisticated kind. But if one is not convinced generally by vericationist arguments aimed at the collapse of other distinctions which seem to be real and graspablefor example, the distinction between truth and ideally warranted assertabilitythen one will not be overly moved by the difculty to which Dennetts thought experiment points. We might just have to accept that our qualia might change in such a fashion that we could not always be certain that they had done so, if the circumstances in which the changes occurred were not favorable to such determination. But this would show neither that the idea of such a change was incoherent, nor that, were we to undergo such changes, we would never be in a good position to think that we had. My second comment is that the whole issue of memory for sensory qualia is a more problematic one than Dennett realizes, as Diana Raffman has usefully emphasized in recent work. Psychological research concerning categorical perception has revealed the existence of denite limits to the ne-grainedness of conceptual representations of sensory data. It appears we can detect or discriminate sensory impressionse.g. shades of color, nuances of toneof a sort much ner than we are able to categorize and label internally. As a result, accurate memory of such impressions is likely impossible, as is any further conceptual processing of them. But of course that does not impugn the fact that we have received those impressions, impressions of a surprisingly high degree of specicity. To make the case concrete, imagine that you are noting a difference in comparing two color patches only a discriminable shade or two apart. Then
Language, Music, and Mind (Cambridge, MA: MIT Press, 1993), and On the Persistence of Phenomenology, in T. Metzinger (eds.), Conscious Experience (Paderborn: Schoningh, 1995): 293308. This effectively takes the wind out of the sails of Dennetts closing moral regarding qualia: it would be a mistake to transform the fact that inevitably there is a limit to our capacity to describe things we experience into the supposition that there are absolutely indescribable properties in our experience (Quining Qualia, 544). We have no need of any illicit transformation in order to arrive at that supposition; it just seems to be empirically true.

Aesthetic Properties, Evaluative Force, and Sensibility

327

imagine that your qualia registrations alter in the direction of lesser renement or greater spread, so that the difference between the patches disappears, with both now appearing more or less the same. By hypothesis, this cannot be equally well interpreted as a case of qualia persistence with alteration in memory of qualia, since there simply are no memories of qualia at the designated level of specicity. Thus, rather than the supposition of qualia being bogus it would appear to be Dennetts thought-experiments that are bogus, insofar as they posit empirical equivalence between hypotheses of qualia shifts and hypotheses of qualia persistence with shifts in memory-of-qualia, given that, at least for highly specic sensory qualia, we simply have no such memories at all. The other main strand of argument against qualia in Dennetts paper is that the doctrine that sharply distinguishes qualia from reactions to qualia is simply not sustainable. Dennett holds the root error in qualia thinking to be the conviction that
what counts as the way the juice tastes to x can be distinguished . . . from what is a mere accompaniment [of that taste] . . . One dimly imagines taking such cases and stripping them down gradually to the essentials, leaving their common residuum, the way things look, sound, feel, taste, smell to various individuals at various times, independently of how those individuals are nonperceptually affected and independently of how they are subsequently disposed to behave or believe . . . the fundamental mistake is supposing there is such a residual property to take seriously . . .

The man who likes the taste of beer, the man who dislikes it, and the man who is indifferent to it, suggests Dennett, do not taste the same taste; and a man who dislikes a smell that he used to like now plausibly smells a different smell. The idea is that the hedonic response to a taste or smell is arguably part of what that taste or smell is for the individual in question, so that the notion of a core, evaluatively neutral, smell or taste, common to the experience of subjects with opposed hedonic reactions to it, is simply a ction. Dennetts thought-experiment involving the taste of cauliower is perhaps his most vivid presentation of the matter:
Imagine now the cauliower cure: someone offers me a pill to cure my loathing for cauliower. He promises that after I swallow this pill cauliower will taste exactly the same to me as it always has, but I will like that taste! Hang on, I might reply. I think you may have just contradicted yourself. But in any event, I take the pill and it works. I become an instant cauliower-appreciator, but if I am asked which of
Quining Qualia, 521.

328

Aesthetic Properties

two possible effects the pill has had on me [i.e. the taste quale has stayed the same but the reaction to it has changed vs. the taste quale has changed] I will be puzzled, and will nd nothing in my experience to shed light on the question. Of course I recognize that the taste is (sort of) the samethe pill hasnt made cauliower taste like chocolate cake, after allbut at the same time my experience is so different now that I resist saying that cauliower tastes the way it used to taste.

It is signicant, I think, that the more compelling of Dennetts thoughtexperiments suggesting that qualia and reactions thereto cannot ultimately be distinguished are ones featuring gustatory qualia, i.e. tastes in the literal sense. But they seem much less compelling when colors, say, are substituted for tastes. Suppose someone has a clear hedonic reaction to certain colors, so that he is, say, tickled pink by pink and made really blue by blue. And then suppose he is suddenly freed of his chromohedonia, courtesy of a little pill, so that he now registers pink and blue with no noticeable affect in either direction. We can imagine him now greeting them inwardly with a mere shrug. Is there any plausibility in the suggestion that the phenomenal appearance of those colors will have changed for him? Is there any reason not to credit the claim he would presumably make, that the colors look the same to him, but now affect him differently as far as his mood is concerned? Consider a timbre one is attending to aurally, rst nding it neither appealing nor unappealing, and then gradually coming to regard it as quite attractive. Is it plausible to doubt that there is a constant sensory impression, a quale, in ones experience as ones attitude toward it undergoes a shift? I suspect that visual and auditory qualia give different results in these thought experimentsones more supportive of the separability of quale and reaction theretothan do gustatory and olfactory ones, for several reasons. With colors and timbres as opposed to tastes and smells, the factor of physical satiety generally does not come into play, and the hedonic effects involved are generally not as pronounced. There is also an obvious sense in which we have more distance on colors or timbres than on smells and tastes, which perhaps enables us to distinguish them more clearly from reactions that accompany them. If we mentally compensate for these differences when we review Dennetts tales of ups and down in the liking for tastes and smells I think we will be less inclined to concede that those ups and downs make for change in those tastes and smells themselves.
Ibid. 535.

Aesthetic Properties, Evaluative Force, and Sensibility

329

VI Fourth worry. Well and good, one might respond to what I have said so far. The real problem about aesthetic attribution, one may insist, still remains. It concerns the fact of continuing and irresolvable differences in aesthetic attributions among even ideal critics, that is, the notable absence of convergence in judgments attributing aesthetic properties to works of art among even optimally prepared and positioned perceivers. For it is that fact, above all, that threatens aesthetic realism as regards attributions to works of art. The thought is basically this. Whatever the truth about the descriptive content of substantive aesthetic terms, that is, whether such content is given by qualitative impressions of distinct character or constellations of low-level perceptual features, and especially if such terms have a signicant evaluative dimension, whether that is a matter of sense or of conversational implication in relevant contexts, we just cannot reasonably expect to nd convergence in the aesthetic judgments of even ideal critics. And that is because of the evident diversity of sensibilities among even ideal critics, as even Hume recognized 250 years ago. Ideal critics, like ordinary folk, sort themselves out into sensibility-types or sensibility-groups, and possibly very many of them. If so, then realism about real-world, or thick, aesthetic properties, that is, properties denoted by substantive aesthetic terms of criticism with their evaluative aspects intact, would appear to be a vanishing prospect. That there might always be more stripped-down, thin, aesthetic properties in the offing that warring critics might be brought to acknowledge would seem to offer only small consolation. For there would be no truth about whether or not a given abstract sculpture was, say, graceful, because some ideal critics, of sensibilities A, B, and C, nd it so, in virtue of reacting positively to the core impression presented by something that was a candidate for being described as graceful, while other ideal critics, of sensibilities D, E, and F, do not, in virtue of reacting negatively to that same core impression.

This is the basic thrust of both Goldman, Aesthetic Value, ch. 1, and John Bender, Realism, Supervenience, and Irresolvable Aesthetic Disputes, Journal of Aesthetics and Art Criticism 54 (1996): 37181. Goldman expresses this worry with special force: The prior issue is whether ideal critics as specied above in terms of their characteristics . . . will agree in their aesthetic judgments . . . if ideal critics disagree in their aesthetic judgments, then the [realist] account will have us ascribing incompatible properties to the same artworks. Works will be both graceful and not graceful,

330

Aesthetic Properties

Furthermore, supervenience, a pillar of property realism, would also be imperiled, since aesthetic properties would no longer supervene on a works intrinsic and relational propertiesits structural features plus its artistic contextbecause differing sensibilities among ideal critics would prevent aesthetic properties from emerging on that basis alone. True, one might reply that supervenience of aesthetic properties would still hold, though in a narrow, sensibility-relativized form. Instead of gracefulness simpliciter we would need to speak of, say, gracefulnessq , which would amount roughly to being disposed to appear graceful to ideal critics of a given sensibility class, comprising sensibilities A, B, and C. But there would be a hollow ring to settling for so qualied a supervenience of properties. At any rate, in order to begin to come to terms with the implications of sensibility diversity for aesthetic realism we need to look at what sensibilities in this context might consist in. We should at the outset recognize the possibility not only of a diversity of sensibilities, but of a diversity of kinds of sensibility. There may, I suspect, very well be two basic kinds of sensibility
powerful and not powerful. The issue of agreement among ideal critics is crucial, since the [realist] account becomes incoherent without such consensus (Aesthetic Value, 289). See Bender, Realism. Anti-realists about aesthetic properties will of course propose a different accommodation with the facts of critical divergence, eschewing talk of supervenience entirely: We can now see how nonrealists will modify the relational account of aesthetic properties . . . to avoid the ascription of incompatible properties that plagues the realist. According to the modied account, when I say that an object has a certain aesthetic property, I am saying that ideal critics who generally share my taste will react in a certain way to its more basic properties (Goldman, Aesthetic Value, 37.) Bender, in the course of arguing against the position on aesthetic supervenience adopted by Goldman in earlier articles, draws a somewhat different conclusion: Goldman seems to think that contrary aesthetic properties of the sort that generate irresolvable disputes, properties with opposed evaluative implications, can each be supervenient upon the nonevaluative base properties of a work . . . I propose to argue that Goldman is mistaken . . . [for Goldman] the truth of our aesthetic judgments is to be construed not as a matter of whether all ideal critics would respond as we have, but of whether ideal critics whose tastes we share would so respond . . . [This] produces a relativized supervenience: relative to a specic standard of taste or set of values there can be no change in a rational critics evaluation of a work without some change in its nonevaluative properties . . . But, I contend, a relativized supervenience of this sort is no supervenience at all. Supervenience is a metaphysical dependency relation asserting that changes in supervenient properties arise only with changes in relevant base properties. The constraint we are left with after relativization is nothing more than the rather trivial, and epistemic, constraint of consistency upon rational judgments (Realism, 3723). It seems to me that the relativization that ultimate nonconvergence among ideal critics induces need not take the form that Bender envisages. For instead of relativity to standards which critics accept we can see it as relativity to sensibilities which partly constitute those critics. The supervenience base would need to expand, of course, to include a particular critical sensibility, and the aesthetic property dened by reference to it would then be a correspondingly much narrower one, implicitly indexed to ideal critics of that stripe. But we would then still have something recognizable as property supervenience, and not just a rational consistency constraint on judgments of any sort.

Aesthetic Properties, Evaluative Force, and Sensibility

331

at play here, which we can label perceptual sensibility and attitudinal sensibility. What I have in mind would parallel the distinction between phenomenal impressions and evaluative reactions in my earlier analysis of evaluatively charged aesthetic properties. A perceptual sensibility would be a disposition to receive phenomenal impressions of certain sorts from various constellations of perceivable non-aesthetic features, while an attitudinal sensibility would be a disposition to react to phenomenal impressions of certain sorts with attitudes of favor or disfavor. There is no need to assume that an attitudinal sensibility is necessarily a xed or inborn matter; it might indeed generally have a strong culturally formed component. Furthermore, we would expect that a persons attitudinal sensibility, in regard to given phenomenal impressions, would clearly be modiable over time, especially insofar as such sensibility was culturally, rather than physiologically, ordained or conditioned. Suppose then, for simplicity, that there were three sensibilities of each type. That gives nine sensibility types with both dimensions taken into account: these would be combined, or perceptual-attitudinal, sensibilities. So in practice, assuming ideality in all other respects, there could still very well be nine distinct proles of aesthetic response, corresponding to nine combined sensibility types, with resulting disagreements in all individual cases. The question naturally arises of telling to what aspect of divergence in sensibility we should attribute any particular disagreement. How could we determine, in a particular case, whether disagreement was rooted in differences in perceptual sensibility, or differences in attitudinal sensibility, or both? What might show, more generally, that judges differed in perceptual sensibility instead of attitudinal sensibility or vice versa? However, though we should certainly be open to the idea of possible diversity in perceptual sensibilities, that is, propensities to receive aesthetic impressions from the same perceptual congurations, it is by no means as clear that there are such among ideal critics as that there is a diversity in attitudinal sensibilities, that is, propensities to like or dislike given aesthetic impressions. So if we can separate descriptive matters from evaluative ones
We may note that the thesis of the ineliminable evaluativity of aesthetic attributions and the thesis of the non-convergence in aesthetic attributions among ideal judges because of irreducible differences in sensibility, though related, only come to the same thing if all differences in sensibility are a matter of differing attitudinal sensibilitiesinvolving different reactions to the same phenomenal impressionsrather than differing perceptual sensibilities, that is, differing dispositions to receive such impressions from the same low-level perceptual features.

332

Aesthetic Properties

in this arena, as I have been suggesting that we in principle can, there may be more hope for ultimate convergence in at least descriptive aesthetic attributions among ideal critics than we might think, even if diversity of perceptual sensibilities is, of course, theoretically possible. But what if that theoretical possibility is realized after all? What if it turns out that there are roughly as many distinct perceptual sensibilities as there are blood types? Well, presumably one would then be well advised to nd ones perceptual as well as attitudinal sensibility group, and stay tuned. But that would not alter the fact that suitably relativized aesthetic properties, at any rate, would still be there for the having and experiencing. There are two questions it is natural to pose at this point. First, why does it matter whether there are groups of perceivers among whom one is likely to nd convergence in aesthetic judgments, given that differences of sensibility clearly do induce divergences, even under ideal conditions and in the last analysis, among aesthetic perceivers? Why should one trouble to identify ones sensibility-group, of any sort? An obvious answer, though perhaps there are others, is that aesthetic recommendations from critics belonging to ones own sensibility-group will be of greater practical worth, having more predictive value as regards ones own aesthetic satisfactions than recommendations from other critics. Second, is belonging to one such sensibility-group better than belonging to another, and if so, how? A positive answer to this query might emerge as a consequence of addressing the crucial though largely overlooked problem raised by Humes celebrated essay on taste, namely, that of explaining why judgments of ideal critics should rationally interest all perceivers, even those far from ideality, and why such perceivers have reason to strive for greater ideality in their own aesthetic dispositions insofar as those dispositions are subject to change. But this is not the place to develop that answer. VII Before concluding this essay I turn briey to a critique which John Bender has offered of my earlier defense of aesthetic realism, one that gives the concerns highlighted under this last rubric a vigorous expression. After providing a
I attempt to do so in Humes Standard of Taste: The Real Problem, Journal of Aesthetics and Art Criticism 60 (2002): 22738 (and Ch. 21, this volume). An answer along similar lines can be found in Goldman, Aesthetic Value.

Aesthetic Properties, Evaluative Force, and Sensibility

333

summary of my attack on one sort of aesthetic antirealism, the main points of which were recounted in the opening section of this essay, Bender notes that I try to block such antirealism, which sees aesthetic properties as simply evaluative reactions to non-aesthetic features, by offering an alternative relational analysis of such properties in terms, roughly, of dispositions or propensities of an artwork to produce in appropriately backgrounded perceivers certain holistic phenomenal impressions. However, Bender is not happy with that analysis, and nds wanting the defense of aesthetic realism it underpins. I here focus on what seem to be his two main criticisms. First, evaluatively neutral aesthetic attributions, Bender claims, are as disputable, and as relative to critical taste, as evaluatively charged aesthetic attributions:
But what if the application of purely descriptive aesthetic terms exhibits the same kind of relativity to the tastes or standards of particular judges [as does application of manifestly evaluative aesthetic terms]? Will there not be irresolvable disputes among critics over the most appropriate way to describe or interpret a works aesthetic content? . . . There can be as much critical indeterminacy concerning whether a musical passage is sad or resigned as there is in judging it disunied or only uninhibited . . . I suggest that if aesthetic properties are dispositions to afford phenomenal impressions or looks, as Levinson claims, these impressions will be variable to a not insignicant degree and will reect the judges sensibilities just as surely as his or her evaluative reactions do.

My response is as follows. Of course there might be as much divergence among ideal critics regarding neutral aesthetic content as there is among them in regard to their evaluative reactions to such content. That is, diversity of what I above called perceptual sensibilities might have as much to do with non-convergence in aesthetic judgments among ideal critics as diversity of attitudinal sensibilities. But it seems there is, at least so far as anyone has shown, little reason to think this is the case. To take Benders illustrative example, it is easy to imagine that ideal critics disagreeing over the application to a stretch of music of charged terms like disunied or uninhibited will never be brought to agree on one rather than the other, because of their differing hedonic reactions to or evaluative attitudes toward the music in question. But it is less easy to imagine, given our